1commit 0c01893774031d8429f0a138048fda7957f15706 2Author: Matt Turner <mattst88@gmail.com> 3Date: Tue Apr 25 14:40:08 2023 -0400 4 5 Bump version for 22.0.0 release 6 7 Signed-off-by: Matt Turner <mattst88@gmail.com> 8 9commit 7a6a34af026f0bef8080b91baf97a729380bca01 10Author: Matthieu Herrb <matthieu.herrb@laas.fr> 11Date: Sun Nov 14 17:38:24 2021 +0100 12 13 Don't set SourceValidate pointer to NULL 14 15 inspired from similar code in amdgpu, fixes a crash when xrandr(1) 16 is invoqued with X server 21.1.1 17 18 Signed-off-by: Matthieu Herrb <matthieu.herrb@laas.fr> 19 20commit 983a779b7a5b2153d995727a0ed3da09732e15d5 21Author: Alan Coopersmith <alan.coopersmith@oracle.com> 22Date: Sat Jul 23 18:39:20 2022 -0700 23 24 radeon_glamor_wrappers.c: Convert from ISO-8859-1 to UTF-8 25 26 Allows Flawfinder to analyze source instead of erroring out on encoding 27 28 Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com> 29 30commit a25ca015e6b46bcf6560abe273cfb529c4bdd287 31Author: Alan Coopersmith <alan.coopersmith@oracle.com> 32Date: Sat Jul 23 18:35:08 2022 -0700 33 34 gitlab CI: enable gitlab's builtin static analysis 35 36 Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com> 37 38commit 7b940097545ea60e5089ae67f593311f136ae271 39Author: Alan Coopersmith <alan.coopersmith@oracle.com> 40Date: Sat Jul 23 18:29:13 2022 -0700 41 42 gitlab CI: enable commit & merge request checks 43 44 Uses ci-fairy from freedesktop/ci-templates 45 46 Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com> 47 48commit 8fc442d6a67a35b031be7e725a0999a7e0403b4f 49Author: Alan Coopersmith <alan.coopersmith@oracle.com> 50Date: Sat Jan 22 09:36:23 2022 -0800 51 52 Fix spelling/wording issues 53 54 Found by using: 55 codespell --builtin clear,rare,usage,informal,code,names 56 57 Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com> 58 59commit 11d549d591c7a6f19d2b39d069e99febbd3113da 60Author: Alan Coopersmith <alan.coopersmith@oracle.com> 61Date: Sat Jan 22 09:24:46 2022 -0800 62 63 Build xz tarballs instead of bzip2 64 65 Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com> 66 67commit 5eba006e4129e8015b822f9e1d2f1e613e252cda 68Author: Michel Dänzer <mdaenzer@redhat.com> 69Date: Tue Feb 2 12:45:54 2021 +0100 70 71 Only include dri.h with older versions of xserver 72 73 Not needed anymore with current versions. 74 75commit ecced3b3c6af501c4c90265f8404b6898f3cf037 76Author: Michel Dänzer <mdaenzer@redhat.com> 77Date: Fri Mar 26 17:16:32 2021 +0100 78 79 Add GitLab CI pipeline 80 81 Based on xf86-video-amdgpu, but applying experience gained in the 82 meantime in other projects and taking advantage of new features 83 available with current versions of GitLab. 84 85commit 77d9ab03ca20e683be4c40ccc879e201b538f7e5 86Author: Michel Dänzer <mdaenzer@redhat.com> 87Date: Fri Mar 26 17:37:53 2021 +0100 88 89 Guard local variable info only used with glamor 90 91 Fixes compiler warning with glamor disabled: 92 93 radeon_dri2.c: In function 'radeon_dri2_exchange_buffers': 94 radeon_dri2.c:732:19: error: unused variable 'info' [-Werror=unused-variable] 95 RADEONInfoPtr info = RADEONPTR(xf86ScreenToScrn(screen)); 96 ^~~~ 97 98commit 3c7c84ed49564907a148ae99b03200e0be350060 99Author: Michel Dänzer <mdaenzer@redhat.com> 100Date: Fri Mar 26 17:42:10 2021 +0100 101 102 Guard local variable priv only used with glamor 103 104 Fixes compile errors with glamor disabled: 105 106 ../../src/radeon_present.c: In function ‘radeon_present_check_flip’: 107 ../../src/radeon_present.c:281:21: error: invalid use of undefined type ‘struct radeon_pixmap’ 108 281 | if (priv && priv->fb_failed) 109 | ^~ 110 ../../src/radeon_present.c:288:19: error: invalid use of undefined type ‘struct radeon_pixmap’ 111 288 | if (priv && !priv->fb_failed) { 112 | ^~ 113 ../../src/radeon_present.c:292:10: error: invalid use of undefined type ‘struct radeon_pixmap’ 114 292 | priv->fb_failed = TRUE; 115 | ^~ 116 117commit 8da3e4561ef82bb78c9a17b8cd8bf139b9cfd680 118Author: Dave Airlie <airlied@redhat.com> 119Date: Mon Jul 13 09:11:28 2020 +1000 120 121 ati: cleanup terminology to use primary/secondary 122 123 The X server changed some API/ABIs here. 124 125 Based on amdgpu patch by Michel 126 127commit 3845392426e9798c1fb2a6fc5e97bf5dfd5c443d 128Author: Niclas Zeising <zeising@daemonic.se> 129Date: Tue Apr 14 17:01:57 2020 +0200 130 131 Fix return value check of drmIoctl() 132 133 When the drmModeSetCursor2() call was replaced with bare drmIoctl() call in 134 92df7097, a bug was introduced. With the use of drmModeSetCursor2(), 135 the return value from drmIoctl() (which calls ioctl()) were mangled, if 136 they were negative, they were replaced by -errno by a wrapper function 137 in xf86drMode.c in libdrm. After replacing drmModeSetCursor2() with the 138 call to drmIoctl(), this mangling no longer happens, and we need to 139 explicitly check if the call to drmIoctl() fails, which is indicated by 140 returning -1, and then why it failed, by checking errno. 141 If the error indicated by errno is EINVAL, then we can't use the 142 DRM_IOCTL_MODE_CURSOR2 ioctl(), and need to fall back to the 143 DRM_IOCTL_MODE_CURSOR ioctl(). 144 145 This bug can manifest itself by an invisible hw cursor on systems where the 146 DRM_IOCTL_MODE_CURSOR2 is not implemented by the graphics driver. 147 148 Credit also to Alexey Dokuchaev for help with developing the fix and 149 testing. 150 151 This fixes #190 152 153 Signed-off-by: Niclas Zeising <zeising@daemonic.se> 154 Reviewed-by: Michel Dänzer <mdaenzer@redhat.com> 155 156commit c0eb5dbd9c1db6b6d5b1574bcd8c584170d7ab54 157Author: Alexey Sheplyakov <asheplyakov@altlinux.org> 158Date: Wed Nov 27 20:50:58 2019 +0400 159 160 Don't crash X server if GPU acceleration is not available 161 162 Commit d1d8e3c8d0a0a0394d395eba171460501745209b causes X server 163 to fail on startup when GPU acceleration is not working (or is 164 disabled). The reason is that `radeon_get_pixmap_bo` function 165 gets called too early (before EXA has been initialized) and 166 fails with an assert: 167 168 #0 __GI_raise (sig=<optimized out>) at ../sysdeps/unix/sysv/linux/raise.c:50 169 #1 0x76ab1c6c in __GI_abort () at abort.c:79 170 #2 0x76ac0b64 in __assert_fail_base (fmt=0x76bfbce4 "%s%s%s:%u: %s%sAssertion `%s' failed.\n%n", assertion=0x7658c80c "key->initialized", file=<optimized out>, line=121, 171 function=0x7658d040 <__PRETTY_FUNCTION__.10607> "dixGetPrivateAddr") at assert.c:92 172 #3 0x76ac0c0c in __GI___assert_fail (assertion=0x7658c80c "key->initialized", file=0x7658c9d0 "../include/privates.h", line=121, 173 function=0x7658d040 <__PRETTY_FUNCTION__.10607> "dixGetPrivateAddr") at assert.c:101 174 #4 0x76579e6c in dixGetPrivateAddr (key=<optimized out>, key=<optimized out>, privates=<optimized out>) at ../include/privates.h:121 175 #5 0x7657a954 in dixGetPrivateAddr (key=<optimized out>, key=<optimized out>, privates=<optimized out>) at exa.c:70 176 #6 dixGetPrivate (key=<optimized out>, privates=<optimized out>) at ../include/privates.h:136 177 #7 exaGetPixmapDriverPrivate (pPix=<optimized out>) at exa.c:68 178 #8 0x7623d460 in radeon_get_pixmap_bo (pPix=0x71c1b8) at radeon.h:804 179 #9 radeon_get_pixmap_handle (pixmap=0x71c1b8, handle=0x7fa22328) at radeon_bo_helper.c:357 180 #10 0x76244458 in radeon_pixmap_get_fb (pix=0x71c1b8) at radeon.h:886 181 #11 drmmode_set_mode_major (crtc=0x691860, mode=0x69191c, rotation=<optimized out>, x=<optimized out>, y=<optimized out>) at drmmode_display.c:918 182 #12 0x762467e8 in drmmode_set_desired_modes (pScrn=0x67c678, drmmode=<optimized out>, set_hw=1) at drmmode_display.c:3128 183 #13 0x0047bfa4 in MapWindow (client=0x669ec8, pWin=0x7206c0) at window.c:2722 184 #14 MapWindow (pWin=0x7206c0, client=0x669ec8) at window.c:2665 185 #15 0x00449650 in dix_main (argc=3, argv=0x7fa22604, envp=<optimized out>) at main.c:247 186 #16 0x76ab2198 in __libc_start_main (main=0x42db10 <main>, argc=3, argv=0x7fa22604, init=<optimized out>, fini=0x606434 <__libc_csu_fini>, rtld_fini=0x77229930 <_dl_fini>, 187 stack_end=0x7fa225e0) at libc-start.c:308 188 #17 0x0042db80 in __start () at ../sysdeps/mips/start.S:110 189 190 Don't call `exaGetPixmapDriverPrivate` if the acceleration (EXA) is not 191 enabled [yet] to avoid the problem. 192 193 Closes: https://gitlab.freedesktop.org/xorg/driver/xf86-video-ati/issues/188 194 Closes: https://bugzilla.altlinux.org/show_bug.cgi?id=37539 195 196commit 4d84cf438e7f1bebf0053035ef0292e9fed257d1 197Author: Michel Dänzer <mdaenzer@redhat.com> 198Date: Fri Nov 29 16:37:32 2019 +0100 199 200 Handle NULL fb_ptr in pixmap_get_fb 201 202 This can happen when HW acceleration is disabled. 203 204 Fixes https://gitlab.freedesktop.org/xorg/driver/xf86-video-ati/issues/188 205 206commit f223035f4ffcff2a9296d1e907a5193f8e8845a3 207Author: Adam Jackson <ajax@redhat.com> 208Date: Tue Feb 4 16:38:06 2020 -0500 209 210 Fix link failure with gcc 10 211 212 Without the 'extern' this looks like a definition not just a 213 declaration, in every file that includes the header. gcc 10 is stricter 214 about this kind of multiple definition. 215 216 Reviewed-by: Michel Dänzer <mdaenzer@redhat.com> 217 218commit b9bd8097e1d2c088b081f1b81799ea3892406214 219Author: Michel Dänzer <mdaenzer@redhat.com> 220Date: Tue Oct 15 17:54:12 2019 +0200 221 222 Bump version for 19.1.0 release 223 224commit 2faaecc69b127248718e759c6c98c84d56dd1b6b 225Author: Michel Dänzer <mdaenzer@redhat.com> 226Date: Fri Sep 20 17:24:19 2019 +0200 227 228 Don't unreference FBs of pixmaps from different screens in LeaveVT 229 230 FindClientResourcesByType finds pixmaps from all screens, but trying to 231 process ones from other screens here makes no sense and likely results 232 in a crash or memory corruption. 233 234 Fixes: 06a465484101 ("Make all active CRTCs scan out an all-black 235 framebuffer in LeaveVT") 236 237commit 2cbbd8648cdd27db8076565943b932ef81337053 238Author: Michel Dänzer <mdaenzer@redhat.com> 239Date: Fri Sep 20 18:47:02 2019 +0200 240 241 Don't set up black scanout buffer if LeaveVT is called from CloseScreen 242 243 Avoids a crash described in 244 https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/merge_requests/43#note_223718 245 246 (Ported from amdgpu commit 5b8bc9fc505c551dcd9b0ed5ab835a49fa4f9fda) 247 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 248 249commit c7ed12cb2ea76999351d7cb87877224bdc0664f0 250Author: Michel Dänzer <michel.daenzer@amd.com> 251Date: Thu Jul 18 18:14:48 2019 +0200 252 253 Don't disable page flipping completely with SW cursor 254 255 Even with SW cursor, page flipping can be used while no X cursor is 256 visible. 257 258 Occurred to me in the context of xorg/xserver#828. 259 (Ported from amdgpu commit 87f41ace4920fd2069794211683659eb25b025a6) 260 261commit 33803c85f761d343aa7300311b8e9489b1a89495 262Author: Michel Dänzer <michel.daenzer@amd.com> 263Date: Tue Jun 25 17:46:23 2019 +0200 264 265 present: Check that we can get a KMS FB for flipping 266 267 This can legitimately fail if the pixmap's storage is shared from 268 another device, e.g. when using PRIME render offloading. 269 270 (Ported from amdgpu commit 7d3fef72e0c871e1677e9e544f4cae5e238b5c52) 271 272commit fee737e82837dc8282a832fc9391ed959c4c3737 273Author: Michel Dänzer <michel.daenzer@amd.com> 274Date: Fri Jun 14 11:23:34 2019 +0200 275 276 Remove dri2_drawable_crtc parameter consider_disabled 277 278 All callers were passing TRUE. 279 280 (Ported from amdgpu commit ea19a5207054bb159fc7fb6d88e0ceb10c3da010) 281 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 282 283commit 2a3f2d2089f603c99be54c98d7033155e771ce7b 284Author: Michel Dänzer <michel.daenzer@amd.com> 285Date: Fri Jun 14 11:21:40 2019 +0200 286 287 dri2: Re-use previous CRTC when possible if pick_best_crtc returns NULL 288 289 This way, the MSC will continue ticking at the rate of (the last mode 290 which was enabled for) that CRTC, instead of the client running 291 unthrottled. 292 293 (Ported from amdgpu commit 3109f088fdbd89c2ee8078625d4f073852492656) 294 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 295 296commit d5f5bc5846ef06c3ecf9e5acf1ca357574f06f5a 297Author: Flora Cui <flora.cui@amd.com> 298Date: Fri Jun 14 11:20:12 2019 +0200 299 300 dri2: reply to client for WaitMSC request in any case 301 302 otherwise client would wait for reply forever and desktop appears hang. 303 304 Signed-off-by: Flora Cui <flora.cui@amd.com> 305 (Ported from amdgpu commit fb06fb814700a47464abd756e1111dcc76d0d776) 306 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 307 308commit f758908db4e71406e5d437d32e43aabd38a63504 309Author: Michel Dänzer <michel.daenzer@amd.com> 310Date: Thu May 9 12:35:37 2019 +0200 311 312 dri3: Always flush glamor before sharing pixmap storage with clients 313 314 Even if glamor_gbm_bo_from_pixmap / glamor_fd_from_pixmap themselves 315 don't trigger any drawing, there could already be unflushed drawing to 316 the pixmap whose storage we share with a client. 317 318 (Ported from amdgpu commit 4b17533fcb30842caf0035ba593b7d986520cc85) 319 Acked-by: Alex Deucher <alexander.deucher@amd.com> 320 321commit d1d8e3c8d0a0a0394d395eba171460501745209b 322Author: Michel Dänzer <michel.daenzer@amd.com> 323Date: Wed Apr 24 12:25:39 2019 +0200 324 325 Retry get_fb_ptr in get_fb 326 327 If get_fb_ptr returns NULL, try again after pixmap_get_handle, it should 328 work then. 329 330 Fixes spurious Present page flipping failures using "normal" pixmaps 331 which aren't shared with direct rendering clients, e.g. with a 332 compositor using the RENDER extension. 333 334 Bugzilla: https://bugs.freedesktop.org/110417 335 (Ported from amdgpu commit bf61e6d7ac1a5754b1026d7f80acf25ef622c491) 336 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 337 338commit 36703f66c3b06875651606a6280d5dc9d9dad51e 339Author: Michel Dänzer <michel.daenzer@amd.com> 340Date: Tue Mar 19 18:01:02 2019 +0100 341 342 Bump version for 19.0.1 release 343 344commit 4407c78bd86da4460ee07a15a365e07d99e0dd27 345Author: Dave Airlie <airlied@redhat.com> 346Date: Thu Mar 14 11:17:32 2019 +0100 347 348 modesetting: add tile property support 349 350 This adds tiling support to the driver, it retrieves the tile info from 351 the kernel and translates it into the server format and exposes the 352 property. 353 354 (Ported from xserver commits 8fb8bbb3062f1a06621ab7030a9e89d5e8367b35 355 and 6abdb54a11dac4e8854ff94ecdcb90a14321ab31) 356 (Ported from amdgpu commit 6ee857726166f495abcd68e4ff60e3a09593d079) 357 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 358 359commit 79bc0e054f37026377d54cac6cd8127d4aa9baca 360Author: Michel Dänzer <michel.daenzer@amd.com> 361Date: Thu Mar 14 11:10:51 2019 +0100 362 363 Use radeon_finish in drmmode_crtc_scanout_update 364 365 radeon_glamor_finish only works if we're using glamor, otherwise it'll 366 crash. 367 368 Fixes: ce7db51020d3 "Cancel pending scanout update in drmmode_crtc_scanout_update" 369 Bug: https://bugs.debian.org/924540 370 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 371 372commit f6cd4a67d7de5d3ff1a6e58a8c83749fc8ffc310 373Author: Michel Dänzer <michel.daenzer@amd.com> 374Date: Fri Mar 8 11:48:10 2019 +0100 375 376 Revert "glamor: Avoid glamor_create_pixmap for pixmaps backing windows" 377 378 This reverts commit 274703087f80342f51fa69c935bb9a1cb0c4ae47. 379 380 Reports of visual corruption were bisected to this, e.g. 381 https://bugs.archlinux.org/task/61941 . I can reproduce this with Turks, 382 but not with Bonaire. I assume it's a Mesa/glamor bug, but let's revert 383 for now. 384 385 Acked-by: Alex Deucher <alexander.deucher@amd.com> 386 387commit 0d132d99e0b750896a78f47d73a8639680495d8c 388Author: Michel Dänzer <michel.daenzer@amd.com> 389Date: Wed Mar 6 17:48:03 2019 +0100 390 391 Bump version for 19.0.0 release 392 393commit c301b8af25d2c2cd49035a4395ebe6c3612df366 394Author: Michel Dänzer <michel.daenzer@amd.com> 395Date: Fri Mar 1 18:28:11 2019 +0100 396 397 dri2: Call drm_queue_handle_deferred in dri2_deferred_event 398 399 drm_queue_handler just puts the event on the signalled list; without 400 calling drm_queue_handle_deferred, actual processing of the event may be 401 delayed indefinitely, e.g. until another event arrives from the kernel. 402 403 This could result in DRI2 clients hanging during DPMS off. 404 405 Fixes: ba83a866af5a "Add radeon_drm_handle_event wrapper for 406 drmHandleEvent" 407 (Ported from amdgpu commit 09be74a3d1dd9604336d9a27f98d132b262dcbaf) 408 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 409 410commit 705020b6247eaa062edc9c88e6ad52f8c5468051 411Author: Michel Dänzer <michel.daenzer@amd.com> 412Date: Fri Mar 1 18:23:30 2019 +0100 413 414 present: Check that flip and screen pixmap pitches match 415 416 If they don't, flipping will result in corrupted display. 417 418 Test case: 419 420 * Run Xorg at 1920x1080 with no window manager 421 * glxgears -geometry 2048x1080 422 423 The Present extension code in xserver 1.21 will check for this. 424 425 (Ported from amdgpu commit a636f42b496b0604ca00a144690ece61d1a88a27) 426 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 427 428commit 15697ee242c30b9ea6775624e8282e0171a113a7 429Author: Michel Dänzer <michel.daenzer@amd.com> 430Date: Mon Jan 28 18:27:10 2019 +0100 431 432 Keep waiting for a pending flip if drm_handle_event returns 0 433 434 drm_wait_pending_flip stopped waiting if drm_handle_event returned 0, 435 but that might have processed only some unrelated DRM events. As long as 436 the flip is pending, we have to keep waiting for its completion event. 437 438 Noticed while working on the previous fix. 439 440 (Ported from amdgpu commit 9045fb310f88780e250e60b80431ca153330e61b) 441 442commit 227123de3d862e691131708b7f55260bee17f2b7 443Author: Michel Dänzer <michel.daenzer@amd.com> 444Date: Mon Jan 28 18:24:41 2019 +0100 445 446 Call drmHandleEvent again if it was interrupted by a signal 447 448 drmHandleEvent can be interrupted by a signal in read(), in which case 449 it doesn't process any events but returns -1, which 450 drm_handle_event propagated to its callers. This could cause the 451 following failure cascade: 452 453 1. drm_wait_pending_flip stopped waiting for a pending flip. 454 2. Its caller cleared drmmode_crtc->flip_pending before the flip 455 completed. 456 3. Another flip was attempted but got an unexpected EBUSY error because 457 the previous flip was still pending. 458 4. TearFree was disabled due to the error. 459 460 The solution is to call drmHandleEvent if it was interrupted by a 461 signal. We can do that in drm_handle_event, because when that is called, 462 either it is known that there are events ready to be processed, or the 463 caller has to wait for events to arrive anyway. 464 465 Bugzilla: https://bugs.freedesktop.org/109364 466 (Ported from amdgpu commit 3ff2cc225f6bc08364ee007fa54e9d0150adaf11) 467 468commit 1bfdccf7639ee2f655dc659cafa63830ba28be85 469Author: Michel Dänzer <michel.daenzer@amd.com> 470Date: Mon Jan 28 18:11:10 2019 +0100 471 472 Only update drmmode_crtc->flip_pending after actually submitting a flip 473 474 And only clear it if it matches the framebuffer of the completed flip 475 being processed. 476 477 Fixes 478 479 (WW) RADEON(0): flip queue failed: Device or resource busy 480 (WW) RADEON(0): Page flip failed: Device or resource busy 481 (EE) RADEON(0): present flip failed 482 483 due to clobbering drmmode_crtc->flip_pending. 484 485 Reproducer: Enable TearFree, run warzone2100 fullscreen, toggle 486 Vertical sync on/off under Video Options. Discovered while investigating 487 https://bugs.freedesktop.org/109364 . 488 489 (Ported from amdgpu commit e72a02ba1d35743fefd939458b9d8cddce86e7f5) 490 491commit dcd3527299c1f6d6faa401c565fa884f4d8f3287 492Author: Michel Dänzer <michel.daenzer@amd.com> 493Date: Mon Jan 28 18:08:35 2019 +0100 494 495 Don't allow TearFree scanout flips to complete in the same vblank period 496 497 We were using a relative target of 0, meaning "complete the flip ASAP". 498 This could result in the flip sometimes, but not always completing in 499 the same vertical blank period where the corresponding drawing occurred, 500 potentially causing judder artifacts with applications updating their 501 window contents synchronized to the display refresh. A good way to test 502 this is the vsynctester.com site in a windowed browser, where the judder 503 results in the large "VSYNC" text intermittently appearing red or cyan 504 instead of the expected gray. 505 506 To avoid this, use a relative target MSC of 1, meaning that if a 507 vertical blank period is in progress, the flip will only complete in the 508 next one. 509 510 Reported by Julian Tempel and Brandon Wright in 511 https://bugs.freedesktop.org/106175 . 512 513 (Ported from amdgpu commit a1b479c7d0066c481af920f297d6af9009dda11e) 514 515commit 274703087f80342f51fa69c935bb9a1cb0c4ae47 516Author: Michel Dänzer <michel.daenzer@amd.com> 517Date: Mon Jan 28 18:06:50 2019 +0100 518 519 glamor: Avoid glamor_create_pixmap for pixmaps backing windows 520 521 If the compositing manager uses direct rendering (as is usually the case 522 these days), the storage of a pixmap allocated by glamor_create_pixmap 523 needs to be reallocated for sharing it with the compositing manager. 524 Instead, allocate pixmap storage which can be shared directly. 525 526 (Ported from amdgpu commit bf326f2ea19daa6c8da23d6788ff301ae70b8e69) 527 528commit 6d1dfe2523e900517bd1e8743c87d6990a82c800 529Author: Michel Dänzer <michel.daenzer@amd.com> 530Date: Mon Jan 28 18:04:41 2019 +0100 531 532 dri2: Flush in dri2_create_buffer2 after calling glamor_set_pixmap_bo 533 534 To make sure the client can't use the shared pixmap storage for direct 535 rendering first, which could produce garbage. 536 537 Bugzilla: https://bugs.freedesktop.org/109235 538 (Ported from amdgpu commit ebd32b1c07208f8dbe853e089f5e4b7c6a7a658a) 539 540commit 77d7abf46446522e686c6b6f1e4857458589ef37 541Author: Michel Dänzer <michel.daenzer@amd.com> 542Date: Mon Jan 28 18:00:20 2019 +0100 543 544 dri3: Flush if necessary in dri3_fd_from_pixmap 545 546 To make sure the client can't use the shared pixmap storage for direct 547 rendering first, which could produce garbage. 548 549 Bugzilla: https://bugs.freedesktop.org/109235 550 (Ported from amdgpu commit d168532ee739f7e33a2798051e64ba445dd3859f) 551 552commit b1c01698f577577e4a88bad0ae08fb5d998e7ebb 553Author: Michel Dänzer <michel.daenzer@amd.com> 554Date: Wed Jan 9 11:26:14 2019 +0100 555 556 Only call drmmode_validate_leases if RandR is enabled 557 558 It would crash if RandR is disabled, e.g. because Xinerama is enabled. 559 560 Bugzilla: https://bugs.freedesktop.org/109230 561 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 562 563commit 38db1bbcfc019c92884c7819a6630c70e543f6b2 564Author: Michel Dänzer <michel.daenzer@amd.com> 565Date: Wed Jan 9 11:33:18 2019 +0100 566 567 Only call drmmode_uevent_init if RandR is enabled 568 569 There's no point in listening for hotplug events if RandR is disabled, 570 as there's no other mechanism for them to be propagated. We were already 571 mostly ignoring them in that case. 572 573 Inspired by 574 https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/commit/1a489142c8e6a4828348cc9afbd0f430d3b1e2d8 575 (via https://bugs.freedesktop.org/109230#c11). 576 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 577 578commit 803f872f7d4b2d80be434bb42ce64dfd295b122c 579Author: Michel Dänzer <michel.daenzer@amd.com> 580Date: Fri Dec 21 18:04:21 2018 +0100 581 582 Use two HW cursor buffers per CRTC 583 584 Switch to the other buffer when xf86_config->cursor changes. Avoids 585 these issues possible when re-using the same buffer: 586 587 * The HW may intermittently display a mix of the old and new cursor 588 images. 589 * If the hotspot changes, the HW may intermittently display the new 590 cursor image at the location corresponding to the old image's hotspot. 591 592 Bugzilla: https://bugs.freedesktop.org/108832 593 (Ported from amdgpu commit 0d60233d26ec70d4e1faa343b438e33829c6d5e4) 594 595commit 91e557f78ad261e76a1829f54722c2c0781742d2 596Author: Michel Dänzer <michel.daenzer@amd.com> 597Date: Fri Dec 21 18:00:09 2018 +0100 598 599 Update cursor position in drmmode_show_cursor if hotspot changed 600 601 The cursor position is updated to be consistent with the new hotspot in 602 the same ioctl call. 603 604 (Ported from amdgpu commit b04697de5270e8e45744a7025c24df1f454a4cf0) 605 606commit 92df709786830d4e30a106dd49d8e0355c50c8f0 607Author: Michel Dänzer <michel.daenzer@amd.com> 608Date: Fri Dec 21 17:54:36 2018 +0100 609 610 Use drmIoctl in drmmode_show_cursor 611 612 This should be functionally equivalent to what drmModeSetCursor(2) do 613 behind the scenes, but allows for new tricks in following changes. 614 615 (Ported from amdgpu commit b344e1559e936046ef02c777fc4f6bcefa3830bc) 616 617commit e14c3d2f86c7be2b5c3d06a47bf0abe954207d0b 618Author: Michel Dänzer <michel.daenzer@amd.com> 619Date: Fri Dec 21 17:38:51 2018 +0100 620 621 Drop RADEONInfoRec::cursor_bo array 622 623 Not needed or even useful for anything. 624 625 (Ported from amdgpu commit e95044e45350870fa7e237860e89ade91ac03550) 626 627commit f66254c171f5a3b052a2a9e0339f17dfb5a60dc2 628Author: Michel Dänzer <michel.daenzer@amd.com> 629Date: Fri Dec 21 12:49:27 2018 +0100 630 631 Automatically try re-enabling TearFree after a flip failed 632 633 Specifically, after both the page flip and vblank ioctls failed, but 634 then the vblank ioctl started working again. This can happen 635 intermittently e.g. when hotplugging a DP display. Previously, TearFree 636 would stay disabled in that case until a modeset was triggered somehow. 637 638 Bugzilla: https://bugs.freedesktop.org/103791 639 (Ported from amdgpu commit bcfa6c258fdf41a9928f8a3c78fc528d0fafee25) 640 641commit ce7db51020d32f17e442338bfd305220feb51630 642Author: Michel Dänzer <michel.daenzer@amd.com> 643Date: Fri Dec 21 12:47:18 2018 +0100 644 645 Cancel pending scanout update in drmmode_crtc_scanout_update 646 647 drmmode_crtc_scanout_update does the equivalent of a scanout update, 648 so no need to do it again. This might also avoid issues if there's a 649 pending scanout update at this point. 650 651 (Ported from amdgpu commit 4e7a24ac5a64e402146953ec5850d13c05742116) 652 653commit 21d65e5b78c8889e363aee8596cd0b0f942fee46 654Author: Michel Dänzer <michel.daenzer@amd.com> 655Date: Fri Dec 21 12:38:04 2018 +0100 656 657 Perform scanout buffer update immediately if drmmode_wait_vblank fails 658 659 Otherwise the damaged screen contents may never be displayed in that 660 case. 661 662 (Ported from amdgpu commit 500fadb16285146e91f62fce3a0ce1360ca684ba) 663 664commit 64942d2c49f9fa1afcc42c07943d076a40963e51 665Author: Michel Dänzer <michel.daenzer@amd.com> 666Date: Thu Dec 20 19:00:08 2018 +0100 667 668 Move deferred vblank events to separate drm_vblank_deferred list 669 670 It was still possible for nested xorg_list_for_each_entry_safe loops 671 to occur over the drm_vblank_signalled list, which could mess up that 672 list. Moving deferred events to a separate list allows processing the 673 drm_vblank_signalled list without xorg_list_for_each_entry_safe. 674 675 Bugzilla: https://bugs.freedesktop.org/108600 676 (Ported from amdgpu commit 51ba6dddee40c3688d4c7b12eabeab516ed153b7) 677 678commit f450632077843a95a6ef269febbfb64a605045ed 679Author: Michel Dänzer <michel.daenzer@amd.com> 680Date: Thu Dec 20 18:48:19 2018 +0100 681 682 Explicitly keep track of whether a DRM event is for a flip or not 683 684 When an async flip is performed, and TearFree is enabled on the CRTC 685 used for timing, we schedule a vblank event for completing the page 686 flip. The DRM event queuing code treated this event like a vblank event, 687 but it needs to be treated like a page flip event. 688 689 (Ported from amdgpu commit e2c7369cae65069aa93eed1c0b678f975ce5c274) 690 691commit 189b6facb3988c00c96d970f8c13ed8d58fa3998 692Author: Michel Dänzer <michel.daenzer@amd.com> 693Date: Thu Dec 20 18:44:24 2018 +0100 694 695 Use drm_abort_one in drm_queue_handler 696 697 At this point, we've already established that e->handler is NULL, no 698 need to check again in drm_queue_handle_one. This also makes it clearer 699 what's happening. 700 701 (Ported from amdgpu commit eda571222f5a6be47f8897e82d85199bb9d95251) 702 703commit 06957e3067b9eb38d30b32d98a07895399f4cb36 704Author: Michel Dänzer <michel.daenzer@amd.com> 705Date: Thu Dec 20 18:40:19 2018 +0100 706 707 glamor: Can work at depth >= 15 with current xserver Git master 708 709 (Ported from amdgpu commit 0734cdf544ffd3f2ac8749ad0e4bf43f8a5cea50) 710 711commit 99ac121770da53196124d80375a5c8edbcf827fa 712Author: Michel Dänzer <michel.daenzer@amd.com> 713Date: Thu Dec 20 18:38:33 2018 +0100 714 715 Skip gamma correction of cursor data if premultiplied R/G/B > alpha 716 717 The un-premultiplied R/G/B values would overflow the gamma LUT, so just 718 pass through the data unchanged, and leave it up to the HW how to 719 interpret such weird premultiplied alpha pixels. 720 721 Bugzilla: https://bugs.freedesktop.org/108355 722 (Ported from amdgpu commit 13c94a373b4858a2d2aa14c22b5f98d53c84c0d9) 723 724commit 0c40a76d1c050d018e6d59bebb5efc9c62be308c 725Author: Michel Dänzer <michel.daenzer@amd.com> 726Date: Thu Dec 20 18:25:21 2018 +0100 727 728 Detect and fix up non-premultiplied cursor data 729 730 X server >= 1.18 already has code for this, but handle it with older X 731 servers as well. 732 733 (Ported from amdgpu commits ad6dfb0124860cf67730bde85867f81d9258c84d & 734 426f9a49655f01863cf4d898f525e5f95984e0c4) 735 736commit 0058fd2ebf4c900b12f129984e98886a7ac84b2f 737Author: Mario Kleiner <mario.kleiner.de@gmail.com> 738Date: Tue Dec 18 11:40:35 2018 +0100 739 740 Fix crash when page flipping in multi-X-Screen/Zaphod mode 741 742 radeon_do_pageflip() indexed the flipdata->fb[] array 743 indexing over config->num_crtc, but the flip completion 744 routines, e.g., drmmode_flip_handler(), index that array 745 via the crtc hw id from drmmode_get_crtc_id(crtc). 746 747 This is mismatched and causes indexing into the wrong 748 array slot at flip completion -> Server crash. 749 750 Always use drmmode_get_crtc_id(crtc) for indexing into 751 the array to fix this. 752 753 Tested on a dual-X-Screen setup with one video output 754 assigned to each X-Screen, page-flipping an OpenGL app 755 on either of both X-Screens. This used to crash when 756 flipping on X-Screen 1, now it doesn't anymore. 757 758 Fixes: 740f0850f1e4 "Store FB for each CRTC in drmmode_flipdata_rec" 759 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 760 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 761 762commit b559fd83b2b026b50fe39a2b69bdc0901fd3330c 763Author: Michel Dänzer <michel.daenzer@amd.com> 764Date: Mon Nov 19 11:38:22 2018 +0100 765 766 Post-release version bump 767 768commit 33cbd449b5305b8e8c0319af8becc1ab1832c101 769Author: Alan Coopersmith <alan.coopersmith@oracle.com> 770Date: Sun Nov 18 12:22:16 2018 -0800 771 772 Update README for gitlab migration 773 774 Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com> 775 776commit 5e6fa5c17a810127f0049816f20db6b871ca77e0 777Author: Michel Dänzer <michel.daenzer@amd.com> 778Date: Wed Oct 24 18:22:05 2018 +0200 779 780 Allow up to six instances in Zaphod mode 781 782 Corresponding to up to six CRTCs being available in the hardware. 783 784 (Ported from amdgpu commit c9d43c1deb9a9cfc41a8d6439caf46d12d220853) 785 786commit c480fd066fe129fa6561fca8c09f037613b753e8 787Author: Michel Dänzer <michel.daenzer@amd.com> 788Date: Wed Oct 24 18:19:42 2018 +0200 789 790 Make wait_pending_flip / handle_deferred symmetric in set_mode_major 791 792 We were always calling the latter, but not always the former, which 793 could result in handling deferred DRM events prematurely. 794 795 (Ported from amdgpu commit 955373a3e69baa241a1f267e96d04ddb902f689f) 796 797commit 86fe8d27b9a3f043e5288ce50eaf1f5fffd24516 798Author: Michel Dänzer <michel.daenzer@amd.com> 799Date: Wed Oct 24 18:14:02 2018 +0200 800 801 Handle pending scanout update in drmmode_crtc_scanout_free 802 803 We have to wait for a pending scanout flip or abort a pending scanout 804 update, otherwise the corresponding event handler will likely crash 805 after drmmode_crtc_scanout_free cleaned up the data structures. 806 807 Fixes crash after VT switch while dedicated scanout pixmaps are enabled 808 for any CRTC. 809 810 (Ported from amdgpu commit 0cd2c337d2c02b8ec2bd994d6124b4aaaad10741) 811 812commit f892d3791219d1041e0cbb1b866e15774004aa18 813Author: Michel Dänzer <michel.daenzer@amd.com> 814Date: Mon Oct 15 17:14:41 2018 +0200 815 816 dri3: Handle radeon_get_pixmap_bo returning NULL 817 818 We were trying to already, but testing the wrong pointer. 819 820 Fixes: b85b7b11f5b5 "Add struct radeon_buffer" 821 Bug: https://bugs.debian.org/910846 822 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 823 824commit 80514b7b1b88f46f3965ff023035dea83b57e652 825Author: Michel Dänzer <michel.daenzer@amd.com> 826Date: Fri Sep 14 17:48:49 2018 +0200 827 828 Bump version for 18.1.0 release 829 830commit 2d58830c3feafc54dccc0b7bf761a466437d4a09 831Author: Michel Dänzer <michel.daenzer@amd.com> 832Date: Wed Sep 5 18:23:29 2018 +0200 833 834 Fix uninitialized use of local variable pitch in radeon_setup_kernel_mem 835 836 Fixes server reset. 837 838 Pointed out by clang: 839 840 ../../src/radeon_kms.c:2721:9: warning: variable 'pitch' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized] 841 if (!info->front_buffer) { 842 ^~~~~~~~~~~~~~~~~~~ 843 ../../src/radeon_kms.c:2765:27: note: uninitialized use occurs here 844 pScrn->displayWidth = pitch / cpp; 845 ^~~~~ 846 847commit 5d5d883496842da84d9418e91cb13454751da625 848Author: Michel Dänzer <michel.daenzer@amd.com> 849Date: Fri Sep 7 18:16:22 2018 +0200 850 851 Bail early from drm_wait_pending_flip if there's no pending flip 852 853 No need to process any events in that case. 854 855 (Ported from amdgpu commit ca5eb9894fff153c0a1df7bdc4a4745713309e27) 856 857commit 4c7d5e50a5e469a541bc463cecb505fe850c0824 858Author: Michel Dänzer <michel.daenzer@amd.com> 859Date: Wed Sep 5 11:29:43 2018 +0200 860 861 Don't use xorg_list_for_each_entry_safe for signalled flips 862 863 drm_wait_pending_flip can get called from drm_handle_event, in which 864 case xorg_list_for_each_entry_safe can end up processing the same entry 865 in both. To avoid this, just process the first list entry until the list 866 is empty. 867 868 (Ported from amdgpu commit 26770be44b89b83bf39c28f2fe284c8cb92ed0c0) 869 870commit 87b9a3e516d19dd1b89a64f6cac990fae53fc1b3 871Author: Michel Dänzer <michel.daenzer@amd.com> 872Date: Wed Sep 5 11:27:25 2018 +0200 873 874 Always delete entry from list in drm_queue_handler 875 876 We left entries without a handler hook in the list, so the list could 877 keep taking longer to process and use up more memory. 878 879 (Ported from amdgpu commit 7eea3e2cd74eed22e982319144e18ae5b1087b78) 880 881commit de88ea2755611bdcb18d91d8234d2ab5be8ff2e9 882Author: Michel Dänzer <michel.daenzer@amd.com> 883Date: Fri Aug 24 11:00:16 2018 +0200 884 885 glamor: Handle ihandle == -1 in radeon_glamor_set_shared_pixmap_backing 886 887commit 84bde2df5b453f8aab35aa18b0cf1f20b8d20488 888Author: Michel Dänzer <michel.daenzer@amd.com> 889Date: Fri Aug 24 10:58:51 2018 +0200 890 891 EXA: Handle ihandle == -1 in RADEONEXASharePixmapBacking 892 893commit 1799680f7bd84e0618f34f4c7486799521ddaf83 894Author: Michel Dänzer <michel.daenzer@amd.com> 895Date: Fri Aug 24 10:56:58 2018 +0200 896 897 Handle ihandle == -1 in radeon_set_shared_pixmap_backing 898 899 It means to stop using the shared pixmap backing. 900 901 Fixes crash when changing PRIME slave output configuration. 902 903commit 3c42bd04800211f64a17640c7ce7c4111ffd4978 904Author: Michel Dänzer <michel.daenzer@amd.com> 905Date: Fri Aug 24 10:55:49 2018 +0200 906 907 EXA: Handle NULL BO pointer in radeon_set_pixmap_bo 908 909commit cba8fe4d64819aaa8ba516aa68dbe6d2aa153046 910Author: Michel Dänzer <michel.daenzer@amd.com> 911Date: Thu Aug 23 18:59:18 2018 +0200 912 913 Use AC_CONFIG_MACRO_DIR instead of AC_CONFIG_MACRO_DIRS 914 915 Older versions of autoconf only supported the former. 916 917 Acked-by: Alex Deucher <alexander.deucher@amd.com> 918 919commit 7b01c10137aba24c8f61dd9b2a19ea257ad24371 920Author: Michel Dänzer <michel.daenzer@amd.com> 921Date: Thu Aug 23 18:42:28 2018 +0200 922 923 Add m4 directory 924 925 Although normally it only warns about it, under some circumstances, 926 aclocal can error out if this directory doesn't exist. 927 928 Reported-by: John Lumby <johnlumby@hotmail.com> 929 Acked-by: Alex Deucher <alexander.deucher@amd.com> 930 931commit 60cd28bbbe4f9db391b4f27c5ff9289d5c2bef40 932Author: Michel Dänzer <michel.daenzer@amd.com> 933Date: Thu Aug 16 18:06:31 2018 +0200 934 935 Remove drmmode_crtc_private_rec::present_vblank_* related code 936 937 Not needed anymore with the more robust mechanisms for preventing nested 938 drmHandleEvent calls introduced in the previous changes. 939 940 (Ported from amdgpu commit 85cd8eef0cbed7b409b07f58d76dacd34aa3ddea) 941 Acked-by: Alex Deucher <alexander.deucher@amd.com> 942 943commit c42f6e2e61d166c8d3ef3fcad175d7050a00288b 944Author: Michel Dänzer <michel.daenzer@amd.com> 945Date: Fri Jul 20 16:56:22 2018 +0200 946 947 Defer vblank event handling while waiting for a pending flip 948 949 This is to avoid submitting more flips while we are waiting for pending 950 ones to complete. 951 952 (Ported from amdgpu commit e52872da69ecc84dafb3355839e35b0383f0d228) 953 Acked-by: Alex Deucher <alexander.deucher@amd.com> 954 955commit ba83a866af5a3784fc4822614375cc081e93197c 956Author: Michel Dänzer <michel.daenzer@amd.com> 957Date: Thu Aug 16 17:44:45 2018 +0200 958 959 Add radeon_drm_handle_event wrapper for drmHandleEvent 960 961 Instead of processing DRM events directly from drmHandleEvent's 962 callbacks, there are three phases: 963 964 1. drmHandleEvent is called, and signalled events are re-queued to 965 _signalled lists from its callbacks. 966 2. Signalled page flip completion events are processed. 967 3. Signalled vblank events are processed. 968 969 This should make sure that we never call drmHandleEvent from one of its 970 callbacks, which would usually result in blocking forever. 971 972 (Ported from amdgpu commit 739181c8d3334ff14b5a607895dfdeb29b0d9020) 973 Acked-by: Alex Deucher <alexander.deucher@amd.com> 974 975commit 93621e408c17dd9e082236c17f051c06558d7f4d 976Author: Michel Dänzer <michel.daenzer@amd.com> 977Date: Thu Aug 16 17:27:06 2018 +0200 978 979 Add radeon_drm_wait_pending_flip function 980 981 Replacing the drmmode_crtc_wait_pending_event macro. 982 983 (Ported from amdgpu commit 6029794e8a35417faf825491a89b85f713c77fc1) 984 Acked-by: Alex Deucher <alexander.deucher@amd.com> 985 986commit ecdf0b7ec9378bc386ce8276f16fb16d09f72812 987Author: Michel Dänzer <michel.daenzer@amd.com> 988Date: Thu Aug 16 17:14:42 2018 +0200 989 990 Move DRM event queue related initialization to radeon_drm_queue_init 991 992 And make radeon_drm_queue_handler not directly accessible outside of 993 radeon_drm_queue.c. 994 995 (Ported from amdgpu commit 0148283984c77f7a6e97026edc3093497547e0a4) 996 Acked-by: Alex Deucher <alexander.deucher@amd.com> 997 998commit 824189b3da9edc33e1a4f5c6130a043da73c1a4c 999Author: Michel Dänzer <michel.daenzer@amd.com> 1000Date: Thu Aug 16 18:22:27 2018 +0200 1001 1002 Use correct FB handle in radeon_do_pageflip 1003 1004 We were always using the handle of the client provided FB, which 1005 prevented RandR transforms from working, and could result in a black 1006 screen. 1007 1008 Bugzilla: https://bugs.freedesktop.org/107519 1009 Fixes: 740f0850f1e4 "Store FB for each CRTC in drmmode_flipdata_rec" 1010 (Ported from amdgpu commit f6cd72e64e85896b6d155bee0930e59771dcb701) 1011 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1012 1013commit 740f0850f1e40403c8dd727e074eae36caeb1f63 1014Author: Michel Dänzer <michel.daenzer@amd.com> 1015Date: Thu Aug 2 18:49:48 2018 +0200 1016 1017 Store FB for each CRTC in drmmode_flipdata_rec 1018 1019 We were only storing the FB provided by the client, but on CRTCs with 1020 TearFree enabled, we use a separate FB. This could cause 1021 drmmode_flip_handler to fail to clear drmmode_crtc->flip_pending, which 1022 could result in a hang when waiting for the pending flip to complete. We 1023 were trying to avoid that by always clearing drmmode_crtc->flip_pending 1024 when TearFree is enabled, but that wasn't reliable, because 1025 drmmode_crtc->tear_free can already be FALSE at this point when 1026 disabling TearFree. 1027 1028 Now that we're keeping track of each CRTC's flip FB separately, 1029 drmmode_flip_handler can reliably clear flip_pending, and we no longer 1030 need the TearFree hack. 1031 1032 (Ported from amdgpu commit 9b6782c821e0bdc53336d98f87ddde752faf7902) 1033 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1034 1035commit ef2a6b818fa47ad571bb0bc105aa8193d51a2190 1036Author: Michel Dänzer <michel.daenzer@amd.com> 1037Date: Thu Aug 2 18:48:04 2018 +0200 1038 1039 glamor: Set RADEON_CREATE_PIXMAP_DRI2 for DRI3 pixmaps 1040 1041 Not doing this resulted in falling back to software for DRI3 client 1042 presentation operations with ShadowPrimary. 1043 1044 (Ported from amdgpu commit 2989d40ef74d9966e8e8df2ef7727b2cc48d4960) 1045 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1046 1047commit db28d35ce9fd07a2a4703f3df0633d4c8291ff9b 1048Author: Michel Dänzer <michel.daenzer@amd.com> 1049Date: Fri Jul 27 11:30:57 2018 +0200 1050 1051 glamor: Use glamor_egl_create_textured_pixmap_from_gbm_bo when possible 1052 1053 Inspired by the modesetting driver. 1054 1055 Fixes screen pixmap corruption with Xorg < 1.20, and as a bonus, 1056 simplifies the code slightly. 1057 1058 Bugzilla: https://bugs.freedesktop.org/107385 1059 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1060 1061commit 7de4ea15880565c6536ed6859fd32ec89ce0b4c0 1062Author: Michel Dänzer <michel.daenzer@amd.com> 1063Date: Fri Jul 27 09:13:12 2018 +0200 1064 1065 Bump version to 18.0.99 1066 1067 Forgot to do this after the 18.0.1 release. 1068 1069commit fed411290b696ecf9f0e799ec4bad054e4a95017 1070Author: Michel Dänzer <michel.daenzer@amd.com> 1071Date: Tue Jul 24 17:28:46 2018 +0200 1072 1073 Use strcpy for RandR output property names 1074 1075 Instead of strncpy with the string length. Avoids new warnings with GCC 1076 8: 1077 1078 ../../src/drmmode_display.c: In function ‘drmmode_output_create_resources’: 1079 ../../src/drmmode_display.c:2240:2: warning: ‘strncpy’ output truncated before terminating nul copying 8 bytes from a string of the same length [-Wstringop-truncation] 1080 strncpy(tearfree_prop->name, "TearFree", 8); 1081 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1082 ../../src/drmmode_display.c:2244:2: warning: ‘strncpy’ output truncated before terminating nul copying 3 bytes from a string of the same length [-Wstringop-truncation] 1083 strncpy(tearfree_prop->enums[0].name, "off", 3); 1084 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1085 ../../src/drmmode_display.c:2245:2: warning: ‘strncpy’ output truncated before terminating nul copying 2 bytes from a string of the same length [-Wstringop-truncation] 1086 strncpy(tearfree_prop->enums[1].name, "on", 2); 1087 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1088 ../../src/drmmode_display.c:2247:2: warning: ‘strncpy’ output truncated before terminating nul copying 4 bytes from a string of the same length [-Wstringop-truncation] 1089 strncpy(tearfree_prop->enums[2].name, "auto", 4); 1090 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1091 1092 (Ported from amdgpu commit f3b2ed37d683f8616a0a31ff63133ddb8fe1a4a3) 1093 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1094 1095commit d258413c956cd46a237d4b5b94a35817ca7ce8f4 1096Author: Michel Dänzer <michel.daenzer@amd.com> 1097Date: Mon Jul 23 18:54:40 2018 +0200 1098 1099 Remove drmmode_terminate_leases 1100 1101 The RandR screen private is already freed when our CloseScreen runs, so 1102 this can't do anything useful. This cleanup has to be done by the X 1103 server itself. 1104 1105 (Ported from amdgpu commit 5f06d6b8ba570b500956ad26fee711d5ac427818) 1106 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1107 1108commit ffdbd46b0abcf8cbe73eb8930ddcd7fd2485d9b5 1109Author: Michel Dänzer <michel.daenzer@amd.com> 1110Date: Mon Jul 23 11:02:32 2018 +0200 1111 1112 Hardcode "non-desktop" RandR property name 1113 1114 It's a bit silly to require current randrproto just for this definition, 1115 which can't really change anyway. 1116 1117 Suggested-by: Qiang Yu <qiang.yu@amd.com> 1118 (Ported from amdgpu commit ae8e02c6fc4ef5d5340b8cd4739e66b19b9e3386) 1119 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1120 1121commit 64bd009d17986c71726724859f100ad40a6e3096 1122Author: Michel Dänzer <michel.daenzer@amd.com> 1123Date: Thu Jul 19 11:33:42 2018 +0200 1124 1125 Don't allocate drmmode_output->props twice 1126 1127 This was accidentally duplicated when porting RandR lease support from 1128 amdgpu. 1129 1130 Trivial. 1131 1132commit 499d2f9d5d301ef1efd4ffc2952677609ef05122 1133Author: Michel Dänzer <michel.daenzer@amd.com> 1134Date: Tue Jul 17 11:43:12 2018 +0200 1135 1136 glamor: Invalidate cached GEM handle in radeon_set_pixmap_bo 1137 1138 We continued using the stale cached handle, causing issues e.g. when 1139 resizing the screen via RandR. 1140 1141 Reported-by: iive on IRC 1142 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1143 1144commit d9a139bc6d453d33593685de67d1425611f3ab6f 1145Author: Michel Dänzer <michel.daenzer@amd.com> 1146Date: Mon Jul 16 12:51:26 2018 +0200 1147 1148 glamor: Fix glamor_block_handler argument in radeon_glamor_finish 1149 1150 I missed this before because this code isn't compiled with current 1151 xserver. 1152 1153 Trivial. 1154 1155commit f533b1f654952cee794de49d28d01947a6571daf 1156Author: Keith Packard <keithp@keithp.com> 1157Date: Wed Jul 11 19:13:28 2018 +0200 1158 1159 Add RandR leases support 1160 1161 Signed-off-by: Keith Packard <keithp@keithp.com> 1162 (Ported from xserver commit e4e3447603b5fd3a38a92c3f972396d1f81168ad) 1163 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 1164 (Ported from amdgpu commit 61040bdfa360975614fb47aa7ea1b3a1abac3427) 1165 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1166 1167commit b4f0f44a9f67f2eafd4a2b5ab919e6ea7fa2acf7 1168Author: Keith Packard <keithp@keithp.com> 1169Date: Wed Jul 11 19:10:20 2018 +0200 1170 1171 modesetting: Create CONNECTOR_ID properties for outputs [v2] 1172 1173 This lets a DRM client map between X outputs and kernel connectors. 1174 1175 v2: 1176 Change CONNECTOR_ID to enum -- Adam Jackson <ajax@nwnk.net> 1177 1178 Signed-off-by: Keith Packard <keithp@keithp.com> 1179 (Ported from xserver commit 023d4aba8d45e9e3630b944ecfb650c081799b96) 1180 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 1181 (Ported from amdgpu commit ab7e39c5a03e24c3ce3ee2f22ada7572bc2d9aa7) 1182 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1183 1184commit 612bda0a5e769f23478b364cf89620222cfac349 1185Author: Keith Packard <keithp@keithp.com> 1186Date: Wed Jul 11 19:07:59 2018 +0200 1187 1188 modesetting: Record non-desktop kernel property at PreInit time 1189 1190 Save any value of the kernel non-desktop property in the xf86Output 1191 structure to avoid non-desktop outputs in the default configuration. 1192 1193 [Also bump randrproto requirement to a version that defines 1194 RR_PROPERTY_NON_DESKTOP - ajax] 1195 1196 Signed-off-by: Keith Packard <keithp@keithp.com> 1197 (Ported from xserver commit b91c787c4cd2d20685db69426c539938c556128a) 1198 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 1199 (Ported from amdgpu commit 14db71a606128c4a207f43298809af279b77e2a8) 1200 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1201 1202commit 4b3e5f81c6032003237bb05c3ab96423c39524fc 1203Author: Michel Dänzer <michel.daenzer@amd.com> 1204Date: Wed Jul 11 19:04:26 2018 +0200 1205 1206 Call drmmode_crtc_gamma_do_set from drmmode_setup_colormap 1207 1208 Instead of from drmmode_set_mode_major. There's no need to re-set the 1209 gamma LUT on every modeset, the kernel should preserve it. 1210 1211 (Ported from amdgpu commit baea4fa492f635cdfe746a84be2e337d9aeae8a9) 1212 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1213 1214commit 16ddc109660dbe813c457765eda66422fb745a3e 1215Author: Michel Dänzer <michel.daenzer@amd.com> 1216Date: Wed Jul 11 19:03:12 2018 +0200 1217 1218 Remove #if 0'd code 1219 1220 This has always been disabled, no need to keep it. 1221 1222 (Ported from amdgpu commit 19a40758be04e1d451a030f452efb49e8aaad541) 1223 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1224 1225commit 447ef0458fac047919d021f2ba1753a647f5c503 1226Author: Michel Dänzer <michel.daenzer@amd.com> 1227Date: Wed Jul 11 19:01:31 2018 +0200 1228 1229 Check dimensions passed to drmmode_xf86crtc_resize 1230 1231 When enabling a secondary GPU output, Xorg can try resizing the screen 1232 beyond the limit advertised by the driver, leading to drmModeAddFB 1233 failing and primary GPU outputs turning off. Check for this and bail 1234 instead. 1235 1236 (Ported from amdgpu commit 940c8b39f79789d4d5ddb8ab8d25a8ae05932756) 1237 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1238 1239commit 6a0c01bbd4ed48c696c38952ee33ce21afec9f91 1240Author: Michel Dänzer <michel.daenzer@amd.com> 1241Date: Wed Jul 11 18:59:58 2018 +0200 1242 1243 Use drmmode_crtc_dpms in drmmode_set_desired_modes 1244 1245 Simplifies the latter slightly. 1246 1247 (Ported from amdgpu commit 74124f2c17dbb4b752707bb7eee398ae099e8a2c) 1248 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1249 1250commit 59441ee3dae15e0e81ed20688b0ba6dba12d7917 1251Author: Michel Dänzer <michel.daenzer@amd.com> 1252Date: Wed Jul 11 18:58:20 2018 +0200 1253 1254 Call drmmode_do_crtc_dpms from drmmode_crtc_dpms as well 1255 1256 Leo pointed out that drmmode_do_crtc_dpms wasn't getting called when 1257 turning off an output with 1258 1259 xrandr --output <output> --off 1260 1261 This meant that the vblank sequence number and timestamp wouldn't be 1262 saved before turning off the CRTC in this case. 1263 1264 Reported-by: Leo (Sunpeng) Li <sunpeng.li@amd.com> 1265 (Ported from amdgpu commit ceeacb455cd058492a493aac954deab8455804b5) 1266 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1267 1268commit 4050b0ad51b1c65945c6474981d1228888738cd4 1269Author: Michel Dänzer <michel.daenzer@amd.com> 1270Date: Wed Jul 11 18:42:58 2018 +0200 1271 1272 Replace 'foo == NULL' with '!foo' 1273 1274 Shorter and sweeter. :) 1275 1276 (Ported from amdgpu commit e8e688f3852fb06b0c34ed5bce47c9493bcd1613) 1277 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1278 1279commit cf8bc72e3473cef2b511e2c938eb00aca82de909 1280Author: Michel Dänzer <michel.daenzer@amd.com> 1281Date: Tue Jul 10 18:57:49 2018 +0200 1282 1283 Wait for pending flips in drmmode_output_set_tear_free 1284 1285 This prevents a nested call to drmHandleEvent, which would hang. 1286 1287 Fixes hangs when disabling TearFree on a CRTC while a DRI3 client is 1288 page flipping. 1289 1290 (Ported from amdgpu commit 04947b83cce3a7782e59dece2c7797cc396c1e05) 1291 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1292 1293commit f01d8cf2bd9681b8f5f0e2eddec0a79614389771 1294Author: Michel Dänzer <michel.daenzer@amd.com> 1295Date: Tue Jul 10 18:53:54 2018 +0200 1296 1297 Refactor drmmode_output_set_tear_free helper 1298 1299 Preparation for the following fix, no functional change intended. 1300 1301 (Ported from amdgpu commit fa30f4601de7a44edfb4a95873bd648946fd4292) 1302 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1303 1304commit ee7e15746148122abf6728fb0f59d6c8ae329e4e 1305Author: Michel Dänzer <michel.daenzer@amd.com> 1306Date: Tue Jul 10 18:53:06 2018 +0200 1307 1308 Set drmmode_crtc->scanout_id = 0 when TearFree is disabled 1309 1310 When disabling TearFree, drmmode_crtc->scanout_id could remain as 1, 1311 but drmmode_set_mode_major would destroy drmmode_crtc->scanout[1], so 1312 scanout_do_update() would keep bailing, and the scanout buffer would 1313 stop being updated. 1314 1315 Fixes freeze after disabling TearFree on a CRTC with active RandR 1316 rotation or other transform. 1317 1318 (Ported from amdgpu commit 7db0c8e9d7586cff4312d4b93684d35de3e6376f) 1319 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1320 1321commit e9d2d149481e2a9c7cba50d43c6a5146124be3f2 1322Author: Michel Dänzer <michel.daenzer@amd.com> 1323Date: Tue Jul 10 18:52:08 2018 +0200 1324 1325 Simplify drmmode_handle_transform 1326 1327 Set crtc->driverIsPerformingTransform for any case we can handle before 1328 calling xf86CrtcRotate. We already clear it afterwards when the latter 1329 clears crtc->transform_in_use. 1330 1331 This should allow our separate scanout buffer mechanism to be used in 1332 more cases. 1333 1334 (Cherry picked from amdgpu commit 8e544b4a0de6717feb4abf00052d57c5b726b5ce) 1335 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1336 1337commit 89d38e976fd6ba6e026cda7c5d05971f1f177f13 1338Author: Michel Dänzer <michel.daenzer@amd.com> 1339Date: Tue Jul 10 18:51:00 2018 +0200 1340 1341 Don't call scanout_flip/update with a legacy RandR scanout buffer 1342 1343 It means we are not using our own scanout buffers. 1344 1345 Fixes crash when TearFree is supposed to be enabled, but 1346 drmmode_handle_transform doesn't set crtc->driverIsPerformingTransform. 1347 1348 Bugzilla: https://bugs.freedesktop.org/105736 1349 (Ported from amdgpu commit 463477661c88cab3a87746499e5838c5b9f9a13b) 1350 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1351 1352commit 9f2409c4d3a9ecad8e769783e307e57a75b1255e 1353Author: Michel Dänzer <michel.daenzer@amd.com> 1354Date: Tue Jul 10 18:49:05 2018 +0200 1355 1356 Simplify drmmode_crtc_scanout_update 1357 1358 Use our own BoxRec for the extents, and RegionEmpty for clearing the 1359 scanout damage region. 1360 1361 (Ported from amdgpu commit 72c3e9c7308fbcdf85708b72f9be14a5f2f8e7b5) 1362 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1363 1364commit e07c38649280b3f4361005bc4c256f2145a72537 1365Author: Michel Dänzer <michel.daenzer@amd.com> 1366Date: Tue Jul 10 18:48:16 2018 +0200 1367 1368 Update RandR CRTC state if set_mode_major fails in set_desired_modes 1369 1370 Without this, RandR would report the CRTC and its outputs as enabled, 1371 even though they were actually off due to the failure. 1372 1373 (Cherry picked from amdgpu commit 4dcda0b48d62944c841cd9540f4ad4c7ac8dee47) 1374 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1375 1376commit 05390ae36a0abb19f1d533ff95a3fceaafdf79c8 1377Author: Michel Dänzer <michel.daenzer@amd.com> 1378Date: Tue Jul 10 18:45:51 2018 +0200 1379 1380 Abort scanout_update_pending event when possible 1381 1382 We don't need to wait for a non-TearFree scanout update before scanning 1383 out from the screen pixmap or before flipping, as the scanout update 1384 won't be visible anyway. Instead, just abort it. 1385 1386 (Ported from amdgpu commit 36d01989cd842588f12fdae5b2cba5fdcf9c91dd) 1387 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1388 1389commit 290291a11598ba9aa594417998502b0e0ac96970 1390Author: Michel Dänzer <michel.daenzer@amd.com> 1391Date: Tue Jul 10 18:43:28 2018 +0200 1392 1393 Track DRM event queue sequence number in scanout_update_pending 1394 1395 Preparation for next change, no behaviour change intended. 1396 (Ported from amdgpu commit 04a5c5f7cfacad8d9ccffe81e388cc3da2036cb5) 1397 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1398 1399commit d14ff6b77824d0e32728869907e88d6866243101 1400Author: Michel Dänzer <michel.daenzer@amd.com> 1401Date: Tue Jul 10 18:41:43 2018 +0200 1402 1403 Ignore RADEON_DRM_QUEUE_ERROR (0) in radeon_drm_abort_entry 1404 1405 This allows a following change to be slightly simpler. 1406 (Ported from amdgpu commit 8fcc3a9b43d3907052a83a96e5a2423afab5ad3f) 1407 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1408 1409commit 731d4b386a55cdb468e37b69d41284150952cf8c 1410Author: Emil Velikov <emil.velikov@collabora.com> 1411Date: Tue Jul 10 18:39:31 2018 +0200 1412 1413 Do not export the DriverRec RADEON 1414 1415 Unused externally and should not be exported. 1416 1417 Signed-off-by: Emil Velikov <emil.velikov@collabora.com> 1418 (Ported from amdgpu commit 7fb8b49895e225b3908c8bd186539de23afe91d1) 1419 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1420 1421commit cc1d0824a8a7ef29f8911b95695f7cb1b4abe9b8 1422Author: Jim Qu <Jim.Qu@amd.com> 1423Date: Tue Jul 10 18:36:42 2018 +0200 1424 1425 Wait for pending scanout update before calling drmmode_crtc_scanout_free 1426 1427 There is a case that when set screen from reverse to normal, the old 1428 scanout damage is freed in modesetting before scanout update handler, 1429 so it causes segment fault issue. 1430 1431 Signed-off-by: Jim Qu <Jim.Qu@amd.com> 1432 1433 [ Michel Dänzer: Only call drmmode_crtc_wait_pending_event before 1434 drmmode_crtc_scanout_free is actually called, slightly tweak commit 1435 message ] 1436 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 1437 (Ported from amdgpu commit 9f6a8905611b5b1d8fcd31bebbc9af7ca1355cc3) 1438 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1439 1440commit 413622bcccc9d36a9c87679f379dc84b21a3c655 1441Author: Michel Dänzer <michel.daenzer@amd.com> 1442Date: Tue Jul 10 17:23:24 2018 +0200 1443 1444 Swap pixmap privates in radeon_dri2_exchange_buffers 1445 1446 Instead of only the BOs. 1447 1448 This matches what amdgpu does, and fixes issues with DRI2 page flipping. 1449 1450 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1451 1452commit 3c4c0213c11d623cba7adbc28dde652694f2f758 1453Author: Michel Dänzer <michel.daenzer@amd.com> 1454Date: Fri Jun 29 17:57:03 2018 +0200 1455 1456 glamor: Use GBM for BO allocation when possible 1457 1458 Inspired by amdgpu. This avoids various issues due to a GEM handle 1459 lifetime conflict between us and Mesa with current glamor. 1460 1461 Bugzilla: https://bugs.freedesktop.org/105381 1462 Tested-by: Konstantin Kharlamov <hi-angel@yandex.ru> 1463 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1464 1465commit b85b7b11f5b5e792f21951b881bd8433d3d70858 1466Author: Michel Dänzer <michel.daenzer@amd.com> 1467Date: Fri May 18 18:49:35 2018 +0200 1468 1469 Add struct radeon_buffer 1470 1471 Inspired by amdgpu, preparation for the following change. For now, this 1472 is mostly a wrapper around struct radeon_bo, no functional change 1473 intended. 1474 1475 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1476 1477commit 37ba075b34130c41bb7a2261bd666af5b29ffaf3 1478Author: Michel Dänzer <michel.daenzer@amd.com> 1479Date: Fri Jun 15 18:23:53 2018 +0200 1480 1481 Refactor radeon_finish helper 1482 1483 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1484 1485commit 6c986e997159ad0086f940294b244fc4c30b61fc 1486Author: Jammy Zhou <Jammy.Zhou@amd.com> 1487Date: Fri Jun 15 18:41:34 2018 +0200 1488 1489 Remove throttling from radeon_dri2_copy_region2 1490 1491 Throttling should be handled by the client-side drivers. 1492 1493 Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com> 1494 (Ported from amdgpu commit 8a34a8149860ac15e83ccdbd8d9a527d8d3e5997) 1495 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1496 1497commit 90b94d40449f665f2d12874598062a5e5e5b64cd 1498Author: Michel Dänzer <michel.daenzer@amd.com> 1499Date: Fri Jun 15 18:27:46 2018 +0200 1500 1501 Move flush from radeon_scanout_do_update to its callers 1502 1503 No functional change intended. 1504 1505 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1506 1507commit 5da2bf43e64f2d702fe05016263f617c10f8f134 1508Author: Michel Dänzer <michel.daenzer@amd.com> 1509Date: Mon May 28 18:35:32 2018 +0200 1510 1511 Factor out radeon_surface_initialize helper 1512 1513 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1514 1515commit ba5d5402b3e2e807d7e47205ac83f930b6c8caf5 1516Author: Michel Dänzer <michel.daenzer@amd.com> 1517Date: Thu May 17 12:36:29 2018 +0200 1518 1519 glamor: Don't store radeon_surfaces in pixmaps 1520 1521 Only EXA needs them. 1522 1523 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1524 1525commit eec4a41925127ae490f0a5156a881a08d521e28e 1526Author: Michel Dänzer <michel.daenzer@amd.com> 1527Date: Thu May 17 18:25:37 2018 +0200 1528 1529 Only initialize libdrm_radeon surface manager for >= R600 1530 1531 Not used with older GPUs. 1532 1533 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1534 1535commit ac18a993a3a68629805b8f272a339e25444c0897 1536Author: Michel Dänzer <michel.daenzer@amd.com> 1537Date: Wed May 23 12:18:36 2018 +0200 1538 1539 EXA: Remove old RADEONEXACreatePixmap hook 1540 1541 Not used by any supported version of xserver. 1542 1543 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1544 1545commit 25836375e0aab17a2e6bd91acb1c19b837ae2109 1546Author: Michel Dänzer <michel.daenzer@amd.com> 1547Date: Thu May 17 18:18:42 2018 +0200 1548 1549 Drop unused drmmode_create_bo_pixmap surface parameter 1550 1551 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1552 1553commit 65c9dfea4e841b7d6f795c7489fede58c5e9631f 1554Author: Michel Dänzer <michel.daenzer@amd.com> 1555Date: Tue May 15 18:59:30 2018 +0200 1556 1557 glamor: Bail CreatePixmap on unsupported pixmap depth 1558 1559 Fixes crash in that case. 1560 1561 Bugzilla: https://bugs.freedesktop.org/106293 1562 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1563 1564commit 3dcfce8d0f495d09d7836caf98ef30d625b78a13 1565Author: Michel Dänzer <michel.daenzer@amd.com> 1566Date: Thu May 17 09:50:50 2018 +0200 1567 1568 Bail from dri2_create_buffer2 if we can't get a pixmap 1569 1570 We would store the NULL pointer and continue, which would lead to a 1571 crash down the road. 1572 1573 Bugzilla: https://bugs.freedesktop.org/106293 1574 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1575 1576commit fdba5373522c008f43e9ca5a9e76e35c6e7ad73d 1577Author: Michel Dänzer <michel.daenzer@amd.com> 1578Date: Thu Mar 15 17:19:59 2018 +0100 1579 1580 Bump version for 18.0.1 release 1581 1582commit e0f5020117d0a6340f4208ec52bd3e3b3aa99a1b 1583Author: Michel Dänzer <michel.daenzer@amd.com> 1584Date: Wed Mar 14 16:54:52 2018 +0100 1585 1586 Pass extents to radeon_scanout_do_update by value 1587 1588 radeon_scanout_extents_intersect could leave the scanout damage region 1589 in an invalid state, triggering debugging checks in pixman: 1590 1591 *** BUG *** 1592 In pixman_region_append_non_o: The expression r->x1 < r->x2 was false 1593 Set a breakpoint on '_pixman_log_error' to debug 1594 1595 (Ported from amdgpu commit 8af989546907ad9fb491d940e1936d3bfc89276b) 1596 1597commit b915e8e6fb956e983c2ce2a5565c20838c85707c 1598Author: Michel Dänzer <michel.daenzer@amd.com> 1599Date: Wed Mar 7 17:28:55 2018 +0100 1600 1601 Wrap the whole miPointerScreenFuncRec, instead of only Set/MoveCursor 1602 1603 We were clobbering entries in mi's global miSpritePointerFuncs struct, 1604 which cannot work correctly with multiple primary screens. Instead, 1605 assign a pointer to our own wrapper struct to PointPriv->spriteFuncs. 1606 1607 Fixes crashes with multiple primary screens. 1608 1609 Fixes: 1fe8ca75974c ("Keep track of how many SW cursors are visible on 1610 each screen") 1611 Reported-by: Mario Kleiner <mario.kleiner.de@gmail.com> 1612 1613commit 7e18aea984e186dd2ab1144470b0c0e832562ad1 1614Author: Michel Dänzer <michel.daenzer@amd.com> 1615Date: Tue Mar 6 18:15:45 2018 +0100 1616 1617 Bump version for 18.0.0 release 1618 1619commit 504b8721b17a672caf1ed3eab087027c02458cab 1620Author: Michel Dänzer <michel.daenzer@amd.com> 1621Date: Mon Mar 5 18:41:44 2018 +0100 1622 1623 Only change Set/MoveCursor hooks from what we expect 1624 1625 Since xf86CursorCloseScreen runs after RADEONCloseScreen_KMS, 1626 PointPriv->spriteFuncs doesn't point to the same struct in the latter as 1627 in RADEONCursorInit_KMS. So we were restoring info->Set/MoveCursor to 1628 the wrong struct. Then in the next server generation, 1629 info->Set/MoveCursor would end up pointing to 1630 drmmode_sprite_set/move_cursor, resulting in an infinite loop if one of 1631 them was called. 1632 1633 To avoid this, only change the Set/MoveCursor hooks if their values 1634 match our expectations, otherwise leave them as is. This is kind of a 1635 hack, but the alternative would be invasive and thus risky changes to 1636 the way we're wrapping CloseScreen, and it's not even clear that can 1637 work without changing xserver code. 1638 1639 Fixes: 1fe8ca75974c ("Keep track of how many SW cursors are visible on 1640 each screen") 1641 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1642 1643commit e0d23092337efe4c1e406c9124ca8d8856863e65 1644Author: Keith Packard <keithp@keithp.com> 1645Date: Fri Mar 2 12:13:43 2018 +0100 1646 1647 modesetting: Update property values at detect and uevent time 1648 1649 We were updating the link-status property when a uevent came in, but 1650 we also want to update the non-desktop property, and potentially 1651 others as well. We also want to check at detect time in case we don't 1652 get a hotplug event. 1653 1654 This patch updates every property provided by the kernel, sending 1655 changes to DIX so it can track things as well. 1656 1657 Signed-off-by: Keith Packard <keithp@keithp.com> 1658 1659 (Ported from amdgpu commit 374cb8fef4fdbb648af089ee80803ec78321f1b2) 1660 1661commit aef8fe6e9e94ebecbbb6adf940160d3a7ab4d5ac 1662Author: Daniel Martin <consume.noise@gmail.com> 1663Date: Fri Mar 2 12:12:09 2018 +0100 1664 1665 modesetting: Reset output_id if drmModeGetConnector failed 1666 1667 If drmModeGetConnector() fails in drmmode_output_detect(), we have to 1668 reset the output_id to -1 too. 1669 1670 Yet another spot leading to a potential NULL dereference when handling 1671 the mode_output member as output_id was != -1. Though, this case should 1672 be very hard to hit. 1673 1674 Signed-off-by: Daniel Martin <consume.noise@gmail.com> 1675 1676 (Ported from amdgpu commit 10054b6c3d9a755b30abb43020121b9631fa296d) 1677 1678commit 1ef12a92ca53cd6fbfed3ef8628616445b007c9a 1679Author: Daniel Martin <consume.noise@gmail.com> 1680Date: Fri Mar 2 12:08:09 2018 +0100 1681 1682 modesetting: Use helper to fetch drmModeProperty(Blob)s 1683 1684 Replace the various loops to lookup drmModeProperty(Blob)s by 1685 introducing helper functions. 1686 1687 Signed-off-by: Daniel Martin <consume.noise@gmail.com> 1688 1689 (Ported from amdgpu commit fb58e06acd6c6bd59de2dbdadbca27eb1dd0025b) 1690 1691commit ace1e3871c39a86becdf473a47fdc096e8ef39f5 1692Author: Michel Dänzer <michel.daenzer@amd.com> 1693Date: Wed Feb 21 17:53:59 2018 +0100 1694 1695 Disable all unused CRTCs before setting desired modes 1696 1697 This might avoid modeset failures in some cases where a CRTC which isn't 1698 used by Xorg was enabled before. 1699 1700 (Ported from amdgpu commit e3aae7a24296f640c0153d1459f3e0820485468a) 1701 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1702 1703commit f2915eb7a4beb6140ebec183aeac02111fc6ba77 1704Author: Christoph Haag <haagch@frickel.club> 1705Date: Thu Mar 1 15:07:00 2018 +0100 1706 1707 fix include order for present.h configure test 1708 1709 xorg-server.h defines _XSERVER64 which is used in X.h to choose the 1710 correct definition of XID 1711 1712 this prevents a failure in the present.h configure test that disables 1713 DRI3 on X.Org 1.20 1714 1715 Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> 1716 1717commit f777e909fd23e063729ecb038997db445f82ef5e 1718Author: Michel Dänzer <michel.daenzer@amd.com> 1719Date: Fri Feb 16 16:55:03 2018 +0100 1720 1721 Don't bail from drmmode_set_desired_modes immediately 1722 1723 If we fail to find or set the mode for a CRTC, keep trying for the 1724 remaining CRTCs, and only return FALSE if we failed for all CRTCs that 1725 should be on. 1726 1727 (Ported from amdgpu commit f5ac5f385f41d1547cfd7ccc8bb35a537a8fffeb) 1728 1729commit 2d63748667d0bcfcd9999e4e41f775eba87897a0 1730Author: Michel Dänzer <michel.daenzer@amd.com> 1731Date: Thu Feb 15 17:47:14 2018 +0100 1732 1733 Simplify depth 30 "kernel too old" error message 1734 1735 We know it's depth 30, no need to have xf86DrvMsg fill it in. 1736 1737 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1738 1739commit 680b4ab43f00b1f635f9359281e13a031d09ad24 1740Author: Michel Dänzer <michel.daenzer@amd.com> 1741Date: Wed Jan 24 12:29:17 2018 +0100 1742 1743 If glamor is too old for depth 30, fall back to EXA or ShadowFB 1744 1745 Instead of not starting up at all. 1746 1747 Corresponding to amdgpu commit 37c7260bdef3a53b0f0295a531f33938e9aad8cf. 1748 1749 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1750 1751commit 16954fab59f95bdca4540c4496b2ec0759c447e9 1752Author: Michel Dänzer <michel.daenzer@amd.com> 1753Date: Wed Feb 14 18:20:04 2018 +0100 1754 1755 Call RADEONFreeRec from RADEONFreeScreen_KMS even if info == NULL 1756 1757 It's safe now. 1758 1759 (Ported from amdgpu commit c9bd1399a13cea2e1331af2c826ca054b88db071) 1760 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1761 1762commit 6b44b8c584f3b63712858be11277842c97d03426 1763Author: Michel Dänzer <michel.daenzer@amd.com> 1764Date: Wed Feb 14 18:06:18 2018 +0100 1765 1766 Don't call RADEONFreeRec from RADEONPreInit_KMS 1767 1768 If the latter fails, Xorg will call RADEONFreeScreen_KMS, which calls 1769 the former. 1770 1771 (Ported from amdgpu commit 103b7285845b786929fb509083c57e074c48f9be) 1772 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1773 1774commit 8a55e66bfd77dcb2553a43e86215b9d3c87b03f0 1775Author: Michel Dänzer <michel.daenzer@amd.com> 1776Date: Wed Feb 14 18:13:16 2018 +0100 1777 1778 Don't use RADEONEntPriv in RADEONFreeRec 1779 1780 It crashes if info == NULL. 1781 1782 (Ported from amdgpu commits fb8444e731765588c0ff1e9053c1c7b73f5f0907 & 1783 cfccf4c4e7e5c73fe4040fabeb1b43283cf29b33) 1784 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1785 1786commit 703cbb7177c6f69264836e975c1921e552fd60a4 1787Author: Michel Dänzer <michel.daenzer@amd.com> 1788Date: Wed Feb 14 17:28:10 2018 +0100 1789 1790 glamor: Allow depth 30 with Xorg >= 1.19.99.1 1791 1792 Corresponding to amdgpu commit 6aee5770fb913713bb1b9a1af8f0d0892a66f21a. 1793 1794 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1795 1796commit 733f606dd6ca8350e6e7f0858bfff5454ddc98ed 1797Author: Michel Dänzer <michel.daenzer@amd.com> 1798Date: Thu Jan 25 16:59:30 2018 +0100 1799 1800 Always use screen depth/bpp for KMS framebuffers 1801 1802 DRI clients can use depth 32 pixmaps while the screen is depth 24, in 1803 which case page flipping would fail. 1804 1805 Reported-by: Mario Kleiner <mario.kleiner.de@gmail.com> 1806 Reviewed-and-Tested-by: Mario Kleiner <mario.kleiner.de@gmail.com> 1807 1808commit 83124426040bc1f2dba74c8eeb5a5a9c7c64e664 1809Author: Mario Kleiner <mario.kleiner.de@gmail.com> 1810Date: Wed Jan 24 18:19:38 2018 +0100 1811 1812 Make XvMC extension initialize at depth 30. 1813 1814 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 1815 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 1816 1817commit 2d171c892d373b45ef9f1ee068648437e2529a78 1818Author: Mario Kleiner <mario.kleiner.de@gmail.com> 1819Date: Wed Jan 24 18:19:37 2018 +0100 1820 1821 Add missing depth 30 -> cpp=4 handling for DRI2. 1822 1823 Missing this apparently didn't cause any ill side effects, only a bit of 1824 confusion when looking at xtrace's of DRI2 clients. 1825 1826 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 1827 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 1828 1829commit 574bfab4bf1fcd95163a8f33cea2889189429d30 1830Author: Mario Kleiner <mario.kleiner.de@gmail.com> 1831Date: Mon Jan 22 03:14:37 2018 +0100 1832 1833 Support exa screen color depth 30 on Linux 3.16 and later. (v2) 1834 1835 Linux 3.16's radeon-kms introduced 10 bpc fb support. 1836 Currently, as of X-Server 1.19, this works with exa accel, 1837 but not yet when using glamor acceleration. 1838 1839 v2: Style fixes, check for glamor first, as suggested 1840 by Michel. 1841 1842 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 1843 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 1844 1845commit ee7e6c25989e4faf1adfa4754b7edd5a9a4722ce 1846Author: Fredrik Höglund <fredrik@kde.org> 1847Date: Mon Jan 22 03:14:36 2018 +0100 1848 1849 exa: Accelerate ARGB2101010 pictures 1850 1851 Tested by Mario on a Radeon HD 4000 series for the r600 1852 exa path, and on Radeon HD 5770 for the evergreen exa 1853 path. 1854 1855 Signed-off-by: Fredrik Höglund <fredrik@kde.org> 1856 Reviewed-and-Tested-by: Mario Kleiner <mario.kleiner.de@gmail.com> 1857 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 1858 1859commit 1f1d4b1fa7d4b22dd8553f7e71251bf17ca7a7b1 1860Author: Mario Kleiner <mario.kleiner.de@gmail.com> 1861Date: Mon Jan 22 03:14:35 2018 +0100 1862 1863 Skip xf86HandleColormaps() at color depth 30. (v2) 1864 1865 The hardware gamma luts get bypassed at color 1866 depth 30 anyway, so skip their setup. 1867 1868 v2: Also don't hook up drmmode_crtc_gamma_set() 1869 to .gamma_set for a screen with depth 30, 1870 so the server can spare itself the effort of 1871 updating and setting unused cluts. 1872 Suggested by Michel. 1873 1874 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 1875 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 1876 1877commit 21f6753462464acfd3c452393328c977a375ce26 1878Author: Mario Kleiner <mario.kleiner.de@gmail.com> 1879Date: Mon Jan 22 03:14:34 2018 +0100 1880 1881 Define per x-screen individual drmmode_crtc_funcs 1882 1883 This allows to en-/disable some functions depending on individual screen 1884 settings. 1885 1886 Prep work for more efficient depth 30 support. 1887 1888 Suggested-by: Michel Dänzer <michel.daenzer@amd.com> 1889 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 1890 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 1891 1892commit 1fe8ca75974c5241c3b556b757f9c966c9ce5002 1893Author: Michel Dänzer <michel.daenzer@amd.com> 1894Date: Wed Dec 27 17:13:30 2017 +0100 1895 1896 Keep track of how many SW cursors are visible on each screen 1897 1898 And use this to determine when we cannot use page flipping for DRI 1899 clients. We previously did this based on whether the HW cursor cannot 1900 be used on at least one CRTC, which had at least two issues: 1901 1902 * Even while the HW cursor cannot be used, no SW cursor may actually be 1903 visible (e.g. because all cursors are disabled), in which case we can 1904 use page flipping for DRI clients anyway 1905 * Even while the HW cursor can be used, there may be SW cursors visible 1906 from non-core pointer devices, in which case we cannot use page 1907 flipping for DRI clients anyway 1908 1909 (Ported from amdgpu commit 69e20839bfeb3ee0b0a732d72de0a32d6c5435fc) 1910 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1911 1912commit b4ce8913ac392df339081f6f3e28bf79a239ee75 1913Author: Michel Dänzer <michel.daenzer@amd.com> 1914Date: Wed Dec 27 17:05:26 2017 +0100 1915 1916 Move cursor related ScreenInit calls into RADEONCursorInit_KMS 1917 1918 And bail if xf86_cursors_init fails. 1919 1920 (Ported from amdgpu commit dfccaa7043ccb157a1f8be7313123792bb7e7001) 1921 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1922 1923commit ee907e9b9563609cb4a4817e17215b50bb1c7177 1924Author: Michel Dänzer <michel.daenzer@amd.com> 1925Date: Fri Dec 1 15:49:11 2017 +0100 1926 1927 Add radeon_dirty_src_drawable helper 1928 1929 Allows tidying up radeon_dirty_src_equals and redisplay_dirty slightly. 1930 1931 (Cherry picked from amdgpu commit 1d65ac395971571094df21ca0408d5972c6b56ec) 1932 Acked-by: Alex Deucher <alexander.deucher@amd.com> 1933 1934commit fccfd75ecf9ccfa628a8ff9c1fbb7f60c0f10ea5 1935Author: Joakim Tjernlund <joakim.tjernlund@infinera.com> 1936Date: Fri Nov 24 17:27:09 2017 +0100 1937 1938 Fix non GLAMOR build 1939 1940 Non GLAMOR builds needs pRADEONEnt so move the USE_GLAMOR define 1941 down to make pRADEONEnt defined in this case. 1942 1943 Signed-off-by: Joakim Tjernlund <joakim.tjernlund@infinera.com> 1944 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 1945 1946commit 15350bd5538c21cd4c6b406d565ca765b7e0218a 1947Author: Michel Dänzer <michel.daenzer@amd.com> 1948Date: Wed Nov 15 18:01:29 2017 +0100 1949 1950 Use correct ScrnInfoPtr in redisplay_dirty 1951 1952 We used the destination pixmap's screen for flushing drawing commands. 1953 But when we are the master screen, the destination pixmap is from the 1954 slave screen. 1955 1956 Fixes crash when the slave screen isn't using the same acceleration 1957 architecture as us. 1958 1959 Bugzilla: https://bugs.freedesktop.org/103613 1960 Fixes: 01b040b4a807 ("Adapt to PixmapDirtyUpdateRec::src being a 1961 DrawablePtr") 1962 (Ported from amdgpu commit 3a4f7422913093ed9e26b73ecd7f9e773478cb1e) 1963 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1964 1965commit 6441210bc53978a2c46c572e9174215b14f9e813 1966Author: Daniel Martin <consume.noise@gmail.com> 1967Date: Fri Oct 20 10:05:35 2017 +0200 1968 1969 modesetting: Check crtc before searching link-status property 1970 1971 No need to lookup the link-status property if we don't have a crtc. 1972 1973 Signed-off-by: Daniel Martin <consume.noise@gmail.com> 1974 (Ported from xserver commit 8d7f7e24261e68459e6f0a865e243473f65fe7ad) 1975 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1976 1977commit cafcb55768b58dc63a8c5682000eba21ad904689 1978Author: Keith Packard <keithp@keithp.com> 1979Date: Mon Sep 25 16:18:22 2017 -0700 1980 1981 modesetting: Skip no-longer-present connectors when resetting BAD links 1982 1983 Outputs may have NULL mode_output (connector) pointers if the 1984 connector disappears while the server is running. Skip these when 1985 resetting outputs with BAD link status. 1986 1987 (Ported from xserver commit 37f4e7651a2fd51efa613a08a1e705553be33e76) 1988 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 1989 1990commit 0cd14777b7e39619f231b841a2d73f3122458030 1991Author: Michel Dänzer <michel.daenzer@amd.com> 1992Date: Fri Oct 20 17:20:30 2017 +0200 1993 1994 Always call drmModeFreeProperty after drmModeGetProperty 1995 1996 We were not doing so in all cases, leaking memory allocated by the 1997 latter. 1998 1999 (Cherry picked from amdgpu commit f6b39bcd45cb06976ba8a3600df77fc471c63995) 2000 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2001 2002commit 84c7d9c1c71d9cee9f50f96671032529b606cf04 2003Author: Michel Dänzer <michel.daenzer@amd.com> 2004Date: Fri Oct 20 17:19:09 2017 +0200 2005 2006 Call TimerFree for timer created in LeaveVT 2007 2008 We were leaking the memory allocated by TimerSet. 2009 2010 (Ported from amdgpu commit 84aad09f18fed6b52b0c073f0bbd675a6de07807) 2011 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2012 2013commit d588015f4f60ce270bcfd8433215fc981a84f143 2014Author: Michel Dänzer <michel.daenzer@amd.com> 2015Date: Fri Oct 20 17:13:36 2017 +0200 2016 2017 Free memory returned by xf86GetEntityInfo 2018 2019 We were leaking it. 2020 2021 (Ported from amdgpu commit cfccf4c4e7e5c73fe4040fabeb1b43283cf29b33) 2022 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2023 2024commit ec90c94881239791d0a66a02fc6d003c73a6c1f3 2025Author: Michel Dänzer <michel.daenzer@amd.com> 2026Date: Fri Oct 20 16:51:16 2017 +0200 2027 2028 Free pRADEONEnt memory in RADEONFreeRec 2029 2030 We were leaking it. 2031 2032 (Inspired by amdgpu commit 9d84934309e4ccd9a43c73d958b8ff10ef2fc990) 2033 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2034 2035commit b93ae9b159790e8a979e9d5d567a0e6fccd7d08a 2036Author: Michel Dänzer <michel.daenzer@amd.com> 2037Date: Thu Oct 19 18:57:45 2017 +0200 2038 2039 Unreference pixmap's FB with EXA as well in radeon_set_pixmap_bo 2040 2041 Not doing so resulted in DRI2 page flips not actually changing the FB 2042 being scanned out, showing intermittent flicker of the "back" buffer 2043 rendering. 2044 2045 Bugzilla: https://bugs.freedesktop.org/102643 2046 Fixes: 55e513b978b2 "Use reference counting for tracking KMS framebuffer 2047 lifetimes" 2048 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2049 2050commit 61113768bd79ec5c2164835e2e048e8f2a42ce9a 2051Author: Darren Salt <devspam@moreofthesa.me.uk> 2052Date: Thu Oct 19 18:17:07 2017 +0200 2053 2054 Clarify when TearFree is automatically enabled. 2055 2056 (Ported from amdgpu commit 2f72be038d22c54620e436af30121dd89f79a003) 2057 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2058 2059commit 7d435354099119234d443b07e2df1c7b9f97cf3c 2060Author: Michel Dänzer <michel.daenzer@amd.com> 2061Date: Wed Oct 4 11:11:33 2017 +0200 2062 2063 Bail if there's a problem with ShadowFB 2064 2065 If we hit a problem while setting up ShadowFB, just carrying on trying 2066 to set up HW acceleration instead is unlikely to work. 2067 2068 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2069 2070commit b71836c9962f8f2eadf0c2188e03faed737c18da 2071Author: Michel Dänzer <michel.daenzer@amd.com> 2072Date: Wed Oct 4 11:05:49 2017 +0200 2073 2074 Post-release version bump 2075 2076commit 34da04daec82077571558ac3fe1ec0c1203a01ad 2077Author: Michel Dänzer <michel.daenzer@amd.com> 2078Date: Tue Oct 3 12:47:14 2017 +0200 2079 2080 Fix VT switching with ShadowFB 2081 2082 We were trying to call acceleration specific functions from LeaveVT. 2083 Instead, memset the scanout buffer to all 0 in LeaveVT and allocate a 2084 new one in EnterVT. 2085 2086 Bugzilla: https://bugs.freedesktop.org/102948 2087 Fixes: 06a465484101 ("Make all active CRTCs scan out an all-black 2088 framebuffer in LeaveVT") 2089 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 2090 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2091 2092commit f9aca8c1a7a423b9e738c88400ae821bdfbcdb1c 2093Author: Michel Dänzer <michel.daenzer@amd.com> 2094Date: Fri Sep 8 17:02:59 2017 +0900 2095 2096 Bump version for 7.10.0 release 2097 2098commit 5cdd334b3402c2431deb3a87a8d04ef590da53ee 2099Author: Michel Dänzer <michel.daenzer@amd.com> 2100Date: Wed Aug 30 18:48:19 2017 +0900 2101 2102 Require xserver >= 1.13 2103 2104 xserver 1.13.0 was released on September 6th, 2012, almost 5 years ago. 2105 2106 This allows cleaning up a bunch of backwards compatibility code. 2107 2108 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2109 Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com> 2110 2111commit 9d9c565c84601f4c6c73ad769f86491088683f7a 2112Author: Michel Dänzer <michel.daenzer@amd.com> 2113Date: Tue Aug 29 18:07:51 2017 +0900 2114 2115 Use a timer for unreferencing the all-black FB 2116 2117 The timer fires 1 second after LeaveVT. This gives the next DRM master 2118 enough time to set up scanout of its own buffers. 2119 2120 Fixes prolonged intermittent black screen when switching from Xorg to 2121 e.g. the GDM Wayland mode login VT. 2122 2123 Fixes: 06a465484101 ("Make all active CRTCs scan out an all-black 2124 framebuffer in LeaveVT") 2125 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2126 2127commit e4a3df19d588a4310fcb889ef34e205d0e92e4d7 2128Author: Michel Dänzer <michel.daenzer@amd.com> 2129Date: Mon Aug 28 17:54:23 2017 +0900 2130 2131 Remove drmmode_scanout_free 2132 2133 Not used anymore. 2134 2135 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2136 2137commit 06a465484101f21e99d3a0a62fb03440bcaff93e 2138Author: Michel Dänzer <michel.daenzer@amd.com> 2139Date: Wed Aug 23 18:00:15 2017 +0900 2140 2141 Make all active CRTCs scan out an all-black framebuffer in LeaveVT 2142 2143 And destroy all other FBs. This is so that other DRM masters can only 2144 get access to this all-black FB, not to any other FB we created, while 2145 we're switched away and not DRM master. 2146 2147 Fixes: 55e513b978b2 ("Use reference counting for tracking KMS 2148 framebuffer lifetimes") 2149 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2150 2151commit 7f0cd68d1b0c132e32ae736371bce3e12ed33c7a 2152Author: Michel Dänzer <michel.daenzer@amd.com> 2153Date: Mon Aug 28 17:53:19 2017 +0900 2154 2155 Create radeon_master_screen helper 2156 2157 Preparatory, no functional change intended yet. 2158 2159 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2160 2161commit 20f6b56fdb74d88086e8e094013fedbb14e50a24 2162Author: Michel Dänzer <michel.daenzer@amd.com> 2163Date: Thu Aug 24 17:10:29 2017 +0900 2164 2165 Create radeon_pixmap_get_fb_ptr helper 2166 2167 Preparatory, no functional change intended yet. 2168 2169 Also inline radeon_pixmap_create_fb into radeon_pixmap_get_fb, since 2170 there's only one call-site anymore. 2171 2172 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2173 2174commit 4bc992c31059eb50e22df4ebf5b92d08411f41ef 2175Author: Michel Dänzer <michel.daenzer@amd.com> 2176Date: Wed Aug 23 17:24:53 2017 +0900 2177 2178 Create drmmode_set_mode helper 2179 2180 Preparatory, no functional change intended yet. 2181 2182 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2183 2184commit 3f6210ca2c8ef60d59efc8139151d3b9838bb875 2185Author: Michel Dänzer <michel.daenzer@amd.com> 2186Date: Thu Jun 22 18:37:26 2017 +0900 2187 2188 Create radeon_pixmap_clear helper 2189 2190 Preparatory, no functional change intended yet. 2191 2192 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2193 2194commit 0472a605e0ec8fec1892bbc3a84698b7ef9c5296 2195Author: Martin Peres <martin.peres@linux.intel.com> 2196Date: Mon Apr 10 16:48:21 2017 +0300 2197 2198 modesetting: re-set the crtc's mode when link-status goes BAD 2199 2200 Despite all the careful planning of the kernel, a link may become 2201 insufficient to handle the currently-set mode. At this point, the 2202 kernel should mark this particular configuration as being broken 2203 and potentially prune the mode before setting the offending connector's 2204 link-status to BAD and send the userspace a hotplug event. This may 2205 happen right after a modeset or later on. 2206 2207 Upon receiving a hot-plug event, we iterate through the connectors to 2208 re-apply the currently-set mode on all the connectors that have a 2209 link-status property set to BAD. The kernel may be able to get the 2210 link to work by dropping to using a lower link bpp (with the same 2211 display bpp). However, the modeset may fail if the kernel has pruned 2212 the mode, so to make users aware of this problem a warning is outputed 2213 in the logs to warn about having a potentially-black display. 2214 2215 This patch does not modify the current behaviour of always propagating 2216 the events to the randr clients. This allows desktop environments to 2217 re-probe the connectors and select a new resolution based on the new 2218 (currated) mode list if a mode disapeared. This behaviour is expected in 2219 order to pass the Display Port compliance tests. 2220 2221 (Ported from xserver commit bcee1b76aa0db8525b491485e90b8740763d7de6) 2222 2223 [ Michel: Bump libdrm dependency to >= 2.4.78 for 2224 DRM_MODE_LINK_STATUS_BAD ] 2225 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2226 2227commit d822a0f47070374ad0c1a97b559bae27724dc52a 2228Author: Michel Dänzer <michel.daenzer@amd.com> 2229Date: Thu Aug 17 16:20:52 2017 +0900 2230 2231 Make radeon_scanout_do_update take a PixmapPtr instead of a DrawablePtr 2232 2233 All callers were already passing in a pixmap. 2234 2235 This allows simplifying the rotated scanout case slightly. 2236 2237 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2238 2239commit 3e24770b1b472fc15df56d06f5f04778c9db63dd 2240Author: Michel Dänzer <michel.daenzer@amd.com> 2241Date: Wed Aug 9 13:02:34 2017 +0900 2242 2243 Use xorg_list_append for the DRM event list 2244 2245 We were adding entries at the start of the list, i.e. the list was 2246 ordered from most recently added to least recently added. However, the 2247 corresponding DRM events are generally expected to arrive in the same 2248 order as they are queued, which means that radeon_drm_queue_alloc would 2249 generally have to traverse the whole list to find the entry 2250 corresponding to an arrived event. Fix this by adding entries at the end 2251 of the list. 2252 2253 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2254 2255commit 36ce7920136c0d723c9397a84e7dd5926a9c7943 2256Author: Michel Dänzer <michel.daenzer@amd.com> 2257Date: Wed Aug 9 19:01:15 2017 +0900 2258 2259 Consolidate radeon_scanout_flip_abort/handler helpers 2260 2261 While at it, make them use crtc->driver_private. 2262 2263 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2264 2265commit d314cbfb228bb4b8762714f98d0c114a8ee3f061 2266Author: Michel Dänzer <michel.daenzer@amd.com> 2267Date: Mon Jul 31 17:12:59 2017 +0900 2268 2269 Always allow DRI2 page flipping with TearFree 2270 2271 Even if TearFree is enabled for the CRTC we're synchronizing to. 2272 2273 v2: 2274 * Move manpage hunk here from previous change. 2275 2276 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> # v1 2277 2278commit 4445765af5b97d0cfd10889fe6d6f58f2ce85659 2279Author: Michel Dänzer <michel.daenzer@amd.com> 2280Date: Thu Feb 9 13:01:03 2017 +0900 2281 2282 Always allow Present page flipping with TearFree 2283 2284 Even if TearFree is active for the the CRTC we're synchronizing to. In 2285 that case, for Present flips synchronized to vertical blank, the other 2286 scanout buffer is immediately synchronized and flipped to during the 2287 target vertical blank period. For Present flips not synchronized to 2288 vertical blank, we simply use the MSC and timestamp values of the last 2289 vertical blank period for timing purposes, and let the normal TearFree 2290 mechanism handle display updates. 2291 2292 v2: 2293 * Move manpage hunk to next change, since TearFree can still prevent 2294 DRI2 page flipping with this change. 2295 2296 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> # v1 2297 2298commit 65e0c5ea1b4adff21d673dbf54af99704c429627 2299Author: Michel Dänzer <michel.daenzer@amd.com> 2300Date: Wed Feb 8 17:52:28 2017 +0900 2301 2302 Pass extents to radeon_scanout_do_update 2303 2304 Preparation for following change, no functional change intended yet. 2305 2306 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2307 2308commit 1443270e52e8562bd8dc3603f301963bd4027cef 2309Author: Michel Dänzer <michel.daenzer@amd.com> 2310Date: Tue Feb 7 18:58:23 2017 +0900 2311 2312 Add source drawable parameter to radeon_scanout_do_update 2313 2314 Preparation for following changes, no functional change intended yet. 2315 2316 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2317 2318commit 7c10ee9c88378d773c0bcf651fdc5d9f2c6dc5e5 2319Author: Michel Dänzer <michel.daenzer@amd.com> 2320Date: Thu Aug 3 18:16:36 2017 +0900 2321 2322 Handle multiple "pending" Present flips 2323 2324 The xserver Present code can submit a flip in response to notifying it 2325 that a vblank event arrived. This can happen before the completion event 2326 of the previous flip is processed. In that case, we were clearing the 2327 drmmode_crtc->flip_pending field prematurely. 2328 2329 Prevent this by only clearing drmmode_crtc->flip_pending when it matches 2330 the framebuffer being scanned out since the flip whose completion event 2331 we're processing. 2332 2333 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2334 2335commit e6d7dc2070f4d21a6900916bb70a31839112882c 2336Author: Michel Dänzer <michel.daenzer@amd.com> 2337Date: Thu Aug 3 17:51:20 2017 +0900 2338 2339 Wait for pending flips synchronously before turning off a CRTC 2340 2341 Allows removing drmmode_clear_pending_flip and the pending_dpms_mode 2342 field and cleaning up the code considerably. 2343 2344 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2345 2346commit f87acdbfb1b0b6d2769764772a52ea8b81675e20 2347Author: Michel Dänzer <michel.daenzer@amd.com> 2348Date: Fri Jul 28 12:28:53 2017 +0900 2349 2350 Create drmmode_crtc_wait_pending_event helper macro 2351 2352 Preparation for following change, no functional change intended yet. 2353 2354 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2355 2356commit 99f1d7a474af3683fe1a66f50c0bb8935478ff0a 2357Author: Michel Dänzer <michel.daenzer@amd.com> 2358Date: Mon Aug 14 12:23:04 2017 +0900 2359 2360 Create drmmode_wait_vblank helper 2361 2362 Allows cleaning up the code considerably. 2363 2364 v2: 2365 * Fix "drmWaiVBlank" typo, add blank line for readability (Slava Abramov) 2366 * Rename in/out sequence parameters to "target_seq" and "result_seq", 2367 hopefully that will be clearer. 2368 2369 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> # v1 2370 2371commit 49cc61ab970ee28d4509b4e2dd0a57165136889f 2372Author: Michel Dänzer <michel.daenzer@amd.com> 2373Date: Tue Aug 1 18:45:53 2017 +0900 2374 2375 Pass reference CRTC to radeon_do_pageflip directly 2376 2377 Simplifies the code slightly. 2378 2379 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2380 2381commit 3f120fa1d5d921656a367751bc079e020e9ab105 2382Author: Michel Dänzer <michel.daenzer@amd.com> 2383Date: Thu Aug 3 17:42:51 2017 +0900 2384 2385 Remove drmmode_crtc->scanout_destroy[] array 2386 2387 No longer necessary since we're reference counting framebuffers. 2388 2389 Reviewed-by: Alex Deucher <alexander.deucher@amd.com 2390 2391commit 01b040b4a80754b8c8a305cbe968416182b54246 2392Author: Michel Dänzer <michel.daenzer@amd.com> 2393Date: Tue Aug 15 17:15:27 2017 +0900 2394 2395 Adapt to PixmapDirtyUpdateRec::src being a DrawablePtr 2396 2397commit 5309bde0c4e28adf2b167191c6d7011a19e31eed 2398Author: Michel Dänzer <michel.daenzer@amd.com> 2399Date: Thu Jul 27 16:11:19 2017 +0900 2400 2401 Allow DRI page flipping when some CRTCs use separate scanout buffers 2402 2403 As long as the CRTC we're synchronizing to doesn't. 2404 2405 v2: 2406 * Remove redundant checks from can_exchange which still prevented DRI2 2407 page flipping 2408 2409 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2410 2411commit 9bc3eef74452d924f9101c024f66ad9b14c404c8 2412Author: Michel Dänzer <michel.daenzer@amd.com> 2413Date: Thu Jul 27 15:46:41 2017 +0900 2414 2415 Add drmmode_crtc_can_flip helper 2416 2417 To reduce code duplication between DRI2 and Present. No functional 2418 change intended yet. 2419 2420 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2421 2422commit c2d26890691ec105858f086b63170ad94c6f7f05 2423Author: Michel Dänzer <michel.daenzer@amd.com> 2424Date: Wed Jul 12 12:13:03 2017 +0900 2425 2426 Use root window (pixmap) instead of screen pixmap for scanout updates 2427 2428 Preparation for following changes, no functional change intended yet. 2429 2430 v2: 2431 * Add drmmode_screen_damage_destroy callback to prevent use-after-free 2432 on server shutdown 2433 2434 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2435 2436commit 377c9dc7e51cc8ebd06816b08682a8098a49508a 2437Author: Michel Dänzer <michel.daenzer@amd.com> 2438Date: Thu Jul 27 15:27:54 2017 +0900 2439 2440 autogen.sh: Pass -f to autoreconf 2441 2442 To ensure that any existing copies of autotools files will be replaced 2443 with the current versions. 2444 2445 (Cherry picked from amdgpu commit a47c0093338d80d84e7033ad15d051925d542ca0) 2446 Acked-by: Alex Deucher <alexander.deucher@amd.com> 2447 2448commit bf0a759e3bced3a93bead033a26871cc340bd309 2449Author: Michel Dänzer <michel.daenzer@amd.com> 2450Date: Thu Jul 27 15:26:23 2017 +0900 2451 2452 Makefile.am: Set ACLOCAL_AMFLAGS = -I m4 2453 2454 Suggested by one of the tools called by autoreconf. 2455 2456 (Cherry picked from amdgpu commit 842bad4b951296ca25f47b50cb358e502bf30ebb) 2457 Acked-by: Alex Deucher <alexander.deucher@amd.com> 2458 2459commit e628f604d032d32971a1ca8670469e88939e3de2 2460Author: Michel Dänzer <michel.daenzer@amd.com> 2461Date: Thu Jul 27 15:24:37 2017 +0900 2462 2463 Add AC_CONFIG_MACRO_DIRS([m4]) to configure.ac 2464 2465 Suggested by one of the tools called by autoreconf. 2466 2467 (Cherry picked from amdgpu commit 227b399badaad9bbef0be5a776ce008d0d243449) 2468 Acked-by: Alex Deucher <alexander.deucher@amd.com> 2469 2470commit 7d7abf99b5441ddb04dbee99bc8fa7abc30d4c46 2471Author: Michel Dänzer <michel.daenzer@amd.com> 2472Date: Wed Jul 26 18:13:35 2017 +0900 2473 2474 Only handle reflection in the driver with Xorg < 1.16 2475 2476 Xorg doesn't handle the hardware cursor correctly in that case for 2477 rotation and general transforms, and we can't force the SW cursor. 2478 2479 Fixes: f2bc882f1c10 ("Handle rotation in the driver also with Xorg 2480 1.12-1.18") 2481 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2482 2483commit 94dc2b80f3ef0b2c17c20501d824fb0447d52e7a 2484Author: Michel Dänzer <michel.daenzer@amd.com> 2485Date: Tue Mar 14 16:57:17 2017 +0900 2486 2487 If a TearFree flip fails, fall back to non-TearFree operation 2488 2489 In order to avoid possible freeze / log file spam in that case. 2490 2491 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99769 2492 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2493 2494commit aff267ee36cc6a703a532f91f82adc1ba1425ff3 2495Author: Michel Dänzer <michel.daenzer@amd.com> 2496Date: Tue Mar 14 16:23:45 2017 +0900 2497 2498 Use drmmode_crtc->scanout_id instead of 0 to check for scanout buffer 2499 2500 Preparation for following change, no functional change intended. 2501 2502 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2503 2504commit df88d51c5005476e6fbfca3f31b54be079ec5aaa 2505Author: Michel Dänzer <michel.daenzer@amd.com> 2506Date: Tue Jun 27 19:15:18 2017 +0900 2507 2508 Use pRADEONEnt->fd exclusively for the DRM file descriptor 2509 2510 This brings us closer to amdgpu. 2511 2512 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2513 2514commit 19626bce4e5e31c863eedb503ea3884ac3f60bea 2515Author: Michel Dänzer <michel.daenzer@amd.com> 2516Date: Tue Jun 27 18:32:28 2017 +0900 2517 2518 Improve drmmode_fb_reference debugging code 2519 2520 If a reference count is <= 0, call FatalError with the call location 2521 (in case it doesn't get resolved in the backtrace printed by 2522 FatalError). 2523 2524 (Ported from amdgpu commit 1b6ff5fd9933c00ec1ec90dfc62e0b531927749b) 2525 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2526 2527commit bc46ffdf71ab3dfa0f95572529e818f2b619d380 2528Author: Michel Dänzer <michel.daenzer@amd.com> 2529Date: Tue Jun 27 18:19:31 2017 +0900 2530 2531 Increase reference count of FB assigned to drmmode_crtc->flip_pending 2532 2533 Otherwise, it could happen that we destroy the FB before the flip 2534 completes, resulting in use-after-free and most likely a crash. 2535 2536 (Ported from amdgpu commit af7221e1c4d2dbdfd488eb0976a835584ea8441c) 2537 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2538 2539commit c9dd28cb0c9c3de676eadac61e727732510f6b9b 2540Author: Michel Dänzer <michel.daenzer@amd.com> 2541Date: Thu Jun 22 12:38:55 2017 +0900 2542 2543 Only call drmmode_scanout_free for non-GPU screens in LeaveVT 2544 2545 Destroying the scanout buffers of GPU screens resulted in a crash when 2546 switching back to the Xorg VT. 2547 2548 Fixes: 4cfa4615f79f ("Use drmmode_crtc_scanout_* helpers for RandR 1.4 2549 scanout pixmaps") 2550 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2551 2552commit 78fad9ca1635cca04b89807e3db029477cf46681 2553Author: Michel Dänzer <michel.daenzer@amd.com> 2554Date: Mon Jun 26 18:58:30 2017 +0900 2555 2556 Include xf86Pci.h for DRICreatePCIBusID with xserver Git master 2557 2558 The declaration has been moved there from dri.h. 2559 2560 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2561 2562commit ef95e3a3cc9189c8906fe5b4d70bfa9f1055d7ce 2563Author: Adam Jackson <ajax@redhat.com> 2564Date: Tue Jun 13 09:32:44 2017 -0400 2565 2566 modesetting: Validate the atom for enum properties 2567 2568 The client could have said anything here, and if what they said doesn't 2569 actually name an atom NameForAtom() will return NULL, and strcmp() will 2570 be unhappy about that. 2571 2572 [copied from xserver d4995a3936ae283b9080fdaa0905daa669ebacfc] 2573 2574 Signed-off-by: Adam Jackson <ajax@redhat.com> 2575 Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> 2576 2577commit d80d01a73c2eaba2e3649b7bc0a3541b3ff782f6 2578Author: Michel Dänzer <michel.daenzer@amd.com> 2579Date: Thu May 18 11:12:34 2017 +0900 2580 2581 Update URLs 2582 2583 * Point to the amd-gfx mailing list 2584 * Specify the component in all bugzilla URLs 2585 * Use https:// for all HTML URLs 2586 2587 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2588 2589commit 2b7d77b90108911777a11ecaa63435552000c958 2590Author: Eric Anholt <eric@anholt.net> 2591Date: Tue May 16 11:04:29 2017 -0700 2592 2593 Use plain glamor_egl_create_textured_screen(). 2594 2595 Since 5064ffab631 (2014), glamor's implementation of _ext just drops the 2596 back_pixmap arg, which we were passing NULL (the default) to anyway. 2597 2598 Signed-off-by: Eric Anholt <eric@anholt.net> 2599 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 2600 2601commit 3a9f7e87c30b9f6334d38dfdd1cb15b28fe37fcb 2602Author: Michel Dänzer <michel.daenzer@amd.com> 2603Date: Thu May 11 18:57:25 2017 +0900 2604 2605 Remove unused struct members from drmmode_display.h 2606 2607 (Ported from amdgpu commit 462ac3341e5bfbded9086d3d9043821d19352b3e) 2608 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2609 2610commit 7dc68e26755466f9056f8c72195ab8690660693d 2611Author: Michel Dänzer <michel.daenzer@amd.com> 2612Date: Tue May 2 11:47:58 2017 +0900 2613 2614 Simplify tracking of PRIME scanout pixmap 2615 2616 Remember the shared pixmap passed to drmmode_set_scanout_pixmap for each 2617 CRTC, and just compare against that. 2618 2619 Fixes leaving stale entries in ScreenRec::pixmap_dirty_list under some 2620 circumstances, which would usually result in use-after-free and a crash 2621 down the line. 2622 2623 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2624 2625commit 55e513b978b2afc52b7cafc5bfcb0d1dc78d75f6 2626Author: Michel Dänzer <michel.daenzer@amd.com> 2627Date: Sat Apr 29 17:05:58 2017 +0900 2628 2629 Use reference counting for tracking KMS framebuffer lifetimes 2630 2631 References are held by the pixmaps corresponding to the FBs (so 2632 the same KMS FB can be reused as long as the pixmap exists) and by the 2633 CRTCs scanning out from them (so a KMS FB is only destroyed once it's 2634 not being scanned out anymore, preventing intermittent black screens and 2635 worse issues due to a CRTC turning off when it should be on). 2636 2637 v2: 2638 * Only increase reference count in drmmode_fb_reference if it was sane 2639 before 2640 * Make drmmode_fb_reference's indentation match the rest of 2641 drmmode_display.h 2642 2643 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2644 2645commit 944391b0052466b71bf9919b56139dc197a7e072 2646Author: Michel Dänzer <michel.daenzer@amd.com> 2647Date: Wed Feb 8 18:14:30 2017 +0900 2648 2649 Pass pixmap instead of handle to radeon_do_pageflip 2650 2651 This brings us in line with amdgpu and prepares for the following 2652 change, no functional change intended. 2653 2654 (Ported from amdgpu commit e463b849f3e9d7b69e64a65619a22e00e78d297b) 2655 2656 v2: 2657 * Be more consistent with the amdgpu code, which should make porting 2658 the following change to amdgpu easier 2659 2660 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2661 2662commit f32c45194ac6f82cbe42d255ed72f857018778e0 2663Author: Michel Dänzer <michel.daenzer@amd.com> 2664Date: Tue May 9 11:18:00 2017 +0900 2665 2666 Apply gamma correction to HW cursor 2667 2668 The display hardware CLUT we're currently using for gamma correction 2669 doesn't affect the HW cursor, so we have to apply it manually when 2670 uploading the HW cursor data. 2671 2672 This currently only works in depth 24/32. 2673 2674 (Ported from amdgpu commit 82fa615f38137add75f9cd4bb49c48dd88de916f) 2675 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2676 2677commit 4c91f36d3058180b5a2d6a23e9b82f5c933d8716 2678Author: Michel Dänzer <michel.daenzer@amd.com> 2679Date: Mon Apr 17 18:32:31 2017 +0900 2680 2681 Don't set modes before RADEONWindowExposures_oneshot is called 2682 2683 The root window contents may be undefined before that, so we don't want 2684 to show anything yet. 2685 2686 Fixes a crash on startup with rotation and virtual resolution set in 2687 xorg.conf. 2688 2689 Bugzilla: https://bugs.freedesktop.org/100276 2690 Fixes: cc9d6b7db9c2 ("Move DPMS check from radeon_scanout_do_update to 2691 radeon_scanout_flip") 2692 (Ported from amdgpu commit 981bac185cfd74ae50dffc28f57cf34623a9595f) 2693 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2694 2695commit d6c29250dbc293c932ca14997eb23399610b106d 2696Author: Daniel Stone <daniels@collabora.com> 2697Date: Fri Apr 7 14:27:58 2017 +0100 2698 2699 Set correct DRM event context version 2700 2701 DRM_EVENT_CONTEXT_VERSION is the latest context version supported by 2702 whatever version of libdrm is present. We were blindly asserting we 2703 supported whatever version that may be, even if we actually didn't. 2704 2705 Set the version as 2, which should be bumped only with the appropriate 2706 version checks. 2707 2708 Signed-off-by: Daniel Stone <daniels@collabora.com> 2709 (Ported from xserver commit 0c8e6ed85810e96d84173a52d628863802a78d82) 2710 v2: Remove second paragraph of commit log, we always initialize 2711 page_flip_handler2 = NULL (Emil Velikov) 2712 Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com> 2713 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> # v1 2714 2715commit 23985d3ef3150ef8cad30b92ef0cd8ba6cfa1f4d 2716Author: Michel Dänzer <michel.daenzer@amd.com> 2717Date: Thu Mar 23 18:06:45 2017 +0900 2718 2719 manpage: Don't put "'" at the beginning of a line 2720 2721 It caused the whole line to be dropped. 2722 2723 Fixes: 58cd1600057e ("Allow toggling TearFree at runtime via output 2724 property") 2725 Reported-by: Andy Furniss <adf.lists@gmail.com> 2726 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (amdgpu commit) 2727 (Ported from amdgpu commit 165b51447643ce37f391f25ca6aecb8d76fabaa3) 2728 2729commit dc30f5e27ef8a8e980adb21c1fea7c4f60c7e896 2730Author: Michel Dänzer <michel.daenzer@amd.com> 2731Date: Thu Mar 23 18:06:23 2017 +0900 2732 2733 Post-release version bump 2734 2735commit 66227060872219c1a0dc84fa4cea6264f7548446 2736Author: Michel Dänzer <michel.daenzer@amd.com> 2737Date: Thu Mar 16 16:38:15 2017 +0900 2738 2739 Bump version for 7.9.0 release 2740 2741commit 9a71445094b728f3d78db8f6808b4782ee19a453 2742Author: Michel Dänzer <michel.daenzer@amd.com> 2743Date: Wed Mar 8 17:42:56 2017 +0900 2744 2745 Pass TRUE to drmmode_set_desired_modes the first time for GPU screens 2746 2747 This is the only place we call drmmode_set_desired_modes for GPU screens 2748 during server startup. Without this change, the display outputs of 2749 secondary GPUs may stay on even while Xorg isn't using them. 2750 2751 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2752 2753commit 67ae5e00a748ad52cf92738d401afff2947b1891 2754Author: Michel Dänzer <michel.daenzer@amd.com> 2755Date: Wed Mar 8 11:20:30 2017 +0900 2756 2757 Skip some initialization steps for GPU screens 2758 2759 Xorg doesn't use the following functionality of GPU screens, so don't 2760 bother initializing it: 2761 2762 * DRI page flipping 2763 * DRI3 / Present / SYNC fences 2764 * XVideo / XvMC 2765 * Root window with background None 2766 2767 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2768 2769commit f0e7948e1c0e984fc27f235f365639e9cf628291 2770Author: Jochen Rollwagen <joro-2013@t-online.de> 2771Date: Sun Mar 5 11:32:38 2017 +0100 2772 2773 Fix build for XServer 1.13 2774 2775 Latest git build stops with the error message 2776 2777 radeon_kms.c: In function 'RADEONWindowExposures_oneshot': 2778 radeon_kms.c:1644:45: error: expected expression before 'RegionPtr' 2779 pScreen->WindowExposures(pWin, pRegion, RegionPtr pBSRegion); 2780 2781 This patch fixes the build. 2782 2783 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 2784 2785commit d63881623f0686a66a2e3e3c1f84e496aa52ec6b 2786Author: Michel Dänzer <michel.daenzer@amd.com> 2787Date: Thu Mar 2 16:22:51 2017 +0900 2788 2789 Don't call radeon_cs_flush_indirect & radeon_bo_wait in drmmode_copy_fb 2790 2791 RADEONWindowExposures_oneshot takes care of it. 2792 2793 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2794 2795commit cc9d6b7db9c2078be1e530a64af6d517c6a42024 2796Author: Michel Dänzer <michel.daenzer@amd.com> 2797Date: Wed Mar 1 17:35:59 2017 +0900 2798 2799 Move DPMS check from radeon_scanout_do_update to radeon_scanout_flip 2800 2801 When radeon_scanout_do_update is called from 2802 drmmode_crtc_scanout_update, drmmode_crtc->pending_dpms_mode may still 2803 be != DPMSModeOn, e.g. during server startup. 2804 2805 Fixes intermittently showing garbage with TearFree enabled. 2806 2807commit 0a12bf1085505017068dfdfd31d23133e51b45b9 2808Author: Michel Dänzer <michel.daenzer@amd.com> 2809Date: Wed Mar 1 18:00:40 2017 +0900 2810 2811 Call drmmode_set_desired_modes from a WindowExposures hook 2812 2813 This is the earliest opportunity where the root window contents are 2814 guaranteed to be initialized, and prevents drmmode_set_mode_major from 2815 getting called before drmmode_set_desired_modes via RADEONUnblank -> 2816 drmmode_crtc_dpms. Also, in contrast to the BlockHandler hook, this is 2817 called when running Xorg with -pogo. 2818 2819 Fixes intermittently showing garbage on server startup or after server 2820 reset. 2821 2822 As a bonus, this avoids trouble due to higher layers (e.g. the tigervnc 2823 Xorg module) calling RADEONBlockHandler_oneshot repeatedly even after 2824 we set pScreen->BlockHandler = RADEONBlockHandler_KMS. 2825 2826 v2: 2827 * Drop spaces between XORG_VERSION_NUMERIC arguments 2828 * Call radeon_bo_wait after radeon_cs_flush_indirect 2829 2830 Bugzilla: https://bugs.freedesktop.org/99457 2831 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1) 2832 2833commit df2d749a4db33298c8ce9f2cfb77c20c5538c9cc 2834Author: Michel Dänzer <michel.daenzer@amd.com> 2835Date: Thu Mar 2 16:05:42 2017 +0900 2836 2837 Fix bogus indentation 2838 2839 Trivial. 2840 2841 Fixes: 58cd1600057e ("Allow toggling TearFree at runtime via output property") 2842 2843commit 9035b6abea557828e672ee455f0c84e43da0906f 2844Author: Michel Dänzer <michel.daenzer@amd.com> 2845Date: Tue Feb 28 18:13:42 2017 +0900 2846 2847 present: Flush before flipping 2848 2849 This isn't necessary for DRI clients, but the Present extension can also 2850 be used for presenting normal pixmaps rendered to via the X11 protocol. 2851 2852 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2853 2854commit 0a4eb0e12f0c9c653cf4cea6fd62e1a507eb261c 2855Author: Michel Dänzer <michel.daenzer@amd.com> 2856Date: Fri Feb 10 12:52:02 2017 +0900 2857 2858 present: Use async flip for unflip if possible 2859 2860 In that case, unflip operations should finish faster in general. 2861 2862 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2863 2864commit 9a951a3e551db58ba50e7a594521ceac54d90615 2865Author: Michel Dänzer <michel.daenzer@amd.com> 2866Date: Tue Feb 28 18:06:55 2017 +0900 2867 2868 present: Also flush before using a flip to unflip 2869 2870 Not doing so might result in intermittently scanning out stale contents 2871 of the screen pixmap. 2872 2873 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2874 2875commit f2bc882f1c1082bed9f496cfab6c8f07a76bc122 2876Author: Michel Dänzer <michel.daenzer@amd.com> 2877Date: Tue Feb 28 17:31:14 2017 +0900 2878 2879 Handle rotation in the driver also with Xorg 1.12-1.18 2880 2881 We cannot use the HW cursor in that case, but in turn we get more 2882 efficient and less teary updates of rotated outputs. 2883 2884 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2885 2886commit ae921a3150f69c38b5b3c88a9e37d54fdf0d5093 2887Author: Michel Dänzer <michel.daenzer@amd.com> 2888Date: Tue Feb 28 17:47:09 2017 +0900 2889 2890 Fold drmmode_crtc_scanout_allocate into drmmode_crtc_scanout_create 2891 2892 Not used anywhere else anymore. 2893 2894 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2895 2896commit 987a34adb319923ad36e2b47a26837248f187c3e 2897Author: Michel Dänzer <michel.daenzer@amd.com> 2898Date: Tue Feb 28 17:17:20 2017 +0900 2899 2900 Call drmmode_crtc_scanout_create in drmmode_crtc_shadow_allocate as well 2901 2902 Calling drmmode_crtc_scanout_allocate in drmmode_crtc_shadow_allocate 2903 resulted in drmmode_crtc_scanout_create called from 2904 drmmode_crtc_shadow_create passing an uninitialized pitch value to 2905 drmmode_create_bo_pixmap. 2906 2907 Fixes issues such as failure to allocate the scanout pixmap or visual 2908 corruption and GPUVM faults when attempting to use rotation with Xorg 2909 <1.19. 2910 2911 Bugzilla: https://bugs.freedesktop.org/99916 2912 Fixes: ea30d856ba5e ("Pass pitch from drmmode_crtc_scanout_allocate to drmmode_create_bo_pixmap") 2913 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2914 2915commit 244d4bc7f8c8f6bc90f49556c0b9344c8aa40295 2916Author: Michel Dänzer <michel.daenzer@amd.com> 2917Date: Thu Feb 9 18:55:04 2017 +0900 2918 2919 present: Wait for screen pixmap BO idle before setting modes for unflip 2920 2921 To make sure the screen pixmap contents are up to date when it starts 2922 being scanned out. 2923 2924 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2925 2926commit 3ff29e5a14451916bc66b4e0028e9a317f0723f8 2927Author: Michel Dänzer <michel.daenzer@amd.com> 2928Date: Thu Feb 9 18:50:40 2017 +0900 2929 2930 present: Only call drmModeRmFB after setting modes for unflip 2931 2932 Fixes display intermittently blanking when a modeset is used for unflip. 2933 2934 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2935 2936commit 0c29deb5a97d9a57e994cc0053c49ddf7aca6ecb 2937Author: Michel Dänzer <michel.daenzer@amd.com> 2938Date: Tue Feb 21 16:56:27 2017 +0900 2939 2940 Use drmmode_crtc_scanout_free in drmmode_fini 2941 2942 We were leaking drmmode_crtc->scanout_damage, which caused trouble on 2943 server reset. Fixes server reset with active separate scanout pixmaps. 2944 2945 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2946 2947commit 58cd1600057e41aade0106d4acf78e23eac6e44f 2948Author: Michel Dänzer <michel.daenzer@amd.com> 2949Date: Thu Dec 1 18:37:57 2016 +0900 2950 2951 Allow toggling TearFree at runtime via output property 2952 2953 Option "TearFree" now sets the default value of the output property. 2954 See the manpage update for details. 2955 2956 TearFree is now enabled by default for outputs using rotation or other 2957 RandR transforms, and for RandR 1.4 slave outputs. 2958 2959 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2960 2961commit 305e2cbf335837a2ab6a24e9ff65815afe038296 2962Author: Michel Dänzer <michel.daenzer@amd.com> 2963Date: Tue Feb 21 15:55:13 2017 +0900 2964 2965 Factor out drmmode_crtc_scanout_update helper 2966 2967 Cleanup in preparation for following change, no functional change 2968 intended. 2969 2970 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2971 2972commit 649644a88347a6d03de68f8c41db03a82deeb23b 2973Author: Michel Dänzer <michel.daenzer@amd.com> 2974Date: Tue Feb 14 17:36:31 2017 +0900 2975 2976 Factor out radeon_prime_dirty_to_crtc helper 2977 2978 Cleanup in preparation for the following change, no functional change 2979 intended. 2980 2981 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2982 2983commit 1351e48efe7a2c28eab447e16f36a00fbd02ae48 2984Author: Michel Dänzer <michel.daenzer@amd.com> 2985Date: Tue Feb 7 18:43:56 2017 +0900 2986 2987 Don't destroy current FB if drmModeAddFB fails 2988 2989 It would probably result in a black screen. 2990 2991 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 2992 2993commit 481394e3c9f9f7d88bb66fe9ae8834c87952a8ab 2994Author: Michel Dänzer <michel.daenzer@amd.com> 2995Date: Tue Feb 7 18:39:05 2017 +0900 2996 2997 Fix flip event data leak if calloc or drmModeAddFB fails 2998 2999 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3000 3001commit 80fc78633962c1ec996b1721bfb757b98dc85bfc 3002Author: Michel Dänzer <michel.daenzer@amd.com> 3003Date: Tue Jan 31 16:25:37 2017 +0900 3004 3005 Don't handle Option "SwapbuffersWait" at all with glamor 3006 3007 It never had any effect with glamor. 3008 3009 v2: Better formatting, no functional change from v1. 3010 3011 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1) 3012 3013commit d611d21d9bf9aaa100ff7a5762e5980634846207 3014Author: Michel Dänzer <michel.daenzer@amd.com> 3015Date: Tue Jan 24 18:05:45 2017 +0900 3016 3017 Enable tiling by default with glamor on PALM 3018 3019 The DFS check is only relevant for EXA. 3020 3021 The lack of tiling prevented DRI3 clients from using page flipping, 3022 resulting in tearing. 3023 3024 Bugzilla: https://bugs.freedesktop.org/99491 3025 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3026 3027commit 55a33205cb9bd8091b4bcd478c2d1d191248681b 3028Author: Michel Dänzer <michel.daenzer@amd.com> 3029Date: Tue Jan 24 18:22:49 2017 +0900 3030 3031 Add 10-radeon.conf xorg.conf.d snippet 3032 3033 This instructs Xorg >= 1.16 to try loading the radeon driver for devices 3034 managed by the radeon kernel driver, even if the ati wrapper driver 3035 isn't available. 3036 3037 Copied from the amdgpu driver. 3038 3039 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3040 3041commit 8e6a4e96b7b27559e186f71b5547abb0a80b96dd 3042Author: Mihail Konev <k.mvc@ya.ru> 3043Date: Thu Jan 26 14:00:22 2017 +1000 3044 3045 autogen: add default patch prefix 3046 3047 Signed-off-by: Mihail Konev <k.mvc@ya.ru> 3048 3049commit e062933b5489705bdcba7a266d14ba6fd3e767ba 3050Author: Emil Velikov <emil.l.velikov@gmail.com> 3051Date: Mon Mar 9 12:00:52 2015 +0000 3052 3053 autogen.sh: use quoted string variables 3054 3055 Place quotes around the $srcdir, $ORIGDIR and $0 variables to prevent 3056 fall-outs, when they contain space. 3057 3058 Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com> 3059 Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net> 3060 Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net> 3061 3062commit 4fe183111f82d412760b175b3c9c8583acba804a 3063Author: Peter Hutterer <peter.hutterer@who-t.net> 3064Date: Tue Jan 24 10:32:07 2017 +1000 3065 3066 autogen.sh: use exec instead of waiting for configure to finish 3067 3068 Syncs the invocation of configure with the one from the server. 3069 3070 Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net> 3071 Reviewed-by: Emil Velikov <emil.velikov@collabora.com> 3072 3073commit a1787e4615a76ab0bb12498be66591c86a6b07ae 3074Author: Michel Dänzer <michel.daenzer@amd.com> 3075Date: Wed Jan 18 19:07:11 2017 +0900 3076 3077 ati: Support loading the amdgpu driver from the ati wrapper 3078 3079 If .../share/X11/xorg.conf.d/10-amdgpu.conf doesn't exist, but the ati 3080 wrapper is loaded, it will otherwise try to use the radeon driver even 3081 for GPUs driven by the amdgpu kernel driver. This can only fail, 3082 potentially in bad ways. 3083 3084 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3085 3086commit 41b82c776b571e1556f300d77c715aebdd87d8e6 3087Author: Michel Dänzer <michel.daenzer@amd.com> 3088Date: Tue Jan 24 18:14:23 2017 +0900 3089 3090 .editorconfig: src/ati.c only uses spaces for indentation 3091 3092 Trivial. 3093 3094commit ea30d856ba5e7274c8ea499293b8b0e721b8e082 3095Author: Michel Dänzer <michel.daenzer@amd.com> 3096Date: Fri Jan 6 17:36:05 2017 +0900 3097 3098 Pass pitch from drmmode_crtc_scanout_allocate to drmmode_create_bo_pixmap 3099 3100 radeon_bo_get_tiling can fail, e.g. for linear BOs. 3101 3102 Fixes TearFree not working in those cases. 3103 3104 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97873 3105 Fixes: 040a7b80e1fc ("Explicitly set the fbcon pixmap pitch again") 3106 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3107 3108commit 720dd07e79b5f5cbbf3ef8bd92300578c485d583 3109Author: Jammy Zhou <Jammy.Zhou@amd.com> 3110Date: Mon Dec 19 17:27:19 2016 +0900 3111 3112 Use render node for DRI3 if available 3113 3114 Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com> 3115 (ported from amdgpu commit ea558e645786b08d75307716036045170e97b43e) 3116 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3117 [ Second attempt, let's see if there's any fallout this time... ] 3118 3119commit 92fc387a6256ce3c95d62142735783ea07162c54 3120Author: Michel Dänzer <michel.daenzer@amd.com> 3121Date: Thu Dec 15 12:44:47 2016 +0900 3122 3123 Simplify drmmode_handle_uevents 3124 3125 No functional change intended. 3126 3127 Reviewed-by: Jim Qu <Jim.Qu@amd.com> 3128 (Cherry picked from amdgpu commit edd276185d42962a13faf9ec9eeebc754ef284e7.) 3129 3130commit 404c64752774d23a2db9aa3134f6ada3155708c9 3131Author: jimqu <Jim.Qu@amd.com> 3132Date: Wed Dec 14 11:46:16 2016 +0900 3133 3134 udev_monitor_receive_device() will block when hotplug monitor 3135 3136 udev_monitor_receive_device() will block and wait for the event of udev 3137 use select() to ensure that this will not block. 3138 3139 Signed-off-by: JimQu <Jim.Qu@amd.com> 3140 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 3141 (Cherry picked from amdgpu commit 732cf4d3a248b288532ad0f3443da49e08dc7507) 3142 3143commit a89c0bced423ce3e09f66edafc35642a7cc03b2f 3144Author: Jochen Rollwagen <joro-2013@t-online.de> 3145Date: Fri Dec 2 10:56:39 2016 +0100 3146 3147 Calculate log base 2 in radeon.h based on clz for all platforms 3148 3149 This commit replaces the inline assembler code (for x86 platforms) and 3150 loop (for non-x86 platforms) in RADEONLog2 with a one-liner version 3151 based on clz (count leading zeroes). 3152 3153 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 3154 3155commit e2942449171fe628a7726e59bcaab65e27d88563 3156Author: Michel Dänzer <michel.daenzer@amd.com> 3157Date: Mon Nov 21 18:47:53 2016 +0900 3158 3159 Call radeon_drm_abort_entry on failure to flip to a scanout pixmap 3160 3161 Fixes leaking the corresponding struct radeon_drm_queue list entry in 3162 that case. 3163 3164 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3165 3166commit 14c3f59f5157885ad8f941f0bad6c0c5e3db12f8 3167Author: Michel Dänzer <michel.daenzer@amd.com> 3168Date: Mon Nov 21 18:45:30 2016 +0900 3169 3170 Call ValidateGC after ChangeClip in radeon_sync_scanout_pixmaps 3171 3172 The wrong order meant that the clipping region wasn't actually applied, 3173 so it always copied the full contents from the other scanout pixmap. 3174 3175 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3176 3177commit a995f5830916a0fee5126263d1bfe48632be3a15 3178Author: Michel Dänzer <michel.daenzer@amd.com> 3179Date: Mon Nov 21 18:39:00 2016 +0900 3180 3181 Fix radeon_scanout_extents_intersect for GPU screens 3182 3183 Fixes incorrect screen updates with TearFree enabled on PRIME slave 3184 outputs which are not located at (0, 0). 3185 3186 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3187 3188commit e543ef3a2fb304cbe3a965fb780632af2e4186f4 3189Author: Michel Dänzer <michel.daenzer@amd.com> 3190Date: Mon Nov 21 18:33:25 2016 +0900 3191 3192 Take current scanout_id into account everywhere involved with TearFree 3193 3194 Fixes various potential issues with TearFree enabled, e.g. outputs 3195 freezing after display configuration changes. 3196 3197 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3198 3199commit f130b10e63f7526360b41aa0918b4940f63f662a 3200Author: Michel Dänzer <michel.daenzer@amd.com> 3201Date: Fri Nov 25 18:12:40 2016 +0900 3202 3203 Add radeon_is_gpu_screen helper 3204 3205 This will hopefully decrease the chance of accidentally breaking the 3206 build against xserver < 1.13 in the future. 3207 3208 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3209 3210commit 487aa62a2a23b86e4ea4714fdfd465c9e513141f 3211Author: Jochen Rollwagen <joro-2013@t-online.de> 3212Date: Thu Nov 24 18:44:01 2016 +0100 3213 3214 fix build for xserver < 1.13 3215 3216 same procedure every few patches..... 3217 3218 Fixes: 13c6bc5e3827 ("Don't install Flush/EventCallback for GPU screens") 3219 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 3220 3221commit 5fea5ef2f07eee4a0f94baab427010b936f1d4b4 3222Author: Michel Dänzer <michel.daenzer@amd.com> 3223Date: Thu Nov 24 18:45:22 2016 +0900 3224 3225 Pass fb_id into drmmode_page_flip_target_absolute/relative 3226 3227 drmmode->fb_id isn't what we need in the TearFree case. Fixes TearFree 3228 freezing with 3229 3230 (WW) RADEON(0): flip queue failed in radeon_scanout_flip: No such file or directory 3231 3232 in the log file. 3233 3234 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98834 3235 Fixes: 1106b2f773ad ("Use DRM_MODE_PAGE_FLIP_TARGET_ABSOLUTE/RELATIVE flags when available") 3236 3237commit 13c6bc5e382765fe567091e1c616c0a26eec04ca 3238Author: Michel Dänzer <michel.daenzer@amd.com> 3239Date: Mon Aug 1 18:11:57 2016 +0900 3240 3241 Don't install Flush/EventCallback for GPU screens 3242 3243 Their purpose is to flush GPU rendering commands corresponding to damage 3244 events, but there can be no damage events corresponding to GPU screen 3245 rendering operations. 3246 3247 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3248 3249commit f11531c99fcd6473f58b4d10efaf3efd84304d8e 3250Author: Michel Dänzer <michel.daenzer@amd.com> 3251Date: Mon Nov 21 18:15:05 2016 +0900 3252 3253 Enable glamor by default with >= R600 and Xorg >= 1.18.3 3254 3255 glamor should now perform at least as well as EXA in general, and this 3256 allows DRI3 to be enabled by default for >= R600. 3257 3258 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3259 3260commit 1106b2f773ad0611c729b27f4c192a26b43ef1e7 3261Author: Michel Dänzer <michel.daenzer@amd.com> 3262Date: Wed Jul 6 15:25:33 2016 +0900 3263 3264 Use DRM_MODE_PAGE_FLIP_TARGET_ABSOLUTE/RELATIVE flags when available 3265 3266 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3267 3268commit aab80d2203efb8569b26dd02a1c651c792e1ef0c 3269Author: Michel Dänzer <michel.daenzer@amd.com> 3270Date: Fri Nov 18 12:11:38 2016 +0900 3271 3272 Post-release version bump 3273 3274commit 5cbe1ee8e499e1b6b2646e341946292721d07e69 3275Author: Michel Dänzer <michel.daenzer@amd.com> 3276Date: Thu Nov 17 11:27:16 2016 +0900 3277 3278 Bump version for 7.8.0 release 3279 3280commit 68023e01842ed7e5c798be77c25993bbe6a5a966 3281Author: Michel Dänzer <michel.daenzer@amd.com> 3282Date: Thu Nov 17 11:32:37 2016 +0900 3283 3284 Update manpage entry for Option "TearFree" 3285 3286 It's now effective for arbitrary transforms as well. 3287 3288commit 9760ef33cba5795eddeda4d5c2fcbe2dcce21689 3289Author: Michel Dänzer <michel.daenzer@amd.com> 3290Date: Tue Nov 8 13:02:43 2016 +0900 3291 3292 Use pRADEONEnt to find both screens of a GPU in radeon_mode_hotplug 3293 3294 Fixes misbehaviour when hotplugging DisplayPort connectors on secondary 3295 GPUs. 3296 3297 Fixes: c801f9f10a5d ("Handle Zaphod mode correctly in radeon_mode_hotplug") 3298 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98626 3299 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3300 3301commit 35bec4937d89b48a79acfcb4f814b7370cb631b2 3302Author: Michel Dänzer <michel.daenzer@amd.com> 3303Date: Tue Nov 8 13:01:58 2016 +0900 3304 3305 Refactor radeon_mode_hotplug 3306 3307 Preparation for the next change, no functional change intended. 3308 3309 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3310 3311commit be9f67339a19a6649eab52aa2e0c8971cd9b4727 3312Author: Michel Dänzer <michel.daenzer@amd.com> 3313Date: Mon Nov 7 16:34:37 2016 +0900 3314 3315 Reindent code in radeon_dri2_create_buffer2 3316 3317 Fixes warning about misleading indentation from recent versions of gcc: 3318 3319 ../../src/radeon_dri2.c: In function ‘radeon_dri2_create_buffer2’: 3320 ../../src/radeon_dri2.c:224:2: warning: this ‘if’ clause does not guard... [-Wmisleading-indentation] 3321 if (flags & RADEON_CREATE_PIXMAP_TILING_MACRO) 3322 ^~ 3323 ../../src/radeon_dri2.c:227:6: note: ...this statement, but the latter is misleadingly indented as if it is guarded by the ‘if’ 3324 if (aligned_width == front_width) 3325 ^~ 3326 3327 No functional change intended. 3328 3329 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3330 3331commit 350a2645a1b127227ff294c0b62d20000d0fd48a 3332Author: Michel Dänzer <michel.daenzer@amd.com> 3333Date: Tue Nov 1 16:30:42 2016 +0900 3334 3335 Check Xorg version at runtime instead of build time in two places 3336 3337 This means that all possible paths can be handled as intended, no matter 3338 which Xorg version the driver happened to be compiled against. 3339 3340 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3341 3342commit 5df36de39952c3a26cb2fbc125f298139a9dd5bc 3343Author: Michel Dänzer <michel.daenzer@amd.com> 3344Date: Fri Aug 26 18:24:08 2016 +0900 3345 3346 Require xserver 1.10 or newer 3347 3348 1.10.0 was released in February 2011. 3349 3350 We've been accidentally requiring 1.10 or newer since 121a6de72da5 ("Keep 3351 track of damage event related flushes per-client v2"). 3352 3353 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3354 3355commit 22b5ce9548393ba2ff73ee234ecd82eeaf0ef6c4 3356Author: Michel Dänzer <michel.daenzer@amd.com> 3357Date: Tue Oct 25 17:28:03 2016 +0900 3358 3359 Consume all available udev events at once 3360 3361 We get multiple udev events for actions like docking a laptop into its 3362 station or plugging a monitor to the station. By consuming as many 3363 events as we can, we reduce the number of output re-evalutions. 3364 3365 It depends on the timing how many events can be consumed at once. 3366 3367 (Inspired by xserver commit 363f4273dd4aec3e26cc57ecb6c20f27e6c813d8) 3368 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3369 3370commit 82d3c8f5500d2a6fb1495e217a0b79c396f1534c 3371Author: Hans de Goede <hdegoede@redhat.com> 3372Date: Tue Oct 25 16:56:40 2016 +0900 3373 3374 PRIME: Fix swapping of provider sink / source capabilities 3375 3376 When a card has import capability it can be an offload _sink_, not a 3377 source and vice versa for export capability. 3378 3379 This went unnoticed sofar because most gpus have both import and export 3380 capability. 3381 3382 Signed-off-by: Hans de Goede <hdegoede@redhat.com> 3383 (Ported from xserver commit 94a1c77259ce39ba59ad87615df39b570ffab435) 3384 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3385 3386commit 6c940446ddadf418ee4959e46fa552b6c1cf6704 3387Author: Michel Dänzer <michel.daenzer@amd.com> 3388Date: Tue Oct 25 16:42:03 2016 +0900 3389 3390 Always call PixmapStopDirtyTracking in drmmode_set_scanout_pixmap 3391 3392 Otherwise, we may leak screen->pixmap_dirty_list entries if 3393 drmmode_set_scanout_pixmap is called repatedly with ppix != NULL, which 3394 can happen from RRReplaceScanoutPixmap. 3395 3396 (Inspired by xserver commit b773a9c8126222e5fed2904d012fbf917a9f22fd) 3397 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3398 3399commit 61df12e2377cbb19a19ca9d5624df8959822da9f 3400Author: Michel Dänzer <michel.daenzer@amd.com> 3401Date: Wed Oct 19 18:55:33 2016 +0900 3402 3403 Don't rely on randr_crtc->scanout_pixmap in drmmode_set_scanout_pixmap 3404 3405 RRReplaceScanoutPixmap may set randr_crtc->scanout_pixmap = NULL before 3406 we get here. 3407 3408 (Inspired by xserver commit f4c37eeee7953df1fe0e3196eda452acf0078e61) 3409 v2: Always return TRUE in the if (!ppix) block. 3410 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1) 3411 3412commit 49cf3b5032a7ce40afe514b7092440e3e19e05aa 3413Author: Michel Dänzer <michel.daenzer@amd.com> 3414Date: Thu Oct 20 15:33:09 2016 +0900 3415 3416 Sayōnara, AM_MAINTAINER_MODE! 3417 3418 If --enable-maintainer-mode got lost from config.status for any reason, 3419 builds would fail in mysterious ways after changing between different 3420 Git commits. 3421 3422 There are more reasons for dropping it in the automake manual: 3423 3424 https://www.gnu.org/software/automake/manual/html_node/maintainer_002dmode.html 3425 3426 I'm not aware of any reason why --disable-maintainer-mode would ever be 3427 useful with this project. 3428 3429 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3430 3431commit 4b17de1cad12e7be12f2c71e5149bfc04564702d 3432Author: Michel Dänzer <michel.daenzer@amd.com> 3433Date: Wed Oct 19 18:09:14 2016 +0900 3434 3435 Remove PCI IDs and bus type from ati_pciids.csv 3436 3437 This cleans up the list of unique chipsets a little further. 3438 3439 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3440 3441commit 1ce1b1656acc6211deb2091ff7f28d51b6daf86b 3442Author: Michel Dänzer <michel.daenzer@amd.com> 3443Date: Wed Oct 19 18:08:36 2016 +0900 3444 3445 Order unique chipsets according to first appearance in ati_pciids.csv 3446 3447 Instead of lexically. This makes it more likely for similar generations 3448 to be close to each other in the list of unique chipsets. 3449 3450 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3451 3452commit 7c66be961de3477fc685eb71b46265adbae923f0 3453Author: Michel Dänzer <michel.daenzer@amd.com> 3454Date: Mon Oct 17 18:38:55 2016 +0900 3455 3456 Enable HW cursor support with PRIME slave output & Xorg > 1.18.99.901 3457 3458 Supported since Xorg 1.18.99.2, but buggy until 1.18.99.901. 3459 3460 (Ported from amdgpu commit 7cc04035c55788261cda89a915c433c2add6cad9) 3461 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3462 3463commit d016f20b14ffe4f90f7457375b2d9f5cf0f5650f 3464Author: Michel Dänzer <michel.daenzer@amd.com> 3465Date: Thu Oct 13 16:33:45 2016 +0900 3466 3467 Restore location of radeon_set_pixmap_bo call in drmmode_create_bo_pixmap 3468 3469 radeon_set_pixmap_bo sets the tiling flags, so it needs to be called 3470 before radeon_get_pixmap_tiling_flags. Fixes a regression with EXA and 3471 TearFree. 3472 3473 Fixes: 8523a733b6a5 ("Propagate failure from radeon_set_pixmap_bo") 3474 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98097 3475 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3476 3477commit 937398f6cd50db105d5e53ab553eb284d44121f4 3478Author: Michel Dänzer <michel.daenzer@amd.com> 3479Date: Thu Oct 6 18:52:37 2016 +0900 3480 3481 Rotate and reflect cursor hotspot position for drmModeSetCursor2 3482 3483 We were always passing the hotspot position in the X screen coordinate 3484 space, but drmModeSetCursor2 needs it in the CRTC coordinate space. The 3485 wrong hotspot position would cause the kernel driver to adjust the 3486 HW cursor position incorrectly when the hotspot position changed. 3487 3488 (Ported from amdgpu commit d42773eb45baff5933730e26878a0b45fcf07b65) 3489 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3490 3491commit 3fc839ff49f01c24eb94d5e1f0ab4065de47bc17 3492Author: Michel Dänzer <michel.daenzer@amd.com> 3493Date: Wed Sep 21 17:50:36 2016 +0900 3494 3495 Only list each unique chipset family once in the log file 3496 3497 (Ported from amdgpu commit 6a1ba044c2b71081e6060d0c096917d6238f2145) 3498 3499commit 9a5eb1915a240e3021ac38a6dacafef81e7bb743 3500Author: Michel Dänzer <michel.daenzer@amd.com> 3501Date: Wed Sep 21 17:48:38 2016 +0900 3502 3503 Clear damage in radeon_scanout_update if it doesn't intersect the CRTC 3504 3505 There's no need to test that same damage again. 3506 3507 (Ported from amdgpu commit a576430526cbc404de64b30e1377a356644e8024) 3508 3509commit 792a30dbe5f7d7815ea14032dbab55b5de8eabca 3510Author: Michel Dänzer <michel.daenzer@amd.com> 3511Date: Wed Sep 21 17:43:57 2016 +0900 3512 3513 Remove w/h parameters from radeon_scanout_extents_intersect 3514 3515 We can use the dimensions of the CRTC's mode instead. 3516 3517 (Ported from amdgpu commit ede7f2bcae63be65e05e3029bfe7c742e5978932) 3518 3519commit 80d794f11a5f047663897e4349da75d89ceff7c7 3520Author: Michel Dänzer <michel.daenzer@amd.com> 3521Date: Wed Sep 21 17:38:57 2016 +0900 3522 3523 Make the dedicated scanout mechanism work with arbitrary transforms 3524 3525 This makes TearFree work with arbitrary transforms, and makes transforms 3526 work better even without TearFree, with xserver >= 1.12. 3527 3528 (Ported from amdgpu commit bf000ea7ef91f5ecb59fc3c1ab8ed9eddcc0841d) 3529 3530commit 09ba7b3b7b1b724aa386dbf2d33f567d7f26326b 3531Author: Michel Dänzer <michel.daenzer@amd.com> 3532Date: Wed Sep 21 17:00:22 2016 +0900 3533 3534 Remove drmmode_load_palette 3535 3536 Not used by any supported version of xserver. 3537 3538 (Ported from amdgpu commits 1091f28e1fa239ee1a973d84a8376fa4a95d7247 3539 and 5a4d3267ac3823fe58b51b0b9075b82375d7180c) 3540 3541commit 730f443955cf2bcedd760c9fef5b11c81eacdd84 3542Author: Michel Dänzer <michel.daenzer@amd.com> 3543Date: Wed Sep 21 16:48:49 2016 +0900 3544 3545 present: Separate checks for flips vs unflips 3546 3547 All unflip checks apply to flips as well, but not vice versa. 3548 3549 (Ported from amdgpu commit 4d506c23c9a628204fa23607931557b07ada3e31) 3550 3551commit 343d28672411a510de21fa57986324ef8bcdd1b6 3552Author: Michel Dänzer <michel.daenzer@amd.com> 3553Date: Wed Sep 21 16:47:17 2016 +0900 3554 3555 dri2: Don't allow flipping when using a dedicated scanout buffer 3556 3557 Fixes issues when mixing rotation and page flipping with current xserver 3558 Git master. 3559 3560 (Ported from amdgpu commit decabd574f90d3df397c80ec931b3fde8a4afb49) 3561 3562commit 83e4781d15d66d6fa0e153eb4cd0e3d9e2ebed9f 3563Author: Michel Dänzer <michel.daenzer@amd.com> 3564Date: Wed Sep 21 16:44:50 2016 +0900 3565 3566 present: Don't allow flipping when using a dedicated scanout buffer 3567 3568 Fixes issues when mixing rotation and page flipping with current xserver 3569 Git master. 3570 3571 (Ported from amdgpu commit 3ed28ce7cd26f89969617ba901ff253091d0d469) 3572 3573commit 8419db3de6157875f9a840773350ecd29cd42d83 3574Author: Michel Dänzer <michel.daenzer@amd.com> 3575Date: Wed Sep 21 16:40:42 2016 +0900 3576 3577 Make sure drmmode_crtc->scanout[] are destroyed when not needed 3578 3579 We failed to do this when going back to scanning out directly from the 3580 screen pixmap. 3581 3582 As a bonus, since we now destroy drmmode_crtc->scanout[] after setting 3583 the new scanout buffer, we may avoid the CRTC turning off intermittently 3584 in this case. 3585 3586 (Ported from amdgpu commit 9c3324715fd395fd486ea341654d78f4f298b97f) 3587 3588commit 956e03d6a6b9478dd8e77c81f426c0d371c5d7a1 3589Author: Michel Dänzer <michel.daenzer@amd.com> 3590Date: Wed Sep 21 16:37:53 2016 +0900 3591 3592 Simplify drmmode_set_mode_major error handling 3593 3594 Initialize ret = FALSE and only set it to TRUE when we've succeeded. 3595 3596 (Ported from amdgpu commit 3bce0519a4008cf87c0e31a7a579e10f5dcdd2f3) 3597 3598commit ace455d1db4db3f9b8b434a3747acfa7e1273db4 3599Author: Hans de Goede <hdegoede@redhat.com> 3600Date: Wed Sep 21 16:31:51 2016 +0900 3601 3602 Only add main fb if necessary 3603 3604 If we're doing reverse-prime; or doing rotation the main fb is not used, 3605 and there is no reason to add it in this case. 3606 3607 Signed-off-by: Hans de Goede <hdegoede@redhat.com> 3608 (Ported from xserver commit 4313122dea0df9affc280ee698e929489061ccc6) 3609 (Ported from amdgpu commit a3ca1500703837cbb8d49c554199a25dea7d5e1e) 3610 3611commit eea370dc4f96e680cc29b416fe62d6e2ea09bc91 3612Author: Hans de Goede <hdegoede@redhat.com> 3613Date: Wed Sep 21 16:27:11 2016 +0900 3614 3615 Remove unnecessary fb addition from drmmode_xf86crtc_resize 3616 3617 drmmode_set_mode_major() is the only user of drmmode->fb_id and will 3618 create it if necessary. 3619 3620 Signed-off-by: Hans de Goede <hdegoede@redhat.com> 3621 (Ported from xserver commit 877453212166fdc912e0d687cdecee11aba563b5) 3622 (Ported from amdgpu commit 9ca1c24235ff5ab2e028333fc326e2eff008c574) 3623 3624commit cf52c76484a4adfa3e0b2271ea0c00f46d32ca48 3625Author: Michel Dänzer <michel.daenzer@amd.com> 3626Date: Tue Sep 13 18:21:37 2016 +0900 3627 3628 Free priv in amdgpu_set_pixmap_bo also if priv->bo == NULL 3629 3630 Fixes memory leak when destroying pixmaps with priv->bo == NULL. 3631 3632 Reported-by: Qiang Yu <qiang.yu@amd.com> 3633 (Ported from amdgpu commit 7f7f9825caf3983902491da27c16d14cd8bf9b7d) 3634 3635commit 50abcc19e29bb50451d031b2c9e99d691d4faf70 3636Author: Michel Dänzer <michel.daenzer@amd.com> 3637Date: Tue Sep 13 18:18:37 2016 +0900 3638 3639 glamor: Fix leak of pixmap private when replacing BO 3640 3641 Reported-by: Qiang Yu <qiang.yu@amd.com> 3642 (Ported from amdgpu commit 397aedafee437c125b8ac1feafb1c3b466842aeb) 3643 3644commit 8523a733b6a5de6116a6332fefc871d4c32652d8 3645Author: Michel Dänzer <michel.daenzer@amd.com> 3646Date: Mon Sep 12 18:54:33 2016 +0900 3647 3648 Propagate failure from radeon_set_pixmap_bo 3649 3650 (Ported from amdgpu commits c315c00e44afc91a7c8e2eab5af836d9643ebb88 3651 and 0d42082108c264568e2aadd15ace70e72388bc65) 3652 3653commit 53be26b00e83f871f0afd39caa5a7a1d6ec4aea1 3654Author: Michel Dänzer <michel.daenzer@amd.com> 3655Date: Wed Aug 31 16:46:56 2016 +0900 3656 3657 Add support for ScreenPtr::SyncSharedPixmap 3658 3659 This allows deferring shared pixmap updates between different drivers. 3660 3661 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3662 3663commit 156b5bc00da2eecb71cad517136f8cd74cf5d2c9 3664Author: Michel Dänzer <michel.daenzer@amd.com> 3665Date: Tue Sep 20 17:50:20 2016 +0900 3666 3667 Only define transform_region function for XF86_CRTC_VERSION >= 4 3668 3669 We're not using it with older xserver. 3670 3671 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3672 3673commit 80cc892ee1ce54fad3cb7dd11bd9df18c359136f 3674Author: Michel Dänzer <michel.daenzer@amd.com> 3675Date: Tue Sep 20 17:45:16 2016 +0900 3676 3677 Use local implementation of RegionDuplicate for older xserver 3678 3679 It was only added in xserver 1.15. Fixes build against older xserver. 3680 3681 Reported-by: Pali Rohár <pali.rohar@gmail.com> 3682 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3683 3684commit 12d30eeb9711bd2b1609d6bbb74c4a1760596f72 3685Author: Qiang Yu <Qiang.Yu@amd.com> 3686Date: Wed Sep 14 17:03:03 2016 +0900 3687 3688 DRI2: Fix radeon_dri2_exchange_buffers width/height copy'n'paste error 3689 3690 Signed-off-by: Qiang Yu <Qiang.Yu@amd.com> 3691 (Ported from amdgpu commit 73c8dc000ad6b2b53ba3aa7155f5e8f6b55623b7) 3692 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3693 3694commit 11cc6843aa3b745aa1361f1a65e465e16696b914 3695Author: Michel Dänzer <michel.daenzer@amd.com> 3696Date: Wed Sep 14 17:01:13 2016 +0900 3697 3698 DRI2: Add interpolated_vblanks in radeon_dri2_get_crtc_msc 3699 3700 We need that in radeon_dri2_drawable_crtc as well for priv->vblank_delta 3701 to work as intended. 3702 3703 radeon_dri2_get_msc was already doing this. 3704 3705 Fixes hangs in some cases when using VDPAU via DRI2 and moving the 3706 window between CRTCs. 3707 3708 (Ported from amdgpu commit abd1a7901c95e4bc78415cf1b7923623b9177152) 3709 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3710 3711commit 380daff43cfefe2a8e4d496eaf65673f2815c878 3712Author: Qiang Yu <Qiang.Yu@amd.com> 3713Date: Wed Sep 14 16:59:38 2016 +0900 3714 3715 Fix radeon_mode_hotplug crash on multi GPU platform. 3716 3717 On multi GPU platform, some screen is created by other GPU DDX. 3718 3719 Signed-off-by: Qiang Yu <Qiang.Yu@amd.com> 3720 (Ported from amdgpu commit 978242977e5dc905e1d5a46b1b0d34b356c7af26) 3721 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3722 3723commit 4a5fa37f74c233c6b9c6a08306688628a8e216e8 3724Author: Michel Dänzer <michel.daenzer@amd.com> 3725Date: Wed Sep 14 16:56:22 2016 +0900 3726 3727 Destroy all dedicated scanout buffers during CloseScreen 3728 3729 Fixes leaking active scanout buffers across a server reset, which also 3730 fixes server reset with glamor and active scanout buffers. 3731 3732 (Ported from amdgpu commit d96dabc71b1b32dc4b422a9633cdd4e0e95da052) 3733 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3734 3735commit cc2555835cd4c5fd2ae4f999a4bf7c18cdb1dda4 3736Author: Michel Dänzer <michel.daenzer@amd.com> 3737Date: Mon Sep 12 18:26:57 2016 +0900 3738 3739 glamor: Reallocate linear pixmap BO if necessary for DRI2 PRIME 3740 3741 Fixes corruption when using DRI2 PRIME render offloading with the master 3742 screen using this driver. 3743 3744 (Ported from amdgpu commit 0007c2f018ba663303d91d847e7c085269a23062) 3745 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3746 3747commit b3e5259e60157fdbdf46ee59b1b78995c2b15f72 3748Author: Michel Dänzer <michel.daenzer@amd.com> 3749Date: Mon Sep 12 18:12:06 2016 +0900 3750 3751 Move DRI2's local fixup_glamor helper to radeon_glamor_set_pixmap_bo 3752 3753 So it can be used outside of the DRI2 code. 3754 3755 (Ported from amdgpu commit 5518bf5d793439b5bab369e5fc18de9a4a3b9dd6) 3756 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3757 3758commit 38632bbd5ff80a9cf8ce584b2bc499d17d15befe 3759Author: Michel Dänzer <michel.daenzer@amd.com> 3760Date: Mon Sep 12 18:05:51 2016 +0900 3761 3762 Consolidate get_drawable_pixmap helper 3763 3764 There were two static helpers for the same purpose. Consolidate them 3765 into a single inline helper which can be used anywhere. 3766 3767 (Ported from amdgpu commit 641f4647b7f51dfd2da330376cd10fa9702b6423) 3768 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3769 3770commit e91858e435672c32f9c4a854b3dec048199d6f7f 3771Author: Michel Dänzer <michel.daenzer@amd.com> 3772Date: Mon Sep 12 18:02:07 2016 +0900 3773 3774 glamor: Fix radeon_glamor_share_pixmap_backing for priv->bo == NULL 3775 3776 Fixes crash when running a compositor and DRI_PRIME client via DRI2. 3777 3778 Reported-by: Qiang Yu <qiang.yu@amd.com> 3779 (Ported from amdgpu commit b36c77695ba77b59a0ccd868454e3af4fc04d5ff) 3780 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3781 3782commit 38797a33117222dadbc89e5f21ed8cd5deef9bea 3783Author: Michel Dänzer <michel.daenzer@amd.com> 3784Date: Fri Sep 2 11:08:28 2016 +0900 3785 3786 Make TearFree effective with PRIME slave scanout 3787 3788 TearFree can now prevent tearing with any possible display 3789 configuration. 3790 3791 Note that there may still be inter-GPU tearing if the primary GPU uses 3792 a different driver. 3793 3794 v2: 3795 * Also test dirty->slave_dst in radeon_prime_scanout_do_update 3796 3797 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> [v1] 3798 3799commit eda1f3df6aaed683036369fe8820da4dac3c2ae2 3800Author: Michel Dänzer <michel.daenzer@amd.com> 3801Date: Thu Sep 1 12:54:13 2016 +0900 3802 3803 Synchronize scanout pixmaps for TearFree 3804 3805 Copy the damaged areas which are still valid in the other scanout pixmap 3806 from there, then only copy the remaining damaged area from the screen 3807 pixmap. 3808 3809 This is slightly more efficient (only needs one Damage record instead of 3810 two, and only needs to copy each screen update across PCIe once with 3811 ShadowPrimary and a discrete GPU), and will be significantly more 3812 efficient for PRIME with the following change. 3813 3814 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3815 3816commit 2f6e5fb15f1a9ce523c85550e493f8bda9d0c00f 3817Author: Michel Dänzer <michel.daenzer@amd.com> 3818Date: Fri Sep 2 17:23:16 2016 +0900 3819 3820 Move up radeon_scanout_extents_intersect 3821 3822 Will be needed higher up by the following changes. No functional change. 3823 3824 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3825 3826commit 5a57005178fc13b6f7e513458ca6dae72a3e5783 3827Author: Michel Dänzer <michel.daenzer@amd.com> 3828Date: Thu Sep 1 17:19:27 2016 +0900 3829 3830 Factor out transform_region helper 3831 3832 While we're at it, fix leaking the memory allocated for xRectangles. 3833 3834 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3835 3836commit 99232f64db52812a843cd616d263d3a6b90eef3d 3837Author: Michel Dänzer <michel.daenzer@amd.com> 3838Date: Tue Dec 1 17:58:13 2015 +0900 3839 3840 Only copy from screen pixmap to shared pixmap on demand for slave scanout 3841 3842 Only copy once for each time we update the corresponding scanout pixmap. 3843 This can significantly reduce the bandwidth usage when there are 3844 frequent updates to the screen pixmap. 3845 3846 This initial implementation only works when both the master and slave 3847 screens use this driver. 3848 3849 v2: 3850 * Reduce churn in radeon_prime_scanout_update_handler 3851 * Clear the correct damage in radeon_dirty_update 3852 3853 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3854 3855commit a92c27484703abc7c410b6ae0e4b8d1efbbb8e6f 3856Author: Michel Dänzer <michel.daenzer@amd.com> 3857Date: Fri Aug 26 18:09:03 2016 +0900 3858 3859 Fix build against xserver < 1.13 3860 3861 pScreen->isGPU was only introduced in 1.13. 3862 3863 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97490 3864 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3865 3866commit 4bd2d01552f18153afa03a8947b22eebf3d67c6b 3867Author: Michel Dänzer <michel.daenzer@amd.com> 3868Date: Fri Aug 26 12:02:38 2016 +0900 3869 3870 Also handle disabled CRTCs in drmmode_clear_pending_flip 3871 3872 If disabling a CRTC had to be deferred due to a pending flip in 3873 drmmode_crtc_dpms, there may no longer be any outputs associated with 3874 the CRTC when we get here. So we have to check for !crtc->enabled and 3875 call drmmode_crtc_dpms in that case as well. 3876 3877 Fixes: 9090309e057d ("Wait for pending flips to complete before turning 3878 off an output or CRTC") 3879 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3880 3881commit a36fdaff40d5b4795a1400c348a80eee94892212 3882Author: Michel Dänzer <michel.daenzer@amd.com> 3883Date: Wed Aug 24 22:52:11 2016 +0900 3884 3885 Don't override crtc parameter value in drmmode_flip_handler/abort 3886 3887 When overriding the crtc parameter value of the last pending CRTC, 3888 drmmode_clear_pending_flip would work on the wrong CRTC, and the last 3889 pending CRTC's flip_pending flag might never get cleared. This would 3890 prevent that CRTC from properly turning off and back on again. 3891 3892 Fixes: 9090309e057d ("Wait for pending flips to complete before turning 3893 off an output or CRTC") 3894 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97392 3895 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3896 3897commit e520ce0ec0adf91ddce5c932d4b3f9477fd49304 3898Author: Michel Dänzer <michel.daenzer@amd.com> 3899Date: Thu Aug 25 18:22:34 2016 +0900 3900 3901 Also call drmmode_clear_pending_flip from radeon_scanout_flip_abort 3902 3903 Not doing so could break DPMS with TearFree. 3904 3905 Reported-and-Tested-by: furkan on IRC 3906 Fixes: 9090309e057d ("Wait for pending flips to complete before turning 3907 off an output or CRTC") 3908 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3909 3910commit b0867063abb197b9134166706d99fcbe5f204bb5 3911Author: Michel Dänzer <michel.daenzer@amd.com> 3912Date: Mon Nov 30 18:54:12 2015 +0900 3913 3914 Track damage accurately for RandR 1.4 slave scanout 3915 3916 This further reduces the PCIe bandwidth usage. 3917 3918 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3919 3920commit ad0a0656dd0e74683e6d7789decba827aa29c221 3921Author: Michel Dänzer <michel.daenzer@amd.com> 3922Date: Fri Nov 27 16:31:21 2015 +0900 3923 3924 Handle RandR 1.4 slave dirty updates via radeon_drm_queue 3925 3926 This reduces PCIe bandwidth usage and tearing. 3927 3928 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3929 3930commit 4cfa4615f79f64062e5e771cd45dd7048f48b4f6 3931Author: Michel Dänzer <michel.daenzer@amd.com> 3932Date: Fri Nov 27 16:53:30 2015 +0900 3933 3934 Use drmmode_crtc_scanout_* helpers for RandR 1.4 scanout pixmaps 3935 3936 This should allow using multiple CRTCs via RandR 1.4 even with xserver 3937 < 1.17. It also simplifies the code a little, and paves the way for 3938 following changes. 3939 3940 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3941 3942commit 121a6de72da5fcf9a32408eff36b2235f3dfbcfe 3943Author: Michel Dänzer <michel.daenzer@amd.com> 3944Date: Mon Jul 11 12:51:46 2016 +0900 3945 3946 Keep track of damage event related flushes per-client v2 3947 3948 This further reduces the compositing slowdown due to flushing overhead, 3949 by only flushing when the X server actually sends XDamageNotify events 3950 to a client, and there hasn't been a flush yet in the meantime. 3951 3952 v2: Use ScreenPrivateKey, fixes invalid memory access with GPU screens 3953 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3954 3955commit 9090309e057dc703d1a5bffd88e6cae14108cfc3 3956Author: Michel Dänzer <michel.daenzer@amd.com> 3957Date: Tue May 10 18:45:30 2016 +0900 3958 3959 Wait for pending flips to complete before turning off an output or CRTC 3960 3961 At least with older kernels, the flip may never complete otherwise, 3962 which can result in us hanging in drmmode_set_mode_major. 3963 3964 Fixes: https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/1577170 3965 3966 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3967 3968commit 9a1afbf61fbb2827c86bd86d295fa0848980d60b 3969Author: Michel Dänzer <michel.daenzer@amd.com> 3970Date: Mon Jul 11 12:22:09 2016 +0900 3971 3972 Use EventCallback to avoid flushing every time in the FlushCallback 3973 3974 We only need to flush for XDamageNotify events. 3975 3976 Significantly reduces compositing slowdown due to flushing overhead, in 3977 particular with glamor. 3978 3979 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3980 3981commit 94fe42f29e0b00a26e810581d6c438ac6d8ecd8a 3982Author: Michel Dänzer <michel.daenzer@amd.com> 3983Date: Tue Jul 12 17:36:27 2016 +0900 3984 3985 Don't enable DRI3 by default with EXA 3986 3987 It doesn't work correctly in all cases, see e.g. 3988 https://bugs.freedesktop.org/show_bug.cgi?id=95475 . I'm not sure this 3989 is fixable, given EXA's architecture. 3990 3991 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 3992 3993commit 6d91fb4fc701895473ff675f440a8eef655e80ca 3994Author: Michel Dänzer <michel.daenzer@amd.com> 3995Date: Tue Jul 12 15:57:28 2016 +0900 3996 3997 Don't enable micro-tiling for scanout buffers on pre-R600 3998 3999 The display engine didn't support it. 4000 4001 Fixes display corruption with options "TearFree" and "ShadowPrimary" 4002 (and rotation or transforms with current xserver) on pre-R600. 4003 4004 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4005 4006commit a37af701768b12d86868a831a79f1e02ee4968cf 4007Author: Michel Dänzer <michel.daenzer@amd.com> 4008Date: Wed Jul 6 12:46:01 2016 +0900 4009 4010 Add explicit RADEON_DRM_QUEUE_ERROR define 4011 4012 Should make the radeon_drm_queue_alloc error handling clearer, and gets 4013 rid of a compile warning about it returning NULL. 4014 4015 Reviewed-by: Alexandre Demers <alexandre.f.demers@gmail.com> 4016 4017commit 024afff27eb9f4c60043214099ddb42bbfb3d289 4018Author: Keith Packard <keithp@keithp.com> 4019Date: Mon Jul 18 21:24:23 2016 -0700 4020 4021 Adapt to video API 22 4022 4023 Deal with drm fd wakeup 4024 4025 Signed-off-by: Keith Packard <keithp@keithp.com> 4026 Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> 4027 4028commit eac4db4f0cb22a23d6e69161dcaacc7d53978aae 4029Author: Adam Jackson <ajax@redhat.com> 4030Date: Tue Jul 19 10:03:56 2016 -0400 4031 4032 Adapt Block/WakeupHandler signature for ABI 23 4033 4034 Signed-off-by: Adam Jackson <ajax@redhat.com> 4035 4036commit 3be841d0ae7d505cef325993205b12d15e98dba9 4037Author: Michel Dänzer <michel.daenzer@amd.com> 4038Date: Fri Jun 24 11:12:49 2016 +0900 4039 4040 Only use RandR APIs if RandR is enabled 4041 4042 Fixes crash with Xinerama enabled, which disables RandR. 4043 4044 Fixes: https://bugs.debian.org/827984 4045 4046 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4047 4048commit 7835558acdce318130ba4a09ef936fd675e3197d 4049Author: Michel Dänzer <michel.daenzer@amd.com> 4050Date: Mon Mar 28 18:49:15 2016 +0900 4051 4052 Adapt to XF86_CRTC_VERSION 7 4053 4054 Now the HW cursor can be used with TearFree rotation. 4055 4056 This also allows always using the separate scanout pixmap mechanism for 4057 rotation, so that should be much smoother even without TearFree enabled. 4058 4059 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4060 4061commit 0945db4d902056bda3d9ad4a4de2dfa685d10a70 4062Author: Tan Hu <tan.hu@zte.com.cn> 4063Date: Fri May 27 17:05:15 2016 +0800 4064 4065 EXA/6xx/7xx: accelerate PictOpOver with component alpha 4066 4067 Subpixel text rendering is typically done with a solid src and 4068 a pixmap mask. Traditionally, this cannot be accelerated in a single 4069 pass and requires two passes [1]. However, we can cheat a little 4070 with a constant blend color. 4071 4072 We can use: 4073 const.A = src.A / src.A 4074 const.R = src.R / src.A 4075 const.G = src.G / src.A 4076 const.B = src.B / src.A 4077 4078 dst.A = const.A * (src.A * mask.A) + (1 - (src.A * mask.A)) * dst.A 4079 dst.R = const.R * (src.A * mask.R) + (1 - (src.A * mask.R)) * dst.R 4080 dst.G = const.G * (src.A * mask.G) + (1 - (src.A * mask.G)) * dst.G 4081 dst.B = const.B * (src.A * mask.B) + (1 - (src.A * mask.B)) * dst.B 4082 4083 This only needs a single source value. src.A is cancelled down in 4084 the right places. 4085 4086 [1] http://anholt.livejournal.com/32058.html 4087 4088 r6xx still be used on some machine, 4089 Ported from commit 4375a6e75e5d41139be7031a0dee58c057ecbd07. 4090 4091 Signed-off-by: Tan Hu <tan.hu@zte.com.cn> 4092 Reviewed-by: Grigori Goronzy <greg@chown.ath.cx> 4093 4094commit 9b9ad669c748f53247e53fa3f3b03a77da5e5cb3 4095Author: Tan Hu <tan.hu@zte.com.cn> 4096Date: Fri May 27 17:05:14 2016 +0800 4097 4098 EXA/6xx/7xx: fast solid pixmap support 4099 4100 Solid pixmaps are currently implemented with scratch pixmaps, which 4101 is slow. This replaces the hack with a proper implementation. The 4102 Composite shader can now either sample a src/mask or use a constant 4103 value. 4104 4105 r6xx still be used on some machine, 4106 Ported from commit 94d0d14914a025525a0766669b556eaa6681def7. 4107 4108 Signed-off-by: Tan Hu <tan.hu@zte.com.cn> 4109 Reviewed-by: Grigori Goronzy <greg@chown.ath.cx> 4110 4111commit aa07b365d7b0610411e118f105e49daff5f5a5cf 4112Author: Michel Dänzer <michel.daenzer@amd.com> 4113Date: Tue May 31 18:16:26 2016 +0900 4114 4115 Add .editorconfig file 4116 4117 Basically a conversion from .dir-locals.el, but also correctly handles 4118 files which predominantly use tabs for indentation. 4119 4120 Also, EditorConfig supports many more editors and IDEs. 4121 4122 Acked-by: Alex Deucher <alexander.deucher@amd.com> 4123 4124commit 040a7b80e1fcbaa93ac17f7113d696d9b853cf8a 4125Author: Michel Dänzer <michel.daenzer@amd.com> 4126Date: Tue Apr 12 18:18:43 2016 +0900 4127 4128 Explicitly set the fbcon pixmap pitch again 4129 4130 The kernel driver returns 0 as the pitch of the fbcon BO via the 4131 DRM_RADEON_GEM_GET_TILING ioctl, so we ended up using an incorrect 4132 pitch in some cases. 4133 4134 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94901 4135 4136 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4137 4138commit 1181b9c582f10b6c523e4b2988e2ce87ecf3d367 4139Author: Michel Dänzer <michel.daenzer@amd.com> 4140Date: Tue May 10 19:02:20 2016 +0900 4141 4142 Enable DRI3 by default when building for Xorg >= 1.18.3 4143 4144 Seems to work well enough in general now. 4145 4146 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4147 4148commit c801f9f10a5d72d935faf21e72f7e7808fb4f05f 4149Author: Michel Dänzer <michel.daenzer@amd.com> 4150Date: Wed Apr 13 11:12:38 2016 +0900 4151 4152 Handle Zaphod mode correctly in radeon_mode_hotplug 4153 4154 We need to scan both screens of the entity for existing connectors, and 4155 enumerate DVI & HDMI connectors consistently regardless of which screen 4156 they're assigned to. 4157 4158 Fixes crash when hot-(un)plugging connectors in Zaphod mode. 4159 4160 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93415 4161 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4162 4163commit 1ca677309720e2f6c953c9e76f5b34c22a4416c6 4164Author: Michel Dänzer <michel.daenzer@amd.com> 4165Date: Tue Mar 15 16:47:35 2016 +0900 4166 4167 present: Support async flips 4168 4169 The xserver Present code only calls radeon_present_flip with 4170 sync_flip=FALSE if radeon_present_screen_init sets 4171 PresentCapabilityAsync, and the latter only sets it if the kernel driver 4172 advertises support for async flips. 4173 4174 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4175 4176commit 90a915c62d012e99193833aecc93974e68880c60 4177Author: Michel Dänzer <michel.daenzer@amd.com> 4178Date: Tue Mar 15 16:42:16 2016 +0900 4179 4180 Add support for async flips to radeon_do_pageflip 4181 4182 Will be used by the next change. No functional change here. 4183 4184 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4185 4186commit ba8b6288c8e6fc4be5d7144ecbe9a1f241881674 4187Author: Qiang Yu <Qiang.Yu@amd.com> 4188Date: Mon Apr 11 16:35:37 2016 +0900 4189 4190 Remove RR_Capability_SinkOutput for GPU without CRTC 4191 4192 Signed-off-by: Qiang Yu <Qiang.Yu@amd.com> 4193 (Ported from amdgpu commit a0bbb373f902e0ffc14570c85faec7e44134f62e) 4194 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4195 4196commit bd4c72c8625996d842824ce4963f2d759fe2954a 4197Author: Tom St Denis <tom.stdenis@amd.com> 4198Date: Fri Apr 8 10:22:11 2016 -0400 4199 4200 dri3: Return NULL from radeon_dri3_pixmap_from_fd if calloc fails. 4201 4202 Signed-off-by: Tom St Denis <tom.stdenis@amd.com> 4203 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 4204 4205commit 3300ea01fbdf770dc084eebbf2854fba35144220 4206Author: Michel Dänzer <michel.daenzer@amd.com> 4207Date: Mon Apr 11 16:11:57 2016 +0900 4208 4209 Post 7.7.0 release version bump 4210 4211commit df6662864b56d1aeeea4c67d9b668e64197afb20 4212Author: Michel Dänzer <michel.daenzer@amd.com> 4213Date: Thu Apr 7 11:44:05 2016 +0900 4214 4215 Bump version for 7.7.0 release 4216 4217commit c12614cfdc5729041c204e3ead1dbfde38011f0c 4218Author: Michel Dänzer <michel.daenzer@amd.com> 4219Date: Mon Apr 4 14:59:41 2016 +0900 4220 4221 Update manpage entry for Option "TearFree" 4222 4223 It's now effective for rotation as well. 4224 4225 (Ported from amdgpu commit faf9d720b7d650f5f1ea657a874d08eac3972e60) 4226 4227 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4228 4229commit 4693b1bd5b5c381e8b7b68a6f7f0c6696d6a68df 4230Author: Michel Dänzer <michel.daenzer@amd.com> 4231Date: Thu Mar 31 17:02:55 2016 +0900 4232 4233 Identify DRM event queue entries by sequence number instead of by pointer 4234 4235 If the memory for an entry was allocated at the same address as that for 4236 a previously cancelled entry, the handler could theoretically be called 4237 prematurely, triggered by the DRM event which was submitted for the 4238 cancelled entry. 4239 4240 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4241 4242commit 83734317e6bdaeebb4462a63f541e73a1d7c2f77 4243Author: Michel Dänzer <michel.daenzer@amd.com> 4244Date: Wed Mar 30 11:44:09 2016 +0900 4245 4246 Update pixmap pitch in radeon_set_pixmap_bo 4247 4248 Stop second guessing it in drmmode_crtc_scanout_create. 4249 4250 Fixes display corruption in some cases with TearFree enabled. 4251 4252 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94751 4253 4254 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4255 4256commit 0b3aac1de9db42bfca545fa331e4985836682ec7 4257Author: Michel Dänzer <michel.daenzer@amd.com> 4258Date: Mon Mar 28 18:43:36 2016 +0900 4259 4260 DRI3: Refuse to open DRM file descriptor for ssh clients (v2) 4261 4262 Fixes hangs when attempting to use DRI3 on display connections forwarded 4263 via SSH. 4264 4265 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93261 4266 4267 v2: Don't do this for Xorg > 1.18.99.1 since the corresponding xserver 4268 change has landed in Git master. 4269 4270 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1) 4271 4272commit 8a6cd4bda05b9569b3dd0a5a75b2cc385b9ecba9 4273Author: Michel Dänzer <michel.daenzer@amd.com> 4274Date: Fri Mar 25 11:50:15 2016 +0900 4275 4276 Revert "Use render node for DRI3 if available" 4277 4278 This reverts commit cd94248ffa7d8fe0b57476f79e7e860dee66d1b0. 4279 4280 It broke VDPAU<->GL interop with DRI3 enabled, because the Gallium VDPAU 4281 code doesn't support DRI3 yet. We can consider re-enabling this once 4282 there is a Mesa release where the Gallium VDPAU code supports DRI3. 4283 4284 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94675 4285 4286 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4287 4288commit ed31cffba0d1bd4b14e5348a1456e4377277059c 4289Author: Michel Dänzer <michel.daenzer@amd.com> 4290Date: Thu Mar 24 18:08:22 2016 +0900 4291 4292 Call RADEONBlockHandler_KMS before setting initial modes 4293 4294 Doing it the other way around meant that there was still a possibility 4295 for the front buffer contents to be uninitialized when they start being 4296 scanned out. 4297 4298 (Ported from amdgpu commit 4a60b4b1851a3cbc2d8ad9048d68eeb6947cf132) 4299 4300 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4301 4302commit e592f32f8b5f5873fcc18b10a69dd5e4ccf11073 4303Author: Michel Dänzer <michel.daenzer@amd.com> 4304Date: Wed Mar 23 18:25:48 2016 +0900 4305 4306 Require xserver 1.9 or newer 4307 4308 1.9.0 was released in August 2010. 4309 4310 We were already unintentionally relying on things not available in 1.8 4311 for at least a year, and nobody has complained. 4312 4313 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4314 4315commit bde466e5d44cad64b4e4eceaa5de80fdbf86356e 4316Author: Michel Dänzer <michel.daenzer@amd.com> 4317Date: Wed Mar 23 18:16:42 2016 +0900 4318 4319 Fix build against older versions of xserver 4320 4321 Also slightly clean up the error handling in radeon_scanout_do_update. 4322 4323 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94614 4324 4325 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4326 4327commit b2b079be0e7ef2704c3df9ee25fe46166e450e83 4328Author: Michel Dänzer <michel.daenzer@amd.com> 4329Date: Wed Mar 23 17:29:47 2016 +0900 4330 4331 Use radeon_alloc_pixmap_bo for allocating scanout BOs 4332 4333 This enables tiling for scanout BOs used for TearFree / ShadowPrimary, 4334 and simplifies the code a little. 4335 4336 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4337 4338commit d21ac4669a8b2cdd4eec5e5a94d1950b7423b8b5 4339Author: Michel Dänzer <michel.daenzer@amd.com> 4340Date: Wed Mar 23 17:27:27 2016 +0900 4341 4342 DRI3 only works with acceleration 4343 4344 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94214 4345 4346 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4347 4348commit eb1a2c57e6c45d7ca57b2e2be643194243b7aae1 4349Author: Michel Dänzer <michel.daenzer@amd.com> 4350Date: Wed Mar 23 17:23:31 2016 +0900 4351 4352 TearFree can only work with acceleration 4353 4354 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4355 4356commit d670c5c9851b4eff21c845d26c7d7e4eb5ee0fa9 4357Author: Michel Dänzer <michel.daenzer@amd.com> 4358Date: Fri Mar 18 16:46:08 2016 +0900 4359 4360 Check for xf86CursorResetCursor 4361 4362 If it's available, Xorg calls it on each mode configuration change. It 4363 does what xf86_reload_cursors does (and more), so we don't need to call 4364 the latter anymore. 4365 4366 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4367 4368commit 7f3d0780ca65a90117c2a61362dbc0899bd9c0b0 4369Author: Michel Dänzer <michel.daenzer@amd.com> 4370Date: Wed Mar 9 18:28:47 2016 +0900 4371 4372 Don't try DRI2/Present flipping while the HW cursor can't be used 4373 4374 Flipping doesn't interact correctly with SW cursor: A flip makes the SW 4375 cursor disappear. It will only appear again when the cursor is moved, 4376 but it will be surrounded by corruption, because the SW cursor code 4377 will restore stale screen contents at the old cursor location before 4378 drawing the cursor at the new location. 4379 4380 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4381 4382commit 3de480e83c0a1824838d662d6d67c9fe85277298 4383Author: Michel Dänzer <michel.daenzer@amd.com> 4384Date: Wed Mar 9 16:44:32 2016 +0900 4385 4386 Factor out HW cursor checking code into drmmode_can_use_hw_cursor 4387 4388 And add a check for RandR 1.4 multihead. 4389 4390 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4391 4392commit 2e0b458e747c7df81dd5f9579a762262af3350e1 4393Author: Ryan Kennedy <Ryan.Kennedy@amd.com> 4394Date: Sun Mar 20 17:57:56 2016 -0400 4395 4396 Use correct tiling in drmmode_create_bo_pixmap 4397 4398 The callers were using 0 for the tiling flags, causing surfaces to be 4399 incorrectly treated as linear. This issue was causing corruption on Cayman 4400 just after login with GNOME3. The kernel was rejecting the cs in 4401 drmmode_copy_fb due to invalid texture pitch. 4402 4403 Signed-off-by: Ryan Kennedy <Ryan.Kennedy@amd.com> 4404 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 4405 4406commit a03271de5ecdaa7790d1316e993c4450b91fe936 4407Author: Michel Dänzer <michel.daenzer@amd.com> 4408Date: Wed Mar 9 18:23:10 2016 +0900 4409 4410 present: Return rotated CRTCs from radeon_present_get_crtc 4411 4412 Sync-to-vblank works fine with rotation. We're still checking for 4413 rotation in radeon_present_check_flip. 4414 4415 Returning NULL from here resulted in the xserver present code falling 4416 back to the fake CRTC running at 1 fps. 4417 4418 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4419 4420commit 40191d82370eb7e58bd34c44966cbf44c3703229 4421Author: Michel Dänzer <michel.daenzer@amd.com> 4422Date: Wed Mar 9 18:15:57 2016 +0900 4423 4424 present: Clear drmmode->fb_id before calling set_mode_major for unflip 4425 4426 Without this, drmmode_set_mode_major may just re-set the FB for the 4427 last flipped BO, in which case the display will probably freeze. 4428 4429 Reproduction recipe: Enable rotation while a fullscreen client is 4430 flipping. 4431 4432 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4433 4434commit 798c4fd16d339b1ad5fd729cc884be084c60e38b 4435Author: Michel Dänzer <michel.daenzer@amd.com> 4436Date: Fri Dec 25 18:57:42 2015 +0900 4437 4438 Make Option "TearFree" effective for rotated/reflected outputs as well (v2) 4439 4440 Support varies by xserver version: 4441 4442 < 1.12: No support for the driver handling rotation/reflection 4443 1.12-1.15: Support for driver handling rotation/reflection, but there's 4444 a bug preventing the HW cursor from being visible everywhere 4445 it should be on rotated outputs, so we can only support 4446 TearFree for reflection. 4447 >= 1.16: While the bug above is still there (fixes pending review), 4448 the driver can force SW cursor for rotated outputs, so we 4449 can support TearFree for rotation as well. 4450 4451 v2: Don't set crtc->driverIsPerformingTransform after xf86CrtcRotate if 4452 it wasn't set before. Fixes breaking rotation with TearFree disabled. 4453 4454 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1) 4455 4456commit eb611a2e4ecce7a1ab85fd72b9b78e3269311dd5 4457Author: Michel Dänzer <michel.daenzer@amd.com> 4458Date: Tue Feb 2 18:04:11 2016 +0900 4459 4460 Consolidate pScreen usage in drmmode_set_mode_major 4461 4462 We were already relying on pScrn->pScreen being non-NULL in some cases, 4463 which is supposedly always true ever since this function is no longer 4464 getting called from ScreenInit. 4465 4466 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4467 4468commit 06602171386e538081c298645fb7ca1a70fe80cc 4469Author: Michel Dänzer <michel.daenzer@amd.com> 4470Date: Thu Feb 4 17:13:57 2016 +0900 4471 4472 Remove check for XF86_CRTC_VERSION 3 4473 4474 We require xserver >= 1.8, which was already at version 3. 4475 4476 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4477 4478commit a88985f5d1e39caca49ceb65678aaa9cb622a0d2 4479Author: Michel Dänzer <michel.daenzer@amd.com> 4480Date: Mon Feb 29 18:12:43 2016 +0900 4481 4482 Deal with modesets and page flips crossing on a CRTC 4483 4484 If we set a mode while a flip is pending, the kernel driver may program 4485 the flip to the hardware after the modeset. If that happens, the hardware 4486 will display the BO from the flip, whereas we will assume it displays the 4487 BO from the modeset. In other words, the display will most likely freeze, 4488 at least until another modeset. 4489 4490 Prevent this condition by waiting for a pending flip to finish before 4491 setting a mode. 4492 4493 Fixes display freezing when setting rotation or a transform with 4494 TearFree enabled. 4495 4496 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4497 4498commit f5d968cbba3c9b7ec202161f2157d8d64778c817 4499Author: Michel Dänzer <michel.daenzer@amd.com> 4500Date: Thu May 21 12:54:31 2015 +0900 4501 4502 Make DRM event queue xf86CrtcPtr based instead of ScrnInfoPtr based 4503 4504 This allows for a minor simplification of the code. 4505 4506 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4507 4508commit e87365117acbd80b7d80fbb5eb30890ef7153291 4509Author: Michel Dänzer <michel.daenzer@amd.com> 4510Date: Thu Feb 25 16:49:16 2016 +0900 4511 4512 DRI2: Also clear dri2_flipping when client disconnects before event 4513 4514 Fixes the following problem: 4515 4516 With DRI3 enabled, run glxgears with LIBGL_DRI3_DISABLE=1, make it 4517 fullscreen and press Escape while it's still fullscreen. This could 4518 result in dri2_flipping not getting cleared, spuriously preventing apps 4519 using DRI3 from flipping. 4520 4521 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4522 4523commit d5dbb07db22d5420c81dfebc060f0dd86e7b8a20 4524Author: Michel Dänzer <michel.daenzer@amd.com> 4525Date: Thu Feb 25 18:02:12 2016 +0900 4526 4527 Remove radeon_scanout_flip_handler 4528 4529 No longer necessary now that radeon_drm_queue_handler can handle 4530 e->handler == NULL. 4531 4532 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4533 4534commit 3989766edde85d1abe7024577b98fc9b007bc02a 4535Author: Michel Dänzer <michel.daenzer@amd.com> 4536Date: Thu Feb 25 16:43:48 2016 +0900 4537 4538 drm_queue: Don't abort events immediately from radeon_drm_abort_client 4539 4540 Keep them around until the DRM event arrives, but then call the abort 4541 functions instead of the handler functions. 4542 4543 This is a prerequisite for the following fix. 4544 4545 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4546 4547commit 7bc089ee9740a3ef91f16e36786d8ed64e165acd 4548Author: jimqu <Jim.Qu@amd.com> 4549Date: Tue Feb 16 17:05:48 2016 +0900 4550 4551 glamor: Return NullPixmap on failure to create shareable pixmap 4552 4553 If we were asked to create a shareable pixmap, it doesn't make sense 4554 to return a pixmap which isn't shareable. Doing so caused trouble down 4555 the line such as a crash with older versions of glamor when trying to 4556 use GLX pixmaps of bpp < 32 via DRI2. 4557 4558 Signed-off-by: JimQu <jim.qu@amd.com> 4559 (ported from amdgpu commit 2fcb7dadd3c71cd405cbbaafc777697538ca9c29) 4560 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4561 4562commit 4cc32031467157ab8788f5c684fb1ac67ae96ff9 4563Author: jimqu <Jim.Qu@amd.com> 4564Date: Tue Feb 16 17:03:24 2016 +0900 4565 4566 Move radeon_glamor_destroy_pixmap before radeon_glamor_create_pixmap 4567 4568 The next commit will call the former from the latter. No functional 4569 change. 4570 4571 Signed-off-by: JimQu <jim.qu@amd.com> 4572 (ported from amdgpu commit 5269a2228bff6023c1a7f3e8534027e1d7addc25) 4573 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4574 4575commit a9b181083d56b5de9b6822d0ed89fcc86de777a6 4576Author: Michel Dänzer <michel.daenzer@amd.com> 4577Date: Tue Feb 16 16:59:40 2016 +0900 4578 4579 Remove pci_dev test from radeon_get_scrninfo 4580 4581 The pci_dev parameter can never be NULL since we only support KMS. 4582 4583 Reported-by: Tom St Denis <tom.stdenis@amd.com> 4584 (ported from amdgpu commit 6e42c58375a4c3229da93c27bbd104af145c6163) 4585 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4586 4587commit 2c7f068b0a0493daae86e5eacdf8d9812fcd28dc 4588Author: Mykola Lysenko <Mykola.Lysenko@amd.com> 4589Date: Mon Feb 15 18:21:39 2016 +0900 4590 4591 Initialize drmmode_crtc dpms_mode to DPMSModeOff 4592 4593 This disables query of disabled pipes for drmWaitVBlank on X start 4594 4595 Signed-off-by: Mykola Lysenko <Mykola.Lysenko@amd.com> 4596 (ported from amdgpu commit 4eb9cedca080b30c57ded349a397620ee7d0cd46) 4597 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4598 4599commit 9358e44f03c95396b7f7a4f7398f5b24d2559bb2 4600Author: Michel Dänzer <michel.daenzer@amd.com> 4601Date: Mon Feb 15 18:20:31 2016 +0900 4602 4603 sync: Check if miSyncShmScreenInit symbol is resolved at runtime 4604 4605 It may be disabled in the Xorg build, either explicitly or because the 4606 xshmfence library isn't available. 4607 4608 (ported from amdgpu commit 1d0b0c1794e65e581a48aa9fb19679d928d82a5d) 4609 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4610 4611commit cd94248ffa7d8fe0b57476f79e7e860dee66d1b0 4612Author: Jammy Zhou <Jammy.Zhou@amd.com> 4613Date: Tue Feb 16 16:36:33 2016 +0900 4614 4615 Use render node for DRI3 if available 4616 4617 Signed-off-by: Jammy Zhou <Jammy.Zhou@amd.com> 4618 (ported from amdgpu commit ea558e645786b08d75307716036045170e97b43e) 4619 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4620 4621commit d41fcccf2f11d0393d252e76bbbe6d233c4ac443 4622Author: Michel Dänzer <michel.daenzer@amd.com> 4623Date: Wed Jan 20 11:20:30 2016 +0900 4624 4625 Set RADEON_GEM_NO_CPU_ACCESS flag for BOs which don't need CPU access 4626 4627 Failing to do this was resulting in the kernel driver unnecessarily 4628 leaving open the possibility of CPU access to those BOs. 4629 4630 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4631 4632commit 9483a3d777919b224f70c3b4d01e4b320a57db31 4633Author: Michel Dänzer <michel.daenzer@amd.com> 4634Date: Fri Jan 15 16:48:46 2016 +0900 4635 4636 Fix RandR CRTC transforms 4637 4638 Currently, Xorg will only transform the cursor as of the first time the 4639 cursor image changes after a transform is set. 4640 4641 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=80678 4642 4643 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4644 4645commit b16856b25086ffb27365ac2249b8da921066ce62 4646Author: Michel Dänzer <michel.daenzer@amd.com> 4647Date: Thu Jan 14 17:32:54 2016 +0900 4648 4649 Set the RandR primary output on startup if Xorg hasn't 4650 4651 Fixes xrandr (XRRGetOutputPrimary) not reporting any output as primary 4652 after startup. 4653 4654 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4655 4656commit c7cf00487cd6d4a5d0f39d5b92ff04f6420d6a32 4657Author: Michel Dänzer <michel.daenzer@amd.com> 4658Date: Fri Nov 27 12:52:28 2015 +0900 4659 4660 Build RandR 1.4 provider name from chipset name and bus ID 4661 4662 Instead of just "radeon", it's now e.g. "KAVERI @ pci:0000:00:01.0". 4663 4664 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4665 4666commit ff9a6b6f079a8419f4e6fadfee778060618bf735 4667Author: Michel Dänzer <michel.daenzer@amd.com> 4668Date: Fri Dec 11 17:21:27 2015 +0900 4669 4670 HAS_DIRTYTRACKING_ROTATION also supports multiple CRTCs 4671 4672 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4673 4674commit 78fbca095ae9887a2d3de48bb07975e2d1126e68 4675Author: Michel Dänzer <michel.daenzer@amd.com> 4676Date: Tue Dec 1 15:38:13 2015 +0900 4677 4678 Load fb module before glamoregl/shadow modules 4679 4680 Fixes unresolved symbols on some systems. 4681 4682 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93105 4683 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4684 4685commit b19417e2fddf4df725951aea5ad5e9558338f59e 4686Author: Michel Dänzer <michel.daenzer@amd.com> 4687Date: Fri Nov 20 18:48:56 2015 +0900 4688 4689 Don't advertise any PRIME offloading capabilities without acceleration 4690 4691 Acceleration is required even for display offloading. Trying to enable 4692 display offloading without acceleration resulted in a crash. 4693 4694 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4695 4696commit 10b7c3def58bb34acc38f076bc230e25b454ab79 4697Author: Michel Dänzer <michel.daenzer@amd.com> 4698Date: Tue Nov 10 16:31:09 2015 +0900 4699 4700 glamor: Deal with glamor_egl_destroy_textured_pixmap being removed 4701 4702 When it's not available, it's safe to call down to the glamor 4703 DestroyPixmap hook instead. 4704 4705 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4706 4707commit 535e5438b2c32f774b9c8c27ee0289b4749548ef 4708Author: Michel Dänzer <michel.daenzer@amd.com> 4709Date: Tue Nov 10 17:31:52 2015 +0900 4710 4711 glamor: Restore all ScreenRec hooks during CloseScreen 4712 4713 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4714 4715commit dfb52777e43f3855476cff812d64de687a199674 4716Author: Michel Dänzer <michel.daenzer@amd.com> 4717Date: Thu Nov 12 13:49:04 2015 +0900 4718 4719 Post 7.6.1 release version bump 4720 4721commit 99cb8c3faf1a4ce368b7500f17a2a7868c15e8e8 4722Author: Michel Dänzer <michel.daenzer@amd.com> 4723Date: Thu Nov 12 13:38:14 2015 +0900 4724 4725 Set version for 7.6.1 release 4726 4727commit 4a2114a717ec8d78cacd8fe158e132d62013faac 4728Author: Michel Dänzer <michel.daenzer@amd.com> 4729Date: Thu Nov 12 13:32:25 2015 +0900 4730 4731 Remove duplicate free(output_ids) call 4732 4733 Fixes double-free regression introduced by commit 875ad48e ("Simplify 4734 drmmode_set_mode_major() and avoid leaking memory."). 4735 4736commit 960a9f37096b53ec9603a99c7b1f3679f915e03f 4737Author: Michel Dänzer <michel.daenzer@amd.com> 4738Date: Thu Nov 12 13:20:18 2015 +0900 4739 4740 Post 7.6.0 release version bump 4741 4742commit 9aae4b71fb402841439e3267fcb7f410d8429c3a 4743Author: Michel Dänzer <michel.daenzer@amd.com> 4744Date: Thu Nov 12 11:56:48 2015 +0900 4745 4746 Bump version for 7.6.0 release 4747 4748commit 98291869ac4a542a0b478920586407ff9d2c8ef0 4749Author: Tom St Denis <tom.stdenis@amd.com> 4750Date: Wed Nov 11 16:01:41 2015 +0900 4751 4752 Clean up radeon_dri2_create_buffer2() 4753 4754 Remove the depth_pixmap variable from the function and clear 4755 out any dead/odd behaviour that results. 4756 4757 Signed-off-by: Tom St Denis <tom.stdenis@amd.com> 4758 (ported from amdgpu commit 6000aef4e2f0a121b94023484406fb6f04688f74) 4759 4760 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4761 4762commit c6fc7e309a8a922f94a1f5f3e8bfb9058cff7ad1 4763Author: Michel Dänzer <michel.daenzer@amd.com> 4764Date: Wed Nov 11 15:59:06 2015 +0900 4765 4766 Properly handle drmModeAddFB failure in drmmode_crtc_scanout_allocate 4767 4768 We were printing an error message, but not propagating the failure. That 4769 would probably lead to trouble down the road. 4770 4771 (ported from amdgpu commit 21e72fb2418b5cc7fc849a9cf951186e209036b0) 4772 4773 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4774 4775commit 4e4f4d53da0539ef9feb8766230a6e9927ae005b 4776Author: Michel Dänzer <michel.daenzer@amd.com> 4777Date: Wed Nov 11 15:57:21 2015 +0900 4778 4779 Eliminate redundant data parameter from drmmode_crtc_scanout_create 4780 4781 drmmode_crtc_scanout_create just needs to call 4782 drmmode_crtc_scanout_allocate when scanout->bo is NULL. 4783 4784 This makes it clearer to the reader / compiler that 4785 drmmode_crtc_scanout_create doesn't dereference scanout->bo when it's 4786 NULL. 4787 4788 (ported from amdgpu commit 8da1d0c870e1081d77925807d6e3bbc61a23f54f) 4789 4790 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4791 4792commit 421a7e797bdd58d83e81af7a6512cc715a3df514 4793Author: Michel Dänzer <michel.daenzer@amd.com> 4794Date: Wed Nov 11 15:51:58 2015 +0900 4795 4796 Don't advertise rotation support without hardware acceleration v2 4797 4798 Rotation currently doesn't work without acceleration (doesn't actually 4799 rotate with Option "NoAccel", crashes with Option "AccelMethod" "none" 4800 or when glamor fails to initialize) and would probably be too slow 4801 anyway. 4802 4803 v2: Also remove now dead code checking for ShadowFB from 4804 drmmode_crtc_scanout_allocate(). 4805 4806 (ported from amdgpu commit dc40582d5ff94d812cbc08f95cf14b80cd0f410d) 4807 4808 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4809 4810commit 875ad48e7b5cdb7beefbf18dddcbee3ed22b5446 4811Author: Tom St Denis <tom.stdenis@amd.com> 4812Date: Wed Nov 11 15:51:19 2015 +0900 4813 4814 Simplify drmmode_set_mode_major() and avoid leaking memory. 4815 4816 The function would leak the memory allocated for output_ids. This 4817 patch addresses that as well as simplifies the logic somewhat. 4818 4819 Signed-off-by: Tom St Denis <tom.stdenis@amd.com> 4820 (ported from amdgpu commit 460560502a1bdf26d06f3c30df46fa9f28ffb9e5) 4821 4822 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4823 4824commit 789d7d6a04cca6b36fb088a074027807ccb8dd61 4825Author: Tom St Denis <tom.stdenis@amd.com> 4826Date: Wed Nov 11 15:48:51 2015 +0900 4827 4828 Clean up allocation in RADEONInitVideo() 4829 4830 The allocation of the adapters should use the correct sizeof (even if 4831 allocating an array of pointers). 4832 4833 Signed-off-by: Tom St Denis <tom.stdenis@amd.com> 4834 (ported from amdgpu commit db3bb2061b9ac16b0922d9afae99874820356a04) 4835 4836 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4837 4838commit d88fa0dd5d37604de8efb05853738cfaca6a3166 4839Author: Tom St Denis <tom.stdenis@amd.com> 4840Date: Wed Nov 11 15:46:50 2015 +0900 4841 4842 Simplify pick best crtc to fold two loops into one 4843 4844 This patch folds the two for loops from radeon_pick_best_crtc() into 4845 one to reduce the LOC and make the routine easier to read. 4846 4847 Signed-off-by: Tom St Denis <tom.stdenis@amd.com> 4848 (ported from amdgpu commit 3055724aef76a624718f26d5f0f9e9d567ffbcfb) 4849 4850 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4851 4852commit dbbcd75b3c80aba77673904d46bca97779fd8a8d 4853Author: Tom St Denis <tom.stdenis@amd.com> 4854Date: Wed Nov 11 12:54:54 2015 +0900 4855 4856 dri2: Avoid calculation with undefined msc value 4857 4858 If the get_msc() call fails for any reason we should avoid updating the 4859 vblank counter delta with undefined data. 4860 4861 Signed-off-by: Tom St Denis <tom.stdenis@amd.com> 4862 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> (minor fixups) 4863 (ported from amdgpu commit 8823c3d4c6db70cff7699b31088f2d92db8faaf4) 4864 4865 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4866 4867commit 7186a8713ba004de4991f21c1a9fc4abc62aeff4 4868Author: Stephen Chandler Paul <cpaul@redhat.com> 4869Date: Fri Oct 23 09:59:36 2015 -0400 4870 4871 Handle failures in setting a CRTC to a DRM mode properly 4872 4873 This fixes a bug where running the card out of PPLL's when hotplugging 4874 another monitor would result in all of the displays going blank and 4875 failing to work properly until X was restarted or the user switched to 4876 another VT. 4877 4878 [Michel Dänzer: Pass errno instead of -ret to strerror()] 4879 4880 Signed-off-by: Stephen Chandler Paul <cpaul@redhat.com> 4881 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 4882 4883commit 548e97b3b7d1e94075a54ca2bb4eb683025098a7 4884Author: Michel Dänzer <michel.daenzer@amd.com> 4885Date: Fri Oct 16 16:26:58 2015 +0900 4886 4887 Call xf86CrtcRotate from initial drmmode_set_desired_modes call 4888 4889 Fixes various problems when rotation is specified in xorg.conf. 4890 4891 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92475 4892 4893 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4894 4895commit ce9914af8d8d5243977023ec7b09c605f9cca8b4 4896Author: Michel Dänzer <michel.daenzer@amd.com> 4897Date: Fri Oct 2 18:07:18 2015 +0900 4898 4899 Only align screen / scanout pixmap height where necessary 4900 4901 When using glamor acceleration, the pixmap's header has to have a height 4902 that matches exactly what the actual height is minus the GPU memory 4903 alignment. Otherwise CRTCs scanning out from the main scanout buffer 4904 (e.g. every CRTC that isn't rotated or transformed in some way) won't 4905 always work. This results in a bug where rotating one monitor in a 4906 multi-monitor setup won't always work properly. Easiest way to reproduce 4907 this: 4908 4909 - Have two monitors (I've gotten this working with a 1920x1080 and 4910 1280x1024, along with two 1920x1080s) 4911 - Rotate one of them from 0° to 90°, then rotate the same monitor from 4912 90° to 180°. The monitor that hasn't been rotated won't properly 4913 update, and will stay on a blank screen 4914 4915 This doesn't seem to make any difference when using EXA for 4916 acceleration. 4917 4918 Compared to Stephen Chandler's patch, this drops the height alignment 4919 in most places and only keeps it where it's really necessary. 4920 4921 Reported-and-Tested-by: Stephen Chandler Paul <cpaul@redhat.com> 4922 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4923 4924commit 0288a4b87b65ba54f37fbeeea3cb32238deee92e 4925Author: Michel Dänzer <michel.daenzer@amd.com> 4926Date: Wed Aug 19 16:54:14 2015 +0900 4927 4928 DRI2: Keep MSC monotonic when moving window between CRTCs 4929 4930 This mirrors the DRI3 implementation in xserver. Fixes VDPAU video 4931 playback hanging when moving the window between CRTCs. 4932 4933 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=66384 4934 4935 (Ported from amdgpu commit 63948ea091a9b324327ade7ec4fc5d67ca7e6f6f) 4936 4937 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4938 4939commit 1d886b526dc49f32dc6744b7a882894bdac4e846 4940Author: Michel Dänzer <michel.daenzer@amd.com> 4941Date: Wed Aug 19 18:11:45 2015 +0900 4942 4943 DRI2: Don't ignore rotated CRTCs in amdgpu_dri2_drawable_crtc 4944 4945 Waiting for vblank interrupts works fine with rotated CRTCs. The only 4946 case we can't handle with rotation is page flipping, which is handled 4947 in can_exchange(). 4948 4949 This fixes gnome-shell hanging on rotation, probably because 4950 amdgpu_dri2_get_msc returned MSC/UST 0 for rotated CRTCs. 4951 4952 (Ported from amdgpu commit 7b3212e33cd36fb6f122774df27b56ec4e1a22b8) 4953 4954 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4955 4956commit 270da55340766074cabff8af4258e29fe2f0fc81 4957Author: Alex Deucher <alexander.deucher@amd.com> 4958Date: Mon Aug 10 15:32:34 2015 -0400 4959 4960 add new OLAND pci id 4961 4962 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 4963 4964commit f123096877f0a0044f3d0315285441f2b5790c29 4965Author: Mario Kleiner <mario.kleiner.de@gmail.com> 4966Date: Mon Aug 10 23:34:39 2015 +0200 4967 4968 Make selection between DRI2 and DRI3 consistent with other drivers. (v2) 4969 4970 Add Option "DRI" to allow selection of maximum DRI level. 4971 4972 This allows the user to select the maximum level of DRI 4973 implementation to use, DRI2 or DRI3. It replaces the old 4974 option "DRI3" which had exactly the same purpose, but 4975 differs from the method used in both intel ddx and nouveau ddx. 4976 Make this consistent before a new stable driver is released. 4977 4978 v2: Retain handling of old Option "DRI3" for backwards 4979 compatibility, but Option "DRI" will take precedence 4980 over "DRI3" if both are provided. 4981 4982 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 4983 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 4984 4985commit 3791fceabf2cb037467dc41c15364e9f9ec1e47e 4986Author: Michel Dänzer <michel.daenzer@amd.com> 4987Date: Thu Aug 6 16:27:01 2015 +0900 4988 4989 Wait for scanout BO initialization to finish before setting mode 4990 4991 This should avoid intermittent artifacts which could sometimes be visible 4992 when setting a new scanout pixmap, e.g. on server startup or when 4993 changing resolutions. 4994 4995 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 4996 4997commit 4e3dfa69e4630df2e0ec0f5b81d61159757c4664 4998Author: Michel Dänzer <michel.daenzer@amd.com> 4999Date: Thu Aug 6 16:16:38 2015 +0900 5000 5001 Only call drmmode_copy_fb (at most) once on server startup 5002 5003 It doesn't make sense to copy the screen contents from console when VT 5004 switching back to Xorg or when Xorg resets. 5005 5006 Fixes intermittent artifacts when VT switching back from console to the 5007 gdm login screen. 5008 5009 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5010 5011commit 09c7cdb923965f9a1ea11d2f449bc02114408938 5012Author: Michel Dänzer <michel.daenzer@amd.com> 5013Date: Thu Jul 30 15:58:54 2015 +0900 5014 5015 glamor: Move declaration of struct radeon_pixmap out of #if/#else blocks 5016 5017 Reviewed-by: Dave Airlie <airlied@redhat.com> 5018 5019commit 936582fde0db461c5c3d78ce4f5f4c93a88a489d 5020Author: Michel Dänzer <michel.daenzer@amd.com> 5021Date: Thu Jul 30 15:56:36 2015 +0900 5022 5023 Remove unused local variable pRADEONEnt 5024 5025 ../../src/radeon_probe.c: In function 'radeon_get_scrninfo': 5026 ../../src/radeon_probe.c:157:22: warning: variable 'pRADEONEnt' set but not used [-Wunused-but-set-variable] 5027 RADEONEntPtr pRADEONEnt; 5028 ^ 5029 5030 Reviewed-by: Dave Airlie <airlied@redhat.com> 5031 5032commit 5510cd6027d2387efdf33575e3bfc424cb11bfd8 5033Author: Dave Airlie <airlied@redhat.com> 5034Date: Mon Jul 27 09:22:57 2015 +1000 5035 5036 radeon: move radeon_pixmap forward declaration into other block 5037 5038 There is already a radeon_pixmap forward decl here, the #else 5039 block is missing one. 5040 5041 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5042 Signed-off-by: Dave Airlie <airlied@redhat.com> 5043 5044commit b32a0a3de84a44b9af4f1ca8be19f10d7fa31b12 5045Author: Dave Airlie <airlied@redhat.com> 5046Date: Wed Jul 22 13:35:07 2015 +1000 5047 5048 radeon: cleanup the entity rec 5049 5050 Some of these were set, some of them were 5051 always opposites, so clean things up. 5052 5053 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5054 Signed-off-by: Dave Airlie <airlied@redhat.com> 5055 5056commit fcb32231a38f9461d12720cbf72f63502197a711 5057Author: Emil Velikov <emil.l.velikov@gmail.com> 5058Date: Tue Jul 21 23:55:20 2015 +0100 5059 5060 Do not link radeon_drv.so against libpciaccess 5061 5062 It's only used/needed by ati_drv.so 5063 5064 Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com> 5065 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 5066 5067commit b6d871bf299c7d0f106c07ee4d8bd3b2337f53cc 5068Author: Dave Airlie <airlied@gmail.com> 5069Date: Thu Jul 9 19:25:36 2015 +1000 5070 5071 radeon: adopt for new X server dirty tracking APIs. 5072 5073 Signed-off-by: Dave Airlie <airlied@redhat.com> 5074 5075commit 95f5d09e3667ded027ae648c97eb4737d8bf67c5 5076Author: Michel Dänzer <michel.daenzer@amd.com> 5077Date: Thu May 28 11:24:42 2015 +0900 5078 5079 present: Handle DPMS off in radeon_present_get_ust_msc 5080 5081 The DRM_IOCTL_WAIT_VBLANK ioctl may return an error during DPMS off, 5082 which would trigger an error message in drmmode_crtc_get_ust_msc. 5083 5084 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5085 5086commit 211862b777d0be251a4662f5dd24f2d400544c09 5087Author: Michel Dänzer <michel.daenzer@amd.com> 5088Date: Fri May 1 18:20:01 2015 +0900 5089 5090 present: Look at all CRTCs to determine if we can flip 5091 5092 Inspired by modesetting driver change by Kenneth Graunke. 5093 5094 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5095 5096commit 802d33e474a82262d9cdf11b03568b0c4929cd0d 5097Author: Michel Dänzer <michel.daenzer@amd.com> 5098Date: Fri May 1 18:54:16 2015 +0900 5099 5100 present: Fall back to modeset for unflip operation 5101 5102 It's not always possible to use the page flip ioctl for this, e.g. 5103 during DPMS off. We were previously just skipping the unflip in that 5104 case, which could result in hangs when setting DPMS off while a 5105 fullscreen Present app is running, e.g. at the GNOME3 lock screen. 5106 5107 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5108 5109commit a8ed62010d5012dfb27773595c446b217f3c00c5 5110Author: Piotr Redlewski <predlewski@gmail.com> 5111Date: Mon Jul 6 19:57:07 2015 +0200 5112 5113 Do not try to enable already enabled CRTCs in DPMS hook 5114 5115 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91227 5116 5117 Signed-off-by: Piotr Redlewski <predlewski@gmail.com> 5118 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 5119 5120commit fc07c370dac4649511eb7322d05a61605a539f39 5121Author: Mario Kleiner <mario.kleiner.de@gmail.com> 5122Date: Sun Jun 28 02:45:58 2015 +0200 5123 5124 Fix inconsistent default eg_tile_split in evergreen_accel.c 5125 5126 eg_tile_split() maps the default: switch-case to tile split 5127 flag 6, like for a tile split size of 4096 Bytes. All other 5128 instances of tile split mappings or reverse mappings in both 5129 the ddx and in Mesa's radeon gallium drivers assign the default: 5130 case to tile split flag 4, consistent with a default of 1024 5131 Bytes. 5132 5133 Make this one outlier consistent with other instances of 5134 mappings in the code to avoid surprises. 5135 5136 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 5137 5138commit ddaba449e8d6fe9fc0d97085e4045843fd8d7af9 5139Author: Piotr Redlewski <predlewski@gmail.com> 5140Date: Wed Jun 24 18:44:39 2015 +0200 5141 5142 Disable CRTCs when disabling the outputs 5143 5144 When turning the outputs off (DPMSModeOff), CRTCs stayed enabled. This led 5145 to higher gpu temperatures than with fglrx driver. 5146 5147 v2: when entering DPMS also disable active CRTCs 5148 v3: use drmmode_set_mode_major() for enabling CRTCs when leaving DPMS 5149 5150 Signed-off-by: Piotr Redlewski <predlewski@gmail.com> 5151 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 5152 5153commit acc11877423ecd81a6e0a7f38466f80e43efee20 5154Author: Michel Dänzer <michel.daenzer@amd.com> 5155Date: Tue Jun 16 17:34:57 2015 +0900 5156 5157 Skip disabled CRTCs in radeon_scanout_(do_)update 5158 5159 The vblank / page flip ioctls don't work as expected for a disabled CRTC. 5160 5161 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5162 5163commit fc9fadaebbc8aead6e030d93a9ccd84561f8f59e 5164Author: Mario Kleiner <mario.kleiner.de@gmail.com> 5165Date: Mon Jun 22 00:09:17 2015 +0200 5166 5167 Don't set TILE_SPLIT flags if surface.tile_split == 0. 5168 5169 On pre-Evergreen hw, libdrm's r6_surface_best() helper 5170 for the surface managers radeon_surface_best() routine 5171 is a no-op and therefore doesn't assign any tile_split 5172 settings to created surfaces, so it leaves 5173 surface.tile_split on its "undefined" value of 0. 5174 5175 Mesa's DRI3/Present backend creates DRI3 Pixmaps via 5176 the DRIImage extension and the radeon gallium driver 5177 implementation of that extension uses the libdrm 5178 surface manager for backing bo creation and treats 5179 an undefined surface.tile_split==0, as returned by 5180 the surface manager for pre-evergreen, as a signal 5181 to not assign any tile_split flags to the DRI3 Pixmaps 5182 bo. 5183 5184 The ddx also uses libdrm surface manager to create the 5185 x-screen pixmap, but so far treated the returned undefined 5186 surface.tile_split==0 by mapping it to eg_tile_split()'s 5187 default tile_split flags, which are different from Mesa's 5188 tiling flags for DRI3 pixmaps. Under DRI3/Present this 5189 causes a mismatch of src pixmap and destination root 5190 pixmaps tiling flags and thereby prevents page flipping 5191 for pixmap presents. 5192 5193 Change the ddx code to treat surface.tile_split==0 the 5194 same way as the radeon gallium driver to avoid mismatched 5195 tiling flags and thereby allow DRI3/Present page-flip to 5196 work on pre-Evergreen hw. 5197 5198 Tested on RV730 and Evergreen "Juniper". 5199 5200 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 5201 5202commit 49f5b0bc301414df049e00d226034e3d6e56421b 5203Author: Michel Dänzer <michel.daenzer@amd.com> 5204Date: Fri May 1 19:02:36 2015 +0900 5205 5206 Don't attempt a DRI2/Present page flip while the other one is flipping 5207 5208 Fixes corrupted display and hangs when switching between DRI2 and DRI3 5209 fullscreen apps, e.g. a compositor using DRI3 and a fullscreen app using 5210 DRI2 or vice versa. 5211 5212 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5213 5214commit afab7839fc15722dbaa7203d00fe7f6ce5336b9d 5215Author: Mario Kleiner <mario.kleiner.de@gmail.com> 5216Date: Wed Jun 10 20:55:08 2015 +0200 5217 5218 Allow/Fix use of multiple ZaphodHead outputs per x-screen. (v2) 5219 5220 Defining multiple ZaphodHead outputs per x-screen in a 5221 multiple x-screen's per gpu configuration caused all 5222 outputs except one per x-screen to go dark, because 5223 there was a fixed mapping x-screen number -> crtc number, 5224 limiting the number of crtc's per x-screen to one. 5225 5226 On a ZaphodHead's setup, be more clever and assign 5227 as many crtc's to a given x-screen as there are 5228 ZaphodHeads defined for that screen, assuming 5229 there are enough unused crtc's available. 5230 5231 Tested on a triple display setup with different combos 5232 of one, two or three ZaphodHeads per one, two or three 5233 x-screens. 5234 5235 This is a port of similar code from xf86-video-nouveau. 5236 5237 v2: Implement suggestions by Michel Dänzer: Less verbose 5238 debug output, more clear warning message on crtc allocation 5239 failure. Move clearing of per gpu assigned_crtc mask to 5240 CloseScreen, indeed testing shows no need for the more 5241 complex new server generation check from v1. 5242 5243 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 5244 (v1) Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5245 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 5246 5247commit d7c82731a8bf3d381bc571b94d80d9bb2dd6e40d 5248Author: Michel Dänzer <michel.daenzer@amd.com> 5249Date: Thu May 21 16:01:16 2015 +0900 5250 5251 DRI2: Fix handling of drmmode_crtc_get_ust_msc return code 5252 5253 Fixes regression introduced in commit 5254 76c2923ac5c7230a8b2f9f8329c308d28b44d9c0 ("DRI2: Split out helper for 5255 getting UST and MSC of a specific CRTC"). 5256 5257 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5258 5259commit e58fc380ccf2a581d28f041fd74b963626ca5404 5260Author: Michel Dänzer <michel.daenzer@amd.com> 5261Date: Thu May 28 10:59:22 2015 +0900 5262 5263 glamor: Only wait for GPU writes in radeon_glamor_prepare_access_cpu_ro 5264 5265 We don't need to wait for GPU reads to finish before reading with the 5266 CPU. 5267 5268 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5269 5270commit 5f82a720374c9c1caebb42bfbeea1f0cf8847d28 5271Author: Michel Dänzer <michel.daenzer@amd.com> 5272Date: Thu May 21 15:46:32 2015 +0900 5273 5274 present: Remove get_drmmode_crtc helper 5275 5276 It was getting the drmmode_crtc_private_ptr in a roundabout way. 5277 5278 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5279 5280commit d64a13ebe0ecd241ee3260dbffd8f4a01e254183 5281Author: Michel Dänzer <michel.daenzer@amd.com> 5282Date: Wed May 20 09:58:38 2015 +0900 5283 5284 Replace SyncFence typedef with declaration of struct _SyncFence 5285 5286 Fixes build failure in some environments: 5287 5288 CC radeon_sync.lo 5289 In file included from radeon_sync.c:28: 5290 /home/tinderbox/xorg-build/include/xorg/misync.h:31: error: redefinition of typedef 'SyncFence' 5291 radeon.h:93: note: previous declaration of 'SyncFence' was here 5292 make[2]: *** [radeon_sync.lo] Error 1 5293 5294 Reported-Tested-and-Acked-by: David Airlie <airlied@redhat.com> 5295 5296commit 818c180c8932233b214a35ba0647af82f7bcec3d 5297Author: Michel Dänzer <michel.daenzer@amd.com> 5298Date: Tue May 12 17:43:57 2015 +0900 5299 5300 glamor: Deal with glamor_glyphs_init being removed from xserver 5301 5302 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5303 5304commit 7c4b78ab10b82c6dba9f72034ff7583859cca63d 5305Author: Alex Deucher <alexander.deucher@amd.com> 5306Date: Tue May 12 13:21:24 2015 -0400 5307 5308 add new bonaire pci id 5309 5310 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 5311 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 5312 5313commit 80eede245d1eda27eaba108b0761a24bfd69aff6 5314Author: Mario Kleiner <mario.kleiner.de@gmail.com> 5315Date: Sat May 9 07:24:36 2015 +0200 5316 5317 present: Fix present notify timestamps and counts. 5318 5319 Invalid (msc,ust)=(0,0) were returned to the server 5320 because a wrong crtc_id was used to specify which 5321 crtc should deliver pageflip completion data. Fix 5322 it in accordance with the dri2 implementation. 5323 5324 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 5325 5326commit e3be8b0a8cf484ff16597413a6172788178e80c8 5327Author: Mario Kleiner <mario.kleiner.de@gmail.com> 5328Date: Sat May 9 07:24:35 2015 +0200 5329 5330 present: Move check for async flips 5331 5332 Check for and reject currently unsupported async_flip 5333 inside radeon_present_check_flip() instead of inside 5334 radeon_present_flip(). 5335 5336 This way the server can detect early that async flips 5337 aren't supported by the ddx and can deal with this 5338 correctly by using its non-vsync'ed CopyRegion fallback. 5339 5340 Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> 5341 5342commit 1584dc545c78e0bce8d4b4b9f26b568e2c211453 5343Author: Michel Dänzer <michel.daenzer@amd.com> 5344Date: Tue Apr 28 17:14:11 2015 +0900 5345 5346 Skip disabled CRTCs in drmmode_set_desired_modes() even if set_hw == FALSE 5347 5348 Not skipping a disabled CRTC results in a crash. 5349 5350 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90187 5351 5352 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5353 5354commit 340baf3804b7d9de8082badfc715a9bee7856e8f 5355Author: Michel Dänzer <michel.daenzer@amd.com> 5356Date: Mon Apr 27 15:07:06 2015 +0900 5357 5358 Don't call radeon_get_pixmap_private() when using EXA 5359 5360 It only works when using glamor. 5361 5362 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90169 5363 5364 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5365 5366commit 80f3d727f93cb6efedd2b39338d2301035965fe2 5367Author: Michel Dänzer <michel.daenzer@amd.com> 5368Date: Wed Apr 22 13:33:15 2015 +0900 5369 5370 On screen resize, clear the new buffer before displaying it 5371 5372 Fixes garbage being intermittently visible during a screen resize. 5373 5374 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=27757#c7 5375 5376 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5377 5378commit 3999bf88cdb192fe2f30b03bd2ed6f6a3f9f9057 5379Author: Michel Dänzer <michel.daenzer@amd.com> 5380Date: Thu Apr 2 18:29:38 2015 +0900 5381 5382 Make drmmode_copy_fb() work with glamor as well 5383 5384 Needed for Xorg -background none. 5385 5386 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5387 5388commit a4a8cdbcc10c1c5f07485a2af9e9e81e490c3e1d 5389Author: Michel Dänzer <michel.daenzer@amd.com> 5390Date: Thu Apr 2 17:46:34 2015 +0900 5391 5392 Update scanout pixmap contents before setting a mode with it 5393 5394 This ensures the scanout pixmaps used for Option "TearFree" and Option 5395 "ShadowPrimary" have been initialized when their initial mode is set. 5396 5397 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5398 5399commit 673e1c7637687c74fc9bdeeeffb7ace0d04b734f 5400Author: Michel Dänzer <michel.daenzer@amd.com> 5401Date: Thu Apr 2 17:54:33 2015 +0900 5402 5403 Defer initial modeset until the first BlockHandler invocation 5404 5405 This ensures that the screen pixmap contents have been initialized when 5406 the initial modes are set. 5407 5408 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=27757 5409 5410 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5411 5412commit 37874a4eeace5df04b02c8fc28f67b824e3f0f5f 5413Author: Michel Dänzer <michel.daenzer@amd.com> 5414Date: Tue Apr 21 17:19:15 2015 +0900 5415 5416 Defer initial drmmode_copy_fb call until root window creation 5417 5418 That's late enough for acceleration to be fully initialized, but still 5419 early enough to set pScreen->canDoBGNoneRoot. 5420 5421 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5422 5423commit 39c497f3efca5ca08343b884f44c93215dcdef31 5424Author: Michel Dänzer <michel.daenzer@amd.com> 5425Date: Thu Apr 2 18:10:42 2015 +0900 5426 5427 Only copy fbcon BO contents if bgNoneRoot is TRUE 5428 5429 Otherwise, the X server will initialize the screen pixmap contents 5430 anyway. 5431 5432 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5433 5434commit 1af044d7eee211fd4b248c236280274a68334da5 5435Author: Michel Dänzer <michel.daenzer@amd.com> 5436Date: Wed Apr 22 18:43:02 2015 +0900 5437 5438 DRI2: Clear old->devPrivate.ptr in fixup_glamor 5439 5440 It doesn't point to the memory of the newly allocated BO. Fixes crash 5441 running piglit with Option "ShadowPrimary" enabled. 5442 5443 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5444 5445commit 43159ef400c3b18b9f4d3e6fa1c4aef2d60d38fe 5446Author: Michel Dänzer <michel.daenzer@amd.com> 5447Date: Mon Apr 20 18:44:36 2015 +0900 5448 5449 Add Option "TearFree" v4 5450 5451 Avoids tearing by flipping between two scanout BOs per (non-rotated) CRTC 5452 5453 v2: 5454 * Fix condition for TearFree log message (Richard Wilbur) 5455 * Log warning message about DRI page flipping being enabled because of 5456 TearFree (or ShadowPrimary) also when building without glamor support 5457 5458 v3: 5459 * Only override fb_id/x/y if all scanout pixmaps have been successfully 5460 allocated 5461 5462 v4: 5463 * Make log warning clearer if drmModePageFlip returns an error 5464 5465 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1) 5466 5467commit ed401f5b4f07375db17ff05e294907ec95fc946d 5468Author: Michel Dänzer <michel.daenzer@amd.com> 5469Date: Fri Apr 3 12:11:00 2015 +0900 5470 5471 glamor: Remove the stride member of struct radeon_pixmap 5472 5473 Its value was always the same as that of the PixmapRec devKind member. 5474 5475 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5476 5477commit ae92d1765fa370a8d94c2856ad6c45d273ec3c69 5478Author: Michel Dänzer <michel.daenzer@amd.com> 5479Date: Wed Mar 18 16:23:24 2015 +0900 5480 5481 glamor: Add Option "ShadowPrimary" v2 5482 5483 When this option is enabled, most pixmaps (including the screen pixmap) 5484 are allocated in system RAM and mostly accessed by the CPU. Changed areas 5485 of the screen pixmap are copied to dedicated per-CRTC scanout pixmaps 5486 regularly, triggered by the vblank interrupt. 5487 5488 v2: 5489 * Set region data pointer to NULL for keeping only the extents 5490 * Move pRegion and pBox local variable declarations closer to their uses 5491 in drmmode_set_mode_major() 5492 5493commit eea79472a84672ee4dc7adc4487cec6a4037048a 5494Author: Michel Dänzer <michel.daenzer@amd.com> 5495Date: Wed Apr 1 15:51:52 2015 +0900 5496 5497 glamor: Add wrappers for the X server rendering hooks 5498 5499 They can choose between using the GPU or CPU for the operation. 5500 5501commit 051d46382656ffc3e6cac1aab3aee7efdf5b623a 5502Author: Michel Dänzer <michel.daenzer@amd.com> 5503Date: Thu Mar 19 17:34:27 2015 +0900 5504 5505 glamor: Add radeon_pixmap parameter to radeon_glamor_create_textured_pixmap 5506 5507commit 2fa021f77372ca93375a3d13a0c43a9089674899 5508Author: Michel Dänzer <michel.daenzer@amd.com> 5509Date: Thu Mar 19 17:38:47 2015 +0900 5510 5511 glamor: Remove unused function radeon_glamor_pixmap_is_offscreen 5512 5513commit e96349ba6281fd18b8bf9c76629128276b065e6c 5514Author: Michel Dänzer <michel.daenzer@amd.com> 5515Date: Wed Apr 1 16:04:13 2015 +0900 5516 5517 Add RADEON_CREATE_PIXMAP_SCANOUT flag 5518 5519 It means that the pixmap is used for scanout exclusively. 5520 5521commit 9be7dd382e86d2b804de81d4e2af7431b2e16843 5522Author: Michel Dänzer <michel.daenzer@amd.com> 5523Date: Wed Apr 1 12:40:16 2015 +0900 5524 5525 Split out struct drmmode_scanout for rotation shadow buffer information 5526 5527 Will be used for other kinds of dedicated scanout buffers as well. 5528 5529commit c32b0530302739f6512755bccf281c2300617376 5530Author: Michel Dänzer <michel.daenzer@amd.com> 5531Date: Thu Mar 19 17:46:48 2015 +0900 5532 5533 Rename scanout_pixmap_x field to prime_pixmap_x 5534 5535 To avoid confusion with upcoming changes. 5536 5537commit 5921ba4ca705a0d919515626088f3948cc4848c1 5538Author: Michel Dänzer <michel.daenzer@amd.com> 5539Date: Tue Mar 31 15:14:52 2015 +0900 5540 5541 present: Don't flip between BOs with different tiling parameters 5542 5543 The kernel driver doesn't handle that correctly yet. 5544 5545 Fixes or at least avoids issues with OpenGL fullscreen apps with DRI3 5546 enabled and using PRIME or with (2D) tiling disabled. 5547 5548 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89720 5549 5550 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5551 5552commit 428e416e7cb04a1e0527da39cfebf70218879a77 5553Author: Michel Dänzer <michel.daenzer@amd.com> 5554Date: Thu Apr 2 10:34:03 2015 +0900 5555 5556 Add radeon_get_pixmap_tiling_flags helper 5557 5558 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5559 5560commit 98fb4199e63fedd4607cddee64bf602d6398df81 5561Author: Michel Dänzer <michel.daenzer@amd.com> 5562Date: Tue Mar 31 12:25:18 2015 +0900 5563 5564 Only enable SYNC extension fences and the Present extension along with DRI3 5565 5566 This avoids some trouble with the Gallium nine state tracker, which uses 5567 the Present extension even when DRI3 is disabled. 5568 5569 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89772 5570 5571 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5572 5573commit f8b0f23e9f4af9f9097ee5e72d53b45173163c41 5574Author: Michel Dänzer <michel.daenzer@amd.com> 5575Date: Fri Mar 27 12:34:55 2015 +0900 5576 5577 DRI2: Use radeon_get_pixmap_handle 5578 5579 Now we can share pixmaps with no struct radeon_bo via DRI2. 5580 5581 Fixes VDPAU video playback freezing when using an OpenGL compositor with 5582 DRI3 enabled. 5583 5584 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89755 5585 5586 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5587 Tested-by: Nick Sarnie <commendsarnex@gmail.com> 5588 5589commit ccbda955ebae1d457d35293833f12791e0f9fb0b 5590Author: Michel Dänzer <michel.daenzer@amd.com> 5591Date: Fri Mar 27 12:16:44 2015 +0900 5592 5593 Move get_pixmap_handle helper to radeon_bo_helper.c 5594 5595 No functional change. 5596 5597 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5598 Tested-by: Nick Sarnie <commendsarnex@gmail.com> 5599 5600commit de5ddd09db82141b263338dcf0c28e01f58268ee 5601Author: Michel Dänzer <michel.daenzer@amd.com> 5602Date: Thu Mar 26 16:33:02 2015 +0900 5603 5604 Move radeon_drm_handler/abort_proc fields to drmmode_flipdata_rec 5605 5606 Their values are the same for all DRM flip ioctl calls within a single 5607 radeon_do_pageflip() call. 5608 5609 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5610 5611commit e8c0f6319fbf4c3ea11e22ab1a68837031bdec8c 5612Author: Michel Dänzer <michel.daenzer@amd.com> 5613Date: Thu Mar 26 16:27:35 2015 +0900 5614 5615 Simplify radeon_do_pageflip() error handling slightly more 5616 5617 We don't need the local variable old_fb_id. 5618 5619 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5620 5621commit 8fc22360d5520469c82092ccb0fcf2af330c573f 5622Author: Michel Dänzer <michel.daenzer@amd.com> 5623Date: Thu Mar 26 15:58:01 2015 +0900 5624 5625 Increase robustness against DRM page flip ioctl failures v3 5626 5627 Centralize cleanup, only clean up things that have been allocated for 5628 the failed ioctl call. 5629 5630 Fixes double-free after a flip ioctl failure. 5631 5632 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89681 5633 5634 v2: Only call drmModeRmFB for flipdata->old_fb_id on receipt of last DRM 5635 page flip event. Fixes Black screen on making glxgears fullscreen with 5636 DRI3 enabled. 5637 v3: Avoid double-free of flipdata in the unlikely case that calloc fails 5638 for flipcarrier, but only for the second or later CRTC. 5639 5640 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v2) 5641 5642commit b16609b453bb1a181198cf27778f205dc23fb642 5643Author: Michel Dänzer <michel.daenzer@amd.com> 5644Date: Thu Mar 26 17:15:21 2015 +0900 5645 5646 glamor: Handle GLAMOR_* flags removed from xserver 5647 5648 The behaviour is the same as when the removed flags were passed in. 5649 5650 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5651 5652commit 391900a670addec39515f924265bfa9f8bfa9ec0 5653Author: Michel Dänzer <michel.daenzer@amd.com> 5654Date: Wed Mar 18 12:56:07 2015 +0900 5655 5656 glamor: Avoid generating GEM flink names for BOs shared via DRI3 v2 5657 5658 We can't create our own struct radeon_bo representation in this case 5659 because destroying that would make the GEM handle inaccessible to glamor 5660 as well. So just get the handle directly via dma-buf. 5661 5662 v2: Close dma-buf file descriptor, pointed out by Axel Davy. 5663 5664 Reviewed-by: Axel Davy <axel.davy@ens.fr> 5665 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5666 5667commit dfad91fffb5bd013785223b42d78886df839eacf 5668Author: Michel Dänzer <michel.daenzer@amd.com> 5669Date: Thu Mar 19 12:37:01 2015 +0900 5670 5671 Present: Add radeon_present_get_pixmap_handle helper 5672 5673 Reviewed-by: Axel Davy <axel.davy@ens.fr> 5674 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5675 5676commit 7b4fc4a677d252d01c2bf80d162bc35814059eaa 5677Author: Michel Dänzer <michel.daenzer@amd.com> 5678Date: Thu Mar 19 12:15:52 2015 +0900 5679 5680 Make radeon_do_pageflip take a BO handle directly 5681 5682 Reviewed-by: Axel Davy <axel.davy@ens.fr> 5683 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5684 5685commit af6076241c0d322b295a4e898407ae2472bd8eb4 5686Author: Michel Dänzer <michel.daenzer@amd.com> 5687Date: Tue Mar 24 13:37:01 2015 +0900 5688 5689 Adapt radeon_sync.c for misyncshm.h vs misync.h as well 5690 5691 This should have been part of the previous commit. Without this, SYNC 5692 extension fences (and by extension DRI3) were disabled because it checked 5693 for HAVE_MISYNC_H, but configure now only defines HAVE_MISYNCSHM_H. 5694 5695commit aa7825eb29cdf6ac9d7b28ad18186807ff384687 5696Author: Michel Dänzer <michel.daenzer@amd.com> 5697Date: Mon Mar 23 10:12:17 2015 +0900 5698 5699 configure.ac: Check for misyncshm.h again 5700 5701 We need this header, which was added after the other misync*.h headers, 5702 along with DRI3 support. 5703 5704commit 6291baaed261e36a63dc001307427fe00ba82259 5705Author: Michel Dänzer <michel.daenzer@amd.com> 5706Date: Thu Mar 19 00:12:06 2015 +0900 5707 5708 EXA: Return NULL from radeon_get_pixmap_bo if there is no driver private 5709 5710 This mirrors what the glamor part of the function does. Fixes a crash 5711 running glxgears_pixmap with DRI3 enabled, reported by "marvin24" on IRC. 5712 5713commit f68d9b5ba0c91a725b5eec9386c61bea8824c299 5714Author: Michel Dänzer <michel.daenzer@amd.com> 5715Date: Wed Mar 18 18:29:27 2015 +0900 5716 5717 dri3: Use screen->CreatePixmap instead of fbCreatePixmap directly 5718 5719 Fixes crash with EXA reported by "marvin24" on IRC. 5720 5721commit 4b0997e56dec0053cb2cb793e0f4ae35055ff7e6 5722Author: Michel Dänzer <michel.daenzer@amd.com> 5723Date: Wed Mar 18 12:30:09 2015 +0900 5724 5725 glamor: Add glamor_fd_from_pixmap define for standalone glamor tree 5726 5727commit af1862a37570fa512a525ab47d72b30400d2e2d6 5728Author: Michel Dänzer <michel.daenzer@amd.com> 5729Date: Wed Mar 18 11:05:40 2015 +0900 5730 5731 Always include misync.h before other misync headers 5732 5733 Older versions of xserver didn't include misync.h from other misync 5734 headers as needed. 5735 5736commit fcd37f65f485291084c174666bd605e215bf1398 5737Author: Michel Dänzer <michel.daenzer@amd.com> 5738Date: Wed Mar 18 10:39:03 2015 +0900 5739 5740 DRI3: Use open hook instead of open_client 5741 5742 We don't need the client pointer, and this allows it to work with older 5743 versions of xserver as well. 5744 5745commit f940fd741b15f03393037c5bb904cd74f012de9d 5746Author: Michel Dänzer <michel.daenzer@amd.com> 5747Date: Wed Mar 18 10:06:47 2015 +0900 5748 5749 Initialize boolean variable before calling xf86GetOptValBool 5750 5751 We were just lucky that it contained 0 in most cases. 5752 5753commit 4a98f60117c387a228d5cbaadb6e298fb4e865df 5754Author: Michel Dänzer <michel.daenzer@amd.com> 5755Date: Tue Mar 17 16:47:11 2015 +0900 5756 5757 Add xorg_list_for_each_entry_safe fallback in radeon_list.h 5758 5759 Fixes another build failure against older versions of xserver reported 5760 by "Pali" on IRC. 5761 5762commit 694e04720b886060fe3eefdce59741f218c8269f 5763Author: Michel Dänzer <michel.daenzer@amd.com> 5764Date: Tue Mar 17 16:32:09 2015 +0900 5765 5766 Simplify includes in radeon_{dri3,present}.c 5767 5768 Just include radeon.h first. Fixes build failures against older versions 5769 of xserver because compat-api.h wasn't getting picked up. Reported by 5770 "Pali" on IRC. 5771 5772commit 64e1e4dbdd3caee6f5d8f6b6c094b4533fa94953 5773Author: Michel Dänzer <michel.daenzer@amd.com> 5774Date: Tue Mar 10 16:53:11 2015 +0900 5775 5776 Add DRI3 support v2 5777 5778 Must be enabled with 5779 5780 Option "DRI3" 5781 5782 in xorg.conf. 5783 5784 v2: Adapt to v2 of patches 11/12. 5785 5786 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5787 5788commit 69ae0194778fe4276895839db92383f63f0b5de4 5789Author: David Heidelberger <david.heidelberger@ixit.cz> 5790Date: Fri Mar 6 17:57:22 2015 +0900 5791 5792 Handle tiling in radeon_set_shared_pixmap_backing 5793 5794 [ Michel Dänzer: Fixups for glamor ] 5795 Signed-off-by: David Heidelberger <david.heidelberger@ixit.cz> 5796 Signed-off-by: Axel Davy <axel.davy@ens.fr> 5797 5798 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5799 5800commit 3c65fb849e1ba9fb6454bcaa55b696548902f3fc 5801Author: Michel Dänzer <michel.daenzer@amd.com> 5802Date: Tue Mar 10 16:52:54 2015 +0900 5803 5804 Add support for the Present extension v2 5805 5806 v2: Fix up for struct radeon_drm_queue -> radeon_drm_queue_entry. 5807 Swapped order of patches 11 & 12 because the Present extension uses 5808 SYNC fences. 5809 5810 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5811 5812commit 8fc9a241ab59ffbcdc178d6415332c88a54e85fe 5813Author: Michel Dänzer <michel.daenzer@amd.com> 5814Date: Thu Mar 5 18:18:56 2015 +0900 5815 5816 Add support for SYNC extension fences v2 5817 5818 v2: Swapped order of patches 11 & 12 because the Present extension uses 5819 SYNC fences. 5820 5821 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5822 5823commit 4a35e2f33d9cdfb608423046391311109f96fb6b 5824Author: Michel Dänzer <michel.daenzer@amd.com> 5825Date: Thu Mar 5 18:34:07 2015 +0900 5826 5827 Fold radeon_glamor_flush into radeon_cs_flush_indirect 5828 5829 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5830 5831commit 4b8adebb80158bcf81ada83bb88517febe931b12 5832Author: Michel Dänzer <michel.daenzer@amd.com> 5833Date: Thu Mar 5 18:40:23 2015 +0900 5834 5835 Move #include "radeon_glamor.h" from radeon.h to where it's needed 5836 5837 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5838 5839commit 76c2923ac5c7230a8b2f9f8329c308d28b44d9c0 5840Author: Michel Dänzer <michel.daenzer@amd.com> 5841Date: Wed Mar 4 10:30:19 2015 +0900 5842 5843 DRI2: Split out helper for getting UST and MSC of a specific CRTC 5844 5845 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5846 5847commit 6c3a721cde9317233072b573f9502348dcd21b16 5848Author: Michel Dänzer <michel.daenzer@amd.com> 5849Date: Tue Mar 10 16:52:18 2015 +0900 5850 5851 DRI2: Use helper functions for DRM event queue management v3 5852 5853 This is mostly in preparation for Present support, but it also simplifies 5854 the DRI2 specific code a little. 5855 5856 v2: Fix up for struct radeon_drm_queue -> radeon_drm_queue_entry. 5857 v3: Removed excess 0s from conversion from microseconds to seconds, 5858 thanks to Richard Wilbur <richard.wilbur@gmail.com> for the catch! 5859 5860 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5861 5862commit c3fa22a479e61d1899fa9d327d9c4e2a7f64b0c1 5863Author: Michel Dänzer <michel.daenzer@amd.com> 5864Date: Wed Mar 11 17:47:59 2015 +0900 5865 5866 DRI2: Move radeon_dri2_flip_event_handler 5867 5868 In preparation for the next change, which will modify it to a static 5869 function which needs to be in the new place. No functional change. 5870 5871 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5872 5873commit 65045112fdc8a9fa36e0e00f46739a6152b775ff 5874Author: Michel Dänzer <michel.daenzer@amd.com> 5875Date: Wed Mar 11 17:33:54 2015 +0900 5876 5877 DRI2: Remove superfluous assignments to *_info->frame 5878 5879 That field is only used for page flipping. 5880 5881 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5882 5883commit ad27f16f308079d06a2b1c788b3cb0947531253a 5884Author: Michel Dänzer <michel.daenzer@amd.com> 5885Date: Wed Mar 11 17:30:11 2015 +0900 5886 5887 DRI2: Simplify blit fallback handling for scheduled swaps 5888 5889 Also use radeon_dri2_schedule_event when possible. 5890 5891 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5892 5893commit b4af8a327ed8420f0ff4ea0f113f4a59406ed4d3 5894Author: Michel Dänzer <michel.daenzer@amd.com> 5895Date: Mon Mar 2 18:59:54 2015 +0900 5896 5897 Add DRM event queue helpers v2 5898 5899 v2: Rename struct radeon_drm_queue to struct radeon_drm_queue_event, 5900 thanks to Richard Wilbur <richard.wilbur@gmail.com> for the suggestion. 5901 Also changed the corresponding parameter and local variable names from 5902 'q' to 'e'. 5903 5904 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5905 5906commit 7c3470f4b659206ed23f761948936ede3a2dba3d 5907Author: Michel Dänzer <michel.daenzer@amd.com> 5908Date: Fri Mar 6 18:51:29 2015 +0900 5909 5910 Move xorg_list backwards compatibility to new radeon_list.h header 5911 5912 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5913 5914commit 7388d0b6c54b9d536fdb161e3aa61b326627b939 5915Author: Michel Dänzer <michel.daenzer@amd.com> 5916Date: Thu Mar 5 17:35:06 2015 +0900 5917 5918 Require at least xserver 1.8 5919 5920 So we can rely on the list.h header. 5921 5922 xserver 1.8 was released in April 2010. 5923 5924 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5925 5926commit b8ec9ed4fe86952763b963c86f0af0dcae69aa6c 5927Author: Jerome Glisse <jglisse@redhat.com> 5928Date: Thu Feb 19 14:47:41 2015 -0500 5929 5930 Proper leak fix, previous leak fix was bogus. 5931 5932 Signed-off-by: Jérôme Glisse <jglisse@redhat.com> 5933 5934commit 63dc36dc49f93cb00111b497ab6805194bc9d240 5935Author: Jerome Glisse <jglisse@redhat.com> 5936Date: Thu Feb 19 11:43:08 2015 -0500 5937 5938 Avoid leaking memory on output. 5939 5940 Signed-off-by: Jérôme Glisse <jglisse@redhat.com> 5941 5942commit c80ea1e3e8c5f155cedadbe4588870fb1f410ab4 5943Author: Rolf Eike Beer <eb@emlix.com> 5944Date: Tue Jan 13 12:11:26 2015 +0100 5945 5946 radeon: remove unneeded inclusion of sarea.h 5947 5948 None of the structs or defines from that header is used. 5949 5950 Signed-off-by: Rolf Eike Beer <eb@emlix.com> 5951 5952commit 04da199231bb3f11cf17f94574a8df05855a7b82 5953Author: Adam Jackson <ajax@redhat.com> 5954Date: Wed Dec 17 14:03:58 2014 -0500 5955 5956 Remove dead accelDFS flag 5957 5958 Signed-off-by: Adam Jackson <ajax@redhat.com> 5959 5960commit c9f8f642fd495937400618a4fc25ecae3f8888fc 5961Author: Michel Dänzer <michel.daenzer@amd.com> 5962Date: Mon Nov 17 12:32:29 2014 +0900 5963 5964 Prefer drmModeSetCursor2 over drmModeSetCursor 5965 5966 The former includes information about the position of the hotspot within 5967 the cursor image. 5968 5969 Copied from xf86-video-modesetting. 5970 5971 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5972 5973commit 2f11dcd43966cf2ee26e61960fd72e6644f5e037 5974Author: Dave Airlie <airlied@redhat.com> 5975Date: Mon Nov 10 13:49:29 2014 +1000 5976 5977 radeon: add support for DP 1.2 display hotplug (v2) 5978 5979 This allows for dynamic creation of conneectors when the 5980 kernel tells us. 5981 5982 v2: fix dpms off crash 5983 5984 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5985 Signed-off-by: Dave Airlie <airlied@redhat.com> 5986 5987commit c88424d1f4aaa78b569e5d44f0b4a47de2f422f4 5988Author: Dave Airlie <airlied@redhat.com> 5989Date: Mon Nov 10 14:17:54 2014 +1000 5990 5991 radeon: move output name creation to its own function 5992 5993 The secondary indent is deliberate to make the next patch more 5994 parseable for mst support. 5995 5996 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 5997 Signed-off-by: Dave Airlie <airlied@redhat.com> 5998 5999commit 32b003cb7657e07d5af6338ad44d768eda87fd33 6000Author: Dave Airlie <airlied@redhat.com> 6001Date: Mon Nov 10 14:12:34 2014 +1000 6002 6003 radeon: stop caching mode resources 6004 6005 This is step one towards MST connector hotplug support, 6006 it stop caching the mode resources structure, and 6007 just passes a pointer to it around. 6008 6009 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6010 Signed-off-by: Dave Airlie <airlied@redhat.com> 6011 6012commit 7c7b38e0b375b6e8853ad2d1092302ea83f6f570 6013Author: David Heidelberger <david.heidelberger@ixit.cz> 6014Date: Sun Oct 12 16:34:21 2014 +0200 6015 6016 radeon/vdpau: don't report VDPAU for < r300 6017 6018 Signed-off-by: David Heidelberger <david.heidelberger@ixit.cz> 6019 6020commit c74de9fec13fac2c836bb2a07ae6f90e1d61e667 6021Author: Michel Dänzer <michel.daenzer@amd.com> 6022Date: Wed Aug 6 11:08:00 2014 +0900 6023 6024 PRIME: Don't advertise offload capabilities when acceleration is disabled 6025 6026 Xorg tends to crash if the user tries to actually use the offload 6027 capabilities with acceleration disabled. 6028 6029 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=57200 6030 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6031 6032commit 636a6382f80412ae4e28abe0cd9f98f84cd52291 6033Author: Michel Dänzer <michel.daenzer@amd.com> 6034Date: Wed Oct 15 12:02:17 2014 +0900 6035 6036 Remove duplicate OPTION_PAGE_FLIP entry 6037 6038 Reported by 'amanual' on IRC 6039 6040commit c854b4479ec5122390ede81b655b259efd9b1800 6041Author: Alex Deucher <alexdeucher@gmail.com> 6042Date: Thu Oct 2 10:22:22 2014 -0400 6043 6044 radeon: bump version post release 6045 6046 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 6047 6048commit 068a59e010ce6bfcd54f5a18cc08c55c54b8618d 6049Author: Alex Deucher <alexander.deucher@amd.com> 6050Date: Wed Oct 1 23:38:52 2014 -0400 6051 6052 radeon: bump version for release 6053 6054 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6055 6056commit 5fee565e420d3efbf72fdf4f63c3d5d93d8ceddb 6057Author: Alex Deucher <alexander.deucher@amd.com> 6058Date: Wed Oct 1 23:36:31 2014 -0400 6059 6060 radeon: require libdrm_radeon 2.4.58 for latest SI/CI pci ids 6061 6062 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6063 6064commit 56c7fb8ada4cd9cf096c6b06a8d368d286d74e68 6065Author: Thomas DeRensis <tderensis@gmail.com> 6066Date: Wed Sep 10 21:20:32 2014 -0400 6067 6068 radeon: fix build warnings regarding const qualifier 6069 6070 Signed-off-by: Thomas DeRensis <tderensis@gmail.com> 6071 6072commit b9a9b244336ed50d7df4f36135f875a3bb6ca948 6073Author: Thomas DeRensis <tderensis@gmail.com> 6074Date: Thu Sep 4 21:45:19 2014 -0400 6075 6076 radeon: fix a leak in radeon_vbo_get_bo() 6077 6078 The dma_bo struct was never freed in the error path. Reported by cppcheck. 6079 6080 Signed-off-by: Thomas DeRensis <tderensis@gmail.com> 6081 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6082 6083commit ba5fcc4d6b6eae3f2c55c6229a53d24d160789c1 6084Author: Alex Deucher <alexander.deucher@amd.com> 6085Date: Thu Aug 21 11:32:45 2014 -0400 6086 6087 radeon/kms: fix mullins pci id 6088 6089 strange errant change from: 6090 39fef269f521c92a2a31c80447e9401bacb3797a 6091 6092 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6093 6094commit efef34930650d6d80f7b527f4cee76d9e5954ace 6095Author: Alex Deucher <alexander.deucher@amd.com> 6096Date: Thu Aug 21 11:30:17 2014 -0400 6097 6098 radeon/kms: add new SI pci ids 6099 6100 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6101 6102commit 39fef269f521c92a2a31c80447e9401bacb3797a 6103Author: Alex Deucher <alexander.deucher@amd.com> 6104Date: Thu Aug 21 11:27:49 2014 -0400 6105 6106 radeon/kms: add new CIK pci ids 6107 6108 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6109 6110commit 7dab6b2fe9cf690b0dae1c127d0374322bd8278e 6111Author: Michel Dänzer <michel.daenzer@amd.com> 6112Date: Wed Aug 20 17:51:35 2014 +0900 6113 6114 Revert "glamor: Set environment variable RADEON_THREAD=0" 6115 6116 This reverts commit 4b5060f357a3cb248c9359c92c1e9c42ef6434c8. 6117 6118 Further testing shows that disabling the thread is only a moderate win in 6119 some cases, but a much bigger loss in some other cases. 6120 6121 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6122 6123commit 1ff383360886c5cee1bb75abcc20675aca0336a6 6124Author: Alex Deucher <alexander.deucher@amd.com> 6125Date: Tue Aug 12 12:27:12 2014 -0400 6126 6127 radeon: fix warnings when building against older xservers 6128 6129 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6130 6131commit 94202cbfbca05a503acdc1cca2f8409d141173af 6132Author: Alex Deucher <alexdeucher@gmail.com> 6133Date: Fri Aug 1 21:55:40 2014 +0200 6134 6135 radeon: enable hawaii accel conditionally (v3) 6136 6137 Only if the kernel has the new CP firmware. 6138 6139 v2: check value of ACCEL_WORKING2 6140 v3 (Andreas Boll): 6141 - check for value 2 or 3 of ACCEL_WORKING2 6142 - update man page 6143 6144 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> (v2) 6145 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6146 Signed-off-by: Andreas Boll <andreas.boll.dev@gmail.com> 6147 6148commit 03930edd49f6b8a8d79910c7be5408b47db9649b 6149Author: Andreas Boll <andreas.boll.dev@gmail.com> 6150Date: Mon Aug 4 16:23:13 2014 +0200 6151 6152 radeon: remove definitions already present in radeon_drm.h 6153 6154 Signed-off-by: Andreas Boll <andreas.boll.dev@gmail.com> 6155 6156commit 91849fba0742ef61ba327e71fc3ce8f754af0a6f 6157Author: Andreas Boll <andreas.boll.dev@gmail.com> 6158Date: Mon Aug 4 16:23:12 2014 +0200 6159 6160 radeon: drop radeon_drm.h 6161 6162 Now we use libdrm's radeon_drm.h. 6163 6164 Signed-off-by: Andreas Boll <andreas.boll.dev@gmail.com> 6165 6166commit b4658901bf1d619f8ff9b5f94344894f935ee6c8 6167Author: Andreas Boll <andreas.boll.dev@gmail.com> 6168Date: Mon Aug 4 16:23:11 2014 +0200 6169 6170 radeon: move RADEON_TILING_{MASK, LINEAR} from radeon_drm.h to radeon.h 6171 6172 This allows us to drop radeon_drm.h from xf86-video-ati and use instead 6173 radeon_drm.h from libdrm. 6174 6175 Signed-off-by: Andreas Boll <andreas.boll.dev@gmail.com> 6176 6177commit 656bae7361c1e018553ef6d6d8c9efad616a4513 6178Author: Andreas Boll <andreas.boll.dev@gmail.com> 6179Date: Mon Aug 4 16:23:10 2014 +0200 6180 6181 radeon: drop redundant radeon_drm.h includes 6182 6183 Already included via radeon.h. 6184 6185 Signed-off-by: Andreas Boll <andreas.boll.dev@gmail.com> 6186 6187commit fbf575cb010e558a87fad0de45660738c8180896 6188Author: Michel Dänzer <michel.daenzer@amd.com> 6189Date: Wed Jul 30 17:03:24 2014 +0900 6190 6191 Add Emacs .dir-local.el file 6192 6193 Based on the one from the Gallium radeon winsys, but enabling tabs for 6194 indentation. 6195 6196commit 4b5060f357a3cb248c9359c92c1e9c42ef6434c8 6197Author: Michel Dänzer <michel.daenzer@amd.com> 6198Date: Thu Jun 19 18:27:59 2014 +0900 6199 6200 glamor: Set environment variable RADEON_THREAD=0 6201 6202 Reviewed-by: Marek Olšák <marek.olsak@amd.com> 6203 6204commit 9b54caf6509a9c02dd17c9c43d6be8f7ddc98054 6205Author: Hans de Goede <hdegoede@redhat.com> 6206Date: Mon Jul 28 15:55:12 2014 +0200 6207 6208 configure: Include xorg-server.h before 6209 glamor.h 6210 6211 glamor.h cannot be included without first including xorg-server.h, this also 6212 applies to including it from configure snippets. 6213 6214 Without this the configure glamor checks fail on systems with the latest 6215 glibc, throwing this error: 6216 6217 In file included from /usr/include/xorg/misc.h:115:0, 6218 from /usr/include/xorg/screenint.h:50, 6219 from /usr/include/xorg/scrnintstr.h:50, 6220 from /usr/include/xorg/glamor.h:32, 6221 from conftest.c:61: 6222 /usr/include/xorg/os.h:579:2: error: expected identifier or '(' before 6223 '__exten 6224 strndup(const char *str, size_t n); 6225 ^ 6226 6227 This is caused by HAVE_STRNDUP not being set (it is set from xorg-server.h), 6228 causing os.h to redefine it. 6229 6230 Signed-off-by: Hans de Goede <hdegoede@redhat.com> 6231 6232commit c4ae0e2cbcc0e2ebf9f13ee92d59b5120254a1dc 6233Author: Michel Dänzer <michel.daenzer@amd.com> 6234Date: Mon Jun 30 10:20:12 2014 +0900 6235 6236 Handle CRTC DPMS from output DPMS hooks 6237 6238 This fixes at least two issues: 6239 6240 The CRTC DPMS hook isn't called after a modeset, so the vertical blank 6241 interrupt emulation code considered the CRTC disabled after a modeset. As 6242 a side effect, page flipping was no longer used after a modeset. 6243 6244 This change also makes sure the vertical blank interrupt emulation code 6245 runs before the hardware CRTC is disabled and after it's enabled from the 6246 output DPMS hook. The wrong order could cause gnome-shell to hang after 6247 a suspend/resume and/or DPMS off/on cycle. 6248 6249 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 6250 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6251 6252commit c0c3cac8613e31e310242695d0955b452f116e25 6253Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6254Date: Wed Jun 25 16:21:57 2014 +0200 6255 6256 bump version post release 6257 6258commit 906a0ec9224146098bb4581486129d2934d36495 6259Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6260Date: Wed Jun 25 16:10:21 2014 +0200 6261 6262 bump version for release 6263 6264commit cc615d06db0332fc6e673b55632bcc7bf957b44b 6265Author: Michel Dänzer <michel.daenzer@amd.com> 6266Date: Wed Jun 4 16:58:53 2014 +0900 6267 6268 Rename Option "NoAccel" to "Accel" 6269 6270 Removes the need for a double negation when forcing acceleration on. 6271 6272 Note that this change is backwards compatible, as the option parser 6273 automagically handles the 'No' prefix. 6274 6275 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6276 6277commit 851b2cf8714618843725f6d067915375485ade9d 6278Author: Adam Jackson <ajax@redhat.com> 6279Date: Wed May 21 09:34:32 2014 -0400 6280 6281 kms: Use own thunk function instead of shadowUpdatePackedWeak 6282 6283 I plan to delete the Weak functions from a future server. 6284 6285 Signed-off-by: Adam Jackson <ajax@redhat.com> 6286 6287commit b2dba2906f0b2284f17f53fd5251ba0f03d52a8b 6288Author: Michel Dänzer <michel.daenzer@amd.com> 6289Date: Thu May 15 16:07:53 2014 +0900 6290 6291 Don't disable acceleration on >= SI on attempts to force EXA 6292 6293 Also make this case clear in the log file: 6294 6295 (WW) RADEON(0): EXA not supported, using glamor 6296 6297 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6298 6299commit be1469cc23aba46daf3293b3d09c5f2e792e7f42 6300Author: Alex Deucher <alexander.deucher@amd.com> 6301Date: Fri May 2 17:48:07 2014 -0400 6302 6303 radeon: enable tiling for mullins 6304 6305 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6306 6307commit 2ae8e4b8d12f5c9bab6655eb8cd3c5c1d5cfb10e 6308Author: Samuel Li <samuel.li@amd.com> 6309Date: Tue Nov 12 15:30:42 2013 -0500 6310 6311 radeon: add Mullins pci ids. 6312 6313 Signed-off-by: Samuel Li <samuel.li@amd.com> 6314 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6315 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6316 6317commit 5c86a3461597647224c33d5190b4531aeeb2655f 6318Author: Samuel Li <samuel.li@amd.com> 6319Date: Thu Apr 17 15:17:28 2014 -0400 6320 6321 radeon: add support for Mullins. 6322 6323 Signed-off-by: Samuel Li <samuel.li@amd.com> 6324 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6325 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6326 6327commit fdb90ffc50acbb7d5ba0598470f9feeac6ce55fc 6328Author: Alex Deucher <alexander.deucher@amd.com> 6329Date: Fri May 2 17:45:45 2014 -0400 6330 6331 radeon: require libdrm_radeon 2.4.54 for mullins support 6332 6333 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6334 6335commit 06e3c8c53ef4bd159f5864eabf726438d008b49a 6336Author: Dave Airlie <airlied@redhat.com> 6337Date: Wed Apr 23 13:39:42 2014 +1000 6338 6339 radeon: fix use-after-free in modesetting cleanup 6340 6341 noticed while looking at something else. 6342 6343 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6344 Signed-off-by: Dave Airlie <airlied@redhat.com> 6345 6346commit dbac18c361f9e514ecb40d0617f9d68b65a542e0 6347Author: Michel Dänzer <michel.daenzer@amd.com> 6348Date: Mon Apr 28 17:51:56 2014 +0900 6349 6350 Revert "Adapt to load_cursor_argb signature change in xserver 1.15.99.902" 6351 6352 This reverts commit 48d3dbc8a0d3bfde88f46e402e530438f9317715. 6353 6354 xserver Git has been updated to be backwards compatible with the 6355 previous API. 6356 6357commit c84230d686c078aac1dc98d82153f8b02521b2e1 6358Author: Michel Dänzer <michel.daenzer@amd.com> 6359Date: Fri Apr 25 09:17:51 2014 +0900 6360 6361 dri2: Handle PRIME for source buffer as well in radeon_dri2_copy_region2 6362 6363 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=77810 6364 6365 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6366 6367commit 48d3dbc8a0d3bfde88f46e402e530438f9317715 6368Author: Michel Dänzer <michel.daenzer@amd.com> 6369Date: Tue Apr 15 17:45:35 2014 +0900 6370 6371 Adapt to load_cursor_argb signature change in xserver 1.15.99.902 6372 6373 Apart from the compiler warning below, not doing this may result in 6374 accidentally using software cursors. 6375 6376 ../../src/drmmode_display.c:808:5: warning: initialization from incompatible pointer type [enabled by default] 6377 .load_cursor_argb = drmmode_load_cursor_argb, 6378 ^ 6379 ../../src/drmmode_display.c:808:5: warning: (near initialization for 'drmmode_crtc_funcs.load_cursor_argb') [enabled by default] 6380 6381 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6382 6383commit 409786a8f780d78a83bf0bddea5d37117ff6fa39 6384Author: Michel Dänzer <michel.daenzer@amd.com> 6385Date: Thu Apr 10 11:43:04 2014 +0900 6386 6387 glamor: Fix test for creating shared pixmaps 6388 6389 The pixmap usage hint is not a bitmask in general. The test for 6390 CREATE_PIXMAP_USAGE_SHARED was incorrectly triggering for a glamor internal 6391 usage hint being added in the xserver tree. 6392 6393 Tested-by: Ed Tomlinson <edtoml@gmail.com> 6394 6395commit aecf1c4e5f4718adcfb85836830d065d3f4f97a5 6396Author: Michel Dänzer <michel.daenzer@amd.com> 6397Date: Thu Apr 10 15:10:56 2014 +0900 6398 6399 dri2: Fix conflicting CreatePixmap usage flag definitions 6400 6401 RADEON_CREATE_PIXMAP_TILING_MICRO_SQUARE was the same as 6402 RADEON_CREATE_PIXMAP_DRI2. 6403 6404 Disambiguate the definitions and rearrange them to try and prevent this 6405 from happening again. 6406 6407 Tested-by: Ed Tomlinson <edtoml@gmail.com> 6408 6409commit b50da3b96c212086cb58501dbe988d64f1f35b6d 6410Author: Hans de Goede <hdegoede@redhat.com> 6411Date: Fri Apr 11 09:44:37 2014 +0200 6412 6413 Fix building on older servers without xf86platformBus.h 6414 6415 Signed-off-by: Hans de Goede <hdegoede@redhat.com> 6416 6417commit ed0cfbb4fe77146b0b38f777bc28f3a4ea6da07f 6418Author: Hans de Goede <hdegoede@redhat.com> 6419Date: Fri Mar 7 13:27:30 2014 +0100 6420 6421 Add support for server managed fds 6422 6423 Signed-off-by: Hans de Goede <hdegoede@redhat.com> 6424 6425commit 3d7861fe112f25874319d4cdc12b745fbcd359cf 6426Author: Hans de Goede <hdegoede@redhat.com> 6427Date: Mon Mar 17 10:38:13 2014 +0100 6428 6429 Add radeon_get_drm_master_fd helper function 6430 6431 This is a preparation patch for adding server-managed-fd support without it 6432 turning into a goto fest. 6433 6434 Signed-off-by: Hans de Goede <hdegoede@redhat.com> 6435 6436commit a63342ad15408071437c80b411d14196f3288aed 6437Author: Hans de Goede <hdegoede@redhat.com> 6438Date: Mon Mar 17 10:36:55 2014 +0100 6439 6440 radeon_open_drm_master get rid of unnecessary goto 6441 6442 Signed-off-by: Hans de Goede <hdegoede@redhat.com> 6443 6444commit bdc412044f6ced056cd57320d1b2ee0d967c2191 6445Author: Michel Dänzer <michel.daenzer@amd.com> 6446Date: Thu Mar 13 16:40:19 2014 +0900 6447 6448 Build against glamor in the xserver tree if available 6449 6450commit 921a153f9964ca452e1241f76c7f7d653f42ceaf 6451Author: Alex Deucher <alexander.deucher@amd.com> 6452Date: Tue Mar 11 12:24:13 2014 -0400 6453 6454 update man page to reflect tiling changes for CI parts 6455 6456 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6457 6458commit 5eee8a4d5c86bb1cc34d8caf2f2b64b53c241fa5 6459Author: Alex Deucher <alexander.deucher@amd.com> 6460Date: Mon Mar 10 16:20:09 2014 -0400 6461 6462 radeon: enable tiling by default on CIK 6463 6464 Now that mesa 10.1 is released, we can enable this by 6465 default for CIK parts. Tiling improves memory bandwidth 6466 utilization. 6467 6468 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6469 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6470 6471commit 691ec3d99c30111a4789830dfccb6eb5d3c40187 6472Author: Alex Deucher <alexander.deucher@amd.com> 6473Date: Mon Mar 10 16:17:34 2014 -0400 6474 6475 radeon: require libdrm 2.4.51 6476 6477 Required for proper tiling support on CIK parts. 6478 6479 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6480 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6481 6482commit 515bcf14d514f9dcaaf30fd0bf1ef6dd6ba9a0cd 6483Author: Michel Dänzer <michel.daenzer@amd.com> 6484Date: Tue Mar 4 12:34:26 2014 +0900 6485 6486 Allow enabling glamor on R500 (and R300) class 3D engines as well. 6487 6488 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75709 6489 6490 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6491 6492commit ea6d0affe52d82becadab6fb1c87f9261b0605a2 6493Author: Michel Dänzer <michel.daenzer@amd.com> 6494Date: Mon Feb 24 13:12:21 2014 +0900 6495 6496 Only log debugging output about initializing colormaps when we're doing so 6497 6498 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6499 6500commit 8da17f30c70f4494ce22ad781a1cee17041812f3 6501Author: Jérôme Glisse <jglisse@redhat.com> 6502Date: Mon Feb 24 14:50:25 2014 -0500 6503 6504 evergreen: fix shader constant upload on ppc 6505 6506 The number of dword we have to swap is (16*4) ie 16 vectors of 6507 4 floats each not 16 floats. Never hit this issue before because 6508 we never had more than 4 constant vector. 6509 6510 Signed-off-by: Jérôme Glisse <jglisse@redhat.com> 6511 6512commit cadb6b493942a84bfeb298751dce0dee39257a06 6513Author: Alex Deucher <alexander.deucher@amd.com> 6514Date: Fri Feb 21 08:33:21 2014 -0500 6515 6516 radeon: don't install colormap handling if there are no crtcs 6517 6518 Fixes a crash on cards with 0 crtcs. 6519 6520 Discussion: 6521 http://lists.freedesktop.org/archives/dri-devel/2014-February/054186.html 6522 6523 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6524 6525commit 7cd972a85705341dd8306eefc558ed9e5def05d7 6526Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6527Date: Thu Feb 20 11:14:03 2014 +0100 6528 6529 return immediately in preinit when called with PROBE_DETECT 6530 6531 This fixes a crash with Xorg -configure. 6532 6533 Bug: 6534 https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/1278046 6535 6536 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6537 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6538 6539commit 8de6f7b2f476d3baa9c7e2bb3544e4bafaad46b7 6540Author: Alex Deucher <alexander.deucher@amd.com> 6541Date: Fri Jan 24 11:04:30 2014 -0500 6542 6543 bump version post release 6544 6545 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6546 6547commit 0333f5bda27dc0ec2edc180c7a4dc9a432f13f97 6548Author: Alex Deucher <alexander.deucher@amd.com> 6549Date: Fri Jan 24 10:19:49 2014 -0500 6550 6551 radeon: bump version for release 6552 6553 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6554 6555commit fc4167f2a85d9cba65078d8fc6f08c7a619ad66e 6556Author: Alex Deucher <alexander.deucher@amd.com> 6557Date: Fri Jan 24 10:17:08 2014 -0500 6558 6559 Require glamor 0.6.0 6560 6561 This is required for Xv support and a number of important 6562 performance improvements. 6563 6564 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6565 6566commit f2a0a5cf6c5a21e2a02280e110a4eb8e6609dace 6567Author: Michel Dänzer <michel.daenzer@amd.com> 6568Date: Wed Jan 22 11:04:42 2014 +0900 6569 6570 Don't require the glamoregl module to be pre-loaded with xserver >= 1.15 6571 6572 The issues with loading it on demand have been fixed in xserver 1.15. 6573 6574 Inspired by Jérôme Glisse on IRC. 6575 6576commit 3213df16d61302148be0088c8f93c6a5a88558f1 6577Author: Michel Dänzer <michel.daenzer@amd.com> 6578Date: Wed Jan 8 11:30:59 2014 +0900 6579 6580 dri2: Make last_vblank_seq local unsigned to match dpms_last_seq 6581 6582 Without this, I was occasionally running into gnome-shell hangs due to 6583 wildly off vblank sequence values. Doesn't seem to happen anymore with 6584 this change. 6585 6586 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6587 6588commit bcc454ea2fb239e13942270faec7801270615b9c 6589Author: Alex Deucher <alexander.deucher@amd.com> 6590Date: Mon Jan 6 09:52:50 2014 -0500 6591 6592 radeon/exa: Always use a scratch surface for UTS to vram 6593 6594 If we don't, we may hit a buffer that crosses the 6595 visible vram boundary resulting in a sigbus when the 6596 CPU accesses the buffer beyond the PCI aperture. 6597 This will introduce an extra copy in certain cases. 6598 6599 This is based on Michel's patch from bug 44099 updated 6600 for all asic families. 6601 6602 Bug: 6603 https://bugs.freedesktop.org/show_bug.cgi?id=44099 6604 6605 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6606 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6607 6608commit 04ef035c9315b4a6fbf1b14720be87cee4099a9f 6609Author: Alex Deucher <alexander.deucher@amd.com> 6610Date: Tue Dec 24 15:14:35 2013 -0500 6611 6612 drm/radeon: fix SUMO2 pci id 6613 6614 0x9649 is sumo2, not sumo. 6615 6616 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6617 6618commit bfbff3b246db509c820df17b8fcf5899882ffcfa 6619Author: Robert Millan <rmh@freebsd.org> 6620Date: Fri Dec 20 11:03:14 2013 +0000 6621 6622 radeon: Restore kernel module load on FreeBSD. 6623 6624 Since the introduction of a call to drmCheckModesettingSupported() 6625 in radeon_kernel_mode_enabled(), with abort condition if such call 6626 fails, the drmOpen() call in radeon_open_drm_master() no longer 6627 takes the responsibility of loading the radeon kernel module. 6628 6629 However at least on FreeBSD (and GNU/kFreeBSD), X is still relied 6630 on to load the modules it needs. This commit restores the old 6631 behaviour of loading kernel modules on these systems. 6632 6633 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=72852 6634 Signed-off-by: Robert Millan <rmh@freebsd.org> 6635 6636commit 796c9a0cb587f528326bede11fa3f3eb7d3edaf1 6637Author: Robert Millan <rmh@freebsd.org> 6638Date: Fri Dec 20 10:55:09 2013 +0000 6639 6640 radeon: Set first parameter of drmOpen() to NULL 6641 6642 Since the introduction of a call to drmCheckModesettingSupported() 6643 in radeon_kernel_mode_enabled(), with abort condition if such call 6644 fails, the first argument to drmOpen() call in radeon_open_drm_master() 6645 has become a no-op red herring. 6646 6647 Such argument (a kernel module name) is supposed to result in load 6648 of specified kernel module. However, this will never happen. The 6649 problem is that if the code containing drmOpen() call is reached, it 6650 means that drmCheckModesettingSupported() check has previously 6651 succeeded, which implies the module is already loaded. 6652 6653 So, drmOpen() will never load a kernel module. But it gives the 6654 impression that it will. 6655 6656 In order to avoid this confusion, this commit replaces it with NULL, 6657 like xf86-video-intel driver does. 6658 6659 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=72849 6660 Signed-off-by: Robert Millan <rmh@freebsd.org> 6661 6662commit d571d6af70ef27efd1ed6420eb892bdde963ed7a 6663Author: Alex Deucher <alexander.deucher@amd.com> 6664Date: Tue Sep 24 11:39:10 2013 -0400 6665 6666 radeon/kms: add Hawaii pci ids 6667 6668 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6669 6670commit e38a92e00d015a6b80a1f3a16d58c61f084b066f 6671Author: Alex Deucher <alexander.deucher@amd.com> 6672Date: Tue Sep 24 11:35:00 2013 -0400 6673 6674 radeon: add support for Hawaii 6675 6676 Disabled by default until the acceleration code stablizes. 6677 6678 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6679 6680commit e4cd0f4392ea11c93088ad429f36eaaf9bcbf505 6681Author: Alex Deucher <alexander.deucher@amd.com> 6682Date: Wed Oct 30 15:21:26 2013 -0400 6683 6684 radeon: enable tiling on SI by default (v2) 6685 6686 Now that mesa 9.2 is out with support for tiling 6687 on SI asics, we can enable it here. Tiling improves 6688 memory bandwidth utilization. 6689 6690 V2: update man page 6691 6692 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6693 6694commit 3b38701a72fa1cad1e4610a2f4330b3da4cc6391 6695Author: Vadim Girlin <vadimgirlin@gmail.com> 6696Date: Fri Nov 1 10:36:39 2013 -0400 6697 6698 radeon: disable 2D tiling on buffers < 128 pixels 6699 6700 Seems to run into alignment problems with certain 6701 card configurations. 6702 6703 bug: 6704 https://bugs.freedesktop.org/show_bug.cgi?id=70675 6705 6706 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6707 6708commit 0c921edf0162fed616cea9d02e168b719243bcd2 6709Author: Jerome Glisse <jglisse@redhat.com> 6710Date: Wed Oct 30 13:01:02 2013 -0400 6711 6712 radeon/glamor: with new pixmap for dri2 drawable no need to create new texture 6713 6714 When creating a new pixmap/bo for dri2 drawable there is no need to create a 6715 new texture, instead the texture associated with the new pixmap should be use 6716 otherwise there is a missmatch between the bo backing the texture for the 6717 drawable and the bo used by dri2 client. 6718 6719 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 6720 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 6721 6722commit f1dc677e79cd7a88d7379a934ebc7d87a3b18805 6723Author: Christopher James Halse Rogers <raof@ubuntu.com> 6724Date: Mon Sep 23 12:25:29 2013 -0700 6725 6726 EXA/evergreen: Paranoia around linear tiling. (v2) 6727 6728 The last two bytes of tiling_mode contain the actual tiling mode; the rest are 6729 extra tiling configuration bits. These configuration bits are not necessarily 6730 zero for a linear buffer, so mask them out before checking for linearity 6731 6732 v2: Also fix up evergreen_textured_videofuncs.c 6733 6734 agd5f: remove trailing whitespace 6735 6736 Signed-off-by: Christopher James Halse Rogers <raof@ubuntu.com> 6737 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6738 6739commit 67fb82a3f0759b171fea21b475a70fa825693570 6740Author: Alex Deucher <alexander.deucher@amd.com> 6741Date: Tue Oct 1 09:35:30 2013 -0400 6742 6743 radeon: fix the non-glamor build harder... 6744 6745 I need to stop pushing patches first thing in the morning. 6746 6747 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6748 6749commit afc0374fdec3e24ece25805724459463e9a19f5e 6750Author: Alex Deucher <alexander.deucher@amd.com> 6751Date: Tue Oct 1 09:32:02 2013 -0400 6752 6753 drm/radeon: fix non-glamor build 6754 6755 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6756 6757commit 2d791370dfc5570eb74d7a1fb3baf4d4c8ecf243 6758Author: Dave Airlie <airlied@redhat.com> 6759Date: Mon Sep 23 07:57:15 2013 +0100 6760 6761 radeon: use glamor Xv support if present. 6762 6763 This creates adaptors using glamor if possible. 6764 6765 Signed-off-by: Dave Airlie <airlied@redhat.com> 6766 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6767 6768commit cbb99f659ee7b18ded0008a606e41ded38c1a194 6769Author: Alex Deucher <alexander.deucher@amd.com> 6770Date: Mon Sep 30 08:55:16 2013 -0400 6771 6772 Revert "radeon: add glamor Xv support (v2)" 6773 6774 This causes problems if glamor doesn't have Xv support 6775 enabled. I just noticed that Dave has a better version, 6776 so use that instead. 6777 6778 This reverts commit 4fc1fa920584ace2c84d75af82d06962d0c84ec8. 6779 6780commit 4fc1fa920584ace2c84d75af82d06962d0c84ec8 6781Author: Alex Deucher <alexander.deucher@amd.com> 6782Date: Thu Sep 26 11:02:55 2013 -0400 6783 6784 radeon: add glamor Xv support (v2) 6785 6786 v2: guard new glamor Xv bits with USE_GLAMOR 6787 6788 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6789 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6790 6791commit 41dfe327ac8740ac2cd84def96b5947224e422e7 6792Author: Alex Deucher <alexander.deucher@amd.com> 6793Date: Thu Sep 26 11:00:49 2013 -0400 6794 6795 radeon: fix limit handling for cards with >4G of ram 6796 6797 We can overflow the 32-bit limit. 6798 6799 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6800 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6801 6802commit c45e728107269c6f51599dad4f6a02ccfef703f1 6803Author: Michel Dänzer <michel.daenzer@amd.com> 6804Date: Wed Sep 18 10:57:52 2013 +0200 6805 6806 DRI2: Install client callback only once 6807 6808 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=60182 6809 Acked-by: Alex Deucher <alexander.deucher@amd.com> 6810 6811commit b955ddd0d41801e4ca0c30a70a5d0b27c3f366c8 6812Author: Alex Deucher <alexander.deucher@amd.com> 6813Date: Mon Sep 16 10:58:22 2013 -0400 6814 6815 glamor: require 0.5.1 or newer 6816 6817 0.5.1 fixes a number of issues. 6818 6819 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6820 6821commit 282587cd0709850e7bacb1d8307065d95dc2c97d 6822Author: Alex Deucher <alexander.deucher@amd.com> 6823Date: Mon Sep 16 09:41:13 2013 -0400 6824 6825 radeon: enable glamor by default (v4) 6826 6827 It's required for newer asics, so enable it by 6828 default. 6829 6830 v2: update the autoconf help. Require --disable-glamor 6831 to build without glamor support. 6832 v3: default to yes rather than auto 6833 v4: fix help text 6834 6835 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6836 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6837 6838commit fdb7563a5cbc736b09c2864b67a93b475c98b2bd 6839Author: Alex Deucher <alexander.deucher@amd.com> 6840Date: Thu Jan 24 21:17:11 2013 -0500 6841 6842 radeon/kms: add berlin pci ids 6843 6844 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6845 6846commit 8927d33f76ee12bc618fecfc59fc7ff1fcedcd5e 6847Author: Mark Kettenis <kettenis@openbsd.org> 6848Date: Thu Aug 22 10:32:46 2013 -0400 6849 6850 Fix shadowfb on big-endian machines 6851 6852 For shadowfb, the framebuffer needs to have the RADEON_TILING_SURFACE 6853 flag set, otherwise the appropriate byte swapping won't happen. 6854 6855 See https://bugs.freedesktop.org/show_bug.cgi?id=66663 6856 6857 Signed-off-by: Mark Kettenis <kettenis@openbsd.org> 6858 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6859 6860commit fa83d3d1636c315bc43dd622d407eb058e3ef976 6861Author: Alex Deucher <alexander.deucher@amd.com> 6862Date: Tue Aug 20 09:29:23 2013 -0400 6863 6864 radeon: disallow glamor on pre-R600 asics 6865 6866 I'm not sure they can handle the shaders properly, especially 6867 only older parts like r300. 6868 6869 This will avoid display corruption problems reported by people 6870 using glamor on older asics by falling back to EXA if they try 6871 and enable glamor. 6872 6873 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6874 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 6875 6876commit d0323622ee9b97a5f246baffbb2c65930a78ed14 6877Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6878Date: Wed Aug 7 11:27:07 2013 +0200 6879 6880 bump version post release 6881 6882 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6883 6884commit 9c97cca5c24409ca8447c99f051a12fd2d494e79 6885Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6886Date: Wed Aug 7 10:48:17 2013 +0200 6887 6888 radeon: bump version for release 6889 6890 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6891 6892commit 16270cfb202ab67dd152644ef019b2f1ee4d0341 6893Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6894Date: Wed Aug 7 10:29:33 2013 +0200 6895 6896 add bicubic_table.py to EXTRA_DIST 6897 6898 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6899 6900commit 2cb9197ca7a337c911f38b5de562a2364b922b86 6901Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6902Date: Wed Aug 7 10:28:52 2013 +0200 6903 6904 kill unused radeon_driver.c 6905 6906 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 6907 6908commit c5cbfcf575b0b4aea6f797558ae974c1453c8e07 6909Author: Alex Deucher <alexander.deucher@amd.com> 6910Date: Tue Jul 30 10:08:25 2013 -0400 6911 6912 drmmode: add support for multi-screen reverse optimus 6913 6914 Initial reverse optimus didn't consider multiple screens, so 6915 this overhauls the code to use the new X server interface, 6916 and allows for multiple outputs on the dGPU to be used with 6917 the iGPU doing the rendering. Ported from Dave's nouveau 6918 patch. 6919 6920 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6921 6922commit 429d5b797769895eb4f5fef816ce4e2f3a342031 6923Author: Dave Airlie <airlied@redhat.com> 6924Date: Tue Jan 8 15:56:37 2013 +1000 6925 6926 radeon: add support for reverse prime (v2) 6927 6928 This adds support for reverse prime configurations 6929 6930 v2: fix compilation with older xservers 6931 6932 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 6933 6934commit 4de9356a2900ae0fb380a2350791ef045629cd05 6935Author: Alex Deucher <alexander.deucher@amd.com> 6936Date: Mon Aug 5 17:57:16 2013 -0400 6937 6938 radeon: fix naming clashes with multiple GPUs (v3) 6939 6940 The compat naming code for UMS causes problems 6941 with multiple GPU as you may end up with the same 6942 output name on multiple GPUs. Adjust the naming on 6943 secondary GPUs to avoid conflicts. 6944 6945 v2: integrate Dave's fixes for nouveau 6946 v3: keep compat with existing naming on primary GPU 6947 6948 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6949 6950commit 2ae6bb18fefddb309920fa69c9b56c3a7f3db7b4 6951Author: Grigori Goronzy <greg@chown.ath.cx> 6952Date: Wed Jul 31 12:01:20 2013 +0200 6953 6954 EXA/evergreen/ni: replace magic number 6955 6956 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 6957 6958commit 6a278369c05a298a4367306d986467a9ceacae8c 6959Author: Raul Fernandes <rgfernandes@gmail.com> 6960Date: Tue Jul 30 09:26:05 2013 -0400 6961 6962 EXA/6xx/7xx: optimize non-overlapping Copy 6963 6964 In case dst and src rectangles of a Copy operation in the same surface 6965 don't overlap, it is safe to skip the scratch surface. This is a 6966 common case. 6967 6968 Based on evergreen/ni patch from Grigori Goronzy. 6969 6970 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 6971 6972commit 4375a6e75e5d41139be7031a0dee58c057ecbd07 6973Author: Grigori Goronzy <greg@chown.ath.cx> 6974Date: Mon Jul 22 02:30:28 2013 +0200 6975 6976 EXA/evergreen/ni: accelerate PictOpOver with component alpha 6977 6978 Subpixel text rendering is typically done with a solid src and a 6979 pixmap mask. Traditionally, this cannot be accelerated in a single 6980 pass and requires two passes [1]. However, we can cheat a little 6981 with a constant blend color. 6982 6983 We can use: 6984 6985 const.A = src.A / src.A 6986 const.R = src.R / src.A 6987 const.G = src.G / src.A 6988 const.B = src.B / src.A 6989 6990 dst.A = const.A * (src.A * mask.A) + (1 - (src.A * mask.A)) * dst.A 6991 dst.R = const.R * (src.A * mask.R) + (1 - (src.A * mask.R)) * dst.R 6992 dst.G = const.G * (src.A * mask.G) + (1 - (src.A * mask.G)) * dst.G 6993 dst.B = const.B * (src.A * mask.B) + (1 - (src.A * mask.B)) * dst.B 6994 6995 This only needs a single source value. src.A is cancelled down in 6996 the right places. 6997 6998 [1] http://anholt.livejournal.com/32058.html 6999 7000commit 94d0d14914a025525a0766669b556eaa6681def7 7001Author: Grigori Goronzy <greg@chown.ath.cx> 7002Date: Thu Jul 18 16:06:23 2013 +0200 7003 7004 EXA/evergreen/ni: fast solid pixmap support 7005 7006 Solid pixmaps are currently implemented with scratch pixmaps, which 7007 is slow. This replaces the hack with a proper implementation. The 7008 Composite shader can now either sample a src/mask or use a constant 7009 value. 7010 7011commit 5bb04351c43a91a1d60348b7293544da05d75e72 7012Author: Grigori Goronzy <greg@chown.ath.cx> 7013Date: Fri Jul 27 17:31:53 2012 +0200 7014 7015 EXA/evergreen/ni: optimize non-overlapping Copy 7016 7017 In case dst and src rectangles of a Copy operation in the same surface 7018 don't overlap, it is safe to skip the scratch surface. This is a 7019 common case. 7020 7021commit c08e09b7bec441c4bf93b4cae4de1260754bf940 7022Author: Grigori Goronzy <greg@chown.ath.cx> 7023Date: Sat May 18 13:46:03 2013 +0200 7024 7025 Fix RADEON_FALLBACK logging 7026 7027commit c16c59f8f9b6aa7a4a6a6465582ad98f02a3606a 7028Author: Mark Kettenis <kettenis@openbsd.org> 7029Date: Sun Jul 7 13:44:13 2013 +0200 7030 7031 Always go through DFS/UTS when byte swapping is needed. 7032 7033 Before commit ef9bfb262db7004bef3704e5d914687e50d3fca4 and 7034 e5bd99faa3b6629a55168386d5dfa936ee4e97ae, byte swapping for the front buffer 7035 used to be done in hardware. Now that this no longer happens we need to let 7036 DFS/UTS ihandle the byte swapping. 7037 7038 See https://bugs.freedesktop.org/show_bug.cgi?id=66663 . 7039 7040 Signed-off-by: Mark Kettenis <kettenis@openbsd.org> 7041 7042commit 1239dbbd8c8d9b55756c1de52cad353171a06522 7043Author: Alex Deucher <alexander.deucher@amd.com> 7044Date: Sun Jul 7 20:22:10 2013 -0400 7045 7046 radeon: bump libdrm requirement to 2.4.46 for CIK support 7047 7048 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7049 7050commit 869e0b187b2d07e637c4482fe50d9f1b0c09f4e4 7051Author: Alex Deucher <alexander.deucher@amd.com> 7052Date: Thu Jan 24 21:12:28 2013 -0500 7053 7054 radeon/kms: add kabini pci ids 7055 7056 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7057 7058commit 0edcefc4c97e4572431c38fe170032470e22e0b7 7059Author: Alex Deucher <alexander.deucher@amd.com> 7060Date: Fri Jun 7 14:56:26 2013 -0400 7061 7062 radeon/kms: add bonaire pci ids 7063 7064 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7065 7066commit ebc32b27af25b23604e725eb50d844a8d26116bb 7067Author: Alex Deucher <alexander.deucher@amd.com> 7068Date: Fri Aug 31 18:17:24 2012 -0400 7069 7070 radeon: update cursor handling for CIK 7071 7072 CIK asics have 128x128 hw cursors 7073 7074 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7075 7076commit 4adaea996454b1ad5185f9c0f37667dbfc266495 7077Author: Alex Deucher <alexander.deucher@amd.com> 7078Date: Fri Jun 7 14:47:36 2013 -0400 7079 7080 radeon: add family ids for CIK 7081 7082 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7083 7084commit 3626ab147b6749a355e7d21710447f13b9796c77 7085Author: Leo Liu <leo.liu@amd.com> 7086Date: Fri Jun 14 10:27:20 2013 -0400 7087 7088 radeon: fix a memory leak in get_modes() callback. 7089 7090 leak happens when looping xrandr prop. 7091 7092 Signed-off-by: Leo Liu <leo.liu@amd.com> 7093 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7094 7095commit c0f7d03cb29166c6c00dd6fb698ba3f70c59dbdf 7096Author: Alex Deucher <alexdeucher@gmail.com> 7097Date: Mon Jun 17 11:38:27 2013 -0400 7098 7099 man: fix up previous commit 7100 7101 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7102 7103commit 03b5edc9dd3859eed2504155970613aaf88df852 7104Author: Eric S. Raymond <esr@thyrsus.com> 7105Date: Fri Jun 7 13:23:14 2013 -0400 7106 7107 Fix some incomplete list entries. 7108 7109 This necessarily involved adding some content, which I was able to 7110 Google for. Please review, correct, and make more specific. 7111 7112 Bodiless .TP entries screw up doclifter's page parsing. Please 7113 don't do that anymore. 7114 7115commit 7ce43639a8e53eefb94e5255a39c1232b6bdbf66 7116Author: Michel Dänzer <michel.daenzer@amd.com> 7117Date: Thu Jun 6 09:58:27 2013 +0200 7118 7119 autogen.sh: Restore passing --enable-maintainer-mode to configure 7120 7121 Looks like this was accidentally dropped when the GNOME Build API was 7122 implemented. 7123 7124commit 365e571d9de0b5979425c34210789afca4ea5f68 7125Author: Samuel Li <samuel.li@amd.com> 7126Date: Wed Mar 20 11:59:58 2013 -0400 7127 7128 radeon: use direct mapping for fast fb access. 7129 7130 Signed-off-by: Samuel Li <samuel.li@amd.com> 7131 7132commit bd2557ea5ef84b975060e929d5ece53ec464336f 7133Author: Ilija Hadzic <ilijahadzic@gmail.com> 7134Date: Wed May 8 22:39:48 2013 -0400 7135 7136 DRI2: add interpolated blanks to frame number in event handlers 7137 7138 The 'frame' argument passed to event handlers is the plain 7139 CRTC vblank counter that stops progressing when the 7140 associated display is in DPMS-off mode. If we have a 7141 DPMS-off period the frame counter and MSC will 7142 diverge, which can cause some higher-level functions 7143 to return incorrect values. 7144 7145 This patch fixes the problem by adding interpolated 7146 vblanks to the frame counter before using it in handler 7147 functions. 7148 7149 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7150 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7151 7152commit 385a92b4fbe8f53b359ef6c463704414d00476fa 7153Author: Ilija Hadzic <ilijahadzic@gmail.com> 7154Date: Wed May 8 22:39:47 2013 -0400 7155 7156 DRI2: support scheduling emulated events with zero delay 7157 7158 Now that we fully emulating a running CRTC through DPMS-off state 7159 it is possible to come up with a zero delay when scheduling 7160 a swap or MSC-wait (e.g., if a call into respective wait function 7161 was entered very late). This patch wraps the TimerSet function 7162 into our own radeon_dri2_schedule_event such that the latter 7163 calls the event right away if zero delay is specified. 7164 7165 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7166 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7167 7168commit e87b52e6ad41ad7a87e43b818d80e7d522d9c68d 7169Author: Ilija Hadzic <ilijahadzic@gmail.com> 7170Date: Wed May 8 22:39:46 2013 -0400 7171 7172 DRI2: make wait_msc and get_msc work on disabled CRTCs 7173 7174 Now that the running CRTC is emulated through DPMS-off 7175 states, wait_msc and get_msc must also work on disabled 7176 CRTCs. When CRTC is disabled, we must extrapolate the 7177 MSC count from present time and last MSC when CRTC was 7178 running and also use timers facility in wait_msc. 7179 7180 v2: CRTC-private now stores frame rate instead of nominal 7181 vblank period. 7182 7183 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7184 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7185 7186commit f1584152bbeda98cc1907359f1c7954a84c8837d 7187Author: Ilija Hadzic <ilijahadzic@gmail.com> 7188Date: Wed May 8 22:39:45 2013 -0400 7189 7190 DRI2: move rename and rework radeon_dri2_deferred_swap 7191 7192 radeon_dri2_deferred_swap will be used to generate 7193 real events (not just fallbacks) so now it needs to 7194 generate real timestamp and frame counter. Also 7195 this function will be used both by schedule_swap 7196 and wait_msc, so give it a more generic name: 7197 radeon_dri2_deferred_event 7198 7199 v2: - Extrapolate the frame number from the time 7200 of actual execution of the function instead 7201 of using the MSC calculated (extrapolated) 7202 at event scheduling time. 7203 - CRTC-private now stores frame rate instead of 7204 nominal vblank period. 7205 7206 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7207 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7208 7209commit 84bce7943b735aee19e26be4d47fdf140564f3f5 7210Author: Ilija Hadzic <ilijahadzic@gmail.com> 7211Date: Wed May 8 22:39:44 2013 -0400 7212 7213 DRI2: track CRTC in event record 7214 7215 Tracking the CRTC associated with an event will save us 7216 some lookups later in event handlers. 7217 7218 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7219 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7220 7221commit e40d5390b3efdea3e02267413350410d8e6a2970 7222Author: Ilija Hadzic <ilijahadzic@gmail.com> 7223Date: Wed May 8 22:39:43 2013 -0400 7224 7225 DRI2: hook up vblank extrapolation to schedule_swap 7226 7227 This patch hooks up swap-scheduling function with 7228 vblank-extrapolation function. Rather than waiting for 7229 fixed time, we calculate exactly how much we should wait 7230 and what we should update target_msc to using 7231 radeon_dri2_extrapolate_msc_delay helper function 7232 and schedule the swap completion using DIX's timer facility. 7233 7234 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7235 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7236 7237commit ffaa5abf207415159cdb28e90da49b95f497ef61 7238Author: Ilija Hadzic <ilijahadzic@gmail.com> 7239Date: Wed May 8 22:39:42 2013 -0400 7240 7241 DRI2: add vblank extrapolation function 7242 7243 Implement a helper function that will be called when emulating 7244 the running CRTC. The function should be called only when CRTC 7245 is in DPMS-off state. It will look at the vblank count and the 7246 time that was recorded last time the CRTC was running and 7247 calculate how long one must wait (from present time) until 7248 the target_msc is reached if the CRTC were running. 7249 7250 v2: - CRTC-private now stores frame rate instead of nominal 7251 vblank period. 7252 - DIX's timer facility can sometimes wake up the scheduled 7253 functions more than a millisecond earlier. To avoid 7254 generating an old MSC, we have to add more margin when 7255 converting the delay in microseconds to milliseconds. 7256 7257 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7258 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7259 7260commit 80ae2291d082b57c70d27a80182f00d760fb3d3a 7261Author: Ilija Hadzic <ilijahadzic@gmail.com> 7262Date: Wed May 8 22:39:41 2013 -0400 7263 7264 DRI2: adjust MSC by the number of interpolated vblanks 7265 7266 Normally, MSC value equals the vblank count of the CRTC 7267 on which the drawable resides. However, we are now interpolating 7268 vblanks through DPMS-off state, so if the CRTC is turned 7269 off at some point, the vlbank count does not change, but 7270 MSC does change by the amount that equals the cumulative 7271 number of interpolated vblanks. 7272 7273 So each time we interact with the CRTC's vblank facility, 7274 we have to adjust the vblank counter: in the request, we 7275 have to subtract the number of interpolated vblanks from 7276 MSC and in the reply we have to add the number of interpolated 7277 vblanks to the MSC. 7278 7279 This patch only makes the MSC correct when CRTC is on. 7280 The subsequent patch will construct the MSC when CRTC is not 7281 running. 7282 7283 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7284 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7285 7286commit 5fd2eb5d12cea32927a9e6c6ce4afd18aa7d046a 7287Author: Ilija Hadzic <ilijahadzic@gmail.com> 7288Date: Wed May 8 22:39:40 2013 -0400 7289 7290 drmmode: calculate interpolated vblanks while in dpms-off state 7291 7292 This adds provisions for interpolating vblanks while the CRTC 7293 is in DPMS-off state. When entering DPMS-off state, we 7294 record the last vblank time, sequence number and frame rate 7295 in CRTC-private structure. 7296 7297 When going back to DPMS-on state we read the current 7298 time and calculate how long we have been off. Then we derive 7299 how many vblanks that would have been had the CRTC remained 7300 running. These are the interpolated vblanks. 7301 7302 Finally, we accumulate the number of interpolated vblanks 7303 in CRTC-private structure to get the number of interpolated 7304 vblanks over the system lifetime. 7305 7306 v2: Track frame rate instead of vblank period. The former 7307 eliminates some roundoff errors. 7308 7309 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7310 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7311 7312commit e41ad30d09be4962cfb1942b0b9f63875dbb2d2a 7313Author: Ilija Hadzic <ilijahadzic@gmail.com> 7314Date: Wed May 8 22:39:39 2013 -0400 7315 7316 drmmode: remove dead code from drmmode_crtc_dpms 7317 7318 We are about to modify this function, so let's clean it 7319 up first. 7320 7321 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7322 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7323 7324commit 373671d2eed30e30b3bdee7e40426cf58c127234 7325Author: Ilija Hadzic <ilijahadzic@gmail.com> 7326Date: Wed May 8 22:39:38 2013 -0400 7327 7328 drmmode: add drmmode_get_current_ust function 7329 7330 The new helper function retrieves current time in the format 7331 that is compatible with vblank timestamps. 7332 7333 v2: - fix an incorrect statement in a comment 7334 - add a #define so that don't depend on libdrm patches that 7335 have not yet been accepted upstream 7336 7337 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7338 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7339 7340commit 006fbbd1d38a089b50ab3197d32815689ed249fa 7341Author: Ilija Hadzic <ilijahadzic@gmail.com> 7342Date: Wed May 8 22:39:37 2013 -0400 7343 7344 DRI2: make populate_vbl_request_type external 7345 7346 We'll need to use this function in other .c files so 7347 remove static annotation from it. Also, add radeon_ 7348 prefix because this function is now part of global namespace, 7349 so it would be nice to know where it is coming from. 7350 7351 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7352 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7353 7354commit 34660d8757ff16acd1686ff2872f4600d92a68bf 7355Author: Ilija Hadzic <ilijahadzic@gmail.com> 7356Date: Wed May 8 22:39:36 2013 -0400 7357 7358 DRI2: change signature of populate_vbl_request_type 7359 7360 This function no longer uses info argument, so nuke it. 7361 7362 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7363 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7364 7365commit ab6e85d1f5dd4657591b314b85fd2453f99c3712 7366Author: Michel Dänzer <michel.daenzer@amd.com> 7367Date: Fri Mar 22 10:51:05 2013 +0100 7368 7369 glamor: Keep old private in radeon_pixmap_set_bo 7370 7371commit 0a84f2760eccf07781ae6c3f55d2d246c155c9ce 7372Author: Alex Deucher <alexander.deucher@amd.com> 7373Date: Mon May 13 17:06:00 2013 -0400 7374 7375 radeon: add Hainan pci ids 7376 7377 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7378 7379commit 6cee1eaebfcef2b2d2c51ec30c31260fd98b564e 7380Author: Alex Deucher <alexander.deucher@amd.com> 7381Date: Mon May 13 17:01:30 2013 -0400 7382 7383 radeon: add hainan chip family 7384 7385 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7386 7387commit 8a9e9b21cbb44c23d37bf18c983e2c3f0fe33e87 7388Author: Michel Dänzer <michel.daenzer@amd.com> 7389Date: Thu Apr 25 17:10:12 2013 +0200 7390 7391 glamor: Fix build against xserver < 1.13 7392 7393 Those didn't define CREATE_PIXMAP_USAGE_SHARED. 7394 7395commit 6216b5aa3793abe71b864170bc9d4d745f027328 7396Author: Alex Deucher <alexander.deucher@amd.com> 7397Date: Thu Apr 25 14:19:37 2013 -0400 7398 7399 radeon: add new richland pci ids 7400 7401 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7402 7403commit fd5903f15b36555bf1ccc1c3865c9fffd499128a 7404Author: Alex Deucher <alexander.deucher@amd.com> 7405Date: Thu Apr 25 14:17:51 2013 -0400 7406 7407 radeon: add new SI pci ids 7408 7409 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7410 7411commit b2826580e4d246d45871b413e4deecb070162577 7412Author: Jerome Glisse <jglisse@redhat.com> 7413Date: Mon Apr 8 13:45:18 2013 -0400 7414 7415 radeonsi: add 2d tiling support 7416 7417 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 7418 7419commit 6e74aacc5e5da3b51744153dad1645caa6ea4ce3 7420Author: Michel Dänzer <michel.daenzer@amd.com> 7421Date: Wed Mar 27 11:57:29 2013 +0100 7422 7423 glamor: Only enable for depth >= 24 7424 7425 It can't handle lower depths yet. 7426 7427 Reported-by: Igor Vagulin <igor.vagulin@gmail.com> 7428 7429commit 33d8408eec806355c2e55726679ec50ef3b769f1 7430Author: Michel Dänzer <michel.daenzer@amd.com> 7431Date: Tue Mar 12 19:03:06 2013 +0100 7432 7433 glamor: Enable by default on SI 7434 7435 We should now handle its major initialization failure cases gracefully. 7436 7437 Reviewed-by: Jerome Glisse <jglisse@redhat.com> 7438 7439commit ad1d8874fab30257ba8ec8bee8ffa42a8fa9b5c0 7440Author: Michel Dänzer <michel.daenzer@amd.com> 7441Date: Fri Mar 15 18:28:27 2013 +0100 7442 7443 glamor: Initial PRIME pixmap sharing hooks. 7444 7445 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=57200 7446 7447 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7448 Tested-by: Christoph Haag <haagch.christoph@googlemail.com> 7449 Tested-by: linedot <linedot@xcpp.org> 7450 7451commit 35b384a00a0c76ea556f55787fccc95ecd51bc0c 7452Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7453Date: Thu Mar 21 15:34:21 2013 +0100 7454 7455 drmmode: handle failure of drmModeGetConnector 7456 7457 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7458 7459commit 751c7259cf23a29fdd2653881e2db63ce3cf1ca5 7460Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7461Date: Thu Mar 21 10:40:42 2013 +0100 7462 7463 Move fd cleanup to correct place. 7464 7465 Oops, just after previous commit I realized that it 7466 needs to be done from RADEONFreeRec, because when preinit fails 7467 info pointer can be null in FreeScreen. 7468 7469 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7470 7471commit 1643b8cd2dc53ed36916e11003590c7037b4ddd6 7472Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7473Date: Thu Mar 21 10:33:19 2013 +0100 7474 7475 radeon: add refcounts to fix up zaphod open/close. 7476 7477 Oops, turns out my previous commits were buggy. 7478 Adding proper refcounts will handle this correctly. 7479 7480 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7481 7482commit a4d1bc43d4d10e2ca1e512c1010e962c587d6c16 7483Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7484Date: Wed Mar 20 17:13:44 2013 +0100 7485 7486 Cleanup drmmode in CloseScreen. 7487 7488 Prevents a use-after-free. 7489 7490 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7491 7492commit 3cb3ab8704813d7c14b9e547a30e37598d70c85a 7493Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7494Date: Wed Mar 20 17:12:29 2013 +0100 7495 7496 Close drm_fd in CloseScreen. 7497 7498 This prevents a small fd leak. 7499 7500 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7501 7502commit 462d8d3b86038438ab854072e546cc7a9efacdd7 7503Author: Michel Dänzer <michel.daenzer@amd.com> 7504Date: Tue Mar 12 18:59:19 2013 +0100 7505 7506 glamor: Bail if the glamoregl module wasn't loaded early 7507 7508 Trying to initialize glamor in that case crashes. 7509 7510 Reviewed-by: Jerome Glisse <jglisse@redhat.com> 7511 7512commit 3e5350bedf50a1a140e0eba08c39ed2451fb9115 7513Author: Alex Deucher <alexander.deucher@amd.com> 7514Date: Fri Mar 8 14:07:42 2013 -0500 7515 7516 radeon: add Richland pci ids 7517 7518 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7519 7520commit bdbf766d8170f6ea6c16b9780ef5a4007730588b 7521Author: Alex Deucher <alexander.deucher@amd.com> 7522Date: Fri Mar 8 14:00:09 2013 -0500 7523 7524 radeon: mark trinity APUs with IGP flag 7525 7526 Was missing on trinity. 7527 7528 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7529 7530commit 5ac335eaae135c0e66cc4643908f41174e2f1123 7531Author: Michel Dänzer <michel.daenzer@amd.com> 7532Date: Fri Mar 8 17:37:41 2013 +0100 7533 7534 Only parse Option "AccelMethod" if glamor was enabled at build time 7535 7536 Will hopefully make it more obvious when it wasn't. 7537 7538commit f1059e3f0934ccf1ae183099dbd4de700a7c81ac 7539Author: Colin Walters <walters@verbum.org> 7540Date: Wed Jan 4 17:37:06 2012 -0500 7541 7542 autogen.sh: Implement GNOME Build API 7543 7544 http://people.gnome.org/~walters/docs/build-api.txt 7545 7546 Signed-off-by: Adam Jackson <ajax@redhat.com> 7547 7548commit 343b01c9bf35b125cd0c3df8db7c01a5fb227bda 7549Author: Alex Deucher <alexander.deucher@amd.com> 7550Date: Thu Jan 24 19:57:23 2013 -0500 7551 7552 radeon/kms: add oland pci ids 7553 7554 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7555 7556commit 9980f1e0e0a5df5fb0c791687e5c98c13089d7d1 7557Author: Alex Deucher <alexander.deucher@amd.com> 7558Date: Thu Jan 24 19:52:04 2013 -0500 7559 7560 radeon: add oland chip family 7561 7562 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7563 7564commit d89e545826969910226cb96e73b64f18e92d2064 7565Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7566Date: Wed Jan 30 17:21:05 2013 +0100 7567 7568 bump version post release 7569 7570 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7571 7572commit 21a4f3eaca98c1c54bffcdf3926c631cec9c409d 7573Author: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7574Date: Wed Jan 30 16:55:27 2013 +0100 7575 7576 radeon: bump version for release 7577 7578 And fixup distcheck.. 7579 7580 Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> 7581 7582commit 102ee4a24e4c9ba0c0b65d3ca8599dc7e0216c1e 7583Author: Alex Deucher <alexdeucher@gmail.com> 7584Date: Fri Jan 11 09:52:32 2013 -0500 7585 7586 radeon/r200: program RE_WIDTH_HEIGHT properly 7587 7588 Values are inclusive. Mesa already sets these 7589 correctly. Also bump EXA/Xv limits from 2047 to 7590 2048. 7591 7592 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7593 7594commit 47689ef8b878a56ea094e21a6090b228e67d605e 7595Author: Michel Dänzer <michel.daenzer@amd.com> 7596Date: Fri Jan 11 15:41:48 2013 +0100 7597 7598 Make radeon_glamor_create_textured_pixmap() return TRUE without USE_GLAMOR. 7599 7600 Otherwise rotation won't work when glamor support isn't built in. 7601 7602 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7603 7604commit 1a8c05fd5559be842d8a038df73de34d4b261c15 7605Author: Michel Dänzer <michel.daenzer@amd.com> 7606Date: Fri Jan 11 13:22:24 2013 +0100 7607 7608 Make CRTC shadow pixmaps usable by glamor. 7609 7610 Fixes rotation with glamor. 7611 7612 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7613 7614commit e5e22d788b103c04d3460a221a3b9dc3be69e0a4 7615Author: Michel Dänzer <michel.daenzer@amd.com> 7616Date: Thu Jan 10 17:00:16 2013 +0100 7617 7618 DRI2/r[345]xx: 16 bit depth buffers use square micro tiling. 7619 7620 Fixes piglit sanity.tests (with 3D drivers which allocate depth buffers from 7621 the X server). 7622 7623 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7624 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7625 7626commit f5d2682584c74587890a9795ece7a8917214a83f 7627Author: Michel Dänzer <michel.daenzer@amd.com> 7628Date: Thu Jan 10 17:00:15 2013 +0100 7629 7630 DRI2: Properly use buffer depth passed in by client. 7631 7632 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7633 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7634 7635commit 4e35b2f530e2ca8c7b7220cacd05c661de43d20d 7636Author: Michel Dänzer <michel.daenzer@amd.com> 7637Date: Thu Jan 10 12:10:52 2013 +0100 7638 7639 Drop support for X servers older than 1.7. 7640 7641 The main purpose is to drop DRI2 compatibility code paths which are getting 7642 awkward to deal with and at this point are probably only tested lightly if at 7643 all. xserver 1.7 was released more than 3 years ago, and 1.6.2 was already 7644 required before. 7645 7646 Mostly mechanical, there's probably potential for more cleanups. 7647 7648 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 7649 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7650 7651commit 857d729eb3672443f89c0516f8226eacb37dad2b 7652Author: Dave Airlie <airlied@redhat.com> 7653Date: Mon Jan 7 14:13:33 2013 +1000 7654 7655 radeon: fix damage reporting for slave pixmaps 7656 7657 This is definitely more correct, not sure if it'll fix any bugs. 7658 7659 Signed-off-by: Dave Airlie <airlied@redhat.com> 7660 7661commit 6981a5c087165b126c15ba0025cffdba218ab652 7662Author: Ilija Hadzic <ihadzic@research.bell-labs.com> 7663Date: Wed Dec 19 10:35:43 2012 -0500 7664 7665 DRI2: limit the swap rate when CRTC is in DPMS-off state 7666 7667 If drawable is displayed on a CRTC and relevant CRTC is in 7668 DPMS off state, defer the swap by a fixed (hard-coded) time. 7669 7670 This patch fixes a bug that caused an application to render 7671 at uncontrolled rate when CRTC goes into DPMS "off" state, 7672 thus thrashing the GPU and CPU and likely offsetting the 7673 power savings achieved by shutting off the display. 7674 7675 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7676 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 7677 7678commit 3657672207322be651cdb94a811337b7c5668c84 7679Author: Ilija Hadzic <ihadzic@research.bell-labs.com> 7680Date: Wed Dec 19 10:35:42 2012 -0500 7681 7682 DRI2: change signature of radeon_dri2_drawable_crtc 7683 7684 Return pointer to the CRTC instead of CRTC ID and expose 7685 consider_disabled option in arguments of this function. 7686 7687 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7688 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 7689 7690commit 61d0aec40e2521488c2fe43e7a6823e5c87d94d7 7691Author: Ilija Hadzic <ihadzic@research.bell-labs.com> 7692Date: Wed Dec 19 10:35:41 2012 -0500 7693 7694 video: add option to include disabled CRTCs in best CRTC search 7695 7696 This patch adds an option called consider_disabled to 7697 radeon_pick_best_crtc function. If this option is set 7698 and searching for best-fit CRTC yields nothing, the search 7699 will be widened to include CRTCs in DPMS "off" state. 7700 The new option is not used yet; it will be in the patches 7701 to follow. 7702 7703 v2: Do not look at the CRTC mode, we only care about its box. 7704 7705 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7706 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 7707 7708commit a60d2152e928a7011fc7c44a885a34c3cdd4f0fe 7709Author: Michel Dänzer <michel.daenzer@amd.com> 7710Date: Thu Dec 27 12:13:57 2012 +0100 7711 7712 glamor: Enable full 2D acceleration for SI as well. 7713 7714 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7715 7716commit 04ebe14fe775d18b534e89b50b90ee293ae1910e 7717Author: Michel Dänzer <michel.daenzer@amd.com> 7718Date: Thu Dec 27 11:54:27 2012 +0100 7719 7720 glamor: Release the drawable after passing to glamor_push_pixels 7721 7722 Ported from xf86-video-intel commit ca5c028c2b4d9bf02002acd484054fe427ea8d09 . 7723 7724 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7725 7726commit cee7d61b144162974238b282d76a70aa2f14cc0d 7727Author: Alex Deucher <alexdeucher@gmail.com> 7728Date: Tue Dec 18 15:49:47 2012 -0500 7729 7730 exa/Xv: fix coordinate limits on AVIVO IGPs 7731 7732 RS6xx asics are r4xx derived, but seem to have r3xx 7733 limitations as far as clipping is concerned. Spotted 7734 by Michel on IRC. 7735 7736 Fixes: 7737 https://bugs.freedesktop.org/show_bug.cgi?id=58469 7738 7739 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7740 7741commit 3c7d024c2671dac541b8f2daed55040f4fd5d62d 7742Author: Michel Dänzer <michel.daenzer@amd.com> 7743Date: Wed Dec 12 10:43:55 2012 +0100 7744 7745 Remove dead option definitions. 7746 7747 These were only used with UMS. 7748 7749 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7750 7751commit bdbd84a6c99943f4e012c92ad9cb920cbd676621 7752Author: Michel Dänzer <michel.daenzer@amd.com> 7753Date: Mon Dec 10 23:35:41 2012 +0100 7754 7755 Remove local variables total_size_bytes. 7756 7757 Not actually used for anything. Refactoring gone wrong? 7758 7759 Also remove a comment that doesn't seem to make sense anymore. 7760 7761 Stumbled over this due to a warning by clang: 7762 7763 ../../src/radeon_kms.c:1562:6: warning: variable 'total_size_bytes' is uninitialized when used here [-Wuninitialized] 7764 total_size_bytes += (64 * 4 * 64); 7765 ^~~~~~~~~~~~~~~~ 7766 ../../src/radeon_kms.c:1557:25: note: initialize the variable 'total_size_bytes' to silence this warning 7767 int total_size_bytes; 7768 ^ 7769 = 0 7770 7771 Reviewed-by: Alex Deucher <aleander.deucher@amd.com> 7772 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7773 7774commit cf0a81547bcde32bdd2b080c0a546a49dd2eb628 7775Author: Michel Dänzer <michel.daenzer@amd.com> 7776Date: Tue Dec 11 11:48:33 2012 +0100 7777 7778 Remove unused local variable 'path'. 7779 7780 ../../src/radeon_probe.c: In function 'radeon_platform_probe': 7781 ../../src/radeon_probe.c:270:11: error: unused variable 'path' [-Werror=unused-variable] 7782 7783 Reviewed-by: Alex Deucher <aleander.deucher@amd.com> 7784 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7785 7786commit d268038a5bdd1655f7533c8dd2edf74fbc82bbee 7787Author: Michel Dänzer <michel.daenzer@amd.com> 7788Date: Mon Jul 2 09:07:14 2012 +0200 7789 7790 Remove dead code flagged by gcc -Wunused-but-set-variable. 7791 7792 Reviewed-by: Alex Deucher <aleander.deucher@amd.com> 7793 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 7794 7795commit 793e1b0ea579cce7489aee57a95f20384d6ecce8 7796Author: Alex Deucher <alexdeucher@gmail.com> 7797Date: Thu Dec 6 09:21:19 2012 -0500 7798 7799 radeon: fix ring count in R300PrepareTexturedVideo 7800 7801 Fixes: 7802 https://bugs.freedesktop.org/show_bug.cgi?id=31364 7803 7804 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7805 7806commit f402805b22e4f4f735924d59c69474b24e623a25 7807Author: Jerome Glisse <jglisse@redhat.com> 7808Date: Fri Nov 30 10:07:51 2012 -0500 7809 7810 radeon: avoid copying over itself ddx fb 7811 7812 In some rare case it seems that the old fb could already be the 7813 ddx fb. As copying from the same bo will trigger a ttm deadlck 7814 detection which will result in infinite kernel/userspace loop. 7815 Avoid doing any copy as anyway it's useless. 7816 7817 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 7818 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 7819 7820commit 60cd6ceaf44b506433aebf6b3a639a17604dfddd 7821Author: Alex Deucher <alexdeucher@gmail.com> 7822Date: Wed Nov 21 18:42:56 2012 -0500 7823 7824 radeon: add new SI pci id 7825 7826 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7827 7828commit 53af6aa99dfb67b5223c28d5129a79694165048d 7829Author: Jerome Glisse <jglisse@redhat.com> 7830Date: Wed Nov 14 12:15:47 2012 -0500 7831 7832 radeon: avoid segfault when pixmap exceed GPU capabilities 7833 7834 We might get a request for a pixmap/drawable that is too big 7835 for GPU capabilities in the dri2 get buffer path. In face of 7836 such things just return NULL to dri2 get buffer request. The 7837 GL driver should then use something like transparent black to 7838 avoid something too ugly on the screen. 7839 7840 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 7841 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 7842 7843commit 71eeb0833489ece745f1e5dc41379ec63161ee81 7844Author: Alex Deucher <alexdeucher@gmail.com> 7845Date: Tue Nov 6 09:43:55 2012 -0500 7846 7847 bump verion post release 7848 7849 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7850 7851commit cf377c991f4c1cd4eee1a8530ba9fdf6f9f63301 7852Author: Alex Deucher <alexdeucher@gmail.com> 7853Date: Tue Nov 6 09:12:39 2012 -0500 7854 7855 radeon: bump version for release 7856 7857 7.0.0, RIP UMS 7858 7859 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7860 7861commit 20bfc652ce40008ea561db2984bccf137409c7fd 7862Author: Ilija Hadzic <ihadzic@research.bell-labs.com> 7863Date: Mon Sep 24 17:19:09 2012 -0400 7864 7865 radeon/radeon_platform_probe: fix Zaphod mode breakage 7866 7867 Using radeon_platform_probe function breaks the Zaphod mode because 7868 it attempts to call xf86AddEntityToScreen multiple times, but nobody 7869 calls xf86SetEntityShared prior to that. Consequently, calls for all 7870 but first device instance fail. 7871 7872 Prior to introduction of platform bus, the logic was that the Probe 7873 function would make the entity sharable, which would cause Xserver 7874 to later make it shared prior to adding it to screen. With the 7875 platform bus loading, add to screen happens in the probe 7876 function so we have to make it shared there. 7877 7878 v2: do not make the entity shared if it was not previously marked 7879 sharable. Should fix: 7880 7881 https://bugs.freedesktop.org/show_bug.cgi?id=56680 7882 https://bugs.freedesktop.org/show_bug.cgi?id=56663 7883 7884 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7885 7886commit b2807b53083af5db0b85645666b0127be16623c7 7887Author: Alex Deucher <alexdeucher@gmail.com> 7888Date: Fri Nov 2 17:57:18 2012 -0400 7889 7890 man: note that 2D tiling requires Mesa 9.0 or newer 7891 7892 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7893 7894commit 1038e29162d869e2ce37f9b8f5f7c8ea37c3e53d 7895Author: Alex Deucher <alexdeucher@gmail.com> 7896Date: Fri Nov 2 10:04:46 2012 -0400 7897 7898 Revert "radeon/radeon_platform_probe: fix Zaphod mode breakage" 7899 7900 This reverts commit 39eac3104c2f08b4d78aab3f88fd104301eb4711. 7901 7902 This breaks non-zaphod configurations with xserver 1.13. 7903 See: 7904 https://bugs.freedesktop.org/show_bug.cgi?id=56680 7905 https://bugs.freedesktop.org/show_bug.cgi?id=56663 7906 7907commit d748894c84e891662cde6d7a5748f74a5aa5175e 7908Author: Alex Deucher <alexdeucher@gmail.com> 7909Date: Fri Nov 2 08:50:16 2012 -0400 7910 7911 man: update man pages for trinity (ARUBA) 7912 7913 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7914 7915commit 6740e8e90b934611c3bb0f9e8732d66e25b8c727 7916Author: Andreas Boll <andreas.boll.dev@gmail.com> 7917Date: Fri Nov 2 13:30:44 2012 +0100 7918 7919 radeon/man: ColorTiling2D is by default on for r6xx-cayman 7920 7921 Signed-off-by: Andreas Boll <andreas.boll.dev@gmail.com> 7922 7923commit 39eac3104c2f08b4d78aab3f88fd104301eb4711 7924Author: Ilija Hadzic <ihadzic@research.bell-labs.com> 7925Date: Mon Sep 24 17:41:32 2012 -0400 7926 7927 radeon/radeon_platform_probe: fix Zaphod mode breakage 7928 7929 Using radeon_platform_probe function breaks the Zaphod mode because 7930 it attempts to call xf86AddEntityToScreen multiple times, but nobody 7931 calls xf86SetEntityShared prior to that. Consequently, calls for all 7932 but first device instance fail. 7933 7934 Prior to introduction of platform bus, the logic was that the Probe 7935 function would make the entity sharable, which would cause Xserver 7936 to later make it shared prior to adding it to screen. With the 7937 platform bus loading, add to screen happens in the probe 7938 function so we have to make it shared there. 7939 7940 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 7941 7942commit 2e96aa641bc8a5c263a5b643961383cdf530d044 7943Author: Marek Olšák <maraeo@gmail.com> 7944Date: Sun Oct 28 12:50:49 2012 -0400 7945 7946 radeon: turn on 2D tiling by default on r6xx-cayman 7947 7948 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7949 7950commit e830c039479539735d0836082abef24f6841c803 7951Author: Marek Olšák <maraeo@gmail.com> 7952Date: Sun Oct 28 12:49:15 2012 -0400 7953 7954 radeon: log if 2D tiling is enabled 7955 7956 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 7957 7958commit ce1b745dcb60dc516ad999756240b78e72a7aa54 7959Author: Alex Deucher <alexander.deucher@amd.com> 7960Date: Tue Oct 16 13:05:46 2012 -0400 7961 7962 radeon: add new pci ids for SI 7963 7964 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 7965 7966commit e3519be285bdab50ef2193f98826a351a4414fae 7967Author: Dave Airlie <airlied@redhat.com> 7968Date: Tue Oct 16 16:09:23 2012 +1000 7969 7970 radeon/dri2: fix the window->pixmap math properly. 7971 7972 Okay kwin did things different, and I really hope this time this is the 7973 correct transformation operation. It works here with gears under kwin 7974 and under mutter and under xcompmgr. 7975 7976 Signed-off-by: Dave Airlie <airlied@redhat.com> 7977 7978commit bd9e2c064d93afbe4ab84febb2a43ae6a95ab17b 7979Author: Dave Airlie <airlied@redhat.com> 7980Date: Thu Oct 11 14:36:11 2012 +1000 7981 7982 radeon/dri2: fix damage reporting on pageflip 7983 7984 when you use MPX or slave usb devices, the damage posting was incorrect, 7985 and resulted in misrendered boxes on the screen. 7986 7987 Fix it by reporting damage correctly. 7988 7989 Signed-off-by: Dave Airlie <airlied@redhat.com> 7990 7991commit e8cb0b721e6ea251f85c799ca0563bfa59a2d37c 7992Author: Adam Jackson <ajax@redhat.com> 7993Date: Wed Sep 26 09:27:54 2012 -0400 7994 7995 Remove mibstore.h 7996 7997 Signed-off-by: Adam Jackson <ajax@redhat.com> 7998 7999commit 8637f772347c958fa3beea6c0dc9c22255db70e5 8000Author: Michel Dänzer <michel.daenzer@amd.com> 8001Date: Thu Sep 13 19:50:35 2012 +0200 8002 8003 glamor: Force acceleration for DRI2 copies. 8004 8005 Improves performance of non-pageflipped 3D apps on SI. 8006 8007 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8008 8009commit 40b87f0e9db2642ca8a70e994fb4d3742e199640 8010Author: Michel Dänzer <michel.daenzer@amd.com> 8011Date: Wed Sep 12 18:44:10 2012 +0200 8012 8013 glamor: Fix DRI2 crash introduced by PRIME changes. 8014 8015 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8016 8017commit 646ffd49d2cf0d3130301e8ad2f49f620bd9ef1b 8018Author: Dave Airlie <airlied@redhat.com> 8019Date: Mon Sep 10 15:31:06 2012 +1000 8020 8021 radeon: fix build against old servers 8022 8023 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=54711 8024 8025 Reported-by: wolput@onsneteindhoven.nl 8026 Signed-off-by: Dave Airlie <airlied@redhat.com> 8027 8028commit 7c7f2775604da61537c24c249884aaa220e06e13 8029Author: Dave Airlie <airlied@redhat.com> 8030Date: Mon Sep 10 11:27:45 2012 +1000 8031 8032 radeon: don't fail to load if we have no modes and gpu driver. 8033 8034 This ensures radeon loads on output less GPUs as a GPU driver. 8035 8036 There are server bugs that also need to be fixed. 8037 8038 Signed-off-by: Dave Airlie <airlied@redhat.com> 8039 8040commit 7fe16dd037ac688726869816888bd77fe4356f8e 8041Author: Dave Airlie <airlied@gmail.com> 8042Date: Wed Sep 5 10:41:51 2012 +1000 8043 8044 radeon: fix typo in glamor paths 8045 8046 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54519 8047 8048 Reported-By: scaty@dcinformatique.com 8049 Signed-off-by: Dave Airlie <airlied@redhat.com> 8050 8051commit cdc985e9190ce54dc6274ebc0e87ab954b15bb0d 8052Author: Dave Airlie <airlied@gmail.com> 8053Date: Tue Sep 4 18:18:50 2012 +1000 8054 8055 radeon: fix build against older servers 8056 8057 Another silly one. 8058 8059 Signed-off-by: Dave Airlie <airlied@redhat.com> 8060 8061commit bd0e9c51407cb99f1fe4f0acfedc37ee0c354040 8062Author: Dave Airlie <airlied@gmail.com> 8063Date: Tue Sep 4 18:17:06 2012 +1000 8064 8065 radeon/dri2: fix build against older servers. 8066 8067 Older servers won't have DRI2UpdatePrime, so fix build. 8068 8069 bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54433 8070 Signed-off-by: Dave Airlie <airlied@redhat.com> 8071 8072commit 75c1f62f44d7b3709c1fa103d59733866967b841 8073Author: Dave Airlie <airlied@redhat.com> 8074Date: Mon Sep 3 13:09:02 2012 +1000 8075 8076 radeon: add capability and provider support 8077 8078 Signed-off-by: Dave Airlie <airlied@redhat.com> 8079 8080commit ca8b61e916998458203dfa4b9d0c997fa7d90ba9 8081Author: Dave Airlie <airlied@redhat.com> 8082Date: Mon Sep 3 13:08:37 2012 +1000 8083 8084 radeon: allocated shared pixmaps as untiled gtt. 8085 8086 Signed-off-by: Dave Airlie <airlied@redhat.com> 8087 8088commit 8c1bf9d8fe3948b72795984e625ef46b2f0bf654 8089Author: Dave Airlie <airlied@redhat.com> 8090Date: Mon Sep 3 13:08:05 2012 +1000 8091 8092 radeon: add pixmap sharing hooks. 8093 8094 This hooks into EXA and the dirty tracking to add sharing and output 8095 offload support. 8096 8097 Signed-off-by: Dave Airlie <airlied@redhat.com> 8098 8099commit 3add8df8122697acfe126d4857f3946ce44b8305 8100Author: Dave Airlie <airlied@redhat.com> 8101Date: Mon Sep 3 13:07:07 2012 +1000 8102 8103 radeon: add dri2 offload support. 8104 8105 This adds the new dri2 interfaces required for radeon to be offloaded to. 8106 8107 Signed-off-by: Dave Airlie <airlied@redhat.com> 8108 8109commit f7502a11c8ef9c453ceb40d26109977116df88c2 8110Author: Dave Airlie <airlied@redhat.com> 8111Date: Mon Sep 3 12:56:56 2012 +1000 8112 8113 radeon: add shared support to pixmaps. 8114 8115 this just adds the interface and shared support to the pixmap. 8116 8117 Signed-off-by: Dave Airlie <airlied@redhat.com> 8118 8119commit f71139a2afe8fffb628331402bf829a6d67c9fff 8120Author: Dave Airlie <airlied@redhat.com> 8121Date: Mon Sep 3 12:37:22 2012 +1000 8122 8123 radeon: add platform bus loading support. 8124 8125 This allows the radeon driver to be loaded via the platform bus mechanism. 8126 8127 Signed-off-by: Dave Airlie <airlied@redhat.com> 8128 8129commit 9911e72ea79e000ee0492d19a434aa98b83b5066 8130Author: Dave Airlie <airlied@redhat.com> 8131Date: Fri Aug 17 17:02:51 2012 +1000 8132 8133 radeon: damage full screen after pageflip 8134 8135 in order for offload to work we need to cause a full 8136 screen damage to be registered after pageflip. 8137 8138 Signed-off-by: Dave Airlie <airlied@redhat.com> 8139 8140commit 0873b8cef21abf1e7286731758f0db1e7f2e7c97 8141Author: Dave Airlie <airlied@redhat.com> 8142Date: Fri Aug 3 16:45:50 2012 +1000 8143 8144 ati: fix stub driver loader to load all drivers properly 8145 8146 Signed-off-by: Dave Airlie <airlied@redhat.com> 8147 8148commit cdf522f5992adf7b76f9153c777e6dddcf7758a6 8149Author: Adam Jackson <ajax@redhat.com> 8150Date: Wed Aug 15 13:38:28 2012 -0400 8151 8152 Implement ->driverFunc 8153 8154 Copied from fbdev, makes it so we can run without iopl. 8155 8156 Signed-off-by: Adam Jackson <ajax@redhat.com> 8157 8158commit 763f387ccac2399f1d58ec112c8bed45196bea5e 8159Author: Adam Jackson <ajax@redhat.com> 8160Date: Wed Aug 15 12:49:10 2012 -0400 8161 8162 Don't bother to include vbe.h 8163 8164 The only VBE code here went away when we dropped UMS. 8165 8166 Signed-off-by: Adam Jackson <ajax@redhat.com> 8167 8168commit aabcf5ea30195274c79ae94c0fe6edc965f5c123 8169Author: Tomas Chvatal <tchvatal@suse.cz> 8170Date: Mon Aug 13 22:58:18 2012 +0200 8171 8172 Add switch for udev to disable automagic detection. 8173 8174 Signed-off-by: Tomas Chvatal <tchvatal@suse.cz> 8175 8176commit aef90993621331f01d17b95a751d4924d77eed0c 8177Author: Alex Deucher <alexdeucher@gmail.com> 8178Date: Mon Aug 6 11:00:45 2012 -0400 8179 8180 radeon: add some new SI pci ids 8181 8182 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8183 8184commit 6ef1ad6a46348d3aecd8d1f5e94431ca2298853c 8185Author: Michel Dänzer <michel.daenzer@amd.com> 8186Date: Fri Jul 13 11:15:25 2012 +0200 8187 8188 Deal more gracefully with DRI2 being unavailable at build or run time. 8189 8190 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8191 8192commit ef8a404391036d8aa814dbda2407c789b8a64b92 8193Author: Michel Dänzer <michel.daenzer@amd.com> 8194Date: Thu Jul 5 20:14:48 2012 +0200 8195 8196 Initial SI support. 8197 8198 Defaults to shadowfb. 3D acceleration is available with glamor. 2D 8199 acceleration is disabled until the radeonsi driver can handle glamor's 8200 shaders. 8201 8202 v2: add chip flags (Alex Deucher) 8203 8204 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8205 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 8206 8207commit e9edd2f5002c642b59f028b3ec076d604ae8ce9d 8208Author: Michel Dänzer <michel.daenzer@amd.com> 8209Date: Wed Jun 20 08:40:07 2012 +0200 8210 8211 Initial glamor support. 8212 8213 Enable at build time with --enable-glamor and runtime with 8214 8215 Option "AccelMethod" "glamor" 8216 8217 The most notable lack of functionality is XVideo. Use something like VDPAU for 8218 now. 8219 8220 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8221 8222commit 9eac8021f3d33a63156f9f5d43a220e88bb3f8db 8223Author: Michel Dänzer <michel.daenzer@amd.com> 8224Date: Wed Jun 27 14:48:47 2012 +0200 8225 8226 EXA: Factor out pixmap BO allocation into a helper function. 8227 8228 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8229 8230commit 060c7836e7f7777bacca4e23f57c5985beab33bc 8231Author: Michel Dänzer <michel.daenzer@amd.com> 8232Date: Wed Jul 4 11:49:18 2012 +0200 8233 8234 Some cosmetic fixups for the radeon manpage. 8235 8236 ATI -> ATI/AMD 8237 PCIE -> PCIe 8238 8239 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8240 8241commit c3a9f64fe40c78331b9340c3832d344f38b11564 8242Author: Michel Dänzer <michel.daenzer@amd.com> 8243Date: Wed Jul 4 11:46:30 2012 +0200 8244 8245 Drop UMS specific parts from the radeon manpage. 8246 8247 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8248 8249commit 5ebebfdb4fadada157d1369f35a8610c0fa72e02 8250Author: Dave Airlie <airlied@redhat.com> 8251Date: Fri Jun 29 15:03:43 2012 +0100 8252 8253 radeon: fix conflict with fix in master. 8254 8255 Signed-off-by: Dave Airlie <airlied@redhat.com> 8256 8257commit ff12ddf613ad42e4566d6e10fc6cf33a4e445736 8258Author: Dave Airlie <airlied@redhat.com> 8259Date: Fri Jun 29 15:00:44 2012 +0100 8260 8261 configure: bump to 6.99.99 8262 8263 this is in prep for 7.0 release. 8264 8265 Signed-off-by: Dave Airlie <airlied@redhat.com> 8266 8267commit 50689ec8dbd4a68527b2ac16cecac298b8d441d0 8268Merge: 6a60fcdf ae682a6a 8269Author: Dave Airlie <airlied@redhat.com> 8270Date: Fri Jun 29 14:59:47 2012 +0100 8271 8272 Merge remote-tracking branch 'origin/kms-only' 8273 8274 This merges the removal of the UMS code and subsequent simplifications. 8275 8276 Hell yes. 8277 8278 * origin/kms-only: (24 commits) 8279 radeon: fix radeonchipsets properly 8280 radeon: fix RADEONChipsets 8281 radeon: migrate remainder of radeon_driver.c to rest of driver 8282 radeon: avoid including large static struct twice. 8283 radeon: drop some more unused macros/inlines 8284 radeon: drop another unused struct member. 8285 radeon: more unused stuff 8286 radeon: drop cardType shouldn't matter to userspace. 8287 radeon: drop radeonGetPixmapOffset 8288 radeon: move more functions into texture video 8289 radeon: drop legacy memory, merge into radeon video 8290 radeon: drop unused txoffset vars. 8291 radeon: drop vb_mc_addr, not needed anymore 8292 radeon: drop offset member of r600 accel object 8293 radeon: move macros into radeon.h 8294 radeon: drop more wrapper macros. 8295 radeon: further macro cleanups. 8296 radeon: make exa copy funcs static. 8297 radeon: drop FUNC_NAME macro and ONCE_ONLY stuff 8298 radeon: drop radeon_commonfuncs. 8299 ... 8300 8301 Conflicts: 8302 configure.ac 8303 8304commit 6a60fcdf060a0f553d8d4f7939c2a05aa04bedfa 8305Author: Dave Airlie <airlied@redhat.com> 8306Date: Fri Jun 29 14:57:58 2012 +0100 8307 8308 configure: bump version after release. 8309 8310 Signed-off-by: Dave Airlie <airlied@redhat.com> 8311 8312commit 008eb209d506ed7195fd544dfa91be9783946fd5 8313Author: Dave Airlie <airlied@redhat.com> 8314Date: Fri Jun 29 14:46:55 2012 +0100 8315 8316 configure: bump version number to 6.14.6 pre-release 8317 8318 Signed-off-by: Dave Airlie <airlied@redhat.com> 8319 8320commit 72fea2635a56c4f572c07fc50fc58f703a21eb4f 8321Author: Dave Airlie <airlied@redhat.com> 8322Date: Fri Jun 29 14:46:17 2012 +0100 8323 8324 radeon: require libdrm 2.4.36 for KMS support. 8325 8326 This is due to some commits to the surface manager that fix bugs. 8327 8328 Signed-off-by: Dave Airlie <airlied@redhat.com> 8329 8330commit 179b035835bfa99283d44ba2d5ee17e686196535 8331Author: Michel Dänzer <michel.daenzer@amd.com> 8332Date: Wed Jun 27 19:26:51 2012 +0200 8333 8334 Fix up displayWidth vs. virtualX confusion in radeon_setup_kernel_mem(). 8335 8336 It was using the pitch (displayWidth) for the virtual width (virtualX). This 8337 prevented using page flipping in some cases, as displayWidth was already 8338 overaligned for virtualX, so the DRI2 front and back buffers ended up having 8339 a different pitch. 8340 8341 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8342 Reviewed-by: Jerome Glisse <jglisse@redhat.com> 8343 8344commit 5dc5012320536f7b1e86a81ed3b823e3626828e1 8345Author: Dave Airlie <airlied@redhat.com> 8346Date: Mon Jun 25 09:23:04 2012 +0100 8347 8348 radeon: dixPrivateKeyRegistered is only in server 1.9 8349 8350 just fallback to the old behaviour on older servers. 8351 8352 should fix: https://bugs.freedesktop.org/show_bug.cgi?id=51388 8353 Review-over-irc: Michel Dänzer <michel.daenzer@amd.com> 8354 Signed-off-by: Dave Airlie <airlied@redhat.com> 8355 8356commit ae682a6a1781ab023ab36e3154d6b3e88da63d64 8357Author: Dave Airlie <airlied@redhat.com> 8358Date: Mon Jun 18 13:11:31 2012 +0100 8359 8360 radeon: fix radeonchipsets properly 8361 8362commit 6514bcdc0fc199b4f3a8b4273bd29e0fe255d232 8363Author: Dave Airlie <airlied@redhat.com> 8364Date: Mon Jun 18 13:02:03 2012 +0100 8365 8366 radeon: fix RADEONChipsets 8367 8368 reported on irc by okias. 8369 8370 Signed-off-by: Dave Airlie <airlied@redhat.com> 8371 8372commit dd4ab5aba9047d522dbcfe8a341c368312e1a471 8373Author: Dave Airlie <airlied@redhat.com> 8374Date: Sat Jun 16 08:47:32 2012 +0100 8375 8376 radeon: migrate remainder of radeon_driver.c to rest of driver 8377 8378 This helps make a few more things static and the driver generally 8379 smaller. 8380 8381 Signed-off-by: Dave Airlie <airlied@redhat.com> 8382 8383commit a19237fad2492d70777167c631f4e6e1fae0e908 8384Author: Dave Airlie <airlied@redhat.com> 8385Date: Sat Jun 16 08:19:35 2012 +0100 8386 8387 radeon: avoid including large static struct twice. 8388 8389 Just extern this in the one file that needs access. 8390 8391 Signed-off-by: Dave Airlie <airlied@redhat.com> 8392 8393commit c8a5c5a0dd9adc46b03b898c9c1719dbbde05ba1 8394Author: Dave Airlie <airlied@redhat.com> 8395Date: Fri Jun 15 19:52:14 2012 +0100 8396 8397 radeon: drop some more unused macros/inlines 8398 8399 These are no longer used anywhere. 8400 8401 Signed-off-by: Dave Airlie <airlied@redhat.com> 8402 8403commit 3097f273443b5b8fce097e2b1e09a17b02d06a40 8404Author: Dave Airlie <airlied@redhat.com> 8405Date: Fri Jun 15 19:45:15 2012 +0100 8406 8407 radeon: drop another unused struct member. 8408 8409 Signed-off-by: Dave Airlie <airlied@redhat.com> 8410 8411commit cb97e75acab84b67b7b7358860788638efc9b344 8412Author: Dave Airlie <airlied@redhat.com> 8413Date: Fri Jun 15 19:38:45 2012 +0100 8414 8415 radeon: more unused stuff 8416 8417 Remove all CurrentLayout stuff. 8418 8419 Signed-off-by: Dave Airlie <airlied@redhat.com> 8420 8421commit 81593e7deb688fa3108a0589c1418459ec0df4de 8422Author: Dave Airlie <airlied@redhat.com> 8423Date: Fri Jun 15 18:52:52 2012 +0100 8424 8425 radeon: drop cardType shouldn't matter to userspace. 8426 8427 Signed-off-by: Dave Airlie <airlied@redhat.com> 8428 8429commit 997b0dddfeb180043ece5eff0ac0d6a1a751f9b5 8430Author: Dave Airlie <airlied@redhat.com> 8431Date: Fri Jun 15 18:43:01 2012 +0100 8432 8433 radeon: drop radeonGetPixmapOffset 8434 8435 no longer needed. 8436 8437 Signed-off-by: Dave Airlie <airlied@redhat.com> 8438 8439commit 639061edb3ac35a24e4d6633c9b6384434dcac64 8440Author: Dave Airlie <airlied@redhat.com> 8441Date: Fri Jun 15 18:38:48 2012 +0100 8442 8443 radeon: move more functions into texture video 8444 8445 make more thing static. 8446 8447 Signed-off-by: Dave Airlie <airlied@redhat.com> 8448 8449commit 85711a0a6a553232b603dbbb2b2d234c35391578 8450Author: Dave Airlie <airlied@redhat.com> 8451Date: Fri Jun 15 17:26:50 2012 +0100 8452 8453 radeon: drop legacy memory, merge into radeon video 8454 8455 This merges these two functions into textured video code. 8456 8457 Signed-off-by: Dave Airlie <airlied@redhat.com> 8458 8459commit 445ca1d66ce7b37762731fe67407cc1b978bb129 8460Author: Dave Airlie <airlied@redhat.com> 8461Date: Fri Jun 15 17:13:48 2012 +0100 8462 8463 radeon: drop unused txoffset vars. 8464 8465 These vars aren't being used anymore. 8466 8467 Signed-off-by: Dave Airlie <airlied@redhat.com> 8468 8469commit 832244de9e6d7aa1b8fbeb29c9a6a86923744d60 8470Author: Dave Airlie <airlied@redhat.com> 8471Date: Fri Jun 15 17:09:00 2012 +0100 8472 8473 radeon: drop vb_mc_addr, not needed anymore 8474 8475 This field is totally unused now, so drop it. 8476 8477 Signed-off-by: Dave Airlie <airlied@redhat.com> 8478 8479commit 6559b6e28bf111b64c3aeb5ec1260acfeb3e0b7f 8480Author: Dave Airlie <airlied@redhat.com> 8481Date: Fri Jun 15 17:05:54 2012 +0100 8482 8483 radeon: drop offset member of r600 accel object 8484 8485 This is always 0 now. 8486 8487 Signed-off-by: Dave Airlie <airlied@redhat.com> 8488 8489commit 857179e538c9752fab239a21312d0fcc80022151 8490Author: Dave Airlie <airlied@redhat.com> 8491Date: Fri Jun 15 16:55:23 2012 +0100 8492 8493 radeon: move macros into radeon.h 8494 8495 Drop separate header file, there isn't many separate macros now. 8496 8497 Signed-off-by: Dave Airlie <airlied@redhat.com> 8498 8499commit 148e889149f5ecc3dc2473f2a9e01572eeb037d3 8500Author: Dave Airlie <airlied@redhat.com> 8501Date: Fri Jun 15 16:52:24 2012 +0100 8502 8503 radeon: drop more wrapper macros. 8504 8505 Signed-off-by: Dave Airlie <airlied@redhat.com> 8506 8507commit ad8629e3323efe484dd1421d94dd934e1bfafe56 8508Author: Dave Airlie <airlied@redhat.com> 8509Date: Fri Jun 15 16:49:13 2012 +0100 8510 8511 radeon: further macro cleanups. 8512 8513 drop some more of the wrapper macros. 8514 8515 Signed-off-by: Dave Airlie <airlied@redhat.com> 8516 8517commit 96644f96f1a2df1d7d1ca8cbdc9a105affdb4f32 8518Author: Dave Airlie <airlied@redhat.com> 8519Date: Fri Jun 15 16:34:53 2012 +0100 8520 8521 radeon: make exa copy funcs static. 8522 8523 These are no longer used outside this area. 8524 8525 Signed-off-by: Dave Airlie <airlied@redhat.com> 8526 8527commit 83904b820b81d226940ec6d663246f243f92e7ba 8528Author: Dave Airlie <airlied@redhat.com> 8529Date: Fri Jun 15 16:30:41 2012 +0100 8530 8531 radeon: drop FUNC_NAME macro and ONCE_ONLY stuff 8532 8533 Drop all the CP vs MMIO macros that name functions. 8534 8535 Signed-off-by: Dave Airlie <airlied@redhat.com> 8536 8537commit b422d8085aab4e01a1ac4ffb247c73fe51aafa0d 8538Author: Dave Airlie <airlied@redhat.com> 8539Date: Fri Jun 15 16:14:54 2012 +0100 8540 8541 radeon: drop radeon_commonfuncs. 8542 8543 modify all the macros and drop it into radeon_accel.c 8544 8545 Signed-off-by: Dave Airlie <airlied@redhat.com> 8546 8547commit 29dd3f84eecd2892a438f004403521de8986a91d 8548Author: Dave Airlie <airlied@redhat.com> 8549Date: Fri Jun 15 16:11:03 2012 +0100 8550 8551 radeon: misc cleanups. 8552 8553 Signed-off-by: Dave Airlie <airlied@redhat.com> 8554 8555commit 678c1f3425c0fa7e98795c8e351da77d04652e02 8556Author: Dave Airlie <airlied@redhat.com> 8557Date: Fri Jun 15 16:06:27 2012 +0100 8558 8559 make pci chipsets static 8560 8561 Signed-off-by: Dave Airlie <airlied@redhat.com> 8562 8563commit 95cbbe999418c72dc407069d31e759d70f49ab16 8564Author: Dave Airlie <airlied@redhat.com> 8565Date: Fri Jun 15 16:01:53 2012 +0100 8566 8567 drop RING_LOCALS/ACCEL_PREAMBLE 8568 8569 these were unused now. 8570 8571 Signed-off-by: Dave Airlie <airlied@redhat.com> 8572 8573commit 18d5ae3bd9075ac1a2ee21b071ac133e2e634b62 8574Author: Dave Airlie <airlied@redhat.com> 8575Date: Fri Jun 15 10:05:03 2012 +0100 8576 8577 radeon: drop all UMS/DRI1/XAA/overlay support. 8578 8579 This overhauls the radeon driver and removes all the old UMS-only code, 8580 it drops all the UMS, DRI1, XAA, overlay Xv, video capture, tv tuners 8581 8582 There are probably a lot more cleanups that will fall out of this afterwards. 8583 8584 So far this is compile/build tested. 8585 8586 Signed-off-by: Dave Airlie <airlied@redhat.com> 8587 8588commit 248e912c487636d7352cfad43c03fc9f19fc2215 8589Author: Michel Dänzer <michel.daenzer@amd.com> 8590Date: Fri Jun 8 12:18:48 2012 +0200 8591 8592 UMS: Fix CRTC DPMS state check. 8593 8594 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8595 8596commit 58883711136fa6671d4f3250d4ee973e2953ebb4 8597Author: Alex Deucher <alexdeucher@gmail.com> 8598Date: Thu Jun 7 11:47:21 2012 -0400 8599 8600 configure: bump version post release 8601 8602 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8603 8604commit 3a9103f855c67bbcd7b97ee2e96836353e0d789d 8605Author: Alex Deucher <alexdeucher@gmail.com> 8606Date: Thu Jun 7 11:27:08 2012 -0400 8607 8608 configure: bump for release 8609 8610 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8611 8612commit f1693ee09e9b251eb75e3e5fe0ee7699fec6fb98 8613Author: Alex Deucher <alexdeucher@gmail.com> 8614Date: Thu Jun 7 11:23:31 2012 -0400 8615 8616 configure: bump libdrm_radeon requirement 8617 8618 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8619 8620commit 77a056c1725c9c44a328ed324687bdf35144e9e2 8621Author: Michel Dänzer <michel.daenzer@amd.com> 8622Date: Wed Jun 6 12:29:37 2012 +0200 8623 8624 Don't page-flip or wait on a CRTC while we're VT-switched away. 8625 8626 We don't know what the CRTC state is. 8627 8628 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8629 8630commit 1d9f1eb7e65c57c52759ccbfa51e8bdd5a10acac 8631Author: Michel Dänzer <michel.daenzer@amd.com> 8632Date: Wed Jun 6 12:16:08 2012 +0200 8633 8634 Don't wait on a CRTC which has been disabled via DPMS (bug #49761). 8635 8636 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=49761 . 8637 8638 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8639 8640commit 4603285aa8efaf2614f15a38254465cec2075f11 8641Author: Alex Deucher <alexdeucher@gmail.com> 8642Date: Tue Jun 5 09:48:59 2012 -0400 8643 8644 radeon: add new PCI ids 8645 8646 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8647 8648commit 9307609420b4b209767d2057b4803bcb16cc1455 8649Author: Dave Airlie <airlied@redhat.com> 8650Date: Sat Jun 2 17:16:34 2012 +0100 8651 8652 radeon: improve smooth startup fallbacks. 8653 8654 If we can't handover the framebuffer, memset it to black. 8655 8656 mostly ported from nouveau. 8657 8658 Signed-off-by: Dave Airlie <airlied@redhat.com> 8659 8660commit 89a4c79a43a2b6963e41d7812e8fe5beedb556af 8661Author: Dave Airlie <airlied@redhat.com> 8662Date: Sat Jun 2 17:14:14 2012 +0100 8663 8664 radeon: fix smooth startup with tiling enabled. 8665 8666 We need to use the surface we worked out when we allocated the front bo, 8667 not work out a new surface from scratch. 8668 8669 This fixes smooth handover from plymouth to gdm on F17. 8670 8671 Signed-off-by: Dave Airlie <airlied@redhat.com> 8672 8673commit c1b9b2c9d71a87a0b98ccdb542da8a30c8c4d99a 8674Author: Dave Airlie <airlied@redhat.com> 8675Date: Fri Jun 1 12:22:06 2012 +0100 8676 8677 radeon: fix compat api for newest planned API. 8678 8679 Signed-off-by: Dave Airlie <airlied@redhat.com> 8680 8681commit 68c51d148b22ee8cbe98b5d5441e3494804ee808 8682Author: Dave Airlie <airlied@redhat.com> 8683Date: Fri Jun 1 12:21:34 2012 +0100 8684 8685 drmmode: drop flags arg to adjust frame. 8686 8687 Signed-off-by: Dave Airlie <airlied@redhat.com> 8688 8689commit b1e9c3081d2b975682a62389aa29063b79f9655c 8690Author: Dave Airlie <airlied@redhat.com> 8691Date: Thu May 24 17:54:41 2012 +0100 8692 8693 radeon: move include down further 8694 8695 fixes tinderbox build. 8696 8697 Signed-off-by: Dave Airlie <airlied@redhat.com> 8698 8699commit 5ed959585f35a4eff2cd803ae222b46d164f210d 8700Author: Dave Airlie <airlied@redhat.com> 8701Date: Thu May 24 08:17:11 2012 +0100 8702 8703 radeon: update compat-api.h for block handler + enable/disable fb. 8704 8705 This updates the compat stuff for the latest block handler code, 8706 and the enable/disable interface. 8707 8708 Signed-off-by: Dave Airlie <airlied@redhat.com> 8709 8710commit 2de17efdcb502f91fb145130c12be3cb164f03c5 8711Author: Dave Airlie <airlied@redhat.com> 8712Date: Thu May 24 08:16:39 2012 +0100 8713 8714 radeon/xvmc: fix build with new API. 8715 8716 This was missing the compat include. 8717 8718 Signed-off-by: Dave Airlie <airlied@redhat.com> 8719 8720commit a1f47768309aa8b14b13f05ba01bb490024c4cbc 8721Author: Dave Airlie <airlied@redhat.com> 8722Date: Thu May 24 07:55:14 2012 +0100 8723 8724 radeon: fix a few more RHDAtomBiosFunc uses 8725 8726 I missed these in my initial search/replace for some reason. 8727 8728 Signed-off-by: Dave Airlie <airlied@redhat.com> 8729 8730commit c26e40cf60eb6a2bffb5ec7cba1f1bbffcbf1d53 8731Author: Dave Airlie <airlied@redhat.com> 8732Date: Wed May 23 12:12:54 2012 +0100 8733 8734 radeon/atombios: rewrite atombios parser code to not use xf86Screens. 8735 8736 Just pass the ScrnInfoPtr around instead. 8737 8738 Signed-off-by: Dave Airlie <airlied@redhat.com> 8739 8740commit 14122214edabd2a574b66577289a7c8732f4e953 8741Author: Dave Airlie <airlied@redhat.com> 8742Date: Wed May 23 11:57:35 2012 +0100 8743 8744 radeon/generic_bus: stop passing scrnIndex 8745 8746 Just pass a pointer to the screen, removes usage of xf86Screens lookup 8747 8748 Signed-off-by: Dave Airlie <airlied@redhat.com> 8749 8750commit 2101e485480e3ffd6a17ed12f5287b8b1e4e0223 8751Author: Dave Airlie <airlied@redhat.com> 8752Date: Tue May 22 16:17:55 2012 +0100 8753 8754 compat for new server API 8755 8756commit 9d2549057ffced996bfaeb8df4901ef57da636c7 8757Author: Dave Airlie <airlied@redhat.com> 8758Date: Wed May 16 15:45:53 2012 +0100 8759 8760 ati: convert to new screen conversion APIs 8761 8762 The compat header takes care of the old server vs new server. 8763 8764 this commit was autogenerated from util/modular/x-driver-screen-scrn-conv.sh 8765 8766 Signed-off-by: Dave Airlie <airlied@redhat.com> 8767 8768commit 80a45bfc26e20d1d481f13cb15c0720d165a8306 8769Author: Dave Airlie <airlied@redhat.com> 8770Date: Wed May 16 15:42:59 2012 +0100 8771 8772 radeon: add compat-api.h 8773 8774commit 84c3900b5179a8d2cc5f925e94ed5b2bbdee1ad9 8775Author: Dave Airlie <airlied@redhat.com> 8776Date: Wed May 23 08:49:45 2012 +0100 8777 8778 radeon/xaa: drop scrnIndex parameter to some functions 8779 8780 This isn't needed, and makes api changes later easier. 8781 8782 Signed-off-by: Dave Airlie <airlied@redhat.com> 8783 8784commit 3fb694b308ebadd1b849836059b6b56bb19385f7 8785Author: Thierry Vignaud <thierry.vignaud@gmail.com> 8786Date: Wed May 16 14:43:53 2012 +0200 8787 8788 UMS/EXA: Add reminder for potential solid picture performance issue. 8789 8790commit 4b9bad959438725e4434e2aa4f142542d8dcc260 8791Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 8792Date: Fri May 11 05:25:32 2012 +0200 8793 8794 radeon: use GB_GR and BG_RG formats for packed yuv video for r600+ 8795 8796 Those formats were invented for exactly that purpose so use them. 8797 This saves some code and also some hw resources (only need one 8798 sampler instead of two for packed yuv). 8799 Only tested on EG. 8800 8801commit 37786e9027b8c8d1f9ec9928915784dd28853766 8802Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 8803Date: Fri May 11 05:25:33 2012 +0200 8804 8805 radeon: avoid rounding errors in texture coords for textured xv on EG+ 8806 8807 make sure the division is done with floats, otherwise the coordinate 8808 can be wrong up to 1 texel. 8809 Particularly visible with clipping and small source scaled up (since one 8810 texel can be a shift of several pixels) but could be seen even unscaled. 8811 Should provide more accurate coords without clipping too depending on the 8812 scale factor probably. 8813 This is a straight port of 688c8a54a00b01e73a11970ad2abe858f8c7c5c4 8814 when I apparently forgot the eg code... 8815 8816commit faea3aafa8c9a7c1bc866ffcd847972c5569c8bc 8817Author: Anisse Astier <anisse@astier.eu> 8818Date: Thu May 10 17:22:59 2012 +0200 8819 8820 Fail more gracefully when drm surface manager can't be initialized 8821 8822 Should make bugs like https://bugs.freedesktop.org/show_bug.cgi?id=48138 8823 easier to diagnose. 8824 8825 [ Michel Dänzer: Appended newline to error message. ] 8826 8827 Signed-off-by: Anisse Astier <anisse@astier.eu> 8828 Singed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8829 8830commit b0b7d8d26fd107df342b5c87b0a38e5bb08101a9 8831Author: Michel Dänzer <michel.daenzer@amd.com> 8832Date: Wed May 9 11:08:49 2012 +0200 8833 8834 EXA/UMS: Synchronize to the GPU before writing solid colour to scratch pixmap. 8835 8836 UMS doesn't do this automagically. It's a big hammer that will probably suck 8837 for performance, but I don't have any better ideas right now. 8838 8839 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8840 8841commit 6bda7ceda645e838723883d133d614def1511d16 8842Author: Michel Dänzer <michel.daenzer@amd.com> 8843Date: Thu May 3 15:07:30 2012 +0200 8844 8845 EXA: Attempt to fix solid picture acceleration with UMS. 8846 8847 Only compile tested, but should fix 8848 https://bugs.freedesktop.org/show_bug.cgi?id=49182 . 8849 8850 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8851 8852commit b33cf9cf9dce5add8aaccab10ba5aaa64247c8ce 8853Author: Alex Deucher <alexdeucher@gmail.com> 8854Date: Wed May 2 09:55:16 2012 -0400 8855 8856 radeon: add MacModel entry for SAM440ep embedded board 8857 8858 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8859 8860commit 0bda305f7ab2a4720b3fea3f318ab2a73be151e5 8861Author: Alex Deucher <alexander.deucher@amd.com> 8862Date: Fri Apr 13 13:28:10 2012 -0400 8863 8864 EXA: Support acceleration of solid pictures on R2xx. 8865 8866 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 8867 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 8868 8869commit a7754b076e355fef3aea082b6e2d3aefbb8e7a9b 8870Author: Alex Deucher <alexander.deucher@amd.com> 8871Date: Fri Apr 13 13:24:46 2012 -0400 8872 8873 EXA: Support acceleration of solid pictures on R1xx. 8874 8875 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 8876 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 8877 8878commit cac2d7ed54918579418fc762558497d3ec547fad 8879Author: Alex Deucher <alexander.deucher@amd.com> 8880Date: Sat Apr 14 08:53:39 2012 -0400 8881 8882 EXA: Support acceleration of solid pictures on Evergreen/NI. 8883 8884 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 8885 Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> 8886 8887commit d88b9700137ee407c483f263bb55c77cd6f92fef 8888Author: Michel Dänzer <michel.daenzer@amd.com> 8889Date: Fri Apr 13 09:35:38 2012 +0200 8890 8891 EXA: Support acceleration of solid pictures on R3xx-R7xx. 8892 8893 Allocate 1x1 scratch pixmaps to hold the solid picture colours. 8894 8895 This works around https://bugs.freedesktop.org/show_bug.cgi?id=47266 and might 8896 improve performance in other cases as well. 8897 8898 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8899 8900commit 66b586b9b9cdaf70f0fcd547b5a04f044d848d44 8901Author: Michel Dänzer <michel.daenzer@amd.com> 8902Date: Mon Apr 16 15:33:36 2012 +0200 8903 8904 RADEONCopySwap: Fix RADEON_HOST_DATA_SWAP_16BIT case. 8905 8906 It was the same code as for RADEON_HOST_DATA_SWAP_32BIT. This caused bus errors 8907 on FreeBSD/PPC, but I'm not sure how it could not cause problems anywhere... 8908 8909 Reported-by: Andreas Tobler <andreast@fgznet.ch> 8910 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8911 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 8912 8913commit 1e656cd5973bbcf85b63406638ede0bccd65e28f 8914Author: Michel Dänzer <michel.daenzer@amd.com> 8915Date: Mon Apr 16 15:44:01 2012 +0200 8916 8917 Make radeon_setup_kernel_mem failures more graceful and verbose. 8918 8919 So that bugs like https://bugs.freedesktop.org/show_bug.cgi?id=48138 can be 8920 diagnosed more easily. 8921 8922 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 8923 8924commit d282719a9c2fb0ee32830aa75b8dfbb9392954ed 8925Author: Jerome Glisse <jglisse@redhat.com> 8926Date: Wed Apr 4 17:08:30 2012 -0400 8927 8928 r6xx-r9xx: force 1D tiling for buffer with height < 64 8929 8930 Due to some old kernel issue, height is 8 aligned insided the ddx 8931 For buffer with height btw 57 & 63 this lead ddx to believe it can 8932 allocate a 2D tiled surface while mesa will not align height and 8933 will assume 1D tiled leading to disagreement and rendering issue. 8934 This patch force buffer with height < 64 to be 1D tiled. 8935 8936 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 8937 8938commit de2419e5509398328903ca61f4ea80852cba6bc4 8939Author: Alex Deucher <alexdeucher@gmail.com> 8940Date: Thu Mar 29 00:28:39 2012 -0400 8941 8942 configure: bump version post release 8943 8944 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8945 8946commit 9425c50e93903fb64d9e569cfdc1e2c35d16ce25 8947Author: Alex Deucher <alexdeucher@gmail.com> 8948Date: Thu Mar 29 00:19:12 2012 -0400 8949 8950 configure: bump version for release 8951 8952 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8953 8954commit dc18d771713ecd893c7d5833da6e0661093161dc 8955Author: Alex Deucher <alexdeucher@gmail.com> 8956Date: Wed Mar 28 17:32:53 2012 -0400 8957 8958 configure: bump libdrm requirement for TN support 8959 8960 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8961 8962commit 03535904a3e1542b3924d0a062c4b022ca196888 8963Author: Alex Deucher <alexdeucher@gmail.com> 8964Date: Tue Mar 27 09:48:28 2012 -0400 8965 8966 radeon: man page updates 8967 8968 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8969 8970commit 6ed191c33caa33c12c2c6dafcba3a5ab1bf4a02f 8971Author: Alex Deucher <alexdeucher@gmail.com> 8972Date: Tue Mar 20 19:57:53 2012 -0400 8973 8974 radeon/kms: add TN pci ids 8975 8976 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8977 8978commit 55d65fcf33eb383e3fbc7a1d469ab68a70a7ab37 8979Author: Alex Deucher <alexdeucher@gmail.com> 8980Date: Tue Mar 20 19:54:57 2012 -0400 8981 8982 radeon/kms: add support for TN (trinity) APUs 8983 8984 - KMS only 8985 - Includes full EXA/Xv support 8986 8987 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 8988 8989commit b5cf9bd693cf4090956add4c33c4fae9c3069a03 8990Author: Marek Olšák <maraeo@gmail.com> 8991Date: Wed Mar 7 11:01:40 2012 -0500 8992 8993 r6xx: initialize SX_MISC 8994 8995 If Mesa set it to 1, the DDX would not render anything = the monitor would 8996 basically freeze. 8997 8998 agd5f: update emit count as well. 8999 9000 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9001 9002commit 355dc4295912c153f5333421594fa90aa119a056 9003Author: Michel Dänzer <michel.daenzer@amd.com> 9004Date: Tue Mar 6 15:52:40 2012 +0100 9005 9006 DRI2: Unreference buffers immediately when event wait info is invalidated. 9007 9008 Deferring this could result in trying to unreference buffers from a previous 9009 server generation, i.e. accessing freed memory. 9010 9011 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9012 Tested-by: Christian König <Christian.koenig@amd.com> 9013 9014commit fe51469b2e02e4d565050bab077985270fb58a9b 9015Author: Michel Dänzer <michel.daenzer@amd.com> 9016Date: Tue Mar 6 15:52:40 2012 +0100 9017 9018 Re-register DRM FD wakeup handler for each server generation. 9019 9020 Fixes hang when trying to use DRI2 swap scheduling after a server reset. 9021 9022 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9023 Tested-by: Christian König <Christian.koenig@amd.com> 9024 9025commit 878454ae8d8e96dd27a19d0b30940d014c4cd7e2 9026Author: Hans Verkuil <hverkuil@xs4all.nl> 9027Date: Fri Feb 24 09:35:39 2012 -0500 9028 9029 Fix ConnectorTable crash in radeon_output.c 9030 9031 The sam440ep PPC board requires a ConnectorTable xorg.conf option, but putting 9032 in that option causes the radeon driver to crash. I finally traced it to a 9033 copy-and-paste bug in radeon_output.c as a result of a major rework in commit 9034 82f12e5a40c1fbcb91910a0f8b725c34fff02aae. 9035 9036 The actual crash occurred in RADEONPrintPortMap(). 9037 9038 Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> 9039 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9040 9041commit 688c8a54a00b01e73a11970ad2abe858f8c7c5c4 9042Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 9043Date: Sat Feb 18 21:12:34 2012 +0100 9044 9045 radeon: avoid rounding errors in texture coords for textured xv 9046 9047 make sure the division is done with floats, otherwise the coordinate 9048 can be wrong up to 1 texel. 9049 Particularly visible with clipping and small source scaled up (since one 9050 texel can be a shift of several pixels) but could be seen even unscaled. 9051 Should provide more accurate coords without clipping too depending on the 9052 scale factor probably. 9053 Changed for r100-r600, though only tested on r300. 9054 9055commit 2778b56252124ef6f636a493d2e1457b43911c37 9056Author: Jerome Glisse <jglisse@redhat.com> 9057Date: Mon Feb 13 20:42:57 2012 -0500 9058 9059 radeon: r6xx-eg use linear general when using scratch bo 9060 9061 In path where we need to use scratch bo as temporary area, 9062 consider it as linear buffer. Not linear aligned. Fix some 9063 case such as in bugs: 9064 9065 https://bugs.freedesktop.org/show_bug.cgi?id=45827 9066 9067 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 9068 9069commit c66ae235700f5efe64eb168327551b8f1d153c9c 9070Author: Michel Dänzer <michel.daenzer@amd.com> 9071Date: Mon Feb 13 10:43:58 2012 +0100 9072 9073 Handle new xorg_list API. 9074 9075 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=45937 9076 9077 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9078 9079commit 7ff277e22c629308915307bbee96eb25ff77f8b9 9080Author: Alex Deucher <alexdeucher@gmail.com> 9081Date: Fri Feb 10 13:04:59 2012 -0500 9082 9083 radeon: fix crash in drmmode_create_bo_pixmap() 9084 9085 Only init surface on r6xx+. Return NULL rather than 9086 FALSE. 9087 9088 Fixes: 9089 https://bugs.freedesktop.org/show_bug.cgi?id=45829 9090 9091 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9092 9093commit 60b949f34df5db05e0e102cc3daa33469aa50cfc 9094Author: Alex Deucher <alexdeucher@gmail.com> 9095Date: Fri Feb 10 13:11:26 2012 -0500 9096 9097 radeon/kms: reusing fd message is not an error 9098 9099 It's standard behavior. 9100 9101 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9102 9103commit e20284409937d784847339b5d466a95012d85940 9104Author: Michel Dänzer <michel.daenzer@amd.com> 9105Date: Fri Feb 3 12:21:59 2012 +0100 9106 9107 EXA/r6xx+: Only set write domain or read domains, not both. (Bug #43893) 9108 9109 Avoids an accounting bug in libdrm_radeon 2.4.31 or older. 9110 9111 See https://bugs.freedesktop.org/show_bug.cgi?id=43893 9112 9113 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9114 9115commit 5a7f64c5170ca424c9dca739662018e30df13413 9116Author: Michel Dänzer <michel.daenzer@amd.com> 9117Date: Wed Feb 8 11:35:35 2012 +0100 9118 9119 Fix UMS build failure. 9120 9121 And some UMS specific warnings. 9122 9123 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9124 9125commit 87f776b966f4200c97a989536d4b71822ae4c0b3 9126Author: Michel Dänzer <michel.daenzer@amd.com> 9127Date: Wed Feb 8 11:10:27 2012 +0100 9128 9129 Remove unused local variable 'height'. 9130 9131 Pointed out by gcc -Wunused-variable. 9132 9133 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9134 9135commit f63262e2f112a348c45f0dcecd891c8b6d9c5ee8 9136Author: Michel Dänzer <michel.daenzer@amd.com> 9137Date: Wed Feb 8 10:28:45 2012 +0100 9138 9139 evergreen: Initialize source surface member for textured video. 9140 9141 Fixes crash reported by Ole Salscheider on IRC. 9142 9143 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9144 9145commit c8f104d38870f14049402bbc14f662c151caeeef 9146Author: Jerome Glisse <jglisse@redhat.com> 9147Date: Tue Feb 7 15:04:37 2012 -0500 9148 9149 radeon: fix tiling for weird resolution 9150 9151 Should also fix xv for some case. 9152 9153 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 9154 9155commit 615033f2b5e3817e335e9d022fc9fdcf8ac8b11a 9156Author: Jerome Glisse <jglisse@redhat.com> 9157Date: Tue Dec 13 11:08:19 2011 -0500 9158 9159 r600-evergreen: use common surface allocator for tiling v11 9160 9161 Use libdrm common surface code so mesa,ddx have same idea 9162 about tiling surface and what their pitch should be and 9163 the alignment constraint. 9164 9165 v2 fix remaining issue add new option to conditionaly enable 9166 v3 fix fbcon copy and r600 exa copy path 9167 v4 fix non tiled path 2D tiling on GPU >= R600, set it to false 9168 as default 9169 v5 adapt to pixel/element size split of libdrm/radeon 9170 v6 update to properly handle falling back to 1d tiled 9171 v6 final fix to tile split value on evergreen and newer 9172 v7 fix default array mode on r6xx, fix height alignment issue 9173 on evergreen 9174 v8 fix tile split value 9175 v9 add stencil tile split support, simplify dri2 for stencil 9176 with evergreen 9177 v10 Try to fix xv path regarding tiling. Adapt to libdrm API 9178 change. Try to fix case where there is no surface which 9179 means non tiled bo. 9180 v11 check for proper libdrm 9181 9182 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 9183 9184commit 36c190671081967bac6fff48aaf66d67b639a48c 9185Author: Michel Dänzer <michel.daenzer@amd.com> 9186Date: Wed Feb 1 13:21:02 2012 +0100 9187 9188 Fix vline range calculations. 9189 9190 The range passed in is in pixmap coordinates, so the CRTC offset needs to be 9191 added to the clamping limits and subtracted from the clamped range for 9192 pre-AVIVO display engines. 9193 9194 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9195 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 9196 9197commit bb0e1531ac6949d38025d7dcb19234fee33b2acf 9198Author: Michel Dänzer <michel.daenzer@amd.com> 9199Date: Wed Feb 1 13:07:11 2012 +0100 9200 9201 Check for empty vline ranges after clamping. 9202 9203 The clamping could turn a previously non-empty range into an empty one. 9204 9205 Also, start == stop means the range is empty. 9206 9207 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9208 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 9209 9210commit 13b3aed4ef9afbcbaea1dcf0ed1acb162b240a3f 9211Author: Alex Deucher <alexdeucher@gmail.com> 9212Date: Tue Jan 10 09:35:09 2012 -0500 9213 9214 EXA/r6xx+: fix rop setting for overlapping copies 9215 9216 Need to use GXCopy for the src to temp copy, then 9217 the original rop for the temp to dest copy. 9218 9219 Noticed by: Frank Huang 9220 9221 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9222 9223commit eb6d769a087b2ed5952f477fc3f0b0625810a287 9224Author: Egbert Eich <eich@suse.de> 9225Date: Tue Nov 15 18:50:56 2011 +0100 9226 9227 DPMS: Split non-modeset CRTC DPMS function. 9228 9229 RADEONRestore() calls crtc->funcs->dpms() after most of the mode setting 9230 subsystems have been restored. This function enables the CRTCs but does 9231 more: it calls DRM pre- and post-modeset ioctls and sets up the palettes 9232 (LUTs). 9233 None of these two things are needed. Accessing the palette registers after 9234 restoring the PLLs can even lead to lockups. 9235 Thus the CRTC DPMS function is split into two parts: one that just enables 9236 /disables the CRTC and one which wraps this function and does the rest. 9237 Now the inner function can be called directly from RADEONRestore() as 9238 there is no need to go thru the RandR hooks in this function while the 9239 RandR hook uses the wrappering function so the full functionality is 9240 preserved from an RandR point of view. 9241 9242 Signed-off-by: Egbert Eich <eich@freedesktop.org> 9243 Reviewed-by: Alex Deucher <alexdeucher@gmail.com> 9244 9245commit ac51e331895b216d288bc7bd108a38b362214668 9246Author: Egbert Eich <eich@suse.de> 9247Date: Mon Nov 14 19:10:01 2011 +0100 9248 9249 UMS: Fix lockups in palette save/restore on pre-AVIVO chips. 9250 9251 The reintroduction of palette save/restore in 5efdf514 causes some 9252 pre-AVIVO chips to lock up. An investigation revealed that accessing 9253 palette registers when the associated PLL is not running is causing 9254 this. With UMS the PLL setup that is saved has been done by the BIOS 9255 typically. 9256 A similar issue was observed when VGA palette save/restore had 9257 been reinitroduced with 80eee856: 9258 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=480312 9259 and has been worked around for Linux without further investigation 9260 by 87e66ce7. 9261 To fix the issue we now 9262 a. introduce 'on-demand' palette saving (ie the palette is 9263 saved before it is first altered). This guarantees that 9264 the palette register are only associated when the associated 9265 CRTC is active and thus the PLLs are powered up and running. 9266 b. move palette restore before PLL restore. 9267 c. eliminate generic VGA palette save/restore which seems to be 9268 unneeded when the palette is restored natively. 9269 It is believed that this caused the behavior described in 9270 https://bugs.freedesktop.org/show_bug.cgi?id=18407#c27 9271 9272 Signed-off-by: Egbert Eich <eich@freedesktop.org> 9273 Reviewed-by: Alex Deucher <alexdeucher@gmail.com> 9274 9275commit 0a8d04eeac95f4db9d03ee31070bd825a7feb0b2 9276Author: Matthieu Herrb <matthieu.herrb@laas.fr> 9277Date: Sun Jan 1 18:27:54 2012 +0100 9278 9279 Update for new vgaHW API. 9280 9281 Signed-off-by: Matthieu Herrb <matthieu.herrb@laas.fr> 9282 Reviewed-by: Mark Kettenis <kettenis@openbsd.org> 9283 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9284 9285commit ae45d7e6d8e6844cd4586c9ee97c21b257fa788f 9286Author: Michel Dänzer <michel.daenzer@amd.com> 9287Date: Wed Dec 28 11:48:36 2011 +0100 9288 9289 DRI2: Can't use page flipping for pixmaps. (bug #42913) 9290 9291 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=42913 . 9292 9293 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9294 9295commit aacbd629b02cbee3f9e6a0ee452b4e3f21376bd3 9296Author: Alex Deucher <alexdeucher@gmail.com> 9297Date: Mon Dec 12 09:32:30 2011 -0500 9298 9299 radeon: add some new pci ids 9300 9301 fixes: 9302 https://bugs.freedesktop.org/show_bug.cgi?id=43739 9303 9304 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9305 9306commit bc54e415e2fd344939c5c788ea0686133a7e2c69 9307Author: Dave Airlie <airlied@redhat.com> 9308Date: Tue Dec 6 15:47:45 2011 +0000 9309 9310 radeon: add original radeon to always tiled. 9311 9312 and actually enable it for M7, previous commit only did one function. 9313 9314 Signed-off-by: Dave Airlie <airlied@redhat.com> 9315 9316commit ba46c7b0cf72d157748981eb3224d5eefb6200aa 9317Author: Dave Airlie <airlied@redhat.com> 9318Date: Tue Dec 6 13:42:49 2011 +0000 9319 9320 radeon: refine always tiled depth check 9321 9322 So it appears the M7 family always tiles its depth buffer also. 9323 9324 Signed-off-by: Dave Airlie <airlied@redhat.com> 9325 9326commit 98b2d5fe1722a43c4bbe7711ed7180a3fb65305f 9327Author: Dave Airlie <airlied@redhat.com> 9328Date: Mon Dec 5 18:44:28 2011 +0000 9329 9330 radeon: r200 depth buffers are always tiled 9331 9332 When we do the allocations we need to make sure the always tiled 9333 nature is taken into account. 9334 9335 Signed-off-by: Dave Airlie <airlied@redhat.com> 9336 9337commit 7dcefc69d9fbceae27cd03083c815e01a19b527e 9338Author: Alex Deucher <alexdeucher@gmail.com> 9339Date: Mon Dec 5 09:21:48 2011 -0500 9340 9341 Xv: Evergreen+ asics support 16k surfaces 9342 9343 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9344 9345commit 72e386d42516e7cd3c2cbf2fffc9174cd3ec8451 9346Author: Alex Deucher <alexander.deucher@amd.com> 9347Date: Wed Nov 30 19:38:35 2011 -0500 9348 9349 radeon: add some new pci ids 9350 9351 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 9352 9353commit 3853c3020d05175ae180b9a188dec7c425bdd0b8 9354Author: Dave Airlie <airlied@redhat.com> 9355Date: Mon Nov 28 18:38:30 2011 +0000 9356 9357 fixup xinerama since 9151f3b1c2ebcc34e63195888ba696f2183ba5e2 9358 9359 since the driver would call RRFirstOutput without checking if randr has 9360 been enabled, and it would crash in privates code. 9361 9362 reported by vereteran on #radeon 9363 9364 Signed-off-by: Dave Airlie <airlied@redhat.com> 9365 Acked-on-irc-by: Michel Dänzer <michel.daenzer@amd.com> 9366 9367commit d669c34f140c000f88c4b4e464e44e6c8694f581 9368Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 9369Date: Mon Nov 21 11:35:40 2011 +1100 9370 9371 ddx/evergreen: Fix endian of ALU constants 9372 9373 The constants are written directly into a buffer object shared with the 9374 card and we "forget" to swap them. This patch fixes it by doing the swap 9375 in evergreen_set_alu_consts() in-place (ie, it modifies the buffer), 9376 which should be fine with the way we use it in the ddx. 9377 9378 This makes everything work fine on my caicos card on a G5 including some 9379 quik tests with Xv, gnome3 shell, etc... 9380 9381 Thanks a lot to Jerome Glisse for holding my hand through debugging that 9382 (and finding the actual bug). 9383 9384 Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> 9385 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9386 9387commit 422bdd4fe6cb728e1dd08a56f6ee2d0f009cbfcb 9388Author: Alex Deucher <alexdeucher@gmail.com> 9389Date: Mon Nov 14 09:39:16 2011 -0500 9390 9391 radeon: add missing FireMV pci id 9392 9393 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9394 9395commit 534fb6e413a909a9d1afd57d1c711844b7c5ddf4 9396Author: Dave Airlie <airlied@redhat.com> 9397Date: Fri Nov 11 10:26:51 2011 +0000 9398 9399 ati: enable bg none when fbcon succeeds and we are built against ABI after 10. 9400 9401 One less patch to keep carrying in Fedora. 9402 9403 Signed-off-by: Dave Airlie <airlied@redhat.com> 9404 9405commit 89452c08048c98fb5cc3dc551b3824be40d52cf2 9406Author: Michel Dänzer <michel.daenzer@amd.com> 9407Date: Tue Nov 8 11:23:11 2011 +0100 9408 9409 UMS: Guard references to PCITAG / pciTag with XSERVER_LIBPCIACCESS (bug #42690) 9410 9411 Should fix https://bugs.freedesktop.org/show_bug.cgi?id=42690 . 9412 9413 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9414 9415commit 5ec34ed95948f7164184551615c1fc4c3eef3b98 9416Author: Ilija Hadzic <ihadzic@research.bell-labs.com> 9417Date: Thu Nov 3 20:16:47 2011 -0400 9418 9419 DRI/DRI2: remove hard-coded limitation to 6 crtcs 9420 9421 DRM's hard limit to the number of CRTCs is 32. ATI DDX unnecessarily 9422 clips this limit to 6 by hard coding initial assumption for 9423 output->possible_crtcs mask to 0x7f (before it gets trimmed down to 9424 what's really possible for a given output) and by allocating only 6 9425 entries for for cursor_bo[] array in RADEONInfoRec. 9426 9427 Fix this and thus allow the ATI DDX to deal with as many CRTCs 9428 as the DRM allows (32), so it is ready if anything with >6 CRTCs 9429 comes out. 9430 9431 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 9432 9433commit 4853ab2cdc3b97948c7cd69eaf4fff54f59774fc 9434Author: Michel Dänzer <michel.daenzer@amd.com> 9435Date: Fri Nov 4 12:15:53 2011 +0100 9436 9437 Turn compile time check into runtime check. 9438 9439 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9440 9441commit bcdb54fe16ebf2e239b84eebf20e8adfe5094bff 9442Author: Alex Deucher <alexdeucher@gmail.com> 9443Date: Thu Oct 20 11:11:35 2011 -0400 9444 9445 check for xserver 1.9.4.901 to enable tiling by default 9446 9447 Previous xservers had a bug in the EXA code which caused 9448 display corruption in some cases. 9449 9450 See: 9451 https://bugs.freedesktop.org/show_bug.cgi?id=33929 9452 9453 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9454 9455commit edde94cba5321e6e51e7fd4d79dde5abc4944495 9456Author: Michel Dänzer <michel.daenzer@amd.com> 9457Date: Wed Nov 2 13:40:05 2011 +0100 9458 9459 Bump version post release. 9460 9461commit 93459f842c2d8dc178a1954b8e05150fcb96ac9a 9462Author: Michel Dänzer <michel.daenzer@amd.com> 9463Date: Wed Nov 2 12:51:15 2011 +0100 9464 9465 Bump version for 6.14.3 release. 9466 9467commit fe3ce559b7f07d39ded39abb38576846d6eb185b 9468Author: Jeremy Huddleston <jeremyhu@apple.com> 9469Date: Tue Nov 1 20:27:03 2011 -0700 9470 9471 Build fix for -Werror=int-to-pointer-cast 9472 9473 Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com> 9474 9475commit d3d6a5da07aaec5961e51c9a8f90c1490ee101b6 9476Author: Tormod Volden <debian.tormod@gmail.com> 9477Date: Sat Oct 29 23:08:46 2011 +0200 9478 9479 radeon: do not include xf86PciInfo.h 9480 9481 We already use atipciids.h instead most places. 9482 9483 Signed-off-by: Tormod Volden <debian.tormod@gmail.com> 9484 9485commit 70da7001e81363ed6ef2c4727c512daf53ae29fe 9486Author: Jeremy Huddleston <jeremyhu@apple.com> 9487Date: Sat Oct 29 20:15:09 2011 -0700 9488 9489 Use malloc/calloc/realloc/free directly 9490 9491 Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com> 9492 9493commit a3bb07efb1757c33d70e2e1928219d12a4dd6498 9494Author: Michel Dänzer <michel.daenzer@amd.com> 9495Date: Tue Oct 25 17:43:58 2011 +0200 9496 9497 EXA >= R6xx / KMS: Avoid running out of CS space at inconvenient times. 9498 9499 Otherwise we may end up with things not properly set up at the beginning of the 9500 next CS. 9501 9502 Fixes http://bugs.debian.org/645007 . 9503 9504 In contrast to the Composite code for < R6xx, this isn't necessary with UMS, 9505 as the draw packet only uses constant space in the indirect buffer, and nothing 9506 else can mess with the 3D state between indirect buffers. 9507 9508 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9509 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 9510 9511commit 23788c4a5e3b6affb9b183e1393edd0e5ca4550e 9512Author: Michel Dänzer <michel.daenzer@amd.com> 9513Date: Thu Oct 27 19:12:22 2011 +0200 9514 9515 EXA < R6xx: Make sure 2D state is re-emitted after running out of CS space. 9516 9517 Otherwise it's basically luck what the 2D state ends up being at the beginning 9518 of the next CS. 9519 9520 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9521 9522commit 6e0e1a821accc6ca95f4134e49b66a6b168c1934 9523Author: Michel Dänzer <michel.daenzer@amd.com> 9524Date: Tue Oct 25 17:39:57 2011 +0200 9525 9526 Make radeon_dri2_create_buffer(s) more robust. (Bug #30047) 9527 9528 In particular, handle and propagate failure to allocate GPU accessible memory, 9529 instead of crashing. Fixes https://bugs.freedesktop.org/show_bug.cgi?id=30047 . 9530 9531 Also take care not to leak resources in error paths. 9532 9533 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9534 9535commit 856583dbca9319c77fed40daa9956e81a0068f9e 9536Author: Jerome Glisse <jglisse@redhat.com> 9537Date: Thu Oct 20 14:17:14 2011 -0400 9538 9539 radeon/kms: fallback to vesa if GPU is not supported by UMS 9540 9541 For GPU not supported by UMS, test in probe so that we properly 9542 fallback to vesa. 9543 9544 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 9545 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 9546 9547commit f6f1b4f7789cfef763c063e671b76b60be8bdd8e 9548Author: Maarten Lankhorst <m.b.lankhorst@gmail.com> 9549Date: Tue Oct 18 14:30:39 2011 +0200 9550 9551 dri2: Add vdpau driver name entry 9552 9553 libvdpau has a driver loading mechanism that looks for a dri2 driver 9554 first before falling back to nvidia, so lets use that. 9555 9556 Allows use of libvdpau_rx00 without having to set things up separately, 9557 similar to the patch to xf86-video-nouveau. 9558 9559 Signed-off-by: Maarten Lankhorst <m.b.lankhorst@gmail.com> 9560 Reviewed-by: Christian König <deathsimple@vodafone.de> 9561 Reviewed-by: Michel Dänzer <michel@daenzer.net> 9562 Tested-by: Michel Dänzer <michel@daenzer.net> 9563 9564commit a6b2bd2d184f10d4c56c4ee17186aedb238a36ec 9565Author: Alex Deucher <alexdeucher@gmail.com> 9566Date: Wed Sep 21 17:00:16 2011 -0400 9567 9568 UMS: fix DDIA enable on some rs690 systems 9569 9570 DVOOutputControl checks the value of of bios scratch reg 3 9571 on some tables and assumes the encoder is already enabled 9572 if the DFP2_ACTIVE bit is set. Clear that bit so the table 9573 sets the DDIA enable bit properly. 9574 9575 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9576 9577commit d78860ba53d9bfcf6c28e1cfd2d970709b5e20fa 9578Author: Michel Dänzer <michel.daenzer@amd.com> 9579Date: Fri Aug 12 11:21:33 2011 +0200 9580 9581 Only call radeon_dri2_close_screen() if DRI2 was enabled. 9582 9583 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9584 9585commit 11330ca5dc61a70fe4507e63230f9133ca22d891 9586Author: Michel Dänzer <michel.daenzer@amd.com> 9587Date: Fri Aug 12 11:21:32 2011 +0200 9588 9589 Remove dead variable remain_size_bytes. 9590 9591 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9592 9593commit fcf0cca9c0ab0f692b222f619aee8f1cdad3b519 9594Author: Michel Dänzer <michel.daenzer@amd.com> 9595Date: Tue Sep 20 12:34:05 2011 +0200 9596 9597 KMS Color Tiling requires xserver which supports EXA_MIXED_PIXMAPS. 9598 9599 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9600 9601commit c96e6fb8a5f5be2319fdb4c431c1ba5279997fe8 9602Author: Alex Deucher <alexdeucher@gmail.com> 9603Date: Sat Sep 17 08:26:12 2011 -0400 9604 9605 man: note that the list of marketing names is non-exhaustive 9606 9607 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9608 9609commit 1a51fce6f6ab169c882a86b936909c0820f27a68 9610Author: Alex Deucher <alexdeucher@gmail.com> 9611Date: Thu Sep 15 17:52:25 2011 -0400 9612 9613 update man page with new marking names 9614 9615 Note that the driver support all currently 9616 shipping asics and the names in the man page 9617 are just a sampling. 9618 9619 Fixes: 9620 https://bugs.freedesktop.org/show_bug.cgi?id=40808 9621 9622 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9623 9624commit 64f237a4cf3ef5bcb3163e88e1447ff275a1eefa 9625Author: Michel Dänzer <michel.daenzer@amd.com> 9626Date: Thu Aug 18 19:11:08 2011 +0200 9627 9628 Convert register ranges for >= r6xx from enums to defines. 9629 9630 Avoids lots of "comparison between 'enum <anonymous>' and 'enum <anonymous>'" 9631 warnings with newer versions of gcc. See 9632 https://bugs.freedesktop.org/show_bug.cgi?id=38238 . 9633 9634 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9635 9636commit 95991fcce45f0dd904f76b9f98f9c7ed5708e4fa 9637Author: Michel Dänzer <michel.daenzer@amd.com> 9638Date: Wed Aug 17 11:10:34 2011 +0200 9639 9640 Bail if we're trying to start up in UMS mode on KMS. 9641 9642 Ideally, the display manager will start the X server again, and everything 9643 will be fine and dandy. But in the worst case, at least we won't hit the 9644 hardware behind the KMS driver's back. 9645 9646 (This change intentionally makes (ab)use of the fact that Bool is defined as 9647 int). 9648 9649 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9650 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 9651 9652commit f95a41b7851565c282d22f8d679db1377428f165 9653Author: Michel Dänzer <michel.daenzer@amd.com> 9654Date: Thu Aug 11 11:22:57 2011 +0200 9655 9656 video: Don't round up bottom/right edge for clipping source width/height. 9657 9658 It's not necessary: If the top/left edge was rounded down, this will be 9659 compensated by the subtraction. 9660 9661 Worse, if the original source width/height is odd, rounding up may result in 9662 reading past the end of the source data. 9663 9664 Fixes http://bugs.debian.org/637258 . 9665 9666 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9667 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> 9668 9669commit 93fc0843a1e31dc9237433bc2bf17df79e956d26 9670Author: Michel Dänzer <michel.daenzer@amd.com> 9671Date: Wed Aug 10 17:44:37 2011 +0200 9672 9673 Change my e-mail address to something that still works, and always will, I hope. 9674 9675commit 9151f3b1c2ebcc34e63195888ba696f2183ba5e2 9676Author: Michel Dänzer <michel.daenzer@amd.com> 9677Date: Tue Aug 9 19:13:26 2011 +0200 9678 9679 Prefer the CRTC of the primary output for synchronization. 9680 9681 See https://bugs.freedesktop.org/show_bug.cgi?id=39696 . 9682 9683 Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> 9684 9685commit 3b9fdc807dd7e52af0576299cefba596040f6f2f 9686Author: Alex Deucher <alexdeucher@gmail.com> 9687Date: Wed Aug 3 16:20:13 2011 -0400 9688 9689 r5xx+: Fix vline setup with crtc offsets 9690 9691 On r5xx+, vline is relative to to the viewport, not 9692 the scanlines. Based on initial patch and investigation 9693 from Herbert Pötzl (Bertl) on IRC. 9694 9695 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9696 9697commit 104b2d7c071f29266b1bc4184a74e9714d14febc 9698Author: Alex Deucher <alexander.deucher@amd.com> 9699Date: Mon Aug 1 10:05:30 2011 -0400 9700 9701 kms: fix possible leak in pageflip code 9702 9703 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 9704 9705commit d29bab632e9ecccba518d4107d52620bf75eb1cf 9706Author: Ville Syrjala <syrjala@sci.fi> 9707Date: Wed May 4 23:51:27 2011 +0300 9708 9709 kms: Move flip_count and co. to a per swap structure 9710 9711 If multiple drawables are doing page flipping, the global drmmode 9712 structure can't be used to keep per swap information. For example 9713 flip_count can increase prematurely due to another swap request, 9714 and then the previous swap request never gets completed, leading to a 9715 stuck client. Move the relevant pieces of data to a strucuture that 9716 gets allocated once per swap request and shared by all involved CRTCs. 9717 9718 Signed-off-by: Ville Syrjala <syrjala@sci.fi> 9719 9720commit 9493563c1ef4b51af0ee8a44cb4e7c5bb280347e 9721Author: Ville Syrjala <syrjala@sci.fi> 9722Date: Wed May 4 23:51:26 2011 +0300 9723 9724 dri2: Update front buffer pixmap and name before exchanging buffers 9725 9726 Buffer exchange assumes that the front buffer pixmap and name 9727 information is accurate. That may not be the case eg. if the window 9728 has been (un)redirected since the buffer was created. 9729 9730 Signed-off-by: Ville Syrjala <syrjala@sci.fi> 9731 9732commit 8c9266ed2da22a510243f9a952c14d4423f48a2b 9733Author: Alex Deucher <alexander.deucher@amd.com> 9734Date: Fri Jul 15 10:44:57 2011 -0400 9735 9736 radeon: add some new NI pci ids 9737 9738 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> 9739 9740commit f59c3b294b0f715fc96e2bbe25893f2b31aa488b 9741Author: Christian König <deathsimple@vodafone.de> 9742Date: Thu Jul 14 11:49:06 2011 +0200 9743 9744 Register XvMC video decoding acceleration 9745 9746commit e8d0d437957b15252dfad775796a3949ed50dbcf 9747Author: Dave Airlie <airlied@redhat.com> 9748Date: Tue Jul 12 11:43:25 2011 -0400 9749 9750 evergreen: Emit SQ_LDS_RESOURCE_MGMT 9751 9752 Avoids rendering problems when compute changes this reg. 9753 9754 Fixes: 9755 https://bugs.freedesktop.org/show_bug.cgi?id=39119 9756 9757 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9758 9759commit 9bb31158466e6168116d841d12c8b4303f11c4a6 9760Author: Alex Deucher <alexdeucher@gmail.com> 9761Date: Wed Jun 22 12:24:28 2011 -0400 9762 9763 evergreen: fix num_banks for 2D tiling config 9764 9765 The field is encoded. 9766 9767 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9768 9769commit 122b471f734aa07427b01d4bec35ff1ac28290b5 9770Author: Alex Deucher <alexdeucher@gmail.com> 9771Date: Fri Jun 17 04:02:16 2011 -0400 9772 9773 dri2: fix copy pasto in a6154c00c64932332e8f6e334661ffd579cfd894 9774 9775 Reported-by: Nils Wallménius <nils.wallmenius@gmail.com> 9776 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9777 9778commit 122bedcbcf45cb583cf51b2fd04ed2805e0ca60b 9779Author: Alex Deucher <alexdeucher@gmail.com> 9780Date: Thu Jun 16 12:57:11 2011 -0400 9781 9782 dri2: missing bit from a6154c00c64932332e8f6e334661ffd579cfd894 9783 9784 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9785 9786commit a6154c00c64932332e8f6e334661ffd579cfd894 9787Author: Alex Deucher <alexdeucher@gmail.com> 9788Date: Thu Jun 16 12:48:43 2011 -0400 9789 9790 dri2/eg+: fix size and alignment of depth/stencil buffers 9791 9792 Base alignment may be 256B or 512B depending on the group 9793 size. Also need to check against front size for virtualX. 9794 9795 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9796 9797commit 41eb1fbb3d9da64feb4a96df7c575e44136b1538 9798Author: Alex Deucher <alexdeucher@gmail.com> 9799Date: Mon Jun 13 12:50:39 2011 -0400 9800 9801 kms/man: update ColorTiling info 9802 9803 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9804 9805commit 481baa5c74271cd7ce38bae3965d2bc4b8809058 9806Author: Alex Deucher <alexdeucher@gmail.com> 9807Date: Mon Jun 13 12:44:07 2011 -0400 9808 9809 kms: enable ColorTiling by default on r6xx-cayman asics 9810 9811 Currently only 1D tiling as 2D tiling still has some corner 9812 cases to fix up. 9813 9814 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9815 9816commit cbcc57b0fa6f581be777bef648f2bf3efe7443ee 9817Author: Alex Deucher <alexdeucher@gmail.com> 9818Date: Mon Apr 4 12:52:00 2011 -0400 9819 9820 radeon: add llano pci ids 9821 9822 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9823 9824commit 36afd1e1055eeadb2396dadcc68b214655bd90a9 9825Author: Alex Deucher <alexdeucher@gmail.com> 9826Date: Tue May 31 16:03:36 2011 -0400 9827 9828 radeon: add support for llano APUs 9829 9830 - KMS only 9831 - Includes full EXA/Xv support 9832 9833 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9834 9835commit 557f46dc2f18734ecf1f18dee7e951e0bf062e63 9836Author: Dave Airlie <airlied@redhat.com> 9837Date: Fri May 27 07:22:08 2011 +1000 9838 9839 bump version after release 9840 9841commit 2fca40ea65d9f2a6f8451c324bb4b82786f34f76 9842Author: Dave Airlie <airlied@redhat.com> 9843Date: Thu May 26 12:52:21 2011 +1000 9844 9845 radeon: bump version for release of 6.14.2 9846 9847commit fab868c5f7d8cafdb0176d2751f216819a5ba66a 9848Author: Alex Deucher <alexdeucher@gmail.com> 9849Date: Wed May 25 01:09:12 2011 -0400 9850 9851 EXA: make evergreen_fix_scissor_coordinates static 9852 9853 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9854 9855commit fe5c42f5155361006b687da824181418f688809f 9856Author: Alex Deucher <alexdeucher@gmail.com> 9857Date: Tue May 24 22:32:01 2011 -0400 9858 9859 cayman: endian fixes for shaders 9860 9861 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9862 9863commit 470ecd02347c32e79316046d01a7d5dad0e2fe99 9864Author: Alex Deucher <alexdeucher@gmail.com> 9865Date: Tue May 24 18:31:47 2011 -0400 9866 9867 EXA/Xv: add workarounds for eg/cayman scissors bugs 9868 9869 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9870 9871commit b913e7ba2b60d47a6660699210cc3cf6f5dc52c2 9872Author: Alex Deucher <alexdeucher@gmail.com> 9873Date: Tue May 24 18:02:52 2011 -0400 9874 9875 Revert "cayman: add scissors workaround." 9876 9877 Needs a proper workaround for a hw bug. 9878 9879 This reverts commit b77d374b0d11f48c33cfffdb4157c4ec4b05ea15. 9880 9881commit 82cb33c3f0e1ba802d7a94f3159b3c5c86cd4043 9882Author: Dave Airlie <airlied@redhat.com> 9883Date: Wed May 18 14:49:17 2011 +1000 9884 9885 cayman: enable all accel 9886 9887commit b77d374b0d11f48c33cfffdb4157c4ec4b05ea15 9888Author: Dave Airlie <airlied@redhat.com> 9889Date: Tue May 24 15:53:58 2011 +1000 9890 9891 cayman: add scissors workaround. 9892 9893 wasted a lot of time getting to this. 9894 9895commit ffeab7a7058298e15294a3b2c740c731e36dda1d 9896Author: Alex Deucher <alexdeucher@gmail.com> 9897Date: Mon Apr 18 18:16:51 2011 -0400 9898 9899 cayman: fix dword counts default state 9900 9901 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9902 9903commit 3cbfae361bf5e779d3364f0f31cfd25bd0f59e65 9904Author: Alex Deucher <alexdeucher@gmail.com> 9905Date: Wed Mar 2 20:48:19 2011 -0500 9906 9907 cayman: add spi state to default state 9908 9909 changed in e3145801b80fd4be4cf770128876e86e89bda66f 9910 9911 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9912 9913commit b8ade97c9d0fa5aacb0e3166868bb72e9bc679a6 9914Author: Alex Deucher <alexdeucher@gmail.com> 9915Date: Wed Mar 2 20:44:19 2011 -0500 9916 9917 cayman: first pass at exa/Xv shaders 9918 9919 Main differences with evergreen: 9920 - 4-way rather than 5-way 9921 - END_OF_PROGRAM bit removed from CF istructions, use 9922 CF_INST_END instead. 9923 - MEGA_FETCH* fields removed from VTX commands 9924 - no more VC, all fetches go through the TC 9925 9926 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9927 9928commit 01b646ed800732985c1638b147716641a99082f9 9929Author: Alex Deucher <alexdeucher@gmail.com> 9930Date: Wed Mar 2 20:39:38 2011 -0500 9931 9932 cayman: add a default state function 9933 9934 The rest of the state functions can be shared 9935 with evergreen. I've noted where there are 9936 differences. 9937 9938 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9939 9940commit 42eecc6f4fb1570769490bdaeac06817c6c36a7e 9941Author: Alex Deucher <alexdeucher@gmail.com> 9942Date: Wed Mar 2 20:20:56 2011 -0500 9943 9944 cayman: add 3D register headers 9945 9946 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9947 9948commit e1d28e011f4a5139cbc778973c63158ed2746716 9949Author: Alex Deucher <alexdeucher@gmail.com> 9950Date: Wed Mar 2 20:13:50 2011 -0500 9951 9952 kms/cayman: stub out exa support 9953 9954 Just fallbacks for now. 9955 9956 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 9957 9958commit 21e44a20b8b1b64079ee77f45aaa5010206ed7b6 9959Author: Michel Dänzer <daenzer@vmware.com> 9960Date: Thu May 12 09:23:38 2011 +0200 9961 9962 UMS: Fix comparison of unsigned variable against < 0. 9963 9964 Pointed out by clang: 9965 9966 ../../src/radeon_crtc.c:242:18: error: comparison of unsigned expression < 0 is always false [-Werror,-Wtautological-compare] 9967 error = error < 0 ? 0xffffffff : error; 9968 ~~~~~ ^ ~ 9969 9970 If a UMS regression is bisected to this commit, the assignment should probably 9971 just be removed, as it's a no-op in the current form. 9972 9973commit 3b893d81982c9381393c92625e308541e0071b05 9974Author: Michel Dänzer <daenzer@vmware.com> 9975Date: Fri May 13 09:10:02 2011 +0200 9976 9977 KMS: Fix output properties logic error. 9978 9979 Pointed out by clang: 9980 9981 ../../src/drmmode_display.c:1023:30: error: use of logical && with constant operand; switch to bitwise & or remove constant [-Werror,-Wconstant-logical-operand] 9982 if (props && (props->flags && DRM_MODE_PROP_ENUM)) { 9983 ^ ~~~~~~~~~~~~~~~~~~ 9984 9985 Reviewed-by: Alex Deucher <alexdeucher@gmail.com> 9986 9987commit f83d58cf5b33686139067f8f898b8e566ba5c253 9988Author: Nicolas Kaiser <nikai@nikai.net> 9989Date: Fri May 13 00:56:31 2011 +0200 9990 9991 man: fix typos 9992 9993 Signed-off-by: Nicolas Kaiser <nikai@nikai.net> 9994 9995commit 90abffbd30f44b9cf76a6e28103ddcb5419b4522 9996Author: Ilija Hadzic <ihadzic@research.bell-labs.com> 9997Date: Fri May 6 09:45:23 2011 -0400 9998 9999 DRI2: fix high-crtc/vblank oversight/bug 10000 10001 improvements to high-crtc handling done in 10002 f0b7d7b449cc77bb2b281d81108507f8bc2e6018 introduced a bug that caused 10003 the populate_vbl_request_type to never use the high-crtc field even 10004 when it should. The reason is that the offending patch put the code 10005 under #ifdef DRM_VBLANK_HIGH_CRTC_MASK which is not visible outside the 10006 enum type, so #else was always taken in compilation type. This patch 10007 fixes it by basing #ifdef on (pre-processor visible) 10008 DRM_VBLANK_HIGH_CRTC_SHIFT constant 10009 10010 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 10011 10012commit 62a4cd180fe884dca24586d453395472516e6496 10013Author: Alex Deucher <alexdeucher@gmail.com> 10014Date: Wed May 4 01:13:55 2011 -0400 10015 10016 fusion: fix tiling enable logic 10017 10018 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10019 10020commit 76638ca687b02d3b1494b9868f817fd4fd892c64 10021Author: Alex Deucher <alexdeucher@gmail.com> 10022Date: Wed May 4 01:06:22 2011 -0400 10023 10024 fusion: enable tiling if DFS works 10025 10026 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10027 10028commit a6d2dba6573a3512d550d7e442bf42ea03012bbc 10029Author: Dave Airlie <airlied@redhat.com> 10030Date: Wed May 4 10:44:43 2011 +1000 10031 10032 radeon: add add hw DFS support for fusion 10033 10034 Fusion had a bug setting up the VM on earlier kernels so we need to work 10035 around that and only enable accel on a new enough kernel. 10036 10037 Signed-off-by: Dave Airlie <airlied@redhat.com> 10038 10039commit 859e052af49e68a826b77a9135c7f067dc331a06 10040Author: Alex Deucher <alexdeucher@gmail.com> 10041Date: Tue May 3 15:15:04 2011 -0400 10042 10043 radeon: add some new pci ids 10044 10045 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10046 10047commit 8f8bbf628c6eed037f57bc8c155f0ecdacbebad1 10048Author: Alex Deucher <alexdeucher@gmail.com> 10049Date: Sun May 1 13:19:15 2011 -0400 10050 10051 man: add cayman to man page 10052 10053 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10054 10055commit 903e90c31cf0319be9297529aa7b8daa1756cf63 10056Author: Alex Deucher <alexdeucher@gmail.com> 10057Date: Wed Apr 20 03:10:08 2011 -0400 10058 10059 EXA/Xv: used cached bo tiling flags for accel setup on 6xx+ 10060 10061 This avoids calling into the kernel for each bo in the accel 10062 code. This is a follow on to: 10063 cc7d1fa39da40a532fcdbe6c7924ca47a879e66a 10064 10065 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10066 10067commit 982c22f16c8eeee9be81779fbfe17d8d3f9b6897 10068Author: Adam Jackson <ajax@redhat.com> 10069Date: Thu Apr 14 16:04:50 2011 -0400 10070 10071 R520: Fix textures larger than 2k 10072 10073 Ported from the equivalent fix in Mesa. 10074 10075commit cc7d1fa39da40a532fcdbe6c7924ca47a879e66a 10076Author: Michel Dänzer <daenzer@vmware.com> 10077Date: Mon Apr 4 17:37:12 2011 +0200 10078 10079 EXA: Cache BO tiling flags. 10080 10081 Calling into the kernel every time is quite expensive, and nobody else should 10082 ever change the tiling flags. 10083 10084 There's still more to do along the same lines for >= R6xx. 10085 10086commit f0b7d7b449cc77bb2b281d81108507f8bc2e6018 10087Author: Michel Dänzer <daenzer@vmware.com> 10088Date: Tue Apr 5 13:36:01 2011 +0200 10089 10090 DRI2: Some cleanups for the scheduling mess. 10091 10092 * Fix build against libdrm that doesn't define *_VBLANK_HIGH_CRTC*. 10093 * If we have more than two CRTCs but can't use DRM_VBLANK_HIGH_CRTC_MASK, don't 10094 enable scheduling in the first place rather than relying on 10095 DRM_VBLANK_SECONDARY magically doing something sensible for higher CRTCs. 10096 * Only set up client state tracking when scheduling is enabled. 10097 * Only declare pRADEONEnt when it's needed, and break long lines. 10098 10099commit 7acf9bc833de539fa2259a051c66a99445a54bc4 10100Author: Alex Deucher <alexdeucher@gmail.com> 10101Date: Mon Apr 4 11:08:37 2011 -0400 10102 10103 radeon: add some additional ontario pci ids 10104 10105 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10106 10107commit 0af6386f85a0f1ba14864a3334164733a10a6cb8 10108Author: Ilija Hadzic <ihadzic@research.bell-labs.com> 10109Date: Thu Mar 24 13:33:27 2011 -0400 10110 10111 xf86-video-ati: (revised #2) add support for vblank on crtc > 1 10112 10113 Hi Alex, 10114 10115 Enclosed is a revised version of two patches sent on Mar 18 and Mar 22, 10116 respectively. Details summarized in these two threads: 10117 http://lists.freedesktop.org/archives/dri-devel/2011-March/009463.html 10118 http://lists.freedesktop.org/archives/dri-devel/2011-March/009582.html 10119 10120 This patch reconciles the DDX with the change in libdrm sent to this list 10121 earlier today. Specifically, it refers to a symbol that has been renamed 10122 from DRM_CAP_HIGH_CRTC to DRM_CAP_VBLANK_HIGH_CRTC. It *supersedes* the 10123 previous patch (i.e. apply it to the master branch as it exists at the 10124 time of this writing, not as an incremental patch to the one sent previously). 10125 10126 Regards, 10127 10128 Ilija 10129 10130 Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com> 10131 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10132 10133commit fe2e0ad3ffa58f40311319c950b842e2928a5740 10134Author: matthew green <mrg@eterna.com.au> 10135Date: Mon Mar 21 12:17:58 2011 -0400 10136 10137 bug fix for r6xx/r7xx UMS 10138 10139 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10140 10141commit 83978ad5fe37581e2b2f3fbd9c073d91b2ae1d50 10142Author: Alex Deucher <alexdeucher@gmail.com> 10143Date: Thu Mar 17 19:07:43 2011 -0400 10144 10145 bump version post release 10146 10147 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10148 10149commit 38d9368e59b2990bf32a028ece2132451b402350 10150Author: Alex Deucher <alexdeucher@gmail.com> 10151Date: Thu Mar 17 18:59:55 2011 -0400 10152 10153 bump for release 10154 10155 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10156 10157commit cdfc007eccc9b130fc08012deef304488eb6e469 10158Author: Dave Airlie <airlied@redhat.com> 10159Date: Wed Mar 16 10:40:42 2011 +1000 10160 10161 radeon/exa: correct function name 10162 10163 this corrects the function name so it matches the contents. 10164 10165 Signed-off-by: Dave Airlie <airlied@redhat.com> 10166 10167commit b93d7658f669f6dc1cfacebcfe955a1e113a537c 10168Author: Dave Airlie <airlied@redhat.com> 10169Date: Wed Mar 16 10:37:22 2011 +1000 10170 10171 radeon/exa: fix scaling check 10172 10173 This caused a regression in firefox, as these numbers are xFixed 10174 values hence 1 is actually 65536. 10175 10176 Should fix: https://bugzilla.redhat.com/show_bug.cgi?id=685084 10177 10178 Signed-off-by: Dave Airlie <airlied@redhat.com> 10179 10180commit 8567b8cd9e136ef69906e02286b3752db2404741 10181Author: Alex Deucher <alexdeucher@gmail.com> 10182Date: Tue Mar 15 13:13:06 2011 -0400 10183 10184 APU: no tiling yet 10185 10186 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10187 10188commit 4d350497012fa31a417ada662006e2d64db2a4b5 10189Author: Dave Airlie <airlied@redhat.com> 10190Date: Tue Mar 15 10:32:19 2011 +1000 10191 10192 radeon: exa shaders don't handle scaling either. 10193 10194 rendercheck tsrccoords test fails. 10195 10196 Signed-off-by: Dave Airlie <airlied@redhat.com> 10197 10198commit 6319a33cb7bd8dba1eaf47bee1e2608cee5d7600 10199Author: Alex Deucher <alexdeucher@gmail.com> 10200Date: Wed Mar 2 19:38:04 2011 -0500 10201 10202 kms: add cayman pci ids 10203 10204 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10205 10206commit fd65def45a57e3242596274d270791df8abc8e29 10207Author: Alex Deucher <alexdeucher@gmail.com> 10208Date: Thu Jan 13 17:02:31 2011 -0500 10209 10210 kms: no accel yet for cayman 10211 10212 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10213 10214commit 585e2a3b803b6d3e070f2427e9cb49bab0949239 10215Author: Alex Deucher <alexdeucher@gmail.com> 10216Date: Thu Jan 13 16:58:01 2011 -0500 10217 10218 kms: add cayman chip family 10219 10220 KMS only. No UMS support for cayman. 10221 10222 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10223 10224commit 6835e22256ba7a74c02504383f63baaba7e87156 10225Author: Alex Deucher <alexdeucher@gmail.com> 10226Date: Wed Mar 2 12:20:45 2011 -0500 10227 10228 r6xx+: truncate point sampled coordinates 10229 10230 By default the hardware rounds texcoords. However, 10231 for point sampled textures, the expected behavior is 10232 to truncate. When we have point sampled textures, 10233 set the truncate bit in the sampler. 10234 10235 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10236 10237commit 8a8862e4165dc5be779bce5e327a31da4826db3b 10238Author: Sedat Dilek <sedat.dilek@googlemail.com> 10239Date: Sat Feb 26 21:04:22 2011 +0100 10240 10241 kms: Fix warning XNFprintf is deprecated (v2) 10242 10243 The first take of my patch missed the breakage against pre-1.10 10244 xservers and now I know XNFasprintf() is not a 1:1 replacement 10245 for XNFprintf(). 10246 The idea for the fix seemed to be so obvious, but as usual "really" 10247 test the patch you are sending to any ML. 10248 I am awfully sorry for the overhasty bad hack. 10249 Thanks to Alan and Alex for the help to fix it correctly, now. 10250 10251 Tested-By: Sedat Dilek <sedat.dilek@gmail.com> 10252 Signed-off-by: Sedat Dilek <sedat.dilek@gmail.com> 10253 10254commit 88304fe9c0ce69012d44a77a5368c35236511dac 10255Author: Alex Deucher <alexdeucher@gmail.com> 10256Date: Mon Feb 28 11:29:19 2011 -0500 10257 10258 kms: don't rotate if acceleration is not enabled 10259 10260 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10261 10262commit 99141761e947e2153ed1a3bc485dc5716a22acb1 10263Author: Alex Deucher <alexdeucher@gmail.com> 10264Date: Mon Feb 28 11:20:44 2011 -0500 10265 10266 kms: don't enable tiling if accel is off 10267 10268 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10269 10270commit 7928c120fb629a782846bcd5734d9fa14cbb099d 10271Author: Dave Airlie <airlied@redhat.com> 10272Date: Mon Feb 28 14:11:57 2011 +1000 10273 10274 xv: fix height alignments for U/V planes 10275 10276 The kernel CS checks showed we were incorrectly aligning the 10277 U/V allocations for when we used them as texture sources. 10278 10279 This should fix 10280 https://bugs.freedesktop.org/show_bug.cgi?id=34567 10281 10282 Signed-off-by: Dave Airlie <airlied@redhat.com> 10283 10284commit e55520c0e582f2a500b3b01c2cb78750e15256ca 10285Author: Michel Dänzer <daenzer@vmware.com> 10286Date: Sat Feb 26 15:27:59 2011 +0100 10287 10288 Revert "kms: Fix warning XNFprintf is deprecated" 10289 10290 This reverts commit cb4dc7ab66016d0c980800f8dbf71bbe993889d3. 10291 10292 Looks like this caused https://bugs.freedesktop.org/attachment.cgi?id=43852 , 10293 but even if it didn't, the following needs to be addressed first: 10294 10295 cc1: warnings being treated as errors 10296 ../../src/radeon_kms.c: In function ‘radeon_open_drm_master’: 10297 ../../src/radeon_kms.c:413: error: passing argument 1 of ‘XNFasprintf’ from incompatible pointer type 10298 /usr/local/include/xorg/Xprintf.h:57: note: expected ‘char **’ but argument is of type ‘char *’ 10299 ../../src/radeon_kms.c:413: error: passing argument 2 of ‘XNFasprintf’ makes pointer from integer without a cast 10300 /usr/local/include/xorg/Xprintf.h:57: note: expected ‘const char * restrict’ but argument is of type ‘uint16_t’ 10301 ../../src/radeon_kms.c:413: error: assignment makes pointer from integer without a cast 10302 make[2]: *** [radeon_kms.lo] Error 1 10303 10304commit cb4dc7ab66016d0c980800f8dbf71bbe993889d3 10305Author: Sedat Dilek <sedat.dilek@gmail.com> 10306Date: Fri Feb 25 17:21:52 2011 -0500 10307 10308 kms: Fix warning XNFprintf is deprecated 10309 10310 Noticed when building ati/radeon ddx against xserver 1.10-rc3. 10311 10312 [ build.log ] 10313 ../../src/radeon_kms.c:412:5: warning: 'XNFprintf' is deprecated (declared at /usr/include/xorg/os.h:273) 10314 10315 agd5f: don't break the build against pre-1.10 xservers. 10316 10317 Signed-off-by: Sedat Dilek <sedat.dilek@gmail.com> 10318 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10319 10320commit ecfdb209afe2aafc378baab8c511f5df7b000270 10321Author: Sedat Dilek <sedat.dilek@googlemail.com> 10322Date: Fri Feb 25 21:48:14 2011 +0100 10323 10324 UMS: Fix build against xserver 1.10-rc3 10325 10326 This issue was introduced due to last minute backout of RandR-1.4 10327 in xserver 1.10-rc3. 10328 10329 Switch to "#ifdef RANDR_14_INTERFACE" as suggested by Keith Packard. 10330 See also <http://lists.x.org/archives/xorg-devel/2011-February/019643.html>. 10331 10332 Note: 10333 The ddx needs a rebuild as the X video driver ABI changed to version 10.0. 10334 10335 Reported-by: Alex Deucher <alexdeucher@gmail.com> 10336 CC: Keith Packard <keithp@keithp.com> 10337 Signed-off-by: Sedat Dilek <sedat.dilek@gmail.com> 10338 10339commit 04293a3a84acd96ccdf8e5e865cf973325ca43a1 10340Author: Alex Deucher <alexdeucher@gmail.com> 10341Date: Thu Feb 24 02:33:41 2011 -0500 10342 10343 dri2: disable pageflipping for transformed displays 10344 10345 Based on Ben's nouveau patch. 10346 10347 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10348 10349commit 91070cfd75d5607c4a72ace780f830f0ddb40e84 10350Author: Alex Deucher <alexdeucher@gmail.com> 10351Date: Wed Feb 23 00:55:28 2011 -0500 10352 10353 Xv: use aligned height to calculate Xv buffer size 10354 10355 Make sure it's large enough for the tiling mode so the 10356 CS checker doesn't complain. Should fix: 10357 https://bugs.freedesktop.org/show_bug.cgi?id=34567 10358 10359 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10360 10361commit acd54a48a6680956ed903fc890204c0c910b6712 10362Author: Mark Kettenis <kettenis@openbsd.org> 10363Date: Sun Feb 20 22:51:04 2011 +0100 10364 10365 Properly restore CLOCK_CNTL_INDEX register 10366 10367 RADEONRestore() restores the CLOCK_CNTL_INDEX register before calling 10368 RADEONRestorePLLRegisters(), which modifies the CLOCK_CNTL_INDEX register 10369 to access the PLL registers. As a result we may end up with the wrong clock 10370 being selected when exiting X. This happens on platforms where the driver 10371 doesn't save and restore the VGA state. 10372 10373 Signed-off-by: Mark Kettenis <kettenis@openbsd.org> 10374 10375commit d7c2ba4286ea32401c0026cb92ec167f7b5753d8 10376Author: Alex Deucher <alexdeucher@gmail.com> 10377Date: Wed Feb 16 18:30:34 2011 -0500 10378 10379 man: add btc chips to the list 10380 10381 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10382 10383commit d0f8bffbbc26818eccdd305ef142387dc4975bba 10384Author: Alex Deucher <alexdeucher@gmail.com> 10385Date: Wed Feb 16 15:09:20 2011 -0500 10386 10387 kms: add tiling support for evergreen/NI 10388 10389 - requires 2.6.37 or newer 10390 - 1D only at the moment 10391 - disabled by default 10392 10393 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10394 10395commit 0471d8412acd82e281a35fc4c6bb2d53b1ff5802 10396Author: Alex Deucher <alexdeucher@gmail.com> 10397Date: Wed Feb 16 15:07:35 2011 -0500 10398 10399 kms: EXA/Xv tiling fixes 10400 10401 - properly set tiling flags for temp surfaces 10402 - fix CB non_disp_tiling bits on evergreen 10403 10404 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10405 10406commit 4d7e1498f7d9eb50e2eddabca193fc27bde24f0e 10407Author: Alex Deucher <alexdeucher@gmail.com> 10408Date: Tue Feb 15 11:24:23 2011 -0500 10409 10410 fix Xv on pre-r6xx asics 10411 10412 broken by: 10413 a9c1358099:"Xv: fix textured video alignment" 10414 10415 Reported by: 10416 Artur Skawina <art.08.09@gmail.com> 10417 10418 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10419 10420commit a9a59717d11af37a2dda5555f6a83c5b65449527 10421Author: Alex Deucher <alexdeucher@gmail.com> 10422Date: Sat Feb 12 05:34:50 2011 -0500 10423 10424 kms: use worst case base/pitch align if we don't have drm tiling info 10425 10426 To avoid CS rejection. 10427 10428 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10429 10430commit 5f9b7fdb4ce27f8c6a64134bb4d5ebeb1c958cca 10431Author: Alex Deucher <alexdeucher@gmail.com> 10432Date: Sat Feb 12 05:19:12 2011 -0500 10433 10434 kms: always get the tiling info even if we aren't tiling 10435 10436 We need the proper group size to calculate alignment so the 10437 ddx and drm agree. 10438 10439 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10440 10441commit 355f0f7a67ce239b7fd2f2e81bcaaa88b2a09f03 10442Author: Alex Deucher <alexdeucher@gmail.com> 10443Date: Sat Feb 12 04:52:52 2011 -0500 10444 10445 kms: fix rotate pitch align 10446 10447 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10448 10449commit f1dc419c989addc4737aed06ec8b8acdb4d40063 10450Author: Alex Deucher <alexdeucher@gmail.com> 10451Date: Fri Feb 11 17:21:10 2011 -0500 10452 10453 kms: evergreen/ni big endian accel support 10454 10455 Based on 6xx/7xx patches from Cédric Cano. 10456 10457 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10458 10459commit e8dc728a549323f1babe337b9d42ad504af1ca39 10460Author: Cédric Cano <ccano@interfaceconcept.com> 10461Date: Fri Feb 11 17:06:47 2011 -0500 10462 10463 ums: atombios endian fixes 10464 10465 agd5f: fix a few more cases. 10466 10467 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10468 10469commit 48ffad957f1dbca909515ffa00629f4caa68706b 10470Author: Cédric Cano <ccano@interfaceconcept.com> 10471Date: Fri Feb 11 17:00:31 2011 -0500 10472 10473 kms: 6xx/7xx big endian accel support 10474 10475 agd5f: minor cleanups 10476 10477 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10478 10479commit 151b22bd7c3b1002a7261538611fb2b468815c86 10480Author: Cédric Cano <ccano@interfaceconcept.com> 10481Date: Fri Feb 11 16:33:16 2011 -0500 10482 10483 kms/6xx+: endian swap cursor uploads 10484 10485 agd5f: remove kms surface for cursor since 10486 we now do the swap in the cursor load to avoid 10487 breaking pre-6xx cards with BE. 10488 10489 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10490 10491commit 309fac041f998e20338d320aabaeda5cc1bb6348 10492Author: Alex Deucher <alexdeucher@gmail.com> 10493Date: Fri Feb 11 14:32:45 2011 -0500 10494 10495 kms/exa/xv: fix alignment checking in accel state setup 10496 10497 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10498 10499commit 25fe3ec4c6dfcdf6ba72a032bc9cb2c84d84029a 10500Author: Alex Deucher <alexdeucher@gmail.com> 10501Date: Fri Feb 11 14:03:56 2011 -0500 10502 10503 kms/exa: UTS/DFS base alignment fixes 10504 10505 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10506 10507commit 3d10278ce511f5dabb68ed86ee43eaaf43585983 10508Author: Alex Deucher <alexdeucher@gmail.com> 10509Date: Fri Feb 11 13:51:11 2011 -0500 10510 10511 Xv: fix textured video alignment harder 10512 10513 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10514 10515commit a9c135809945fb372ce390f06918b9fd8e455b3f 10516Author: Alex Deucher <alexdeucher@gmail.com> 10517Date: Fri Feb 11 13:21:43 2011 -0500 10518 10519 Xv: fix textured video alignment 10520 10521 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10522 10523commit e5bd99faa3b6629a55168386d5dfa936ee4e97ae 10524Author: Alex Deucher <alexdeucher@gmail.com> 10525Date: Fri Feb 11 12:14:55 2011 -0500 10526 10527 kms: remove RADEON_TILING_SURFACE flag on front bo 10528 10529 Should have been removed in: 10530 ef9bfb262db7004bef3704e5d914687e50d3fca4 10531 10532 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10533 10534commit 92ffce1892d24a98d789aad57a4230cadb0c812a 10535Author: Alex Deucher <alexdeucher@gmail.com> 10536Date: Fri Feb 11 03:12:43 2011 -0500 10537 10538 EXA/r6xx+: properly account for height alignment in copy temp buffer 10539 10540 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10541 10542commit 39104c6e8461cf49c1bb03a18858ad75a9d98b46 10543Author: Alex Deucher <alexdeucher@gmail.com> 10544Date: Thu Feb 10 14:31:43 2011 -0500 10545 10546 remove EVERGREENSetAccelState() 10547 10548 It was a duplicate of the R600 variant. 10549 10550 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10551 10552commit 7ac3a2e0bcdadff7c7172a9f833f526b526da16b 10553Author: Alex Deucher <alexdeucher@gmail.com> 10554Date: Thu Feb 10 14:24:50 2011 -0500 10555 10556 6xx+: switch to linear aligned rather than linear general 10557 10558 linear aligned is supposedly more performant, but more 10559 importantly, linear general only works on the CB without 10560 slices. The texture blocks technically don't support 10561 linear general although, I think linear general gets 10562 upgraded to linear aligned in the hw which is why it 10563 currently works. 10564 10565 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10566 10567commit e3145801b80fd4be4cf770128876e86e89bda66f 10568Author: Alex Deucher <alexdeucher@gmail.com> 10569Date: Thu Feb 10 14:14:55 2011 -0500 10570 10571 evergreen/NI: consolidate spi setup 10572 10573 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10574 10575commit be67ded05621aff9c85525372fd119071d3278ec 10576Author: Alex Deucher <alexdeucher@gmail.com> 10577Date: Thu Feb 10 14:06:38 2011 -0500 10578 10579 6xx/7xx: consolidate spi setup 10580 10581 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10582 10583commit 2c5ae1724307e0dba5d0306fe27c1e15a7390a2f 10584Author: Alex Deucher <alexdeucher@gmail.com> 10585Date: Thu Feb 10 13:47:53 2011 -0500 10586 10587 evergreen/ni: consolidate CB state handling 10588 10589 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10590 10591commit c52d817b51e13447802fe338be2d247ffba2a669 10592Author: Alex Deucher <alexdeucher@gmail.com> 10593Date: Thu Feb 10 13:35:02 2011 -0500 10594 10595 6xx/7xx: consolidate remaining CB state 10596 10597 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10598 10599commit 5f9bc127d302404432b631e3e774192950f57b42 10600Author: Alex Deucher <alexdeucher@gmail.com> 10601Date: Thu Feb 10 03:28:01 2011 -0500 10602 10603 kms/r6xx+: clean up pitch/height alignment in EXA UTS/DFS 10604 10605commit a4899db96029acde6cd400fc0541693a487898e7 10606Author: Alex Deucher <alexdeucher@gmail.com> 10607Date: Sat Feb 5 14:11:42 2011 -0500 10608 10609 UMS: fix spelling in error message 10610 10611 Noticed by: Warren Block <wblock@wonkity.com> 10612 10613 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10614 10615commit 84450ea9221f32296c1f786a6aa126a08bb9756f 10616Author: Alex Deucher <alexdeucher@gmail.com> 10617Date: Sat Feb 5 14:05:48 2011 -0500 10618 10619 UMS/DCE3.2: fix segfault 10620 10621 atombios_pick_dig_encoder() needs to be called 10622 with a valid crtc assigned. 10623 10624 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10625 10626commit a27b5dbd9fba689f1814c0888d1b7459b6c6b1a1 10627Author: Alex Deucher <alexdeucher@gmail.com> 10628Date: Thu Feb 3 19:06:22 2011 -0500 10629 10630 Bump version post release 10631 10632 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10633 10634commit 0a1a0513a61f392580bde39cca4880f2c19abc8d 10635Author: Alex Deucher <alexdeucher@gmail.com> 10636Date: Thu Feb 3 18:09:17 2011 -0500 10637 10638 bump version for release 10639 10640 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10641 10642commit ef9bfb262db7004bef3704e5d914687e50d3fca4 10643Author: Alex Deucher <alexdeucher@gmail.com> 10644Date: Tue Feb 1 19:24:44 2011 -0500 10645 10646 kms/pre-6xx: fix pageflipping with tiling 10647 10648 Use UTS/DFS to tile/untile as appropriate for sw access on 10649 pre-6xx. 10650 10651 Fixes: 10652 https://bugs.freedesktop.org/show_bug.cgi?id=33738 10653 10654 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10655 10656commit bb16dd77321e5a64a3cb2d2ca9982117799ac1a8 10657Author: Alex Deucher <alexdeucher@gmail.com> 10658Date: Tue Feb 1 19:20:00 2011 -0500 10659 10660 kms/r6xx+: fix tiling and pageflipping harder 10661 10662 Thanks for Michel for final fix. Fixes: 10663 https://bugs.freedesktop.org/show_bug.cgi?id=33738 10664 for r6xx+ asics. A similar approach for pre-r6xx 10665 asics is pending. 10666 10667 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10668 10669commit 9c124f781049ef11a8b19894a29b7f62975b3011 10670Author: Alex Deucher <alexdeucher@gmail.com> 10671Date: Tue Feb 1 18:59:40 2011 -0500 10672 10673 radeon: remove device 0x4243 10674 10675 0x4243 is a PCI bridge, not a gpu. 10676 10677 See: 10678 https://bugs.freedesktop.org/show_bug.cgi?id=33815 10679 10680 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10681 10682commit 66eb81b62e5ae8e1d7bd44ed8a179e5ec1ca69af 10683Author: Michel Dänzer <daenzer@vmware.com> 10684Date: Mon Jan 31 12:26:20 2011 +0100 10685 10686 UMS: Slightly improve xserver version check. 10687 10688 This will ensure the xserver definitions necessary to compile the new code are 10689 there. xf86CrtcSetMode will be unresolved at runtime with a few xserver 10690 commits, but I'm afraid this is the best we can do. 10691 10692 See https://bugs.freedesktop.org/show_bug.cgi?id=33719 . 10693 10694commit 3dc28c86eb57cac819e4ed650acfe1f7df6ef565 10695Author: Alex Deucher <alexdeucher@gmail.com> 10696Date: Thu Jan 27 13:09:26 2011 -0500 10697 10698 ums: fix build against xserver >= 1.10 10699 10700 xf86CrtcSetMode was removed in: 10701 http://cgit.freedesktop.org/xorg/xserver/commit/?id=8b35118c03590a7ad3786d3284bafb3f40fcb8cc 10702 10703commit f58e1354b78bf6b70120bddfe1566da3b0723f72 10704Author: Jerome Glisse <jglisse@redhat.com> 10705Date: Wed Jan 26 16:13:30 2011 -0500 10706 10707 dp: fix displayport support by syncing with KMS code 10708 10709 Warning the dp clock value are divided by 10 in ddx (10 times 10710 bigger than kernel value) this is somethings very picky. 10711 10712 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 10713 10714commit 0f90133ac61c5e1d80b6fa9256e079349940029f 10715Author: Alex Deucher <alexdeucher@gmail.com> 10716Date: Mon Jan 24 12:40:22 2011 -0500 10717 10718 vbo: warn if we add a bo with a pointer to the wait list 10719 10720commit 42529603ecf86fcfd0a8f3495d9db9ce1ee9b4c3 10721Author: Alex Deucher <alexdeucher@gmail.com> 10722Date: Mon Jan 24 12:30:56 2011 -0500 10723 10724 vbo: always flush the cbuf bo in flush_indirect() 10725 10726 Always flush the cbuf in case we end up with a cbuf mapped in 10727 Prepare*(), but never end up issuing a draw call since the cbuf 10728 may be in use by multiple ops. 10729 10730 The CS check for the cbuf bo is no longer necessay because 10731 the cbuf bo is NULL at this point due to the radeon_vbo_put() 10732 just prior. 10733 10734 Fixes: 10735 https://bugs.freedesktop.org/show_bug.cgi?id=32188 10736 10737commit 9b0e03e98ab739efb4031f81cc4a1a50b3d87a42 10738Author: Alex Deucher <alexdeucher@gmail.com> 10739Date: Mon Jan 24 00:13:25 2011 -0500 10740 10741 vbo: avoid extra cs revalidation in vbo alloc 10742 10743 Make sure we have a valid bo before revalidating. 10744 10745commit fadee0409a8e13b78bbccb83dd70f590fee23d57 10746Author: Alex Deucher <alexdeucher@gmail.com> 10747Date: Fri Jan 21 17:30:28 2011 -0500 10748 10749 vbo: balance radeon_vbo_get() and radeon_vbo_put() 10750 10751 bo_ref() and bo_map() in radeon_vbo_get() 10752 bo_unmap() and bo_unref() in radeon_vbo_put() 10753 10754 rather than doing the bo_map() separately in 10755 radeon_vbo_space(). 10756 10757commit af7d81625a8cf873e6efc881489b3eda9861bd03 10758Author: Alex Deucher <alexdeucher@gmail.com> 10759Date: Fri Jan 21 17:01:25 2011 -0500 10760 10761 evergreen: use vb_offset rather than vb_start_op for cbuf offset 10762 10763commit 4817fac5f728b777939e2e2bee16b842c9e1367b 10764Author: Alex Deucher <alexdeucher@gmail.com> 10765Date: Fri Jan 21 15:59:32 2011 -0500 10766 10767 vbo: remove useless radeon_vb_discard() 10768 10769 We already reset vb_start_op to -1 in the 10770 UMS/KMS ib discard functions. 10771 10772commit 0a03f03a65aad925ba2d9c76b1d3356184607bf9 10773Author: Alex Deucher <alexdeucher@gmail.com> 10774Date: Thu Jan 20 17:07:32 2011 -0500 10775 10776 evergreen/ni: fix Xv artifacts 10777 10778 Port of the 6xx/7xx fix to evergreen. Bad texture size 10779 for texture cache flush. 10780 10781commit f41cf83750ba9a2e0797fdb21ea9104b85ce53aa 10782Author: Marton Balint <cus@fazekas.hu> 10783Date: Thu Jan 20 17:04:56 2011 -0500 10784 10785 r6xx/7xx: fix Xv artifacts 10786 10787 bad textures size for cache flushes. 10788 10789 Fixes: 10790 https://bugs.freedesktop.org/show_bug.cgi?id=22007 10791 10792commit 57fbddfc21d8c6794f378489b764cc2a0ad4a48c 10793Author: Michel Dänzer <daenzer@vmware.com> 10794Date: Tue Jan 18 16:07:52 2011 +0100 10795 10796 Fix crash freeing KMS video memory. 10797 10798 Where's that brown paper bag? :} 10799 10800commit 6548bb9836253c586023ffe5ad1497ddabaa50fc 10801Author: Michel Dänzer <daenzer@vmware.com> 10802Date: Tue Jan 18 10:23:41 2011 +0100 10803 10804 Fix KMS textured video leaks (bug #33193). 10805 10806 v2: Fix radeon_legacy_free_memory() argument type error pointed out by 10807 Marton Balint, refactor video memory freeing logic into helper function. 10808 10809 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=33193 . 10810 10811commit edc3496b55577ee8509ddd9188e6f2bcdf7169a1 10812Author: Dave Airlie <airlied@redhat.com> 10813Date: Mon Jan 17 07:55:29 2011 +1000 10814 10815 radeon: add back flush in block handler. 10816 10817 this is also needed for certain things like stipple rendering. 10818 10819 Signed-off-by: Dave Airlie <airlied@redhat.com> 10820 10821commit 63d7dece3f4be8b14012b3a2bedd850831c0437b 10822Author: Alex Deucher <alexdeucher@gmail.com> 10823Date: Thu Jan 13 11:24:26 2011 -0500 10824 10825 rs880 fix typo in HD 4250 string 10826 10827 Noticed by Nigel Taylor 10828 10829 Fixes: 10830 https://bugs.freedesktop.org/show_bug.cgi?id=33057 10831 10832commit c5b3db18d888552328e9718ea022794fc5bde352 10833Author: Alex Deucher <alexdeucher@gmail.com> 10834Date: Tue Jan 11 16:21:18 2011 -0500 10835 10836 kms: fix pitch aligment for scanout 10837 10838 Display has slightly stricter pitch alignment requirements 10839 than other blocks. Factor that in when aligning pitch. 10840 10841 Fixes: 10842 https://bugs.freedesktop.org/show_bug.cgi?id=32997 10843 10844commit bbd7adce889359b5eb3239b73e904b3ede283e12 10845Author: Alex Deucher <alexdeucher@gmail.com> 10846Date: Tue Jan 11 15:41:03 2011 -0500 10847 10848 radeon: fix yet another pitch align 10849 10850commit af2e6d7d2f1b3d8f8f6b0acfb2b7b0cfaff7bcdb 10851Author: Alex Deucher <alexdeucher@gmail.com> 10852Date: Tue Jan 11 14:42:17 2011 -0500 10853 10854 radeon: fix pitch align in pageflip code 10855 10856commit e5d0a400d08da2358fac9c2ad12042f125525736 10857Merge: 0e432dff e27e9b4e 10858Author: Alex Deucher <alexdeucher@gmail.com> 10859Date: Mon Jan 10 18:16:49 2011 -0500 10860 10861 Merge branch 'kms-pflip' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-ati 10862 10863commit 0e432dff9e06a183acaeb20db29cbd03ff0f4b82 10864Author: Alex Deucher <alexdeucher@gmail.com> 10865Date: Thu Jan 6 20:56:45 2011 -0500 10866 10867 NI: add pci ids 10868 10869commit 34d61785b9bc13287bd7ab8bdd8a1b99a3df4440 10870Author: Alex Deucher <alexdeucher@gmail.com> 10871Date: Tue Nov 23 20:54:57 2010 -0500 10872 10873 Add EXA/Xv acceleration support for NI chips 10874 10875commit 97322c00faf08461b941edf97efe86d8b082a0ce 10876Author: Alex Deucher <alexdeucher@gmail.com> 10877Date: Wed Nov 24 12:18:04 2010 -0500 10878 10879 Bail on NI cards with UMS 10880 10881 NI cards require KMS. 10882 10883commit c678b79f99238473df6ff1bedbae840950cdf88c 10884Author: Alex Deucher <alexdeucher@gmail.com> 10885Date: Tue Nov 23 20:45:25 2010 -0500 10886 10887 add NI family ids 10888 10889commit e27e9b4e50ad42885ad2e25be897cdf29aa59712 10890Author: Mario Kleiner <mario.kleiner@tuebingen.mpg.de> 10891Date: Sun Dec 19 03:56:17 2010 +0100 10892 10893 ddx/ati: Bugfix for pageflip consistency check. 10894 10895 target_msc wasn't passed from vblank event handler 10896 to pageflip routine, due to a missing initalization. 10897 Now fixed. 10898 10899 Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de> 10900 10901commit 30c5f217f91ed1ee15b5fff0002521d98682efb7 10902Author: Alex Deucher <alexdeucher@gmail.com> 10903Date: Fri Dec 17 13:40:40 2010 -0500 10904 10905 exa/dfs/ontario: always use the CPU for DFS for now 10906 10907 There are issues with rendering to gart memory that haven't 10908 been sorted out completely yet. 10909 10910commit e97079c1b621433b4aa3ecfdd2aa0884aac5df76 10911Author: Alex Deucher <alexdeucher@gmail.com> 10912Date: Thu Dec 16 18:49:55 2010 -0500 10913 10914 DCE4.1: fix DIGx routing 10915 10916 Works more like DCE4.0. With only two DIGx blocks 10917 the routing is hardcoded based on link. 10918 This fixes blank screen issues when changing crtc 10919 routing due to incorrect crtc to dig mapping. 10920 10921 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 10922 10923commit 3b1fb12dec7df4e9434b2c51568ac8d997638922 10924Author: Alex Deucher <alexdeucher@gmail.com> 10925Date: Sun Dec 12 19:16:47 2010 -0500 10926 10927 man: minor formatting fixes 10928 10929commit b1af52e522158788a9613d6f1a2309a3987ad59f 10930Author: Alex Deucher <alexdeucher@gmail.com> 10931Date: Tue Dec 7 19:17:09 2010 -0500 10932 10933 man: add rs880 info 10934 10935commit f9bbb26dd97254b66de11bb2abd821aa293ecba5 10936Author: Matt Turner <mattst88@gmail.com> 10937Date: Thu Dec 2 17:00:11 2010 -0500 10938 10939 Add missing call to vbeFree (bug 4417) 10940 10941 Also kill unneeded vbe.h include from radeon_bios.c. 10942 10943 Signed-off-by: Matt Turner <mattst88@gmail.com> 10944 10945commit 9fc16ff3f8806971cb495e8838af1e5f9149d34f 10946Author: Alex Deucher <alexdeucher@gmail.com> 10947Date: Fri Dec 3 02:04:28 2010 -0500 10948 10949 radeon: remove duplicated define 10950 10951 xfixedtofloat macro was defined several times 10952 10953commit f074762fff4253a457cb48519bb33a2e2c90f8b9 10954Author: Alex Deucher <alexdeucher@gmail.com> 10955Date: Fri Dec 3 01:25:41 2010 -0500 10956 10957 radeon/exa: fallback for non-affine transforms 10958 10959 We need to implement a texture lookup with perspective 10960 divide for non-affine transforms. For now just fallback. 10961 Should fix: 10962 https://bugs.freedesktop.org/show_bug.cgi?id=31799 10963 although it appears either EXA or the xserver gets this 10964 wrong too. 10965 10966commit 774b114baa932078f8bf91ea302850a8c54c2c07 10967Author: Nicolas Joly <njoly@pasteur.fr> 10968Date: Thu Dec 2 16:02:44 2010 -0500 10969 10970 ati: remove leftover word in ati.man 10971 10972 fixes: 10973 https://bugs.freedesktop.org/show_bug.cgi?id=28297 10974 10975commit 69639ef377a9d6701cdef902f8a1c5e0b58cf833 10976Author: Alex Deucher <alexdeucher@gmail.com> 10977Date: Thu Dec 2 02:12:24 2010 -0500 10978 10979 radeon/kms: pageflipping man page updates 10980 10981commit f48af8a6cfa1ac665f07b8f9712e94b77bc4f5e9 10982Author: Mario Kleiner <mario.kleiner@tuebingen.mpg.de> 10983Date: Tue Nov 23 04:41:53 2010 +0100 10984 10985 ddx/ati: Fix reporting of pageflip completion events on multi-head. 10986 10987 When a drawable is page-flipped on multiple crtc's (fullscreen 10988 drawable on mirror-mode or multi-head x-screen), only one pageflip 10989 event is finally delivered, after the last participating crtc signals 10990 flip completion, this to avoid visual corruption. 10991 10992 Old code returned vblank count and timestamps of flip completion 10993 of this last crtc, instead of the values of the "master crtc", the 10994 one that was used for initially scheduling/triggering the pagflip 10995 via vblank events. (master = radeon_dri2_drawable_crtc(drawable)) 10996 10997 This patch makes sure that the pageflip completion values of the 10998 "master" crtc are returned, otherwise client applications will 10999 get confused by the random (msc, ust) values returned by whichever 11000 crtc was the last to complete its flip. Without this, the returned 11001 values change randomly and jump forward and backward in time and 11002 count. 11003 11004 The patch also implements a consistency check on returned vblank 11005 count values of pageflip completion. Impossible values are detected, 11006 a x-warning is logged and returned (msc,ust) values are marked invalid, 11007 so clients could perform error handling. Such a warning would indicate 11008 bugs in the pageflip completion routine of future kms drivers or the 11009 ddx to aid driver debugging. 11010 11011 Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de> 11012 11013commit 122536ee0aeb1eef1a9d80d5e464dcb423dc2837 11014Author: Mario Kleiner <mario.kleiner@tuebingen.mpg.de> 11015Date: Mon Nov 22 04:11:07 2010 +0100 11016 11017 ddx/ati: Add option "SwapbuffersWait" to control vsync of DRI2 swaps. 11018 11019 A new optional kms driver option "SwapbuffersWait" is defined 11020 for xorg.conf, which defaults to "on". If "on", DRI2 bufferswaps 11021 will be synchronized to vsync, otherwise not. 11022 11023 This currently only affects copy-swaps, not pageflipped swaps. 11024 It also requires a swap_interval setting of zero by the OpenGL 11025 client. 11026 11027 Ideally, we'd provide a way for dri2 to pass the current swap 11028 interval to the ddx so we could change this dynamically. 11029 11030 Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de> 11031 11032commit 0de680730294bd623f6b3e189faa7b88a09d3a2a 11033Author: Alex Deucher <alexdeucher@gmail.com> 11034Date: Wed Nov 17 17:39:39 2010 -0500 11035 11036 radeon/kms: add pageflip support 11037 11038 requires radeon drm 2.8.0 or higher 11039 11040 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11041 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 11042 11043commit fccdca8db34010f566bd068c74cdef0f4a8cb7f5 11044Author: Alex Deucher <alexdeucher@gmail.com> 11045Date: Wed Nov 17 17:37:25 2010 -0500 11046 11047 radeon/kms: allow tiled front buffer on 6xx/7xx 11048 11049 Use UTS/DFS to tile/untile as appropriate for sw access. 11050 Also enables pageflipping with tiling enabled. 11051 11052commit 035f7f3ab529ca19b853066792af8a23d08a0f53 11053Author: Dave Airlie <airlied@redhat.com> 11054Date: Wed Dec 1 10:23:14 2010 +1000 11055 11056 radeon: add some missing math.h and assert.h includes 11057 11058commit 0f924e83d7be6fca7e024acdf973089efdf3b627 11059Author: Dave Airlie <airlied@redhat.com> 11060Date: Wed Dec 1 10:23:33 2010 +1000 11061 11062 radeon: make sure EXA is off when EXA is disabled with --disable-exa 11063 11064commit e142e55c56d8440d7185b594c696ec9eeb699cda 11065Author: Alex Deucher <alexdeucher@gmail.com> 11066Date: Tue Nov 30 15:51:21 2010 -0500 11067 11068 evergreen: set default group_bytes to 256 11069 11070 512 seems to cause issues on certain cards. 11071 11072 Fixes: 11073 https://bugs.freedesktop.org/show_bug.cgi?id=31724 11074 11075commit 90f831361844f1b80b3f6bb718ff5ac584d73d48 11076Author: Alex Deucher <alexdeucher@gmail.com> 11077Date: Mon Nov 29 18:09:05 2010 -0500 11078 11079 evergreen: use dot4 for transforms 11080 11081commit 3cae361d0448b6e231c80f53d64bdbbdd74dc4cf 11082Author: Alex Deucher <alexdeucher@gmail.com> 11083Date: Mon Nov 29 17:44:47 2010 -0500 11084 11085 6xx/7xx: clean up gpr/const handling in shaders 11086 11087commit d9bcac516f2a810acb300b29169e56a2df0b47ac 11088Author: Alex Deucher <alexdeucher@gmail.com> 11089Date: Mon Nov 29 17:23:30 2010 -0500 11090 11091 r6xx/r7xx use dot4 for transforms 11092 11093commit 5d3f33729be0639cef17372345b2dab6127e39d9 11094Author: Alex Deucher <alexdeucher@gmail.com> 11095Date: Wed Nov 24 10:59:09 2010 -0500 11096 11097 radeon/kms: fix output enumeration 11098 11099 If a system has a mix of DVI-I, and DVI-D connectors, or a mix 11100 of HDMI-A and HDMI-B connectors, the enumeration was per 11101 connector type, but the naming was the same for multiple 11102 connectors (DVI-[0-n] for DVI-I, -D, -A, and HDMI-[0-n] 11103 for HDMI-A, and -B), you'd end up with multiple connectors 11104 with the same name. Fix this by tracking the number of 11105 connectors across similar connector types. 11106 11107commit dc470b5382ffa27bd4a673b21fb6b978d3716ca0 11108Author: Alex Deucher <alexdeucher@gmail.com> 11109Date: Tue Nov 23 01:46:15 2010 -0500 11110 11111 radeon/man: divide driver options into UMS and KMS 11112 11113commit 4cbb7100df4b983d77c20a77431cf2f5101bde36 11114Author: Alex Deucher <alexdeucher@gmail.com> 11115Date: Tue Nov 23 01:22:04 2010 -0500 11116 11117 update man page to include ontario 11118 11119commit 718dfd443897d70d97fc19afddfbcfe1e39a9c59 11120Author: Alex Deucher <alexdeucher@gmail.com> 11121Date: Mon Nov 22 13:11:50 2010 -0500 11122 11123 Ontario: add workaround for small pixmap issues with DFS 11124 11125 On past asics these kind a problems were usually due to a bad 11126 backend map config. I suspect in this case, and the previous 11127 cases it may be related to using linear buffers rather than 11128 tiled. 11129 11130 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11131 11132commit 61c97b0812c820564c7bc9d64a998c109bcd46b3 11133Author: Alex Deucher <alexdeucher@gmail.com> 11134Date: Mon Nov 22 13:10:16 2010 -0500 11135 11136 Add EXA/Xv acceleration support for Ontario Fusion APUs 11137 11138 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11139 11140commit 0659f35dfbeda7dd3ff58da1deb6484fabadc332 11141Author: Alex Deucher <alexdeucher@gmail.com> 11142Date: Mon Nov 22 13:07:00 2010 -0500 11143 11144 Add Ontario fusion APU pci ids 11145 11146 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11147 11148commit fa0cec331e6c94cebf82270dcd6c179b15851c5e 11149Author: Alex Deucher <alexdeucher@gmail.com> 11150Date: Mon Nov 22 12:56:45 2010 -0500 11151 11152 ontario: add UMS modesetting support 11153 11154 Accel not enabled yet. 11155 11156 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11157 11158commit 7ad1f01573e930aa6238eaeb0e11708ae548dfce 11159Author: Jerome Glisse <jglisse@redhat.com> 11160Date: Fri Nov 19 10:27:40 2010 -0500 11161 11162 evergreen: fix typo for fetch resource offset 11163 11164 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 11165 11166commit 3455a3b58532ea3ad901a317126968ab6cbb21b7 11167Author: Alex Deucher <alexdeucher@gmail.com> 11168Date: Wed Nov 17 17:32:41 2010 -0500 11169 11170 radeon/kms: fix buffer base alignment for tiling 11171 11172 On r6xx+, 2D tiling can require larger than 4k base alignment. 11173 11174commit beb7fecd0191e38fb238134ba612985062cf9770 11175Author: Dave Airlie <airlied@redhat.com> 11176Date: Mon Nov 15 11:59:58 2010 +1000 11177 11178 dce4: improve the VT switching code. 11179 11180 add in lots more blocks of regs to save/restore 11181 11182commit 51e51f864a691e51b5193d102398ced667e0beb1 11183Author: Dave Airlie <airlied@redhat.com> 11184Date: Thu Nov 11 18:48:40 2010 +1000 11185 11186 evergreen: remove warnings since last commit 11187 11188commit abbf73ee990512ac16ca77e8bb23288495e1f9f4 11189Author: Dave Airlie <airlied@redhat.com> 11190Date: Thu Nov 11 09:38:15 2010 +1000 11191 11192 evergreen: add UMS VT switch support. 11193 11194 This isn't perfect, but it brings back text VTs here on the 11195 DAC and DVI outputs. 11196 11197commit e1dfaf93d06bc5eafdbc2e1823d19204ce8f242c 11198Author: Dave Airlie <airlied@redhat.com> 11199Date: Thu Nov 11 09:15:27 2010 +1000 11200 11201 avivo: use arrays to store the crtc/pll info for save/restore 11202 11203 this is just prep work for evergreen VT save/restore 11204 11205commit 0c2834e67df6e143a7d15f373faaddca1dda6b18 11206Author: Michel Dänzer <daenzer@vmware.com> 11207Date: Fri Nov 5 09:30:40 2010 +0100 11208 11209 Fix incorrect CS size in one textured video path. (Bug #31364) 11210 11211 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=31364 . 11212 11213commit 2a7b080727f61559cf8444c93a5088f4155a8fdf 11214Author: Alex Deucher <alexdeucher@gmail.com> 11215Date: Fri Oct 29 16:56:45 2010 -0400 11216 11217 radeon/kms: more alignment fixes 11218 11219commit bf60af579382a0d48f7a65a4cec88759cc8b683c 11220Author: Dave Airlie <airlied@redhat.com> 11221Date: Thu Oct 28 13:06:54 2010 +1000 11222 11223 dri2: reference count the client privates key/callback 11224 11225 This lets multi-screen work better, but still having issues after server 11226 recycle, but it doesn't crash at least. 11227 11228commit fb22d0c06a7dc42216230e198ff443d8035e9d21 11229Author: Dave Airlie <airlied@redhat.com> 11230Date: Thu Oct 28 11:10:03 2010 +1000 11231 11232 radeon: only register block handler once per fd. 11233 11234 should fix https://bugs.freedesktop.org/show_bug.cgi?id=29726 11235 11236 the problem is of course the second head instance tries to access the 11237 fd and fails, however I think this might break syncing on the second 11238 head but not sure, but its better than just hanging up the X server 11239 11240commit 040140b5d7cbeaf5d83687bc87e4a236ff9f0cbd 11241Author: Alex Deucher <alexdeucher@gmail.com> 11242Date: Wed Oct 27 20:16:17 2010 -0400 11243 11244 drm/radeon: unify fb bo alignment harder 11245 11246 More duplicated paths discoved... 11247 11248commit f07f9b7b61c05f2de1d61bb0e2f71bd017c8d36a 11249Author: Alex Deucher <alexdeucher@gmail.com> 11250Date: Wed Oct 27 12:37:42 2010 -0400 11251 11252 kms/radeon: unify fb bo alignment handling 11253 11254 Previously there were 3 different paths with what should 11255 have had duplicated code: 11256 - EXACreatePixmap2 11257 - Initial front buffer creation 11258 - Randr resize 11259 11260 This patch attempts to unify the alignment across all 3. 11261 11262 This may fix tiling issues in some cases and should make 11263 buffer pitches match for pageflipping. 11264 11265commit d31046ba6c8eee9b7decc3875697d37c38bc38f3 11266Author: Michel Dänzer <daenzer@vmware.com> 11267Date: Wed Oct 27 12:12:04 2010 +0200 11268 11269 Re-use result of GetScreenPixmap call. 11270 11271 Fixes compile warning due to local variable ppix being unused when building 11272 against current xserver Git. 11273 11274commit 4dd7f835c17b00707f1d8d4e36a24380ba52761e 11275Author: Michel Dänzer <daenzer@vmware.com> 11276Date: Wed Oct 27 12:09:03 2010 +0200 11277 11278 Fix another stray xfree() call. 11279 11280commit f815e9c3c61b1ed73e7dd9383587efcdd8dfc07f 11281Author: Michel Dänzer <daenzer@vmware.com> 11282Date: Wed Oct 27 11:29:12 2010 +0200 11283 11284 Don't try to delete DRI2 event list entries that were never added. (Bug #31086) 11285 11286 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=31086 . 11287 11288commit df92245e1fd0ead2581b675cd96ac8bcc8cb65d7 11289Author: Alex Deucher <alexdeucher@gmail.com> 11290Date: Sun Oct 24 14:27:27 2010 -0400 11291 11292 r100: add some missing bits in SE_COORD_FMT 11293 11294 Texture coordinates work fine with or without these, 11295 but this should be more correct I think although 11296 I don't think it matters since we aren't sending w 11297 anyway. 11298 11299commit fc7aa6e3bbb86eaa2f34b827629f7e82cde65ac7 11300Author: Alex Deucher <alexdeucher@gmail.com> 11301Date: Sun Oct 24 11:56:48 2010 -0400 11302 11303 UMS/radeon: fix typo in restore palette 11304 11305commit 5efdf514ab0e5439114e8dd7a49105812155eb69 11306Author: Alex Deucher <alexdeucher@gmail.com> 11307Date: Sun Oct 24 11:45:19 2010 -0400 11308 11309 UMS/radeon: fix screen dimming on VT switch 11310 11311 Save and restore the palettes on VT switch. The restore 11312 has to be done after the vga restore to work properly as 11313 determined by Jonathan Kollasch. 11314 11315 Fixes: 11316 https://bugs.freedesktop.org/show_bug.cgi?id=18407 11317 11318commit 1e3ede62c39307d0210cb2044badd619a4f44fec 11319Author: Alex Deucher <alexdeucher@gmail.com> 11320Date: Sun Oct 24 11:24:07 2010 -0400 11321 11322 UMS/atom: default to DVI mode if we don't have monitor hdmi info 11323 11324 Fixes: 11325 https://bugs.freedesktop.org/show_bug.cgi?id=30330 11326 11327commit 51f26a7f01b68e373867c03881868ca8830664d6 11328Author: Mark Schreiber <mark7@alumni.cmu.edu> 11329Date: Sat Apr 10 10:55:08 2010 -0700 11330 11331 Correct spelling and grammar 11332 11333commit 0f575984594e6977b4a7cb794dad8c6e56c850f1 11334Author: Alex Deucher <alexdeucher@gmail.com> 11335Date: Tue Oct 19 19:21:39 2010 -0400 11336 11337 Xv: limit overlay to 2047x2047 11338 11339 Fixes: 11340 https://bugs.freedesktop.org/show_bug.cgi?id=15391 11341 11342commit 0f184630cd4760199430320215e4280438fc4ce5 11343Author: Alex Deucher <alexdeucher@gmail.com> 11344Date: Tue Oct 19 17:13:06 2010 -0400 11345 11346 radeon: fix compilation on xserver >= 1.10 11347 11348 Fixes: 11349 https://bugs.freedesktop.org/show_bug.cgi?id=30451 11350 11351 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11352 11353commit 179b0e15924ae2757efaece6488ae835ba6436e3 11354Author: Dave Airlie <airlied@redhat.com> 11355Date: Tue Oct 19 13:37:22 2010 +1000 11356 11357 r600: for 2d tiles we need to align to channels no banks. 11358 11359commit 809cacbc23ab37aa7952665f9b2e37c03a34f173 11360Author: Adam Jackson <ajax@redhat.com> 11361Date: Mon Oct 18 12:32:11 2010 -0400 11362 11363 kms: Let the server know if we got more than 1 EDID block 11364 11365 Otherwise things like xf86MonitorIsHDMI() won't work right. 11366 11367 Signed-off-by: Adam Jackson <ajax@redhat.com> 11368 11369commit 042037e485396981cf4a420c247104ba5d016ca0 11370Author: Alex Deucher <alexdeucher@gmail.com> 11371Date: Thu Oct 14 17:00:51 2010 -0400 11372 11373 evergreen: requires same cursor fix as avivo chips 11374 11375commit 0f37458c253aec46d98db6661c4def711df8fc17 11376Author: Adam Jackson <ajax@redhat.com> 11377Date: Thu Oct 14 14:37:34 2010 -0400 11378 11379 Don't allow the config file to override BIOS location 11380 11381 Signed-off-by: Adam Jackson <ajax@redhat.com> 11382 11383commit 2918df3a312fc600e83f84c298b9653d2eb3a8d5 11384Author: Alex Deucher <alexdeucher@gmail.com> 11385Date: Wed Oct 13 17:16:04 2010 -0400 11386 11387 evergreen: increase XV_CRTC Xv attribute range 11388 11389 Cover all 6 possible crtcs. 11390 11391commit 0d1f9fd03d0196dda5c1ce34b2e68b007c1874f2 11392Author: Alex Deucher <alexdeucher@gmail.com> 11393Date: Sun Oct 10 12:06:22 2010 -0400 11394 11395 DCE3+: UMS modesetting fixes 11396 11397 - add mising LVTMA case statement for DCE3.0 dig encoder 11398 - some DCE4 systems have EN/DISABLE_OUTPUT actions 11399 11400commit acf4024aa0ef192355b2bd8281507b1c2e6b7de4 11401Author: Alex Deucher <alexdeucher@gmail.com> 11402Date: Fri Oct 8 16:30:36 2010 -0400 11403 11404 ATOM DDC fixes for UMS 11405 11406 - Header size was already subtraced from table size 11407 - Only hw capable ddc pads are shared with aux 11408 11409commit 74fd2b91477106a26a2d9fb4b11c885910996041 11410Author: Alex Deucher <alexdeucher@gmail.com> 11411Date: Thu Oct 7 17:33:00 2010 -0400 11412 11413 DCE3.2 UMS: fix duallink 11414 11415 Should fix: 11416 https://bugs.freedesktop.org/show_bug.cgi?id=30686 11417 11418commit 8c544272f218fd8bd93a898b96afedc2498b27bb 11419Author: Alex Deucher <alexdeucher@gmail.com> 11420Date: Thu Oct 7 17:25:47 2010 -0400 11421 11422 fix the non-kms build 11423 11424 Should fix: 11425 https://bugs.freedesktop.org/show_bug.cgi?id=30685 11426 11427commit 81360adffb2a66b9a95a38671f9227a9718c9841 11428Author: Oldřich Jedlička <oldium.pro@seznam.cz> 11429Date: Mon Sep 6 21:14:52 2010 +0200 11430 11431 radeon: proper DRI2 pending events handling when client gone. (v6) 11432 11433 Properly handle asynchronous DRI2 events for disconnected clients. 11434 Track client's pending requests and mark them as invalid when the 11435 client disconnects. 11436 11437 This is based on the version from Alban Browaeys in bug #29065. 11438 11439 v1 (Alban Browaeys): Based upon a detailed explanation from Oldřich 11440 Jedlička and comments from Christopher James Halse Rogers. 11441 on http://lists.x.org/archives/xorg-driver-ati/2010-August/016780.html . 11442 11443 v2: Updated version to apply on master. Removed unnecessary 11444 client_index field from _DRI2FrameEvent. Added freeing/removing from 11445 list to failed paths of radeon_dri2_schedule_wait_msc and 11446 radeon_dri2_schedule_swap. 11447 11448 v3: Adopt to older xorg-server that doesn't have dixRegisterPrivateKey. 11449 11450 v4: Conditional include of list.h, unreachable return removed. 11451 11452 v5: Distribute list.h as xorg_list.h, remove xorg-server version check. 11453 Use the version from xorg-server when available (checked in 11454 configure.ac). 11455 11456 v6: Removed xorg_list.h, made DRI2 scheduling features dependent on 11457 list.h presence. 11458 11459commit adee138f007e26307f1aab7f8fe066150c12e55d 11460Author: Jesse Adkins <jesserayadkins@gmail.com> 11461Date: Tue Sep 28 13:29:50 2010 -0700 11462 11463 Purge cvs tags. 11464 11465 Signed-off-by: Jesse Adkins <jesserayadkins@gmail.com> 11466 Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com> 11467 11468commit f1bf9cbb471d60ee4b0422bc4e896e3b86f4d75f 11469Author: Alex Deucher <alexdeucher@gmail.com> 11470Date: Wed Oct 6 11:44:42 2010 -0400 11471 11472 radeon: man page updates 11473 11474 - tv-out is not supported on carding using 11475 the rage theatre chip for tv-out 11476 - add evergreen cards to the list 11477 11478commit 0921ecc1c751df0dd56e0b1d0b78ab53d7164904 11479Author: Alex Deucher <alexdeucher@gmail.com> 11480Date: Mon Oct 4 12:34:47 2010 -0400 11481 11482 evergreen: use EXACreatePixmap2 if available 11483 11484commit 5bdb6434975584eef90eb9e5955b9c2a14b7f327 11485Author: Alex Deucher <alexdeucher@gmail.com> 11486Date: Mon Oct 4 12:31:51 2010 -0400 11487 11488 evergreen: port Karl's UTS/DFS changes to evergreen 11489 11490commit be8f45cbd313b68ad663f303c64edb4525b8f981 11491Merge: e843faf3 ff5f466e 11492Author: Alex Deucher <alexdeucher@gmail.com> 11493Date: Mon Oct 4 12:36:56 2010 -0400 11494 11495 Merge branch 'evergreen_accel' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-ati 11496 11497commit ff5f466e6ca0ee541a255facff6c9198976dfd52 11498Author: Alex Deucher <alexdeucher@gmail.com> 11499Date: Fri Oct 1 15:46:43 2010 -0400 11500 11501 evergreen/dri2: fix depth allocation for depth+stencil 11502 11503 evergreen uses separate allocations for depth and stencil, 11504 so to handle that, create a depth buffer large enough to 11505 handle both. This is required for using the stencil 11506 buffer in mesa. 11507 11508commit e843faf355c864beab81e74f0e39f8ad53d4c2bf 11509Author: Alex Deucher <alexdeucher@gmail.com> 11510Date: Thu Sep 30 19:30:35 2010 -0400 11511 11512 radeon: fix fbLocation for >32 bit MC addresses 11513 11514 If the fbLocation was at an address >32 bits, we'd fail. 11515 Change fbLocation to uint64_t and properly cast when needed. 11516 11517commit 886febc882053e09294225e85b102f965041b62b 11518Author: Alex Deucher <alexdeucher@gmail.com> 11519Date: Thu Sep 30 19:20:17 2010 -0400 11520 11521 r6xx: fix bad mask when setting up HDP_NONSURFACE_BASE 11522 11523 This fails for MC addresses >32 bits 11524 11525commit 16589b862feb796e15dc6e471d25604a4f5e7c1c 11526Author: Alex Deucher <alexdeucher@gmail.com> 11527Date: Mon Sep 27 18:28:43 2010 -0400 11528 11529 bump version post release 11530 11531commit cc5005af61f45a3552f7358dc5aa711e42f5af54 11532Author: Alex Deucher <alexdeucher@gmail.com> 11533Date: Mon Sep 27 18:20:53 2010 -0400 11534 11535 bump version for release 11536 11537commit 7f8820fcec8c90bf2f823170bd08a23e8b4ff7af 11538Author: Michael Cree <mcree@orcon.net.nz> 11539Date: Mon Sep 27 13:39:12 2010 -0400 11540 11541 Fix some unaligned 32bit accesses in the AtomBios code. 11542 11543 On the Alpha architecture unaligned 32bit accesses incur a software 11544 trap to the kernel and pollute the kernel logs. Fixed by use of the 11545 ldl_u() interface. 11546 11547 Signed-off-by: Michael Cree <mcree@orcon.net.nz> 11548 11549commit c4f834cdfbe96aa47ac5fb039f9dd7aa9730c8a3 11550Author: Nicolas Reinecke <nr@das-labor.org> 11551Date: Mon Sep 27 13:33:55 2010 -0400 11552 11553 radeon: Convert remaining x(c)alloc/xfree to m/calloc/free. 11554 11555 Fixes deprecation warnings missed out by 11556 f7a91ece264af9f3fd2fc18e99aefcda93ce9f5c 11557 11558commit f8fb9312d791af1f77020e8c2d35bb30841ed9aa 11559Author: Karl Tomlinson <karlt+@karlt.net> 11560Date: Sun Aug 22 22:46:33 2010 +1200 11561 11562 RADEONPrepareAccess_CS: fallback to DFS when pixmap is in VRAM 11563 11564 This avoids costly CPU VRAM reads and lets EXA manage a system memory cache 11565 of the portions of pixmaps needed for unaccelerated operations. 11566 11567 https://bugs.freedesktop.org/show_bug.cgi?id=27139 11568 11569commit 35c4ff936601ee083f51510a5192fb97d622a483 11570Author: Karl Tomlinson <karlt+@karlt.net> 11571Date: Sun Aug 22 22:28:06 2010 +1200 11572 11573 radeon: complete UTS and DFS even when a scratch BO is not necessary 11574 11575 Turns on the big-endian paths even for little-endian systems, and adds 11576 similar paths to the r6xx/r7xx functions. 11577 11578 This makes UTS and DFS reliable, which will let PrepareAccess (with 11579 mixed pixmaps) choose to fail based on whether the pixmap is in VRAM 11580 (to avoid CPU reads). 11581 11582commit d46381a3a6bf10903803f5acaa7aa0ce06373b96 11583Author: Karl Tomlinson <karlt+@karlt.net> 11584Date: Sun Aug 22 21:02:45 2010 +1200 11585 11586 radeon: complete big endian UTS and DFS even when scratch allocation fails. 11587 11588 On big endian systems, PrepareAccess will fail when byte-swapping is 11589 required so UploadToScreen and DownloadFromScreen cannot rely on 11590 fallback to PrepareAccess. 11591 11592 When scratch BO space allocation fails, this patch merely adds simple 11593 fallback to direct CPU access without any GPU blit. This sometimes 11594 requires a CS flush even in UploadToScreen. 11595 (No allocation retry after a flush is added here.) 11596 11597commit 4ced4e1eff67946e306c0c67c9ed59dd5f3c4ba9 11598Author: Karl Tomlinson <karlt+@karlt.net> 11599Date: Sun Aug 22 20:04:42 2010 +1200 11600 11601 RADEONDownloadFromScreenCS: flush CS writes before mapping BO for read 11602 11603 If unflushed CS operations write to the pixmap BO, then these need to be 11604 flushed before mapping the BO for read. This currently only affects big 11605 endian systems and only when the operation writes to the GTT domain. 11606 11607commit a4eef8faffbb1ea2f742273ee855f4e6f992e5c8 11608Author: Karl Tomlinson <karlt+@karlt.net> 11609Date: Sat Aug 21 22:29:34 2010 +1200 11610 11611 FinishAccess_CS: set bo_mapped to FALSE on unmap 11612 11613 This is actually only necessary when PrepareAccess may behave differently on 11614 different calls with the same pixmap, which currently doesn't happen. 11615 11616 However resetting bo_mapped is necessary to let PrepareAccess (with mixed 11617 pixmaps) choose to fail based on whether the pixmap is in VRAM (to avoid CPU 11618 reads). 11619 11620commit bfebe039af0c0282d04eb6234b6e6d1e02097146 11621Author: Karl Tomlinson <karlt+@karlt.net> 11622Date: Sat Aug 21 21:44:39 2010 +1200 11623 11624 DownloadFromScreenCS: download via a scratch BO if pixmap domain is unknown 11625 11626 radeon_bo_is_busy() may return without setting the domain out-parameter. 11627 If this happens, then download via a scratch GTT BO to avoid CPU VRAM read. 11628 11629commit e26a59e9db8067882327f872e3d2d760ce4c66f3 11630Author: Mathieu Bérard <mathieu@mberard.eu> 11631Date: Thu Sep 9 19:32:38 2010 -0400 11632 11633 Fix NULL possible deref in evergreen_cp_wait_vline_sync 11634 11635commit b90cb61ccb0f4f80e0627141f223354a9371d47d 11636Author: Alex Deucher <alexdeucher@gmail.com> 11637Date: Tue Sep 7 11:51:16 2010 -0400 11638 11639 radeon: set interlaced and doublescan enabled for randr outputs 11640 11641 interlaced used to work without setting these parameters. Changes 11642 in the xserver seem to require them now. 11643 11644 Should fix: 11645 https://bugs.freedesktop.org/show_bug.cgi?id=29591 11646 11647commit 7cc0d3311f23ad569af004e0e4a0e2efbac107e5 11648Author: Alex Deucher <alexdeucher@gmail.com> 11649Date: Fri Sep 3 00:32:27 2010 -0400 11650 11651 evergreen: fix Xv 11652 11653 VS const buffer offset was wrong. 11654 11655 fixes: 11656 https://bugs.freedesktop.org/show_bug.cgi?id=29788 11657 11658commit 7c6ce4e62693f446d7d3c8a86502ccc03c0e55b0 11659Author: Alex Deucher <alexdeucher@gmail.com> 11660Date: Thu Sep 2 20:34:56 2010 -0400 11661 11662 evergreen: add additional default state 11663 11664 Note, you also need a drm patch to fix the GPU hangs: 11665 drm/radeon/kms/evergreen: fix gpu hangs in userspace accel code 11666 11667 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11668 11669commit 2b98ec1f7e931019a4ab699a56d5dfaa395946fb 11670Author: Alex Deucher <alexdeucher@gmail.com> 11671Date: Wed Sep 1 13:24:19 2010 -0400 11672 11673 r6xx/r7xx: fix typos 11674 11675 some stray - signs 11676 11677commit 966ac1be81da76c8aa4ea46b63f3ca5358a2c021 11678Author: Alex Deucher <alexdeucher@gmail.com> 11679Date: Fri Aug 27 18:22:21 2010 -0400 11680 11681 evergreen: work around bad data in some i2c tables 11682 11683 The 7th entry in a lot of evergreen i2c gpio tables is partially 11684 zeroed. Fix the entry. 11685 11686 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11687 11688commit 91f707d308d4bbf16c3d62d046cf280fef5a8f4c 11689Author: Heikki Lindholm <holin@iki.fi> 11690Date: Fri Aug 27 02:26:24 2010 -0400 11691 11692 xv: fix non-kms/non-dri Xv column ordering on big endian systems 11693 11694 Column order is wrong on big endian systems, primarly because of a 11695 bits / bytes mix up with the bpp variable. Fix tested with r100 and 11696 r300, screen depth 16 and 32 with YV12 and YUY2 (overlay, textured video), 11697 RGBA and RGBT (overlay). 11698 11699 Should fix: https://bugs.freedesktop.org/show_bug.cgi?id=29041 11700 11701 Signed-off-by: Heikki Lindholm <holin@iki.fi> 11702 11703commit 6a2c8587a4e05a8be2a2e975a6660942cfe115d6 11704Author: Christopher James Halse Rogers <christopher.halse.rogers@canonical.com> 11705Date: Fri Aug 27 13:14:33 2010 +1000 11706 11707 dri2: Reference count DRI2 buffers 11708 11709 When a client calls ScheduleSwap we set up a kernel callback when the 11710 relevent vblank event occurs. However, it's possible for the client 11711 to go away between calling ScheduleSwap and the vblank event, 11712 resulting in the buffers being destroyed before they're passed to 11713 radeon_dri2_frame_event_handler. 11714 11715 Add reference-counting to the buffers and take a reference in 11716 radeon_dri2_schedule_swap to ensure the buffers won't be destroyed 11717 before the vblank event is dealt with. 11718 11719 This parallels the approach taken by the Intel DDX in commit 11720 0d2392d44aae95d6b571d98f7ec323cf672a687f. 11721 11722 Fixes: http://bugs.freedesktop.org/show_bug.cgi?id=29065 11723 11724 v2: Don't write completion events to the client if it has quit. 11725 v3: Don't try to unref the NULL buffers from a DRI2_WAITMSC event. 11726 Take a ref in schedule_swap earlier, so the offscreen fallback 11727 doesn't incorrectly destroy the buffers. 11728 11729 Signed-off-by: Christopher James Halse Rogers <christopher.halse.rogers@canonical.com> 11730 Signed-off-by: Dave Airlie <airlied@redhat.com> 11731 11732commit e9928fe036e9382fd7bc353f3f05531445f08977 11733Author: Dave Airlie <airlied@redhat.com> 11734Date: Wed Aug 25 10:42:39 2010 +1000 11735 11736 remove explicit buffer submit from copy region 11737 11738 port of 0be3e95c844247746742805830860ace9f546d99 from intel driver. 11739 11740 Remove explicit batchbuffer submit in DRI2 copyregion 11741 11742 Now that we submit from the flush callback chain, we know we'll always 11743 submit before the client receives the reply or event that blocks it from 11744 rendering the next frame. 11745 11746 Signed-off-by: Dave Airlie <airlied@redhat.com> 11747 11748commit 9f13049ddf06f6f2138851a548cfb82f12a52f42 11749Author: Dave Airlie <airlied@redhat.com> 11750Date: Wed Aug 25 08:56:37 2010 +1000 11751 11752 radeon: add correct flushing for direct rendered 11753 11754 this is a port of 69d65f9184006eac790efcff78a0e425160e95aa from the Intel 11755 driver. 11756 11757 Submit batch buffers from flush callback chain 11758 11759 There are a few cases where the server will flush client output buffers 11760 but our block handler only catches the most common (before going into select 11761 If the server flushes client buffers before we submit our batch buffer, 11762 the client may receive a damage event for rendering that hasn't happened yet 11763 11764 Instead, we can hook into the flush callback chain, which the server will 11765 invoke just before flushing output. This lets us submit batch buffers 11766 before sending out events, preserving ordering. 11767 11768 Fixes 28438: [bisected] incorrect character in gnome-terminal under compiz 11769 https://bugs.freedesktop.org/show_bug.cgi?id=28438 11770 11771 Signed-off-by: Kristian Høgsberg <krh@bitplanet.net> 11772 11773 Signed-off-by: Dave Airlie <airlied@redhat.com> 11774 11775commit 141cbc283fddeb67c2a6c47a0f0f5c2aa2bfb430 11776Author: Alex Deucher <alexdeucher@gmail.com> 11777Date: Tue Aug 24 18:01:05 2010 -0400 11778 11779 evergreen Xv: remove stray line that caused a segfault 11780 11781 fixes: 11782 https://bugs.freedesktop.org/show_bug.cgi?id=29788 11783 11784commit f170dddd3538a587f2363ef5fa10a4b484e762da 11785Author: Alex Deucher <alexdeucher@gmail.com> 11786Date: Mon Aug 23 18:58:12 2010 -0400 11787 11788 evergreen: use vbo pool for constant buffers 11789 11790commit eede93b057bbdddcde5da3220a3f8c6f73784a7e 11791Author: Alex Deucher <alexdeucher@gmail.com> 11792Date: Mon Aug 23 13:16:27 2010 -0400 11793 11794 radeon: move vbo data to a separate struct 11795 11796 this way we can share the vbo code with const buffers 11797 11798commit d8abf27dbd14f5eb746c5e8b8b1436ad292d8ec6 11799Author: Alex Deucher <alexdeucher@gmail.com> 11800Date: Mon Aug 23 01:31:27 2010 -0400 11801 11802 evergreen: fix dword counts in default state 11803 11804commit 5a7c9d94733a0db1d3565447acc9f0e751db5950 11805Author: Alex Deucher <alexdeucher@gmail.com> 11806Date: Mon Aug 23 00:43:48 2010 -0400 11807 11808 radeon: fix legacy lvds dpms sequence 11809 11810 Take from my kms commit. 11811 11812 Should fix: 11813 https://bugs.freedesktop.org/show_bug.cgi?id=19459 11814 11815 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11816 11817commit 6930d2c981221757b1e11ef194809f085753a611 11818Author: Alex Deucher <alexdeucher@gmail.com> 11819Date: Fri Aug 20 16:55:21 2010 -0400 11820 11821 Add initial EXA and Xv support for evergreen 11822 11823 Based on the r6xx/r7xx code updated for evergreen. 11824 Still causes GPU hangs in some cases. We haven't 11825 tracked down why yet. Might be related to constant 11826 buffer persistence. 11827 11828 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11829 11830commit bdd41fecdb19c83c6c7b793016b61d38065dcd13 11831Author: Alex Deucher <alexdeucher@gmail.com> 11832Date: Fri Aug 20 01:17:13 2010 -0400 11833 11834 evergreen: set encoder type to DVI for HDMI 11835 11836 Fixes: 11837 http://bugs.freedesktop.org/show_bug.cgi?id=27452 11838 11839 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11840 11841commit ad8ea1f6e5fcb4f163622cf5eab953ea812b5829 11842Author: Alex Deucher <alexdeucher@gmail.com> 11843Date: Thu Aug 19 17:04:35 2010 -0400 11844 11845 DCE3+: switch pads to ddc mode when doing i2c 11846 11847 The pins for ddc and aux are shared so you need to switch the 11848 mode when doing ddc. The ProcessAuxChannel table already sets 11849 the pin mode to DP. This should fix unreliable ddc issues 11850 on DP ports using non-DP monitors. 11851 11852 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11853 11854commit fd686668289258ffaf6b81057545e50612aac6a8 11855Author: Dave Airlie <airlied@redhat.com> 11856Date: Thu Aug 12 12:59:18 2010 +1000 11857 11858 radeon: fixup non-kms build 11859 11860commit 5a9865d90c23c4ce0f46d380ea9119eac87a99eb 11861Author: Dave Airlie <airlied@redhat.com> 11862Date: Thu Aug 12 11:18:06 2010 +1000 11863 11864 evergreen: fix multi-head cursor support 11865 11866commit 8351bb9f085fde5dc47b115375efcc61adc23859 11867Author: Dave Airlie <airlied@redhat.com> 11868Date: Tue Aug 10 15:05:45 2010 +1000 11869 11870 radeon: take 8/10 encoding into a/c on displayport link 11871 11872commit b91e0efcb24eea32e6978c146c5409eeeeac0a62 11873Author: Dave Airlie <airlied@redhat.com> 11874Date: Tue Aug 10 15:05:25 2010 +1000 11875 11876 dce32: remove rmx workaround 11877 11878commit 94bc1b7156cd0866566dc44a823c7e051bb45175 11879Author: Dave Airlie <airlied@redhat.com> 11880Date: Tue Aug 10 13:24:52 2010 +1000 11881 11882 atombios: fixup set crtc source like KMS 11883 11884 This removes a bunch of strict aliasing warnings and fixes the 11885 codepaths up like the latest KMS code, including a workaround for a bug 11886 on evergreen. 11887 11888commit 9bc716eb62d4e0eed2902b92437a42634eef6ba1 11889Author: Dave Airlie <airlied@redhat.com> 11890Date: Tue Aug 10 13:23:21 2010 +1000 11891 11892 atombios: move adjust pixel clock around to follow KMS code flow 11893 11894 This reworks the pixel clock adjusting code to follow the KMS style, 11895 also fixes warnings in this code. 11896 11897commit 31de43bf9d9eb93cc2b2150474ea7404beabe49d 11898Author: Dave Airlie <airlied@redhat.com> 11899Date: Tue Aug 10 13:21:39 2010 +1000 11900 11901 displayport: retry on timeout 11902 11903 this is ported from KMS 11904 11905commit 9ef67335583d36080d227e8bce1966afe08e0486 11906Author: Dave Airlie <airlied@redhat.com> 11907Date: Tue Aug 10 13:21:01 2010 +1000 11908 11909 evergreen: don't call YUV table on evergreen 11910 11911commit 1cce55e8ba43e7958cb67147aeaeb068826ab99f 11912Author: Dave Airlie <airlied@redhat.com> 11913Date: Tue Aug 10 13:20:13 2010 +1000 11914 11915 evergreen: add support to parse firmware info for ext dp clk 11916 11917commit bbffd67d3296344e8735b007cdee83146d38369c 11918Author: Dave Airlie <airlied@redhat.com> 11919Date: Tue Aug 10 13:14:54 2010 +1000 11920 11921 atombios: realign digital transmitter/encoder setup with kms 11922 11923commit 6244153467665f5007e2fc7786b4bcc4b0b96030 11924Author: Dave Airlie <airlied@redhat.com> 11925Date: Tue Aug 10 12:57:22 2010 +1000 11926 11927 update atombios.h to latest from kernel 11928 11929commit bb7c77ca75e857f90791b0dd1c04c8e2f19d0e3c 11930Author: Alex Deucher <alexdeucher@gmail.com> 11931Date: Mon Aug 9 22:18:53 2010 -0400 11932 11933 atom: upstream parser update 11934 11935 fixes digital output problems on evergreen asics 11936 11937commit cc0a167ff2db9b097891883ded147af1d67e4407 11938Author: Alex Deucher <alexdeucher@gmail.com> 11939Date: Tue May 25 18:17:15 2010 -0400 11940 11941 r6xx/r7xx: default to micro (1D) tiling for now 11942 11943 SW de-tiling used in the r600c 3D driver has issues with 11944 certain configurations. 11945 11946 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11947 11948commit b5bfdbd70d9671250957ccd41dfc8818850d257e 11949Author: Alex Deucher <alexdeucher@gmail.com> 11950Date: Thu Aug 5 17:26:28 2010 -0400 11951 11952 r6xx/r7xx: add support for tiling with kms (v3) 11953 11954 Requires radeon drm 2.6.0 and updated mesa. 11955 11956 v2: - fix lockup due to not emiting DB_DEPTH_INFO 11957 https://bugs.freedesktop.org/show_bug.cgi?id=28342 11958 - fix drm minor version to deal with evergreen accel 11959 v3: rebase on current ddx git 11960 11961 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 11962 11963commit a3c59c6f6be7067421e348142da0ca13428dcd57 11964Author: Alex Deucher <alexdeucher@gmail.com> 11965Date: Wed Aug 4 12:07:36 2010 -0400 11966 11967 radeon: add new pci ids 11968 11969commit a456587b77ae357750179a50f8db2a17c0f2738e 11970Author: Alex Deucher <alexdeucher@gmail.com> 11971Date: Mon Aug 2 14:24:41 2010 -0400 11972 11973 r6xx/r7xx: move syrface sync emit to the functions that emit surface info 11974 11975 reduces code duplication. 11976 11977commit 8eba977cab1878ba247da8160771d41194d8014f 11978Author: Alex Deucher <alexdeucher@gmail.com> 11979Date: Thu Jun 10 13:34:59 2010 -0400 11980 11981 EXA: move more common helpers to exa_shared 11982 11983commit 71c1a2704af23b61439cee5ce784f7fe267a8a26 11984Author: Alex Deucher <alexdeucher@gmail.com> 11985Date: Thu Jun 10 11:52:43 2010 -0400 11986 11987 remove unused reg headers from radeon_exa_shared.c 11988 11989commit d73aef78919005369af1b60df138439b4b6105c3 11990Author: Alex Deucher <alexdeucher@gmail.com> 11991Date: Mon Aug 2 13:27:12 2010 -0400 11992 11993 r6xx/r7xx: set VGT_MAX_VTX_INDX to a larger value 11994 11995commit dacaf5d827b58c39f9e5a7ac0530f9ea6e257347 11996Author: Dave Airlie <airlied@redhat.com> 11997Date: Mon Aug 2 08:33:51 2010 +1000 11998 11999 fix make distcheck 12000 12001commit c79ce215a01b45fc63b483da167ae37ec7aefad6 12002Author: Dave Airlie <airlied@redhat.com> 12003Date: Sun Aug 1 16:51:48 2010 +1000 12004 12005 radeon/r600: restructure exa + vbo to provide more sharing 12006 12007 This is a precursor for r300/500 vbo support. 12008 12009 Signed-off-by: Dave Airlie <airlied@redhat.com> 12010 12011commit 82254b59268140c4102ae3cd713743ae2be15c00 12012Author: Alex Deucher <alexdeucher@gmail.com> 12013Date: Fri Jul 30 17:15:05 2010 -0400 12014 12015 r6xx/r7xx: unify composite mask and non-mask pixel shader 12016 12017commit 1c17f3a192f644e8e38b5cfb1470f49434bfba27 12018Author: Alex Deucher <alexdeucher@gmail.com> 12019Date: Fri Jul 30 16:34:54 2010 -0400 12020 12021 r6xx/r7xx: clean up composite vertex shader 12022 12023 keep CF, ALU, Fetch instructions in separate groups 12024 12025commit f9d6c0de231357f96e2e0de71e6c9221bcb36bd4 12026Author: Gaetan Nadon <memsize@videotron.ca> 12027Date: Fri Jul 23 13:28:42 2010 -0400 12028 12029 The local copy of the modes code is no longer required. 12030 12031 The server 1.2 as shipped in the tarball on the web does not contain the 12032 modes code. It was added just after and found in git branch server-1.2-branch. 12033 12034 The modes code was initially included in version ati 6.8.0 and fails to compile 12035 with server 1.2 as it requires randr 1.2. The modes code is included in server 12036 versions 1.3 and later, so there is no need to provide an unknown version of 12037 the modes code in the ati driver tarball. It will never be used. 12038 12039 This patch makes the ati driver requiring server 1.3 or later. 12040 Version 6.8.0 configures and builds ok on server 1.3 12041 Master branch post 6.13.1 configures and builds ok on server 1.3 12042 12043 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12044 12045commit 0028419acb0762eeb950de5fe702c93e70301612 12046Author: Gaetan Nadon <memsize@videotron.ca> 12047Date: Sat Jul 24 10:51:18 2010 -0400 12048 12049 config: add AM_PROG_CC_C_O for per-target compilation flags 12050 12051 Per-target compilation flags (theatre200_drv_la_CFLAGS) are required 12052 when multiple targets which require different compiler flags, 12053 are build in the same makefile. 12054 12055 Automake issues a command with -c and -o flags which not all compilers 12056 support. The object fles are prefixed with theatre200_drv_la. 12057 The macro AM_PROG_CC_C_O must then be used to provide this feature 12058 on compilers that do not have it. If not, a warning is issued at make time. 12059 12060 This macros checks for compiler support and if missing, uses a "compile" 12061 script it generates in the package root directory. 12062 12063 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12064 12065commit aca0a8669b538d58f018f95c9b22e6b3ec1ffe50 12066Author: Gaetan Nadon <memsize@videotron.ca> 12067Date: Wed Jul 21 16:49:04 2010 -0400 12068 12069 config: add comments for main statements 12070 12071commit 66e614f7115efeec237b3b916d9637e8b3e8985c 12072Author: Gaetan Nadon <memsize@videotron.ca> 12073Date: Wed Jul 21 16:07:00 2010 -0400 12074 12075 config: replace deprecated use of AC_OUTPUT with AC_CONFIG_FILES 12076 12077 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12078 12079commit 16e5510c90ef1ba2bbaab78d18943f080b86d809 12080Author: Gaetan Nadon <memsize@videotron.ca> 12081Date: Wed Jul 21 14:37:41 2010 -0400 12082 12083 config: replace deprecated AC_HELP_STRING with AS_HELP_STRING 12084 12085 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12086 12087commit b36d8c09e91382f4cfa71635374ec88f5b676d1c 12088Author: Gaetan Nadon <memsize@videotron.ca> 12089Date: Wed Jul 21 14:05:22 2010 -0400 12090 12091 config: replace deprecated AM_CONFIG_HEADER with AC_CONFIG_HEADERS 12092 12093 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12094 12095commit cd9351b04c2d6982b28c647a63d550eb3e1937eb 12096Author: Alex Deucher <alexdeucher@gmail.com> 12097Date: Wed Jul 21 13:48:24 2010 -0400 12098 12099 r6xx/r7xx: group op variable state 12100 12101 Group the op variable state into one emit block, re-order 12102 to reduce dwords emitted. 12103 12104commit 5f838c664e8010f4e51afecd4100d73a96fe1209 12105Author: Gaetan Nadon <memsize@videotron.ca> 12106Date: Wed Jul 21 09:27:42 2010 -0400 12107 12108 config: complete AC_INIT m4 quoting 12109 12110 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12111 12112commit 48ec2e65c268b426ab9a3e214d174447cf5b5936 12113Author: Gaetan Nadon <memsize@videotron.ca> 12114Date: Wed Jul 21 08:33:38 2010 -0400 12115 12116 config: remove unrequired AC_SUBST for LIBDRM and LIBUDEV 12117 12118 These macros are called by the PKG_CHECK_MODULES macro. 12119 12120 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12121 12122commit 1aabb7eb7d8f06c7481151145db3b9a722ce4ef0 12123Author: Gaetan Nadon <memsize@videotron.ca> 12124Date: Tue Jul 20 21:54:11 2010 -0400 12125 12126 config: remove unrequired AC_SUBST([DRI_CFLAGS]) 12127 12128 This macro is called by PKG_CHECK_MODULES 12129 12130 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12131 12132commit 8f92b349821a3ee5ed8df55273d905b9605385aa 12133Author: Gaetan Nadon <memsize@videotron.ca> 12134Date: Tue Jul 20 21:44:57 2010 -0400 12135 12136 config: remove unrequired AC_SUBST([XORG_CFLAGS]) 12137 12138 This macro is called by PKG_CHECK_MODULES 12139 12140 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12141 12142commit 5b483b832f9c4a5b92ffb7f72a470669201f1fba 12143Author: Gaetan Nadon <memsize@videotron.ca> 12144Date: Tue Jul 20 20:24:42 2010 -0400 12145 12146 config: remove unrequired AC_HEADER_STDC 12147 12148 Autoconf says: 12149 "This macro is obsolescent, as current systems have conforming 12150 header files. New programs need not use this macro". 12151 12152 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12153 12154commit 6574e3a16eb3631ee7e00ee60a8c9ba95c8b84ef 12155Author: Gaetan Nadon <memsize@videotron.ca> 12156Date: Tue Jul 20 19:41:30 2010 -0400 12157 12158 config: remove AC_PROG_CC as it overrides AC_PROG_C_C99 12159 12160 XORG_STRICT_OPTION from XORG_DEFAULT_OPTIONS calls 12161 AC_PROG_C_C99. This sets gcc with -std=gnu99. 12162 If AC_PROG_CC macro is called afterwards, it resets CC to gcc. 12163 12164 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12165 12166commit 593eff2924c2ad161d8b490fbf6d7e433fbe2a80 12167Author: Gaetan Nadon <memsize@videotron.ca> 12168Date: Tue Jul 20 18:45:18 2010 -0400 12169 12170 config: update AC_PREREQ statement to 2.60 12171 12172 Unrelated to the previous patches, the new value simply reflects 12173 the reality that the minimum level for autoconf to configure 12174 all x.org modules is 2.60 dated June 2006. 12175 12176 ftp://ftp.gnu.org/gnu/autoconf/autoconf-2.60.tar.gz 12177 12178 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12179 12180commit c2ab6ffc25aa6759cbbb4c1fbbd4a136b38983bf 12181Author: Gaetan Nadon <memsize@videotron.ca> 12182Date: Tue Jul 20 16:15:29 2010 -0400 12183 12184 config: upgrade to util-macros 1.8 for additional man page support 12185 12186 Use MAN_SUBST now supplied in XORG_MANPAGE_SECTIONS 12187 The value of MAN_SUBST is the same for all X.Org packages. 12188 12189commit cdeb1949c820242f05a8897d3ddd0718f204dacf 12190Author: Jerome Glisse <jglisse@redhat.com> 12191Date: Thu Jul 15 16:21:41 2010 -0400 12192 12193 kms: don't call cursor helper if using software cursor 12194 12195 Fix : 12196 https://bugzilla.redhat.com/show_bug.cgi?id=601713 12197 https://bugzilla.redhat.com/show_bug.cgi?id=598358 12198 12199 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 12200 12201commit 06691376b1ee963c711420edaf5a03eab6f5658f 12202Author: Dave Airlie <airlied@redhat.com> 12203Date: Wed Jul 7 13:15:03 2010 +1000 12204 12205 fix build on non-kms 12206 12207commit 052cf0169ae70d5448af6dc4db840b2fc195569b 12208Author: Dave Airlie <airlied@redhat.com> 12209Date: Wed Jul 7 11:10:46 2010 +1000 12210 12211 configure.ac: bump version post release 12212 12213commit ad999e633ff41d27eed9d2c6535e163a7181b0bd 12214Author: Dave Airlie <airlied@redhat.com> 12215Date: Wed Jul 7 10:49:22 2010 +1000 12216 12217 set version for release 12218 12219commit 37b348059b1c15d7b381cd3df3db52bd9ee6613e 12220Author: Alex Deucher <alexdeucher@gmail.com> 12221Date: Wed Jun 30 12:56:48 2010 -0400 12222 12223 remove rv100 quirk 12224 12225 Some RV100 cards with 2 VGA ports show up with DVI+VGA, however 12226 some boards with DVI+VGA have the same subsystem ids. Better 12227 to have a VGA port show up as DVI than having a non-useable 12228 DVI port. 12229 12230 reported by DHR in irc. 12231 12232 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12233 12234commit c8ea942bd0e9d1c055d50b94440aa4dae425f20b 12235Author: Cooper Yuan <cooperyuan@gmail.com> 12236Date: Tue Jun 29 20:34:57 2010 +0800 12237 12238 Remove HDP_SOFT_RESET function, there is no need to reset HDP block. 12239 12240 This commit can fix an issue reported on DELL server, system gets hang 12241 during soft resetting while another application tries to access PCI 12242 configuration space. 12243 12244commit 139b38bf67ec10d876cc56df833541d497ae4fa4 12245Author: Alex Deucher <alexdeucher@gmail.com> 12246Date: Thu Jun 24 14:55:09 2010 -0400 12247 12248 r6xx shader: use ADDR() for CF_DWORD0 12249 12250 no change in functionality 12251 12252commit 801e83227a59a29eea425ea612083bbf2b536c30 12253Author: Wolfram <bugzilla1@malloc.de> 12254Date: Mon Jun 21 18:59:19 2010 -0400 12255 12256 r6xx/r7xx: fix ums cmd buffer leak 12257 12258 Fixes: 12259 https://bugs.freedesktop.org/show_bug.cgi?id=27957 12260 12261commit b13d719080b75fc6db4d15d2d323b8fce8b7ad06 12262Author: Alex Deucher <alexdeucher@gmail.com> 12263Date: Mon Jun 21 18:49:44 2010 -0400 12264 12265 r6xx/r7xx: fix miscount in state emit 12266 12267commit c3c5c8e2cc91b51a24effdffb85281216eed731d 12268Author: Alex Deucher <alexdeucher@gmail.com> 12269Date: Mon Jun 21 14:30:17 2010 -0400 12270 12271 r6xx/r7xx accel: add back some additional default state 12272 12273 This adds back everything removed in c29157bbf5b0dd26857675282ab094082fbaed0d 12274 except CB_FOG_*, CB_CLEAR_* and the VPORT transforms. Those shouldn't 12275 be needed as we aren't using fog or viewport transforms. We probably don't 12276 need all the state that was added back either but I can't reproduce any 12277 problems here, so it's hard to say which parts are problematic. 12278 12279 Should fix: 12280 https://bugs.freedesktop.org/show_bug.cgi?id=28629 12281 and several corruption reports on #radeon. 12282 12283 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12284 12285commit 800cb2088fec698d0626063a9ab198ff534938c0 12286Author: Michel Dänzer <daenzer@vmware.com> 12287Date: Mon Jun 21 08:15:14 2010 +0200 12288 12289 DRI2: Fix up confusion between windows and pixmaps. 12290 12291 Fixes crashes with xserver master, where looking up a pixmap private on a 12292 window no longer happens to work. 12293 12294commit f7a91ece264af9f3fd2fc18e99aefcda93ce9f5c 12295Author: Michel Dänzer <daenzer@vmware.com> 12296Date: Mon Jun 14 08:52:16 2010 +0200 12297 12298 Convert x(c)alloc/xfree to m/calloc/free. 12299 12300 Fixes deprecation warnings with xserver master and should also work with older 12301 xservers. 12302 12303commit ea37d24b1b6d4cbcf73e680846de25b72af216e3 12304Author: Dave Airlie <airlied@redhat.com> 12305Date: Mon Jun 21 13:55:15 2010 +1000 12306 12307 radeon: fix support for 1.9 server master. 12308 12309 This moves pixmap and mode set into CSR where its allowed. Should work fine on 12310 old servers also. 12311 12312 Signed-off-by: Dave Airlie <airlied@redhat.com> 12313 12314commit fdd8ecafd054f65842351aee6ee6fba7af6613b2 12315Author: Alex Deucher <alexdeucher@gmail.com> 12316Date: Wed Jun 16 19:02:10 2010 -0400 12317 12318 r6xx/r7xx: macro safety fixes 12319 12320commit 4651d77211b508cb6b76931807780e317f232220 12321Author: Alex Deucher <alexdeucher@gmail.com> 12322Date: Wed Jun 16 12:28:36 2010 -0400 12323 12324 radeon: fix depth 16 with ums 12325 12326 Fixes: 12327 https://bugs.freedesktop.org/show_bug.cgi?id=28494 12328 https://bugzilla.redhat.com/show_bug.cgi?id=554967 12329 12330 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12331 12332commit 1e1d6a515428b6884fea586d180346fc74ef75a1 12333Author: Alex Deucher <alexdeucher@gmail.com> 12334Date: Wed Jun 16 12:20:03 2010 -0400 12335 12336 r3xx-r5xx Xv: disable bicubic filtering by default 12337 12338 - makes Xv more consistent with r1xx/r2xx/r6xx/r7xx 12339 - Xv attributes like brightness, contrast, hue, etc. only work 12340 when bicubic is disabled. 12341 - avoids performance issues on some systems when sampling from textures 12342 in GART with kms. 12343 - It can be re-enabled with using xvattr 12344 12345 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12346 12347commit c29157bbf5b0dd26857675282ab094082fbaed0d 12348Author: Alex Deucher <alexdeucher@gmail.com> 12349Date: Tue Jun 15 19:39:42 2010 -0400 12350 12351 r6xx/r7xx: remove unnecessary state emit 12352 12353 No need to emit state that doesn't impact that we use for EXA/Xv. 12354 12355 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12356 12357commit 35280545fcfb911f189d7657bd1040a28450fe7b 12358Author: Alex Deucher <alexdeucher@gmail.com> 12359Date: Tue Jun 15 19:21:32 2010 -0400 12360 12361 r6xx/r7xx: reorg default and clipping state emit 12362 12363 Group ordered registers to save command buffer space. 12364 Reduces the default and clipping state from 256 to 160 dwords. 12365 12366 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12367 12368commit a2528a734c1d4e8639c49e5d222e3630a93ffbfd 12369Author: Alex Deucher <alexdeucher@gmail.com> 12370Date: Tue Jun 15 17:00:34 2010 -0400 12371 12372 r6xx/r7xx accel: remove some duplicate emits and minor clean up 12373 12374 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12375 12376commit 5f093357f18eb9bea641394ab86a92a1766d8f2e 12377Author: Gaetan Nadon <memsize@videotron.ca> 12378Date: Sat Jun 12 15:11:46 2010 -0400 12379 12380 COPYING: replace stub file with actual copyright notices 12381 12382 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 12383 12384commit 0c2118586d3edeecc2473b5d685472df4b5e70fa 12385Author: Dave Airlie <airlied@redhat.com> 12386Date: Fri Jun 11 15:53:18 2010 +1000 12387 12388 radeon: fixup last fix, use CURSOR_WIDTH not RADEON_CURSOR_WIDTH. 12389 12390commit 9a117f768cb8261327fd3c324da9c98875785cc1 12391Author: Dave Airlie <airlied@redhat.com> 12392Date: Fri Jun 11 14:28:24 2010 +1000 12393 12394 radeon: set proper stride for cursor in tiling flags. 12395 12396 definitely incorrect, but hope it doesn't break anything. 12397 12398 Signed-off-by: Dave Airlie <airlied@redhat.com> 12399 12400commit b6346ede94f0d0b11ee04770cf52508cb0a5e6c6 12401Author: Dave Airlie <airlied@redhat.com> 12402Date: Thu Jun 10 20:13:58 2010 -0400 12403 12404 drmmode: fix big endian issue with properties. 12405 12406 On a power machine with an rn50, this was causing load detection to be turned off after the first X run, subsequent X runs would find nothing connected. 12407 12408 Signed-off-by: Dave Airlie <airlied@redhat.com> 12409 12410commit 1971dc6d758dea13b9fb6d1c8f516c165628b5e5 12411Author: Alex Deucher <alexdeucher@gmail.com> 12412Date: Tue Jun 8 11:18:28 2010 -0400 12413 12414 radeon: fix rn50 cloning with kms 12415 12416 Since they only have one crtc sometimes the xserver doesn't assign 12417 a crtc to one of the outputs even though both outputs have common modes 12418 which results in only one monitor being enabled. Assign a crtc in 12419 preinit so that both outputs light up. 12420 12421 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12422 12423commit 426114b4a99d37b394efe3336968bb0ab9b6e9ae 12424Author: Dave Airlie <airlied@redhat.com> 12425Date: Tue Jun 8 11:34:35 2010 +1000 12426 12427 xv: fix Xv on M6/RV100 under KMS. 12428 12429 pRADEONEnt->HasCRTC2 wasn't setup under KMS. 12430 12431 Signed-off-by: Dave Airlie <airlied@redhat.com> 12432 12433commit f64bf0de8e2de7c1bf9cc0c614603dd23c9060ad 12434Author: Alex Deucher <alexdeucher@gmail.com> 12435Date: Thu Jun 3 14:48:21 2010 -0400 12436 12437 evergreen: explicitly disable accel on evergreen 12438 12439 Previously we relied on the drm not having accel enabled 12440 to make sure evergreen used shadowfb, now we when we enable 12441 accel in the drm, we need to make sure the ddx doesn't try 12442 and use it until it's implemented. 12443 12444 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12445 12446commit 428125c095b18c2a2864c1aef24ac0f384b6be54 12447Author: Marek Olšák <maraeo@gmail.com> 12448Date: Sun May 16 20:11:17 2010 +0000 12449 12450 r3xx-r5xx: enable color tiling by default on KMS 12451 12452 Signed-off-by: Marek Olšák <maraeo@gmail.com> 12453 12454commit 6712ce4f4715c8ce7c2fcddb52ca9b103bdd634c 12455Author: Marek Olšák <maraeo@gmail.com> 12456Date: Sun May 16 20:09:31 2010 +0000 12457 12458 rs740: follow macro-switch too 12459 12460 Signed-off-by: Marek Olšák <maraeo@gmail.com> 12461 12462commit 842fa162e9d3bbad2bc44d3732bbc8e5a54402ea 12463Author: Alex Deucher <alexdeucher@gmail.com> 12464Date: Tue May 18 11:21:01 2010 -0400 12465 12466 dri2: use radeon_pick_best_crtc() 12467 12468 It's already exposed and used by exa and Xv and has 12469 the same functionality. radeon_covering_crtc() can 12470 be dropped eventually. 12471 12472commit 64e6858aaf5d0e39ecc9f9804689012f3a38660a 12473Author: Alex Deucher <alexdeucher@gmail.com> 12474Date: Mon May 17 12:50:44 2010 -0400 12475 12476 r2xx exa: fix typo in 27c3326863deae36bc35e2c3b73fffd400208ff1 12477 12478 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12479 12480commit 27c3326863deae36bc35e2c3b73fffd400208ff1 12481Author: Alex Deucher <alexdeucher@gmail.com> 12482Date: Mon May 17 09:53:32 2010 -0400 12483 12484 r1xx/r2xx EXA: handle RepeatNone properly with transforms 12485 12486 Fallback with unsupported ops like r3xx, r6xx. 12487 12488 Should fix fdo bug 28142. 12489 12490 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 12491 12492commit 30591320ec46e491ba20904cc64f3405b51c6505 12493Author: Jerome Glisse <jglisse@redhat.com> 12494Date: Fri May 7 17:35:09 2010 +0200 12495 12496 kms: add support for the MSC swap & sync API 12497 12498 This patch is mostly a port over of Intel ddx code for 12499 MSC support. It needs a radeon KMS module with version 12500 2.4 which has the query for hw crtc id. 12501 12502 Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> 12503 Signed-off-by: Jerome Glisse <jglisse@redhat.com> 12504 12505commit 766024dcc61c83490540910ce752f9bfe6dddba4 12506Author: Marek Olšák <maraeo@gmail.com> 12507Date: Mon May 3 22:56:27 2010 +0200 12508 12509 r3xx-r5xx: fix texturing with small macrotiled pixmaps 12510 12511 Pixmaps smaller than a macrotile cannot be used as textures because hardware 12512 automatically switches to macro-linear and therefore sampled pixels are 12513 messed up. This behavior is known as MACRO_SWITCH. 12514 12515 The only sane workaround seems to be not using macrotiling for small pixmaps. 12516 12517 The function RADEONMacroSwitch has been ported from r300g and implements 12518 MACRO_SWITCH the same way it's implemented in hardware. It's been well tested 12519 in r300g. 12520 12521 This commit also fixes blit-based framebuffer reads, which are used for tiled 12522 surfaces in r300g, when ColorTiling is enabled. 12523 12524 Signed-off-by: Marek Olšák <maraeo@gmail.com> 12525 12526commit 6aedd57f6c2f08d0151a8bd1c5893d40d3db709a 12527Author: Alex Deucher <alexdeucher@gmail.com> 12528Date: Tue May 4 14:30:52 2010 -0400 12529 12530 radeon: fix pll_out_min default for pre-avivo 12531 12532 Fixes fdo bug 27960 12533 12534commit b2ecd1bb507ed4e0d9ba6ebd498c51da2e8136fa 12535Author: Michel Dänzer <daenzer@vmware.com> 12536Date: Fri Apr 23 18:28:22 2010 +0200 12537 12538 FB/MMIO_cnt should be ints, not pointers. 12539 12540 Pointed out by compiler warnings. 12541 12542 https://bugs.freedesktop.org/show_bug.cgi?id=27817 12543 12544commit 1ad41054989f7c3edec373ccf09dceda6d7c94b4 12545Author: Alex Deucher <alexdeucher@gmail.com> 12546Date: Thu Apr 22 16:48:21 2010 -0400 12547 12548 r3xx: fix gb pipe setup for SE cards 12549 12550 Noticed by Tormod Volden 12551 12552commit 59e287d7c484b4addd4a06d013670577639c7ae2 12553Author: Alex Deucher <alexdeucher@gmail.com> 12554Date: Thu Apr 22 02:05:40 2010 -0400 12555 12556 r3xx-r5xx: fix vertex units 12557 12558 Noticed by Tormod Volden. 12559 12560 RV3xx is 2, RV560,RV570 is 8 12561 12562commit 2a6b409496f26da0436972b5feae6ea035dde08d 12563Author: Owain Ainsworth <zerooa@googlemail.com> 12564Date: Mon Apr 19 14:37:33 2010 +0100 12565 12566 Reference count shared driver mappings. 12567 12568 With MMIO it wasn't *such* a bit deal if we leaked the smallish mapping. 12569 with FB it could be a larger deal. So instead of worrying about this, 12570 reference count the mappings in the entity structure and unmap them when 12571 no one cares anymore. 12572 12573 Prompted by a discussion with airlied 12574 12575 Signed-off-by: Dave Airlie <airlied@redhat.com> 12576 12577commit 761f0de5556e46f166280476185977f720efe586 12578Author: Owain G. Ainsworth <zerooa@googlemail.com> 12579Date: Mon Apr 19 04:05:36 2010 +0100 12580 12581 Make consistent use of fbOffset and share fb mappings. 12582 12583 What we were doing previously was mapping the framebuffer for zaphod for 12584 only this driver instances chunk, however, fbOffset was (rightly) set to 12585 the offset into the whole framebuffer we were using. 12586 12587 Since in some cases we did operations on the FB virtual address + 12588 fbOffset (for example zeroing the framebuffer on entervt) we were 12589 actually pissing all over ourselves in those cases. 12590 12591 Fix this by implementing shared fb mappings like we do for MMIO already, 12592 and whenever we wish to refer to our area of FB space we always use 12593 fbOffset. Fixes zaphod for some users on r600 chipsets, my 4870 is still 12594 behaving strangely on screen 0, but I suspect that is another bug. 12595 12596 Once calculation (in PreInitAccel) is now wrong because of this, however 12597 dri on zaphod does now happen so this is irrelavent, add a comment to 12598 that effect. 12599 12600 Signed-off-by: Dave Airlie <airlied@redhat.com> 12601 12602commit 2059d628c5fb03222a62502cc1b5724bf296a7b3 12603Author: Owain G. Ainsworth <zerooa@googlemail.com> 12604Date: Mon Apr 19 04:05:35 2010 +0100 12605 12606 Kill per-context SAREA support. 12607 12608 It has never been used, and since the world is changing it almost 12609 certainly never will be. Good riddance. 12610 12611 Signed-off-by: Dave Airlie <airlied@redhat.com> 12612 12613commit c4b46f56279cc18c8114235ca41cfb41ffede67f 12614Author: Owain G. Ainsworth <zerooa@googlemail.com> 12615Date: Mon Apr 19 04:05:34 2010 +0100 12616 12617 Remove remnants of the ShowCache option. 12618 12619 The rest of it died a long time ago. 12620 12621 Signed-off-by: Dave Airlie <airlied@redhat.com> 12622 12623commit 03b34c5fb12c4df180d4df113d2b654f6afd7230 12624Author: Dave Airlie <airlied@redhat.com> 12625Date: Mon Apr 19 17:58:07 2010 +1000 12626 12627 radeon: add FireMV 2400 pci id. 12628 12629 From RH#581927 - thanks to Carsten Clasohm 12630 12631 Signed-off-by: Dave Airlie <airlied@redhat.com> 12632 12633commit 6c07816025f70e522986614c29c300ea13e5d932 12634Author: Alex Deucher <alexdeucher@gmail.com> 12635Date: Fri Apr 16 18:50:17 2010 -0400 12636 12637 r5xx texvid: deal with large numbers of verts 12638 12639 should fix fdo bug 25884 12640 12641commit 5ee320a6b1b4c65fe592c8ac4d1981799242d59e 12642Author: Alex Deucher <alexdeucher@gmail.com> 12643Date: Fri Apr 16 18:42:56 2010 -0400 12644 12645 r3xx texvid: deal with large numbers of verts 12646 12647 should fix fdo bug 25884 12648 12649commit 47af3f4f266232517486238917d82fc5ca9c82e6 12650Author: Alex Deucher <alexdeucher@gmail.com> 12651Date: Fri Apr 16 20:50:18 2010 -0400 12652 12653 r2xx texvid: deal with large numbers of verts 12654 12655 should fix fdo bug 25884 12656 12657commit 57577d5cd0641b7cad02242478699bcfece59227 12658Author: Alex Deucher <alexdeucher@gmail.com> 12659Date: Fri Apr 16 20:40:56 2010 -0400 12660 12661 r1xx texvid: deal with large numbers of verts 12662 12663 should fix fdo bug 25884 12664 12665commit 31a888e78fa403f2bddacee098a29f36eaa969cb 12666Author: Alex Deucher <alexdeucher@gmail.com> 12667Date: Fri Apr 16 15:52:59 2010 -0400 12668 12669 r5xx textured video: split into prepare and draw functions 12670 12671commit f7ebe4a127965bb9e46d20c60331bcc4dab76fc5 12672Author: Alex Deucher <alexdeucher@gmail.com> 12673Date: Fri Apr 16 15:46:24 2010 -0400 12674 12675 r3xx textured video: split into prepare and draw functions 12676 12677commit 10799b4ce0c18c5a7d92d688549e8e777344f111 12678Author: Alex Deucher <alexdeucher@gmail.com> 12679Date: Fri Apr 16 15:38:49 2010 -0400 12680 12681 r2xx textured video: split into prepare and draw functions 12682 12683commit bd1cf42201a1f918080bf34786a6de327cb31b5f 12684Author: Alex Deucher <alexdeucher@gmail.com> 12685Date: Fri Apr 16 15:31:40 2010 -0400 12686 12687 r1xx textured video: split into prepare and draw functions 12688 12689commit c1136f94b80346065893f8a43c0fbf60c8c9b057 12690Author: Alex Deucher <alexdeucher@gmail.com> 12691Date: Fri Apr 16 15:23:14 2010 -0400 12692 12693 Xv: track vtx_count, is_planar in port priv 12694 12695commit 2f680b631357661511456f07b7b4ba1aa2cbf30c 12696Author: Alex Deucher <alexdeucher@gmail.com> 12697Date: Fri Apr 16 13:43:41 2010 -0400 12698 12699 radeon/kms: fix possible crtc mask for evergreen 12700 12701commit 4656f5dff1ed72fa2c7a1484305f2aef7b65ff2b 12702Author: Alex Deucher <alexdeucher@gmail.com> 12703Date: Fri Apr 16 13:36:25 2010 -0400 12704 12705 radeon/kms: fix crash when using more than two heads 12706 12707 fixes fdo bug 27692 12708 12709commit a69e749d0562887af6bd236c38802472e54640c4 12710Author: Dave Airlie <airlied@redhat.com> 12711Date: Fri Apr 16 13:58:54 2010 +1000 12712 12713 kms: add uevent support. 12714 12715 When we get a hotplug event from the kernel we should notify the client side to reconfigure displays. 12716 12717 based on work by ajax in intel driver. 12718 12719 Signed-off-by: Dave Airlie <airlied@redhat.com> 12720 12721commit 22a46dddd375b2b9399e12fdf168fa5292ff17a4 12722Author: Alex Deucher <alexdeucher@gmail.com> 12723Date: Tue Apr 13 11:08:59 2010 -0400 12724 12725 radeon: add support for pll algo selection 12726 12727 tv-out on atom systems is very particular about it's 12728 dividers. force it to use the old algo. 12729 12730 Should fix fdo bug 27593. 12731 12732commit c1b817c45477c21234abaaebe78feb7ea4fd92b1 12733Author: Alex Deucher <alexdeucher@gmail.com> 12734Date: Mon Apr 12 16:32:04 2010 -0400 12735 12736 atom: disable TV encoder when VGA is in use 12737 12738 Switching between TV and VGA caused problems on some 12739 systems since the TV encoder was left enabled. 12740 12741 Should fix fdo bug 25520. 12742 12743commit 878814aeab7031ff6169fcaf2be869b7595c8699 12744Author: Alex Deucher <alexdeucher@gmail.com> 12745Date: Mon Apr 12 11:16:14 2010 -0400 12746 12747 Xv overlay: set scalerwidth to 1920 on r3xx/r4xx chips 12748 12749 reported on phoronix forums. 12750 12751commit 276fe739f1caa95778213054375a65288b80f320 12752Author: Alex Deucher <alexdeucher@gmail.com> 12753Date: Fri Apr 9 20:03:09 2010 -0400 12754 12755 radeon: add some new pci ids 12756 12757commit eb5665688ef9b52f03f61546351d0848cab54740 12758Author: Alex Deucher <alexdeucher@gmail.com> 12759Date: Thu Apr 8 00:31:52 2010 -0400 12760 12761 radeon: don't setup Xv on rn50 12762 12763 It has no overlay or 3D engine. 12764 12765 fixes fdo bug 27528 12766 12767commit 3c6c6afcbfc6f2c582e13a7c37fe2014b0875b22 12768Author: Alex Deucher <alexdeucher@gmail.com> 12769Date: Mon Apr 5 12:12:56 2010 -0400 12770 12771 change version post release 12772 12773commit fb7911912e60b2cdbc2152b96847775b767ca3ef 12774Author: Alex Deucher <alexdeucher@gmail.com> 12775Date: Mon Apr 5 11:59:24 2010 -0400 12776 12777 bump version for release 12778 12779commit 476a1c6e8b42807b897b8c6f8550ec42b2d5c10e 12780Author: Alex Deucher <alexdeucher@gmail.com> 12781Date: Thu Apr 1 12:42:56 2010 -0400 12782 12783 radeon: fix possible gpio i2c table overrun 12784 12785 GPIO_I2C_INFO does not always have ATOM_MAX_SUPPORTED_DEVICE 12786 entries. Should fix Novell bug 589022. 12787 12788commit 6baa96c44ca93b88acf5233335cee233e59d5af4 12789Author: Dave Airlie <airlied@redhat.com> 12790Date: Thu Apr 1 11:24:32 2010 +1000 12791 12792 r600/exa: further cleanup, use the object struct in the accel state. 12793 12794 This cleans up the accel state objects as well. 12795 12796 Signed-off-by: Dave Airlie <airlied@redhat.com> 12797 12798commit 87f49a24019967552c5478d6af89ba438b4261c8 12799Author: Dave Airlie <airlied@redhat.com> 12800Date: Thu Apr 1 10:52:58 2010 +1000 12801 12802 r600: cleanup wasteful variables. 12803 12804 Just assign directly to the structs. 12805 12806 Signed-off-by: Dave Airlie <airlied@redhat.com> 12807 12808commit 478319e55d52fce241d040912fea952071e3b742 12809Author: Dave Airlie <airlied@redhat.com> 12810Date: Thu Apr 1 10:45:52 2010 +1000 12811 12812 r600: reduce function call overhead. 12813 12814 Create a small accel object that can be used to reduce the amount 12815 of parameters passed to SetAccelState. This can be cleaned up a lot more. 12816 12817 Signed-off-by: Dave Airlie <airlied@redhat.com> 12818 12819commit bc93395b3eb5e3511c1b62af90693269f4fa6e13 12820Author: Alex Deucher <alexdeucher@gmail.com> 12821Date: Wed Mar 31 22:30:24 2010 -0400 12822 12823 r6xx+ EXA: fix swapped domains in kms UTS 12824 12825 Noticed by Dave. 12826 12827 Should fix fdo bug 27284 12828 12829commit 3a33b1a92c328b690bdc1aafa0011e4326109815 12830Author: Owain G. Ainsworth <zerooa@googlemail.com> 12831Date: Tue Mar 30 17:14:02 2010 +0100 12832 12833 Add support for MacModel autodetection on OpenBSD. 12834 12835 This allows radeon-equipped macs to work without an xorg.conf on openbsd. 12836 Patch originally written by Mark Kettenis (kettenis@openbsd.org) 12837 12838commit c7eeda8c3f3514ba95ebf2893fbe124bf526b3df 12839Author: Alex Deucher <alexdeucher@gmail.com> 12840Date: Fri Mar 26 01:38:36 2010 -0400 12841 12842 radeon: add support eDP connectors with and LCD device tag 12843 12844 should fix fdo bug 27322 12845 12846commit ff323e11485fa5030ac278c836bf80a535c22249 12847Author: Alex Deucher <alexdeucher@gmail.com> 12848Date: Fri Mar 26 01:24:16 2010 -0400 12849 12850 radeon: fix name string for eDP 12851 12852commit 53ac06161eb2b8cffb1b88e24a9a21cfd12e5883 12853Author: Alex Deucher <alexdeucher@gmail.com> 12854Date: Tue Mar 23 13:34:38 2010 -0400 12855 12856 r6xx+ EXA/Xv: add a R600SetAccelState function 12857 12858 This moves CS bo checking and alignment checks into 12859 a central location. It also cleans up the code. 12860 12861commit 7a044472dfea7cf05ba4c2b87be30e94e2ae0b62 12862Author: Alex Deucher <alexdeucher@gmail.com> 12863Date: Tue Mar 23 01:27:22 2010 -0400 12864 12865 r6xx+ EXA: always use a temp surface for overlapping copies 12866 12867 The scanline based fallback code was complex and added a lot of 12868 overhead. It also didn't work with kms. 12869 12870commit d33bddaedae81126ab7f0023af1c6443bb0b4c4f 12871Author: Alex Deucher <alexdeucher@gmail.com> 12872Date: Mon Mar 22 18:15:23 2010 -0400 12873 12874 r6xx+ EXA: always use the accel_state state in DoPrepareCopy 12875 12876commit dda3f5a99e7a2dc5d57860f4d07df3498e1e21df 12877Author: Alex Deucher <alexdeucher@gmail.com> 12878Date: Mon Mar 22 18:05:36 2010 -0400 12879 12880 r6xx EXA/Xv: track src/dst domains 12881 12882 Much of the code is shared, so track the src/dst 12883 domains so we make sure the uses consistent domains 12884 for each bo. 12885 12886commit 5c256808cb5fea955eea96ffe9196473715156aa 12887Author: Alex Deucher <alexdeucher@gmail.com> 12888Date: Wed Mar 17 23:47:50 2010 -0400 12889 12890 XAA: disable render accel 12891 12892 It's been reported broken for a while. Should fix 12893 fdo bug 27151, others. 12894 12895commit a0683be5cc082bdbdd3bc4e9b52f39f423650946 12896Author: Dave Airlie <airlied@redhat.com> 12897Date: Thu Mar 18 12:36:25 2010 +1000 12898 12899 radeon: avoid using DRI1 init path on DRI2 driver. 12900 12901 I was playing with multi-seat and found this code, fixed 12902 it up to be sane and more DRI2 like. 12903 12904 Signed-off-by: Dave Airlie <airlied@redhat.com> 12905 12906commit 080a5414593e9b59ed923f26aa6057747b0c868f 12907Author: Alex Deucher <alexdeucher@gmail.com> 12908Date: Tue Mar 16 12:33:39 2010 -0400 12909 12910 kms: fix ums naming compat for DisplayPort 12911 12912commit 819b4015349b5d8c5ffa5f979097599774fce5bb 12913Author: Alex Deucher <alexdeucher@gmail.com> 12914Date: Mon Mar 15 13:47:29 2010 -0400 12915 12916 dump version for rc release 12917 12918commit d3482a947e3731be4ed0b00b4e3079470700dc4c 12919Author: Michael Cree <mcree@orcon.net.nz> 12920Date: Fri Mar 12 22:23:31 2010 +1300 12921 12922 Fix some word accesses in AtomBios to work on all architectures. 12923 12924 The UINT16LE_TO_CPU(), etc., macros are used in the AtomBios code to 12925 fix up endian issues but they do not address bad alignment or assist 12926 architectures that cannot perform hardware byte or word accesses. 12927 This patch inserts use of the ldw_u(), etc., interface of the Xserver 12928 into certain AtomBios accesses to address alignment issues. 12929 12930 This resolves Debian bug 572311, namely that the driver when compiled 12931 for generic Alpha architecture (i.e. doesn't use the byte-word extension) 12932 resulted in no display output on certain Radeon cards. 12933 12934 Signed-off-by: Michael Cree <mcree@orcon.net.nz> 12935 12936commit 488c9fd8300505cc6c0c2f8f0f00849f27cc5d63 12937Author: Alex Deucher <alexdeucher@gmail.com> 12938Date: Mon Mar 15 12:25:57 2010 -0400 12939 12940 r6xx/r7xx: fix domain handling in accel code 12941 12942 Noticed by Pauli and Michel on IRC. 12943 12944 Improves GetImage performace by a factor of ~10. 12945 12946commit 2ace2591d92fb6d3ce7a6453edb04b36a6c49a32 12947Author: Alex Deucher <alexdeucher@gmail.com> 12948Date: Mon Mar 15 12:03:58 2010 -0400 12949 12950 radeon: remove some leftover debugging output 12951 12952commit 67e81c8f17ddde6eba633d2a5aef528e1d598d89 12953Author: Andrzej Hajda <andrzej.hajda@wp.pl> 12954Date: Wed Mar 10 18:19:35 2010 -0500 12955 12956 radeon: add support for pal on legacy IGP chips 12957 12958 Based on my initial non-working patch. 12959 12960 Fixes some element of fdo bug 12007 12961 12962commit 3a44f1cb0d2bb748692b1024003de8ee88ca77a5 12963Author: Alex Deucher <alexdeucher@gmail.com> 12964Date: Tue Mar 9 09:44:01 2010 -0500 12965 12966 atom: i2c gpio fixes 12967 12968 Basically a port of my kms patch. This allows us 12969 to remove some quirks. 12970 12971commit d85cb40f516c67305e818302bec7ee817df4144c 12972Author: Matt Turner <mattst88@gmail.com> 12973Date: Sun Mar 7 14:24:35 2010 -0500 12974 12975 Don't check for Xinerama. 12976 12977 It doesn't seem to be used anywhere, so don't require it. 12978 12979 CC: Jerome Glisse <jglisse@redhat.com> 12980 CC: Alex Deucher <alexdeucher@gmail.com> 12981 CC: Dave Airlie <airlied@redhat.com> 12982 Signed-off-by: Matt Turner <mattst88@gmail.com> 12983 12984commit e7b41f8cb082ed462d29bf3fc440072424cbd852 12985Author: Alex Deucher <alexdeucher@gmail.com> 12986Date: Fri Mar 5 19:16:11 2010 -0500 12987 12988 radeon: disable frac fb div with new pll code 12989 12990 fixes fdo bug 26897 12991 12992commit 14aff767490c253cbcdd411f812e50b91673119e 12993Author: Alex Deucher <alexdeucher@gmail.com> 12994Date: Wed Mar 3 13:31:19 2010 -0500 12995 12996 radeon: add new RS880 pci id 12997 12998commit e6dc886634b38e4a36af7b5f0b23299d5acd7244 12999Author: Dave Airlie <airlied@redhat.com> 13000Date: Tue Mar 2 10:25:15 2010 +1000 13001 13002 radeon: bump configure.ac 13003 13004commit 4975658f05c387b39b3e96a292a7683f17645c2c 13005Author: Dave Airlie <airlied@redhat.com> 13006Date: Sat Feb 27 16:47:19 2010 +1000 13007 13008 pciids: hopefully fix HP 13009 13010commit e76b90b399c3cc0f0998c0209300c46f97505498 13011Author: Alex Deucher <alexdeucher@gmail.com> 13012Date: Fri Feb 26 15:01:28 2010 -0500 13013 13014 rv740: disable dfs workaround for drm 1.32+ 13015 13016 rv740 pipe setup was fixed in 2.6.33 13017 13018commit 2de0af5f5b806f9dbfdb1e9b6a5cf96d9433961d 13019Author: Matt Turner <mattst88@gmail.com> 13020Date: Wed Feb 24 22:46:28 2010 -0500 13021 13022 Use RADEON_ALIGN instead of open coding it. 13023 13024 Also fix some RADEON_ALIGN(x, 63), which would return incorrect results 13025 for odd x. Though this shouldn't happen, it's still not right. You 13026 wouldn't ever write (x + 62) & ~62 which is clearly wrong (and what it 13027 expands to). 13028 13029 CC: Jerome Glisse <jglisse@redhat.com> 13030 CC: Alex Deucher <alexdeucher@gmail.com> 13031 CC: Dave Airlie <airlied@redhat.com> 13032 Signed-off-by: Matt Turner <mattst88@gmail.com> 13033 13034commit c7e81d2f3a372e0d5f751dd0c5091aec2b56d936 13035Author: Matt Turner <mattst88@gmail.com> 13036Date: Wed Feb 24 22:46:27 2010 -0500 13037 13038 Use/define RADEON_GPU_PAGE_SIZE instead of sprinkling 4096 everywhere. 13039 13040 Also, define RADEON_BUFFER_ALIGN in terms of it, and replace some 13041 RADEON_ALIGN(x, RADEON_BUFFER_ALIGN) with RADEON_ALIGN(x, 13042 RADEON_GPU_PAGE_SIZE) since this is really what was intended. 13043 13044 CC: Jerome Glisse <jglisse@redhat.com> 13045 CC: Alex Deucher <alexdeucher@gmail.com> 13046 CC: Dave Airlie <airlied@redhat.com> 13047 Signed-off-by: Matt Turner <mattst88@gmail.com> 13048 13049commit c0a5c9403dff254e1669df606a4193794270aeff 13050Author: Dave Airlie <airlied@redhat.com> 13051Date: Fri Feb 26 19:17:28 2010 +1000 13052 13053 pci: add HP vendor id 13054 13055commit 3970cc5563900f08ba5236b23f765c926616c345 13056Author: Cooper Yuan <cooperyuan@gmail.com> 13057Date: Fri Feb 26 16:14:12 2010 +0800 13058 13059 add more triple-head server ID 13060 13061commit b499eee7506f7478649e0000e9f8fcd00ac9d88d 13062Author: James Le Cuirot <chewi@aura-online.co.uk> 13063Date: Thu Feb 25 21:13:18 2010 -0500 13064 13065 clean up kms zaphod handling 13066 13067 Taken from James patch on bug 24523 13068 13069 agd5f: adapt to the current code 13070 13071commit 433c8617341f5768255826435a2b09afba684f02 13072Author: Alex Deucher <alexdeucher@gmail.com> 13073Date: Thu Feb 25 21:03:13 2010 -0500 13074 13075 update ZaphodHeads option 13076 13077 Now that the screen section mapping is fixed, 13078 Make the option per-instance and allow multiple 13079 outputs to be specified; e.g., DVI-0 and HDMI-0 13080 associated with instance 0 and LVDS and VGA-0 13081 associated with instance 1. 13082 13083commit f3e20b01c77f02bdf3483ceee27a1f00d240c33d 13084Author: James Le Cuirot <chewi@aura-online.co.uk> 13085Date: Thu Feb 25 19:10:19 2010 -0500 13086 13087 zaphod: Fix entity instance assignment. 13088 13089 Each screen needs to have a pointer back to its 13090 device's entity instance, Currently Screen1 is 13091 pointing to Screen0's instance so when it comes 13092 to match up the screen sections with the detected 13093 screens, Screen0 gets chosen in both cases. 13094 13095 See bug 24523 13096 13097commit 1b7e9a2e50f77819b3aff4e37ba39eaec69ff932 13098Author: Alex Deucher <alexdeucher@gmail.com> 13099Date: Tue Feb 23 11:39:55 2010 -0500 13100 13101 radeon: fixes for zaphodheads option 13102 13103 Needed for systems with more than two outputs. 13104 Both KMS and non-KMS zaphod work on systems with 13105 more than two outputs now. 13106 13107commit bd83e5f1c75b8c00b43846c7225568917fc9a8ce 13108Author: Dave Airlie <airlied@redhat.com> 13109Date: Tue Feb 23 19:27:24 2010 +1000 13110 13111 zaphod: hopefully fix kms + zaphod 13112 13113commit 0c3468d812e3790ce03d9e76779ae81e7b7b82d5 13114Author: Alex Deucher <alexdeucher@gmail.com> 13115Date: Mon Feb 22 17:34:51 2010 -0500 13116 13117 AVIVO: set frac fb divider pll flag 13118 13119 matches previous behavior 13120 13121commit 212e152536c4325e6799018891d9aee132681f48 13122Author: Alex Deucher <alexdeucher@gmail.com> 13123Date: Mon Feb 22 17:27:24 2010 -0500 13124 13125 radeon: update new pll algo 13126 13127 - add support for pre-avivo chips 13128 - add support for fixed post/ref dividers 13129 - add support for non-fractional fb dividers 13130 13131 By default avivo chips use the new algo and 13132 pre-avivo chips use the old algo. Use the 13133 "NewPLL" option to toggle between them (set to 13134 TRUE for the new algo, FALSE for the old). 13135 13136commit e68d3a3890fc81c51f2006b5548da1e8756ad2fd 13137Author: Alex Deucher <alexdeucher@gmail.com> 13138Date: Thu Feb 18 14:02:00 2010 -0500 13139 13140 RS600: add connector quirk 13141 13142 System lists DVI port as HDMI. 13143 13144 fixes fdo bug 26605 13145 13146commit a3b730eceb522c7ac1ef3dd6f6c7d773118d03f7 13147Author: Jerome Glisse <jglisse@redhat.com> 13148Date: Thu Feb 18 14:13:48 2010 +0100 13149 13150 r6xx/kms: when reseting BO force default state emission 13151 13152 In KMS world each cs need to fully initialize the 3D engine 13153 when we were erasing cs in IBDiscard for r6xx we weren't 13154 forcing reemission of default state. This patch force this 13155 emission. Fix : 13156 13157 http://bugs.freedesktop.org/show_bug.cgi?id=26603 13158 http://bugzilla.kernel.org/show_bug.cgi?id=15284 13159 13160commit 579cdcf9b4e38c791a497b747a055fc0a07d8dd6 13161Author: Alex Deucher <alexdeucher@gmail.com> 13162Date: Wed Feb 17 12:22:48 2010 -0500 13163 13164 radeon: add ZaphodHeads option 13165 13166 Allows users that want to use zaphod mode to select 13167 which xrandr outputs are assigned to which head. E.g., 13168 13169 Option "ZaphodHeads" "LVDS,VGA-0" 13170 will assign LVDS to the first zaphod driver instance 13171 and VGA-0 to the second instance. 13172 13173commit 47136fa347d1756523239746b4c74cd5278a1118 13174Author: Michel Dänzer <daenzer@vmware.com> 13175Date: Mon Feb 15 16:44:39 2010 +0100 13176 13177 EXA: Fix order of coordinates passed to radeon_pick_best_crtc(). 13178 13179 The wrong order prevented tearing avoidance from working with EXA/DRI2. 13180 13181commit 6e95905f759b844373be62ec513c78153ed00222 13182Author: Michel Dänzer <daenzer@vmware.com> 13183Date: Mon Feb 15 17:02:16 2010 +0100 13184 13185 Allocate BOs for 'video offscreen surfaces' and 'TV-in' in VRAM. 13186 13187 I think these can only really work in VRAM, though right now there's probably 13188 nothing to prevent them from getting evicted... 13189 13190commit 78e7047c5235b09858b66dd3688d39aaa27d7589 13191Author: Pauli Nieminen <suokkos@gmail.com> 13192Date: Mon Feb 15 13:40:37 2010 +0200 13193 13194 Allocate Xv buffers to GTT. 13195 13196 KMS doesn't have acceleration for upload to vram. memcpy/memmove to VRAM 13197 directly is very slow (40M/s in benchmark) which causes visible problems 13198 to video. 13199 13200 Allocating video buffer in GTT will give good performance (350-450M/s) 13201 for memmove operation. This is nice performance boost for Xv under KMS. 13202 13203 Signed-off-by: Pauli Nieminen <suokkos@gmail.com> 13204 13205commit 3ec25e59854b6b03ad763bc374d3475a50f562d8 13206Author: Pauli Nieminen <suokkos@gmail.com> 13207Date: Mon Feb 15 13:45:28 2010 +0200 13208 13209 Use memcpy instead of memove for RADEONSwapCopy. 13210 13211 memcpy has about 25% better performance than memmove when 13212 destination is GTT (wc caching). Changing memmove to memcpy 13213 in SwapCopy will reduce CPU time spent moving data between 13214 host and GPU. 13215 13216 memcpy will specially help Xv when playing high resolution 13217 videos. 13218 13219 The swap path doesn't support overlapping memory copies either 13220 so extra safety is not worth the performance hit in commonly 13221 used code path. 13222 13223 Signed-off-by: Pauli Nieminen <suokkos@gmail.com> 13224 13225 [ Michel Dänzer: Fixed up whitespace ] 13226 13227commit 221ef11b31756deb7134801730e76c040e841f5c 13228Author: Alex Deucher <alexdeucher@gmail.com> 13229Date: Fri Feb 12 14:15:10 2010 -0500 13230 13231 r6xx/r7xx: implement EXA vline support 13232 13233commit 8ad4025affe1fb2f417e3a3031d74f83be5df253 13234Author: Alex Deucher <alexdeucher@gmail.com> 13235Date: Fri Feb 12 11:34:47 2010 -0500 13236 13237 radeon: consolidate crtc selector for vline wait 13238 13239 Use the Xv version as it takes into account the area 13240 covered by the op rather than just picking the largest 13241 crtc area. 13242 13243commit 4f9d1714a7382594b834d446bbe502663f6a2778 13244Author: Alex Deucher <alexdeucher@gmail.com> 13245Date: Fri Feb 5 12:27:51 2010 -0500 13246 13247 r500: fix relocs for Xv planar video 13248 13249 fixes fdo bug 26445 13250 13251commit b7ca1ab11ac0e4e4be5a1e7789ae8633b460a2d4 13252Author: Dave Airlie <airlied@redhat.com> 13253Date: Fri Feb 5 13:37:20 2010 +1000 13254 13255 rs400/rs480: mc idle bit is bit 2 like original radeon 13256 13257 backport from a KMS fix, the rs400/480 mc idle is bit 2 not bit 4. 13258 13259 Signed-off-by: Dave Airlie <airlied@redhat.com> 13260 13261commit 97c387e269672b146a95b45fbef5c8c18e473e36 13262Author: Gaetan Nadon <memsize@videotron.ca> 13263Date: Fri Jan 29 16:08:17 2010 -0500 13264 13265 config: remove dead LINUXDOC macro usage 13266 13267 This module does not generate LINUXDOC documentation. 13268 13269 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 13270 13271commit 05fa2dee0c63d56fd523ee1ebab479bcb9f5e89d 13272Author: Alex Deucher <alexdeucher@gmail.com> 13273Date: Thu Feb 4 13:13:58 2010 -0500 13274 13275 evergreen: minor clean up in transmitter setup 13276 13277 both link and encoder must be set for linkb 13278 13279commit 76eea5e02776f7beeb8f4491a374bd36792eda92 13280Author: Alex Deucher <alexdeucher@gmail.com> 13281Date: Thu Feb 4 10:52:14 2010 -0500 13282 13283 r600: reduce number of cache flushes 13284 13285 We don't need to flush so often. Next step 13286 would be to move the flushing to the drm and only 13287 flush after each command buffer rather than each 13288 draw. 13289 13290commit 8d63d70f7ebaf9d250f0449d3720ef47516c05df 13291Author: Alex Deucher <alexdeucher@gmail.com> 13292Date: Wed Feb 3 15:22:22 2010 -0500 13293 13294 evergreen: use external clock source for DP PHY 13295 13296 DP CRTC clock always comes from DCPLL. This frees 13297 up PPll1/2 for non-DP-mode PHYs and CRTCs 13298 13299commit ecbc26431914216a8b207e81451282ea07c8b92f 13300Author: Alex Deucher <alexdeucher@gmail.com> 13301Date: Wed Feb 3 03:52:07 2010 -0500 13302 13303 evergreen: blank/unblank DP in dpms calls 13304 13305commit 710a2fd07235349f084ec40626cc28fbae523d27 13306Author: Alex Deucher <alexdeucher@gmail.com> 13307Date: Wed Feb 3 03:00:24 2010 -0500 13308 13309 evergreen: pll fixes 13310 13311 SetPixelClock needs to be called for DP outputs with a 13312 valid PLL. 13313 13314commit 9ab2377ee23c5dc360ba135a77aa6b181a1862e6 13315Author: Alex Deucher <alexdeucher@gmail.com> 13316Date: Tue Feb 2 15:15:36 2010 -0500 13317 13318 evergreen: DP requires coherent mode 13319 13320commit 634da80056525a48ec17ffb81e2fb59b168bea9e 13321Author: Alex Deucher <alexdeucher@gmail.com> 13322Date: Tue Feb 2 15:09:57 2010 -0500 13323 13324 evergreen: fix units on frac_fb_div 13325 13326commit a887818f491f6c7315c56c4e0d0b702c4c6aa4ac 13327Author: Alex Deucher <alexdeucher@gmail.com> 13328Date: Mon Feb 1 11:01:47 2010 -0500 13329 13330 evergreen: add pci ids 13331 13332commit cb2772b69480268c059766c4f6b209ce590ede0e 13333Author: Alex Deucher <alexdeucher@gmail.com> 13334Date: Fri Jan 29 12:59:46 2010 -0500 13335 13336 evergreen: add atombios crtc/pll functions 13337 13338commit bd8e04cb7b39f38b6958273582a9b324a9f0759a 13339Author: Alex Deucher <alexdeucher@gmail.com> 13340Date: Mon Feb 1 10:07:43 2010 -0500 13341 13342 evergreen: add atom support for digital outputs 13343 13344 analog is already supported by the existing code. 13345 13346commit 8ad40d3c32ad5b91725bd37fcade6bed504df421 13347Author: Alex Deucher <alexdeucher@gmail.com> 13348Date: Tue Jan 26 15:39:44 2010 -0500 13349 13350 evergreen: add crtc set base/format support 13351 13352commit 2c9cf0a07ff9e5a4989861bc2fdfd71d841013a8 13353Author: Alex Deucher <alexdeucher@gmail.com> 13354Date: Thu Dec 10 15:29:14 2009 -0500 13355 13356 evergreen: add lut support 13357 13358commit 65246545c3dd9bfef73e8a0f200bd5909b89a167 13359Author: Alex Deucher <alexdeucher@gmail.com> 13360Date: Tue Jan 26 12:10:02 2010 -0500 13361 13362 evergreen: add hw cursor support 13363 13364commit c05cad56b69d239fa2e69905d15f4f08b9db4c55 13365Author: Alex Deucher <alexdeucher@gmail.com> 13366Date: Thu Jan 28 01:28:52 2010 -0500 13367 13368 evergreen: add base asic support 13369 13370commit ed63e1b1abe8810b5da6b4140892337eef08a9ea 13371Author: Alex Deucher <alexdeucher@gmail.com> 13372Date: Fri Jan 29 12:42:33 2010 -0500 13373 13374 evergreen: add register and utility defines 13375 13376commit 019260ec4a9f9fbf2ac63a2ca3314aa308471f7e 13377Author: Alex Deucher <alexdeucher@gmail.com> 13378Date: Thu Dec 10 14:57:39 2009 -0500 13379 13380 evergreen: add chip enums 13381 13382commit e86ae564d8b94d2cc8d3c7cceb452a197ab6387a 13383Author: Alex Deucher <alexdeucher@gmail.com> 13384Date: Mon Feb 1 12:03:23 2010 -0500 13385 13386 atom: update to the latest upstream atombios.h changes 13387 13388commit 77b13a02c70842a58e0590d0243f0ae016c5a640 13389Author: Alex Deucher <alexdeucher@gmail.com> 13390Date: Mon Feb 1 12:16:20 2010 -0500 13391 13392 r500: fix CS section size missmatch in Xv code 13393 13394commit 5c11264517d1b9808415d47e9dedc345610f0f86 13395Author: Dave Airlie <airlied@linux.ie> 13396Date: Fri Jan 29 15:10:33 2010 +1000 13397 13398 radeon: further digital encoder cleanups. 13399 13400 this just cleans up a few more bits of the digital encoder setup. 13401 13402 Signed-off-by: Dave Airlie <airlied@redhat.com> 13403 13404commit 872a11fa9efb0b76cef83089ea1bb80792b94930 13405Author: Alex Deucher <alexdeucher@gmail.com> 13406Date: Thu Jan 28 01:19:23 2010 -0500 13407 13408 radeon: clean up dig encoder handling 13409 13410 - switch the var name to dig_encoder 13411 - quiet coherent messages 13412 - clean up dig encoder selection 13413 13414 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 13415 13416commit c1a73ff51b371ecc6f57a81a33f644344b4e3f70 13417Author: Dave Airlie <airlied@linux.ie> 13418Date: Thu Jan 28 12:58:29 2010 +1000 13419 13420 radeon: add support for picking the digitial encoder 13421 13422 Testing on a W500 laptop found the LVDS + DisplayPort would conflict 13423 over the choice of digital encoder, and only one would work. 13424 13425 This patch fixes the DCE3 case where LVTMA requires the second digital 13426 encoder and lets DP/DVI use the first in most cases. 13427 13428 Signed-off-by: Dave Airlie <airlied@redhat.com> 13429 13430commit 95b56275d80dcee48a7927df124602c34cd72235 13431Author: Alex Deucher <alexdeucher@gmail.com> 13432Date: Wed Jan 27 15:48:25 2010 -0500 13433 13434 avivo: add some hotplug detect regs 13435 13436commit 5a4327f7784361933484895c9af751ccfa242d48 13437Author: Alex Deucher <alexdeucher@gmail.com> 13438Date: Tue Jan 26 16:06:00 2010 -0500 13439 13440 radeon: minor pll updates 13441 13442 add new fixed post divider option 13443 13444commit cbd642af7c76469d701471daea0c8d167567ccf8 13445Author: Alex Deucher <alexdeucher@gmail.com> 13446Date: Tue Jan 26 12:03:45 2010 -0500 13447 13448 radeon: use mmio bar size rather than hardcoded number for register ops 13449 13450 newer asics have larger mmio bars 13451 13452commit e5933fd74923a07aa401bdb4ecd8d1d641148440 13453Author: Alex Deucher <alexdeucher@gmail.com> 13454Date: Mon Jan 25 13:18:47 2010 -0500 13455 13456 KMS: fix no accel option 13457 13458 NO_ACCEL wasn't handled correctly with KMS. 13459 13460 fixes fdo bug 26198 13461 13462commit ba99a82ced1a6101365876723d22bcfec4607537 13463Author: Dave Airlie <airlied@redhat.com> 13464Date: Mon Jan 25 13:50:13 2010 +1000 13465 13466 radeon/kms: only enable EXA pixmaps under kms if we support render accel 13467 13468 I've seen RN50s with 64MB of RAM that are slow as molasses, this 13469 should fix them. 13470 13471 RH bug 556400 13472 13473 Signed-off-by: Dave Airlie <airlied@redhat.com> 13474 13475commit de40d36264364434541377c7f65f3836dc514fc0 13476Author: Kusanagi Kouichi <slash@ac.auone-net.jp> 13477Date: Fri Jan 22 17:18:51 2010 +0900 13478 13479 Remove unnecessary checks. 13480 13481 Signed-off-by: Kusanagi Kouichi <slash@ac.auone-net.jp> 13482 13483commit 30a19b75cc82b8e04c45e6684b84f9a4ccc0505b 13484Author: Alan Coopersmith <alan.coopersmith@sun.com> 13485Date: Tue Jan 19 14:54:04 2010 -0800 13486 13487 Fix configure with --disable-dri 13488 13489 AM_CONDITIONAL calls can't be made inside if statements, since 13490 automake won't know what to do when the if isn't true. 13491 13492 Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com> 13493 13494commit bd011784c0609f1fd03dcb784744c8f5cdef33da 13495Author: Alex Deucher <alexdeucher@gmail.com> 13496Date: Tue Jan 19 11:59:38 2010 -0500 13497 13498 ATOM: Upstream parser updates 13499 13500commit 95d63e408cc88b6934bec84a0b1ef94dfe8bee7b 13501Author: Jerome Glisse <jglisse@redhat.com> 13502Date: Mon Jan 18 10:46:50 2010 +0100 13503 13504 r6xx/r7xx: emit relocation for FRAG & TILE buffer 13505 13506 FRAG & TILE buffer are unused but still they need 13507 to be associated with a valid relocation so that 13508 userspace can't try to abuse them to overwritte 13509 GART and then try to write anywhere in system 13510 memory. 13511 13512commit bbaf71fb46e3e0acdadcb91e9b1c73396bd9aa8e 13513Author: Alan Coopersmith <alan.coopersmith@sun.com> 13514Date: Fri Jan 15 15:09:27 2010 -0800 13515 13516 Update Sun license notices to current X.Org standard form 13517 13518 Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com> 13519 13520commit 3d158716a60d5a113a541ea2f680b81a1be41ad2 13521Author: Dave Airlie <airlied@redhat.com> 13522Date: Thu Jan 14 11:21:33 2010 +1000 13523 13524 displayport: fix DDC on DVI->DP convertors. 13525 13526 Seems we have to not do auxch DDC if we aren't talking to a DP 13527 sink. 13528 13529commit 6a363f68415d37c302151581f2a86855dba39b67 13530Author: Alex Deucher <alexdeucher@gmail.com> 13531Date: Mon Jan 11 15:13:45 2010 -0500 13532 13533 rv100: fix typo in fix for bug 25992 13534 13535 Noticed by Maarten Maathuis. 13536 13537commit 7b01e1ee29f681bf1735ecded6445d12beeb52d8 13538Author: Alex Deucher <alexdeucher@gmail.com> 13539Date: Mon Jan 11 12:40:06 2010 -0500 13540 13541 rv100: reject modes >135 Mhz with DVI 13542 13543 Due to heat issues. fixes bug 25992 13544 13545 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 13546 13547commit 74da9a66a420a40bb5d562f07e97ce6ac2eb010f 13548Author: Alex Deucher <alexdeucher@gmail.com> 13549Date: Fri Jan 8 16:51:43 2010 -0500 13550 13551 radeon/kms: add new strings for eDP and TV 13552 13553commit 61977f61f516b54d1fd9df34d0f0892fbf5227d7 13554Author: Alex Deucher <alexdeucher@gmail.com> 13555Date: Thu Jan 7 02:15:31 2010 -0500 13556 13557 radeon: fallback to i2c for edid if aux fails 13558 13559 In case of a DP->DVI adapter for example. 13560 13561commit 958d073869404f60e56dc0cc70b3e7de85904694 13562Author: Alex Deucher <alexdeucher@gmail.com> 13563Date: Thu Jan 7 02:12:00 2010 -0500 13564 13565 radeon: add initial support for eDP (embedded DisplayPort) 13566 13567 Should fix fdo bug 25931 13568 13569commit 944ae73f7cd5f9a1622b38e06a58565812cb102f 13570Author: Alex Deucher <alexdeucher@gmail.com> 13571Date: Thu Jan 7 01:52:35 2010 -0500 13572 13573 radeon: updated ObjectID.h 13574 13575commit 48aa5064aff4b9adf768e480df2312d4375e9c40 13576Author: Alex Deucher <alexdeucher@gmail.com> 13577Date: Mon Jan 4 11:05:44 2010 -0500 13578 13579 radeon: fix LVDS power sequence on Mac cards 13580 13581 Noticed by John R. Dunning. Fix taken from radeonfb. 13582 13583 I'm not sure if this sequence would be useful on any PC 13584 laptops or not so make it mac specific for now. 13585 13586commit 36bd69affc996c92c40b7360a7fbaa1a3a46abfd 13587Author: Gaetan Nadon <memsize@videotron.ca> 13588Date: Wed Dec 16 15:52:18 2009 -0500 13589 13590 configure.ac: remove unused sdkdir=$(pkg-config...) statement 13591 13592 The sdkdir variable isn't use, so remove the statement. 13593 13594 Acked-by: Dan Nicholson <dbn.lists@gmail.com> 13595 13596 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 13597 13598commit 9d0f3af7278dc939fd4e6f3ea69d9f488a9fbed7 13599Author: Alex Deucher <alexdeucher@gmail.com> 13600Date: Tue Dec 22 08:45:27 2009 -0500 13601 13602 radeon: add cvt timing if we only have panel w/h 13603 13604 fixes mac laptops without an edid 13605 13606 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 13607 13608commit 6e1f5553c6d7e3b5d089af2e3d587efe95936855 13609Author: Alex Deucher <alexdeucher@gmail.com> 13610Date: Mon Dec 21 17:32:40 2009 -0500 13611 13612 ATOM: add new power table defs 13613 13614commit 4b05c47ac657f9a93d76221269761ed64c81f716 13615Author: Dave Airlie <airlied@redhat.com> 13616Date: Thu Dec 17 14:30:23 2009 +1000 13617 13618 radeon: use new libdrm api 13619 13620commit 1d620800bd79290967fa487decf798e318c45f25 13621Author: Ingmar Vanhassel <ingmar@exherbo.org> 13622Date: Sun Dec 20 21:04:56 2009 +0100 13623 13624 Replace shave with automake silent-rules 13625 13626 Which are enabled by default in Xorg's util-macros 1.3 if automake-1.11 13627 is used. 13628 13629 Signed-off-by: Ingmar Vanhassel <ingmar@exherbo.org> 13630 13631commit 52f6a2ff682ad1f8e26ef1cd7d380493dcd57752 13632Author: Gaetan Nadon <memsize@videotron.ca> 13633Date: Wed Dec 16 15:29:53 2009 -0500 13634 13635 src/Makefile.am: missing 3 header files in tarball 13636 13637 radeon_dummy_bufmgr.h 13638 radeon_vbo.h 13639 simple_list.h 13640 Replace leading spaces with tab for radeon_atomwrapper.h 13641 13642 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 13643 13644commit 3a30210d50b27f8772fc5045133940246764fce9 13645Author: Matthias Hopf <mhopf@suse.de> 13646Date: Tue Dec 15 10:53:48 2009 -0500 13647 13648 fix 200M freezes on VT switch if CRTC is disabled 13649 13650 It appears that RS4xx chips need to have the crtc 13651 enabled when the timing is programmed. 13652 13653 agd5f: minor fixes/cleanup of the original patch 13654 13655commit 299d395bd3f294239dee58ab7d607d7d2c657f61 13656Author: Corbin Simpson <MostAwesomeDude@gmail.com> 13657Date: Sun Dec 13 14:51:17 2009 -0800 13658 13659 dri: be more useful if disabling DRI due to KMS/UMS mismatch. 13660 13661 It's only fair. 13662 13663 Compile-tested only. 13664 13665 Signed-off-by: Dave Airlie <airlied@redhat.com> 13666 13667commit 0e5c9d87b5d7e0751df71cc8958ca5ccaed25104 13668Author: Alex Deucher <alexdeucher@gmail.com> 13669Date: Thu Dec 10 14:27:43 2009 -0500 13670 13671 ATOM: split set base/format into separate functions 13672 13673 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 13674 13675commit 46630da5fd6f45bb8ea150b870162997480d69c7 13676Author: Alex Deucher <alexdeucher@gmail.com> 13677Date: Wed Dec 9 12:55:25 2009 -0500 13678 13679 AVIVO: add new PLL code 13680 13681 This should hopefully help the problems with flickering 13682 and blinking monitors reported on some systems. If there 13683 are problems, the old PLL algorithm can be selected with: 13684 Option "NewPLL" "FALSE" 13685 in the device section of your X config. 13686 13687 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 13688 13689commit f082b1693d6f7f763ccf5a8436a89890ca2c6129 13690Author: Matthijs Kooijman <matthijs@stdin.nl> 13691Date: Thu Dec 3 12:10:14 2009 -0500 13692 13693 radeon: fix crtc2 dpms 13694 13695 noticed by Matthijs Kooijman on fdo bug 22140 13696 13697commit 3a96fbf3b84522cf1ba1b176e82b662222331c14 13698Author: Alex Deucher <alexdeucher@gmail.com> 13699Date: Wed Dec 2 18:02:33 2009 -0500 13700 13701 radeon: only read RADEON_PPLL_REF_DIV for ref div on pre-avivo 13702 13703 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 13704 13705commit f03450796d2e9247a1228c4e2abb1dfad7aecddf 13706Author: Kusanagi Kouichi <slash@ac.auone-net.jp> 13707Date: Wed Dec 2 01:36:37 2009 -0500 13708 13709 radeon: Lift hardcoded limit from RADEONQueryImageAttributes 13710 13711 The dimension of an XvImage is limited to 2048 x 2048 even if an adaptor 13712 supports larger image. 13713 XvCreateImage and XvShmCreateImage lower the width or height of an image. 13714 XvPutImage and XvShmPutImage return BadValue. 13715 The cause is that 2048 is hardcoded in RADEONQueryImageAttributes. 13716 13717commit efbc2c80ab02879edf3b7b3d65b16c45ddce5017 13718Merge: 88a50a30 a612813c 13719Author: Dave Airlie <airlied@redhat.com> 13720Date: Wed Dec 2 14:30:36 2009 +1000 13721 13722 Merge remote branch 'origin/displayport' 13723 13724commit 88a50a30df11a06263209340a42251851f8e2334 13725Author: Dave Airlie <airlied@redhat.com> 13726Date: Mon Nov 30 15:32:12 2009 +1000 13727 13728 r600: fix multi-operation in single batch support. 13729 13730 This ports the mesa DMA buffer handling with the 3 lists, 13731 13732 Signed-off-by: Dave Airlie <airlied@redhat.com> 13733 13734commit b2597deea3a3953ff50d54ff37e3c043eac409f4 13735Author: Kusanagi Kouichi <slash@ac.auone-net.jp> 13736Date: Mon Nov 30 11:10:50 2009 -0500 13737 13738 R600: fix some warnings 13739 13740commit 4d90dc3cb248e9d61c0c490bba80f6727fafd318 13741Author: Alex Deucher <alexdeucher@gmail.com> 13742Date: Mon Nov 30 11:05:38 2009 -0500 13743 13744 atom: pull misc mode info for lvds panel mode from bios tables 13745 13746 sync polarity, etc. This will likely fix LVDS problems 13747 on some laptops. 13748 13749 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 13750 13751commit a612813c5f329f00a9271e7bb69abd2630a49403 13752Author: Alex Deucher <alexdeucher@gmail.com> 13753Date: Fri Nov 27 17:20:30 2009 -0500 13754 13755 radeon: clean up DP code 13756 13757 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 13758 13759commit a8dbf7c23481501987971a9c0b6cb0760f86127f 13760Author: Dave Airlie <airlied@redhat.com> 13761Date: Fri Nov 27 10:16:00 2009 +1000 13762 13763 r600/accel: cleanup and merge a lot of the accel functions 13764 13765 All of the drawing ops were the exact same modulo the vtx size, 13766 this along with the vertex buffer wrapping code could all be consolidated 13767 into a smaller set of functions. 13768 13769 This also adds 2 VBO which we switch between, and merges a #define to 13770 enable the multiple operations in one CS under KMS mode. 13771 13772 Multi-operation still isn't working though. 13773 13774 Signed-off-by: Dave Airlie <airlied@redhat.com> 13775 13776commit 971e463b2cfdc1ba1c78a24fb439d33dd9155dfc 13777Author: Alex Deucher <alexdeucher@gmail.com> 13778Date: Thu Nov 26 12:45:15 2009 -0500 13779 13780 dce3: CV1/TV1OutputControl tables didn't go away until dce3.2 13781 13782 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 13783 13784commit 95385f0906f371a1ed6e60eb3e597e699c7a3222 13785Author: Jerome Glisse <jglisse@redhat.com> 13786Date: Thu Nov 26 12:21:21 2009 +0100 13787 13788 kms: Fix resizing when acceleration is disabled 13789 13790 When acceleration is disabled we need to reallocate 13791 a new shadow framebuffer and we should also avoid 13792 calling any EXA function as EXA is disabled in such 13793 case. 13794 13795commit 8195385d3cedf70d89ca576371af6bca5045bbd2 13796Author: Dave Airlie <airlied@redhat.com> 13797Date: Thu Nov 26 15:43:54 2009 +1000 13798 13799 dce3: no CV1 or TV1 Output Control tables anymore. 13800 13801 According to the DCE3 docs we should only use DAC1/2 not 13802 CV1/TV OutputControl, also my rv730 bios doesn't have any 13803 CV1/TV tables. 13804 13805 Signed-off-by: Dave Airlie <airlied@redhat.com> 13806 13807commit 381b14d5422131ea8ff942ab06a328f6beca0d3f 13808Author: Dave Airlie <airlied@redhat.com> 13809Date: Thu Nov 26 15:28:15 2009 +1000 13810 13811 atombios: CV mode retrival was broken. 13812 13813 a) crev and frev are reversed 13814 b) my rv730 bios only has one mode in it, so 13815 bounds check the table size. 13816 13817commit a67a13c315a9235be052662e2a3270686cccc0f0 13818Author: Dave Airlie <airlied@redhat.com> 13819Date: Thu Nov 26 13:41:36 2009 +1000 13820 13821 radeon: fix typo in unused as of yet CV code 13822 13823commit cd9bfb5892eebef6dff14ac414dff074db5ddee4 13824Author: Dave Airlie <airlied@redhat.com> 13825Date: Thu Nov 26 06:43:13 2009 +1000 13826 13827 fixes displayport for me here 13828 13829commit 256de862a3c017dcb4ed0d8689294a552eda0d18 13830Author: Dave Airlie <airlied@redhat.com> 13831Date: Thu Nov 26 06:35:39 2009 +1000 13832 13833 displayport: cleanup call dp from dpms on instead. 13834 13835commit 05551295c5e0946745163f17e5c1d3d41b94bcbf 13836Author: Kristian Høgsberg <krh@bitplanet.net> 13837Date: Tue Nov 24 21:09:03 2009 -0500 13838 13839 dri2: Use drmGetDeviceNameFromFd() instead of open coded loop. 13840 13841commit 0061c4db1d3aecdca13efb5133ab8784dc37df95 13842Author: Dave Airlie <airlied@redhat.com> 13843Date: Wed Nov 25 15:39:10 2009 +1000 13844 13845 Revert "r600: enable multiple operations in one CS" 13846 13847 turn this off for now, on my rv635 desktop, I started getting blanks 13848 in places in firefox and the odd bit of font corruption, need to 13849 track that down. 13850 13851 This reverts commit 985a065518b1d33599de33f7fe082d3302db58a6. 13852 13853commit 8b28534bcf877557a5681fa3b4f107c74615d3de 13854Author: Dave Airlie <airlied@redhat.com> 13855Date: Wed Nov 25 13:24:44 2009 +1000 13856 13857 radeon/exa: change option to determine exa pixmap usage. 13858 13859 This moves to a boolean instead of using VRAM sizing. 13860 13861 as per Michel's suggestions on list. 13862 13863 Signed-off-by: Dave Airlie <airlied@redhat.com> 13864 13865commit f0acb16fa19844453adc6db3399977fba7c8a0db 13866Author: Dave Airlie <airlied@redhat.com> 13867Date: Wed Nov 25 12:04:47 2009 +1000 13868 13869 r600/xv: drop inited 3d false in xv code 13870 13871commit 985a065518b1d33599de33f7fe082d3302db58a6 13872Author: Dave Airlie <airlied@redhat.com> 13873Date: Wed Nov 25 11:54:08 2009 +1000 13874 13875 r600: enable multiple operations in one CS 13876 13877 This switches on multiple ops in a single CS under KMS/DRI2. 13878 13879 It gets for on a Pentium D 3 + rv740 from 330,000 to 500,000 13880 with x11perf -aa10text. 13881 13882 It also knocks a couple of seconds of gtkperf -a 13883 13884 Signed-off-by: Dave Airlie <airlied@redhat.com> 13885 13886commit af816ac752820255f245793b53a7cca5a4a49cd4 13887Author: Dave Airlie <airlied@redhat.com> 13888Date: Wed Nov 25 11:53:07 2009 +1000 13889 13890 r600: fixup problems with EXA operation reset for multiple ops 13891 13892 To put multiple ops into one CS, you can't just discard the whole 13893 IB. This add supports for reset the CS cdw to the correct place 13894 after an op discards. 13895 13896 Still doesn't enable the final accel bits. 13897 13898commit 3d8dcbc29323a3c644100bec13aa93f024653bd3 13899Author: Dave Airlie <airlied@redhat.com> 13900Date: Wed Nov 25 11:43:57 2009 +1000 13901 13902 kms: flush cs on close screen. 13903 13904 this is needed for server recycle. 13905 13906 Signed-off-by: Dave Airlie <airlied@redhat.com> 13907 13908commit 19f1a357944f9d8f4567a2691a68067ec033ccb7 13909Author: Dave Airlie <airlied@redhat.com> 13910Date: Wed Nov 25 11:41:52 2009 +1000 13911 13912 radeon: fix check for no work in operation 13913 13914commit 3a460a14b9603159f10d89da27b559c36a184e27 13915Author: Dave Airlie <airlied@redhat.com> 13916Date: Wed Nov 25 10:33:17 2009 +1000 13917 13918 r600: refactor code to help future acceleration speedups. 13919 13920 This changes the vertex buffer index to be an offset, and 13921 records the start of the vb for each operation and uses 13922 that to set the operations up. 13923 13924 This still flushes after each operation to make sure we have 13925 no regressions in non-kms/kms cases. 13926 13927 Signed-off-by: Dave Airlie <airlied@redhat.com> 13928 13929commit 797a3f0c71c94477eec565ea2c95553c6f66d9fd 13930Author: Dave Airlie <airlied@redhat.com> 13931Date: Wed Nov 25 10:07:59 2009 +1000 13932 13933 kms: allow prepare copy to fail without dying 13934 13935commit ba76acc76e0d0a51a1c488e4d5494a71e155a05b 13936Author: Dave Airlie <airlied@redhat.com> 13937Date: Tue Nov 24 15:38:47 2009 +1000 13938 13939 radeon/kms: drop special r600 indirect flush 13940 13941 This just merges it with the main kms cs flush 13942 13943commit 3c527d752d3cced219c6113acb932511b16e35fb 13944Author: Dave Airlie <airlied@redhat.com> 13945Date: Tue Nov 24 15:11:12 2009 +1000 13946 13947 r600: enable mixed pixmaps 13948 13949commit b9eb2380df5145c85b2ef694a04628c055c47429 13950Author: Alex Deucher <alexdeucher@gmail.com> 13951Date: Mon Nov 23 12:12:02 2009 -0500 13952 13953 DCE3.2: fix uniphy2 dvi issues 13954 13955 In some cases the atom transmitter table sets the 13956 golden value of this reg differently which some monitors 13957 don't like. I haven't had time to dig further, so this 13958 works around it for now. 13959 13960 Fixes fdo bug 24313 13961 13962 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 13963 13964commit d0dd5122d8d75b7c4a72df07dc7f562b2e75dd5e 13965Author: Gaetan Nadon <memsize@videotron.ca> 13966Date: Mon Nov 23 09:25:05 2009 -0500 13967 13968 Makefile.am: add ChangeLog and INSTALL on MAINTAINERCLEANFILES 13969 13970 Now that the INSTALL file is generated. 13971 Allows running make maintainer-clean. 13972 13973commit 3d7888aa0493f46d826f4d80dea36f1f642f3c6b 13974Author: Dave Airlie <airlied@redhat.com> 13975Date: Mon Nov 23 15:26:15 2009 +1000 13976 13977 radeon: fix warning in printf 13978 13979commit 3a1a8b7b73424a769baf917bba2223ea6515b596 13980Author: Dave Airlie <airlied@redhat.com> 13981Date: Mon Nov 23 15:07:10 2009 +1000 13982 13983 radeon: surround mixed pixmaps usage with ifdef check for old servers 13984 13985commit e20af9c94982ec6487dae42c141d33cf6a7a2eb4 13986Author: Dave Airlie <airlied@redhat.com> 13987Date: Fri Nov 20 12:15:02 2009 +1000 13988 13989 radeon: avoid using hw pixmaps when we have little VRAM. 13990 13991 This patch returns NULL for pixmap creation when we are using 13992 mixed pixmaps and the pixmap has a size. 13993 13994 The size check is necessary for the front buffer. 13995 13996 We add a flag to force pixmap creation for certain pixmaps 13997 that need to be hw, like the DRI2 and Xv ones. 13998 13999 Idea from Michel and workarounds from Ben Skeggs. 14000 14001 v2: add Option "EXALowVRAM" to allow configuring this, value in MBs. 14002 14003 Signed-off-by: Dave Airlie <airlied@redhat.com> 14004 14005commit 543338842ba7fe7da4902947d386ea1cabd1cdf0 14006Author: Alex Deucher <alexdeucher@gmail.com> 14007Date: Fri Nov 20 16:38:43 2009 -0500 14008 14009 dp fix timing 14010 14011 aux channel atom command table delay is 10 usec units 14012 this gets the link training working, but alas, no image. 14013 14014commit dd3eab848cf352bb96c3d01fe6028d8a4a8e451e 14015Author: Alex Deucher <alexdeucher@gmail.com> 14016Date: Fri Nov 20 01:06:49 2009 -0500 14017 14018 r300: render target limit is 2560 14019 14020 limit was increased to 4021 in r4xx. 14021 14022 fixes fdo bug 25191. 14023 14024 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14025 14026commit df35d40d4ec260cdab57a9b36e4e96b2a6c2a827 14027Author: Alex Deucher <alexdeucher@gmail.com> 14028Date: Thu Nov 19 20:01:42 2009 -0500 14029 14030 DP: hack that makes DP work. 14031 14032 Something's wrong with the status check. 14033 14034commit f42412ec4dbd21f562a7fe4239c0a1f17ef1b8f2 14035Author: Alex Deucher <alexdeucher@gmail.com> 14036Date: Thu Nov 19 19:40:54 2009 -0500 14037 14038 wip 14039 14040commit eb9bc133fc426e67b397e661bfd22bf62009d9d3 14041Author: Dave Airlie <airlied@redhat.com> 14042Date: Fri Nov 20 09:23:31 2009 +1000 14043 14044 kms: recalculate the flush limits after screen resize. 14045 14046 When we resize the front buffer we need to reduce the flush limits 14047 appropriately. 14048 14049 Signed-off-by: Dave Airlie <airlied@redhat.com> 14050 14051commit f7f58ef4c042e492618665a6c5555e8e67387ab3 14052Author: Dave Airlie <airlied@redhat.com> 14053Date: Fri Nov 20 09:22:39 2009 +1000 14054 14055 radeon: r100/r200 have a 2047 scissor limit 14056 14057 We were overflowing this in my case with a 2704 width desktop, 14058 so videos were stopping around 700 bytes across the screen. 14059 14060 can I haz shatter already? 14061 14062 Signed-off-by: Dave Airlie <airlied@redhat.com> 14063 14064commit 54c09778d79de1516ee511d6d7d5b54dd4bf9130 14065Author: Alex Deucher <alexdeucher@gmail.com> 14066Date: Thu Nov 19 13:32:30 2009 -0500 14067 14068 DP wip 14069 14070 - use coherent 14071 - fix up lane/clock numbers 14072 14073commit 02d017a0552db2c8a48506c5b26c48a1a2ab75b6 14074Author: Gaetan Nadon <memsize@videotron.ca> 14075Date: Wed Oct 28 14:41:41 2009 -0400 14076 14077 INSTALL, NEWS, README or AUTHORS files are missing/incorrect #24206 14078 14079 Automake 'foreign' option is specified in configure.ac. 14080 Remove from Makefile.am 14081 14082commit 396ee9b7e7e3b1dd64532270c4e32cdb33857596 14083Author: Gaetan Nadon <memsize@videotron.ca> 14084Date: Wed Oct 28 14:09:09 2009 -0400 14085 14086 INSTALL, NEWS, README or AUTHORS files are missing/incorrect #24206 14087 14088 Add missing INSTALL file. Use standard GNU file on building tarball 14089 README may have been updated 14090 Remove AUTHORS file as it is empty and no content available yet. 14091 Remove NEWS file as it is empty and no content available yet. 14092 14093commit 1dcda0c4d5d506e18a7989ff05f2ad8569a73f23 14094Author: Gaetan Nadon <memsize@videotron.ca> 14095Date: Mon Oct 26 12:54:21 2009 -0400 14096 14097 Several driver modules do not have a ChangeLog target in Makefile.am #23814 14098 14099 The git generated ChangeLog replaces the hand written one. 14100 Update configure.ac to xorg-macros level 1.3. 14101 Use XORG_DEFAULT_OPTIONS which replaces four XORG_* macros 14102 Update Makefile.am to add ChangeLog target if missing 14103 Remove ChangeLog from EXTRA_DIST or *CLEAN variables 14104 This is a pre-req for the INSTALL_CMD 14105 14106commit 5b517bcc353e7111d2365a213505f27cfae30924 14107Author: Gaetan Nadon <memsize@videotron.ca> 14108Date: Thu Oct 22 12:34:16 2009 -0400 14109 14110 .gitignore: use common defaults with custom section # 24239 14111 14112 Using common defaults will reduce errors and maintenance. 14113 Only the very small or inexistent custom section need periodic maintenance 14114 when the structure of the component changes. Do not edit defaults. 14115 14116commit 605c869076985b69350c893cf2bbcc90de8dc6bf 14117Author: Alex Deucher <alexdeucher@gmail.com> 14118Date: Wed Nov 18 01:51:12 2009 -0500 14119 14120 more dp cleanup 14121 14122 - clean up handling of linkb, etc. 14123 - add support for ATOM_TRANSMITTER_ACTION_SETUP_VSEMPH 14124 to existing transmitter control function 14125 - fix lots of warnings 14126 14127commit bce1911f66f446b793125ac8cd5463bf6292629e 14128Author: Alex Deucher <alexdeucher@gmail.com> 14129Date: Wed Nov 18 00:04:33 2009 -0500 14130 14131 radeon: pull previous displayport work into master 14132 14133commit 437113124bbd6fb166825169eabec4dfde900dd9 14134Author: Alex Deucher <alexdeucher@gmail.com> 14135Date: Tue Nov 17 14:16:22 2009 -0500 14136 14137 radeon: deal with connectors sourced to the same encoder 14138 14139 Some systems have multiple connectors connected to the 14140 same encoder; e.g., DVI and HDMI connected to the same 14141 encoder with the same ddc line. Since we expose 14142 connectors as xrandr outputs, randr treats them separately 14143 which results in it trying to source the same encoder to 14144 different crtcs. If we have an HDMI and DVI-D port on the 14145 same encoder, pick the one to be considered connected 14146 based on the edid (HDMI if edid indicates HDMI, DVI 14147 otherwise). Also, don't turn off (dpms) encoders that are 14148 shared. 14149 14150 Fixes fdo bug 21767. 14151 14152 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14153 14154commit 9b8605d4e2bf7703c7848f3e2022a8c7328a3acd 14155Author: Alex Deucher <alexdeucher@gmail.com> 14156Date: Tue Nov 17 11:45:09 2009 -0500 14157 14158 radeon: fix potential memory leak in ddc code 14159 14160 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14161 14162commit ce8299962003de572122561a6f7f61eaccf633b2 14163Author: Dave Airlie <airlied@redhat.com> 14164Date: Tue Nov 17 15:23:30 2009 +1000 14165 14166 kms: check for 0 mask for cloned outputs. 14167 14168 Signed-off-by: Dave Airlie <airlied@redhat.com> 14169 14170commit 06d68f7fa3870f9601e837e1834d33e8395008b1 14171Author: Dave Airlie <airlied@redhat.com> 14172Date: Tue Nov 17 15:10:04 2009 +1000 14173 14174 kms: add cloning support. 14175 14176 We have to translate from the kernel encoder cloning to the randr 14177 connector cloning, we do this by constructing an encoder mask per output 14178 and an encoder cloning mask per output then comparing them. 14179 14180 Signed-off-by: Dave Airlie <airlied@redhat.com> 14181 14182commit 7587ce16ab0380337fe801f457c1d1d9b4141cc5 14183Author: Alex Deucher <alexdeucher@gmail.com> 14184Date: Mon Nov 16 19:16:03 2009 -0500 14185 14186 r600: remove un-needed format conversions 14187 14188 we do the normalization and xforms in the vertex 14189 shader. 14190 14191 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14192 14193commit ba605b5c9ff48765f2b66960aa4cc81b6235d012 14194Author: Alex Deucher <alexdeucher@gmail.com> 14195Date: Mon Nov 16 19:00:52 2009 -0500 14196 14197 r600: fix num format in vtx fetch 14198 14199 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14200 14201commit 0c4710c67a2fee2061fc3da43c9f908585693cfa 14202Author: Alex Deucher <alexdeucher@gmail.com> 14203Date: Thu Nov 12 11:11:17 2009 -0500 14204 14205 radeon: man page updates 14206 14207commit 2af2744c80f3307e6d84edba9479be5bd36e1d03 14208Author: Alex Deucher <alexdeucher@gmail.com> 14209Date: Wed Nov 11 13:12:24 2009 -0500 14210 14211 radeon: fix support for external tmds on some legacy chips 14212 14213 Should fix bug 11801. 14214 14215 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14216 14217commit f17657fdc83d8f4c0386d2c7dade98de5b94acbe 14218Author: Alex Deucher <alexdeucher@gmail.com> 14219Date: Mon Nov 9 16:59:11 2009 -0500 14220 14221 r600/r700: typo, fix mask of DB_ALPHA_TO_MASK 14222 14223 noticed by glisse. 14224 14225commit 0ee7763fb33bf7fbc4ae1e8372cbc74578bfb720 14226Author: Alex Deucher <alexdeucher@gmail.com> 14227Date: Mon Nov 9 15:09:34 2009 -0500 14228 14229 IGP: some IGP chips report as AGP 14230 14231 Set bus type appropriately. fixes bug 25002 14232 14233 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14234 14235commit da1fcddaade58e371c7bee555ce32944c689f810 14236Author: Alex Deucher <alexdeucher@gmail.com> 14237Date: Thu Nov 5 14:13:49 2009 -0500 14238 14239 radeon: properly fix unused label warnings 14240 14241commit 6e496f46fdc90d0a6cef41e9d45a6458aaed1eca 14242Author: Alex Deucher <alexdeucher@gmail.com> 14243Date: Thu Nov 5 14:11:55 2009 -0500 14244 14245 Revert "Fix warning about unused defined labels in radeon_exa_funcs.c." 14246 14247 This reverts commit b6ac42cce2156880ac18b32e15acbff40b2f53e9. 14248 14249 This patch is very broken. 14250 14251commit b6ac42cce2156880ac18b32e15acbff40b2f53e9 14252Author: Tomáš Chvátal <scarabeus@gentoo.org> 14253Date: Sat Oct 31 23:07:07 2009 +0100 14254 14255 Fix warning about unused defined labels in radeon_exa_funcs.c. 14256 14257 Signed-off-by: Tomáš Chvátal <scarabeus@gentoo.org> 14258 14259commit 87d7235790866f9c19ef08972d5237f09d940cd9 14260Author: Alex Deucher <alexdeucher@gmail.com> 14261Date: Thu Nov 5 10:23:03 2009 -0500 14262 14263 EXA: fallback if no pMaskPicture->pDrawable 14264 14265 A solid or gradient mask could be used for blending 14266 the source picture onto the destination picture. 14267 14268 Fixes fdo bug 24838 14269 14270 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14271 14272commit 873897c54429e230a23011a9dd2f1069f4eb7752 14273Author: Alex Deucher <alexdeucher@gmail.com> 14274Date: Thu Nov 5 00:51:11 2009 -0500 14275 14276 DCE3+: call transmitter init on mode set 14277 14278 Generally this is done at post, but might not always 14279 be done with softboot or for connectors on docking 14280 stations. 14281 14282 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14283 14284commit 3028374488cc0f34942ac372c8d05cf15898a613 14285Author: Alex Deucher <alexdeucher@gmail.com> 14286Date: Wed Nov 4 18:48:03 2009 -0500 14287 14288 ATOM/DCE3+: fix up usPixelClock calculation for Transmitter tables 14289 14290 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14291 14292commit 52973e8f2c4104de60d00c0e5c29a31f77bf3de7 14293Author: Alex Deucher <alexdeucher@gmail.com> 14294Date: Sun Nov 1 13:58:03 2009 -0500 14295 14296 radeon/atom: fix tv-out on r5xx 14297 14298 fixes bug 24796 14299 14300 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14301 14302commit 19b4f5f9b647cd470616b8ec1d34f1e4afa357ad 14303Author: Alex Deucher <alexdeucher@gmail.com> 14304Date: Tue Oct 27 11:55:20 2009 -0400 14305 14306 ATOM: fix up DVO for DCE 3.x 14307 14308 AdjustDisplayPll takes care of this for us. 14309 14310commit 5a0019126a57138ee506d9a66738c9e8b75cbb96 14311Author: Alex Deucher <alexdeucher@gmail.com> 14312Date: Tue Oct 27 11:39:06 2009 -0400 14313 14314 radeon: fix DVO on AVIVO chips 14315 14316 DVO generally requires 2x ppll on AVIVO cards due to the 14317 way the tmds chip is wired up. Because of this, disable 14318 cloning of DVO outputs. 14319 14320 fixes fdo bug 21857. 14321 14322 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14323 14324commit 2a6eeec72d6bb04a32225883f431c1d2e8cff123 14325Author: Alex Deucher <alexdeucher@gmail.com> 14326Date: Tue Oct 27 11:18:44 2009 -0400 14327 14328 atom: loosen pll min output limits 14329 14330 Limiting the pll output range is a good thing generally as 14331 it limits the number of possible pll combinations for a given 14332 frequency presumably to the ones that work best on each card. 14333 That's why the limits are in the bios tables. However, certain 14334 duallink DVI monitors seem to like pll combinations that would 14335 be limited by this at least on pre-DCE 3.0 r6xx hardware. This 14336 might need to be adjusted per family or per clock range in the 14337 future. 14338 14339 Fixes fdo bug 24727. 14340 14341 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14342 14343commit f0d9d80fee4176eaba9435a9539f29d0eefe2a87 14344Author: Luca Tettamanti <kronos.it@gmail.com> 14345Date: Sat Oct 24 16:45:23 2009 -0400 14346 14347 Disable color tiling on r600/r700 14348 14349 Not supported yet. 14350 14351commit ac499c3c0ea92f83ebd8127bb8cb5de625de6c94 14352Author: Alex Deucher <alexdeucher@gmail.com> 14353Date: Fri Oct 23 10:48:55 2009 -0400 14354 14355 radeon: clamp the internal FB map to the aperture size v2 14356 14357 We don't use the invisible memory yet and on cards with 14358 large amounts of vram this can cause the top of GART 14359 calculation to overflow. 14360 14361 Fixes bug fdo bug 24301: 14362 http://bugs.freedesktop.org/show_bug.cgi?id=24301 14363 14364 v2: only clamp cards with more than 512 MB. This seems 14365 to cause problems on some older cards due to the way the 14366 drm and ddx set up the internal memory map. 14367 14368commit e57b54daf318ff59315a1c9ed21934f288c5c782 14369Author: Alex Deucher <alexdeucher@gmail.com> 14370Date: Thu Oct 22 16:07:58 2009 -0400 14371 14372 R7xx: program additional CUR/GRPH regs for 40 bit addresses 14373 14374 The *_HIGH regs are reversed. The secondary ones are in the 14375 primary block and vice versa. 14376 14377commit 4cf06dfba617529291ce4b4c306c4fc1bba110ee 14378Author: Alex Deucher <alexdeucher@gmail.com> 14379Date: Thu Oct 22 02:08:45 2009 -0400 14380 14381 ATOM: use indirect addressing for regs >= 0x10000 14382 14383commit 66b194a78c470cb3978f310828dd96c3f3e96944 14384Author: Alex Deucher <alexdeucher@gmail.com> 14385Date: Wed Oct 21 15:58:18 2009 -0400 14386 14387 ATOM: fix up get clock info 14388 14389 Newer revisions of the atom firmware table have 14390 changed. This helps select better pll dividers 14391 in some cases. Noticed by Mathias Froehlich. 14392 14393commit bd89b7501f294ac645390ef144df569953c81dc4 14394Author: Alex Deucher <alexdeucher@gmail.com> 14395Date: Sun Oct 18 16:21:36 2009 -0400 14396 14397 radeon: fixup CustomEDID option 14398 14399 - add support for analog option to force edid to analog 14400 mode 14401 - fix compilation on older xservers 14402 - updated man page 14403 14404commit 255ac8d54e3dc12e579e05edb41b40946a00eab8 14405Author: Alex Deucher <alexdeucher@gmail.com> 14406Date: Tue Oct 13 12:27:52 2009 -0400 14407 14408 radeon: add XV_CRTC attribute for textured video 14409 14410 This attribute allows the user to override which 14411 crtc is synced with when XV_VSYNC is enabled. This 14412 is useful for clone modes where the user can might want 14413 to override the default. 14414 14415commit 801d900add4731f3c424ebb780ad2dbd7a1c6a25 14416Author: Alex Deucher <alexdeucher@gmail.com> 14417Date: Mon Oct 12 10:00:57 2009 -0400 14418 14419 ATOM: minor cleanups to pll setup 14420 14421commit f4407962cd7b272e0860319f11f6a6583ef226c2 14422Author: Csillag Kristof <csillag.kristof@united-consult.hu> 14423Date: Fri Oct 9 18:00:09 2009 -0400 14424 14425 Fix some issues with CustomEDID support 14426 14427 - Fixes a memory allocation problem introduced with the previous 14428 version of the CustomEDID patch (now in GIT), and 14429 - Enhances the syntax of the CustomEDID option so that I can force 14430 digital output, even if the EDID was acquired over an analog link. 14431 14432commit de55995e82c3875f70b6394fff440d695d062113 14433Author: Alex Deucher <alexdeucher@gmail.com> 14434Date: Fri Oct 9 11:07:30 2009 -0400 14435 14436 r600 EXA: fix up mask reg mixup 14437 14438commit a168caf5349b757873ad89d977772feaa7fb6e58 14439Author: Alex Deucher <alexdeucher@gmail.com> 14440Date: Fri Oct 9 10:32:26 2009 -0400 14441 14442 r600 EXA: fix some compile warnings 14443 14444commit fb8dda2928bd0def2b34c8c8e414129e650d6eb5 14445Author: Alex Deucher <alexdeucher@gmail.com> 14446Date: Fri Oct 9 10:30:21 2009 -0400 14447 14448 r600 EXA: properly check planemask 14449 14450 We only support per-component masks. 14451 14452commit 1b25bad2d8ebe481ffb4b0fd3a85ab636bd4fec2 14453Author: Dave Airlie <airlied@linux.ie> 14454Date: Fri Oct 9 20:09:52 2009 +1000 14455 14456 Revert "radeon: clamp the internal FB map to the aperture size" 14457 14458 This reverts commit 5f846360c46f5a989f5d0fde6d251cdbd61d4968. 14459 14460 Numerous reports of system hangs since this, I'm guessing 14461 some sort of conflict with the drm memory setup. This code 14462 has always been fragile between kernel/userspace drivers. 14463 14464commit a956d478043b44c49aed39c7aa3d576ef32cadc5 14465Author: Dave Airlie <airlied@linux.ie> 14466Date: Fri Oct 9 20:08:31 2009 +1000 14467 14468 Revert "kms: setup colormap and gamma correctly." 14469 14470 This reverts commit 60d9685abddccec17c1a9a5ec48cbe9c92543e0f. 14471 14472 which breaks colormap on 1.6 X server 14473 14474 Conflicts: 14475 14476 src/drmmode_display.c 14477 14478commit a493feb94c0d84aaf5bfd84c6940d7f03974dd86 14479Author: Dave Airlie <airlied@redhat.com> 14480Date: Fri Oct 9 15:05:25 2009 +1000 14481 14482 kms: don't use scratch pixmaps at all 14483 14484 use our own wrapper to allocate a pixmap for wrapping a bo in. 14485 14486commit 0caffbd6fd6b356b9b7d715c015285eaa91e3e36 14487Author: Dave Airlie <airlied@redhat.com> 14488Date: Fri Oct 9 15:04:56 2009 +1000 14489 14490 drmmode: reload cursors on modeset 14491 14492commit 02e12ae6be7bc1976f63848fa2854d320d5ab36e 14493Author: Alex Deucher <alexdeucher@gmail.com> 14494Date: Wed Oct 7 19:56:58 2009 -0400 14495 14496 ATOM: rework crtc modeset 14497 14498 - clean up tv timing handling 14499 - unify SetCRTCTiming and SetCRTCDTDTiming interfaces 14500 14501commit d499eeaf22f77a1294e99aa38a50aa6810bb684a 14502Author: Alex Deucher <alexdeucher@gmail.com> 14503Date: Wed Oct 7 17:43:11 2009 -0400 14504 14505 ATOM: reorder crtc dpms based on bios recommendations 14506 14507commit 4ccd2a21f9c50515b4246f35454f76ef49006c76 14508Author: Alex Deucher <alexdeucher@gmail.com> 14509Date: Wed Oct 7 16:30:03 2009 -0400 14510 14511 ATOM: add support for AdjustDisplayPll table 14512 14513 Depending on the output and clock, this table will 14514 adjust the pixelclock accordingly. 14515 14516commit 5f846360c46f5a989f5d0fde6d251cdbd61d4968 14517Author: Alex Deucher <alexdeucher@gmail.com> 14518Date: Wed Oct 7 17:01:55 2009 -0400 14519 14520 radeon: clamp the internal FB map to the aperture size 14521 14522 We don't use the invisible memory yet and on cards with 14523 large amounts of vram this can cause the top of GART 14524 calculation to overflow. 14525 14526 Fixes bug fdo bug 24301: 14527 http://bugs.freedesktop.org/show_bug.cgi?id=24301 14528 14529commit 186bf719c394116341abff471fb4b4803650da9a 14530Author: Dave Airlie <airlied@redhat.com> 14531Date: Wed Oct 7 17:23:36 2009 +1000 14532 14533 radeon: fix rotation since mixed pixmaps. 14534 14535 Passing null in here stops the mixed pixmap code doing it wrong, 14536 but I'm not 100% sure how it is meant to work, we have the 14537 same issue with the real front bo by the looks of it. 14538 14539commit e08411af1aa8c7d4233ba593b84360397cdbb307 14540Author: Roel Kluin <roel.kluin@gmail.com> 14541Date: Tue Oct 6 18:31:17 2009 -0400 14542 14543 radeon: Fix duplicated bit settings 14544 14545 [agd5f: adapted from kms patch] 14546 14547 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14548 14549commit e59ae08270711512e64b70f79b6476cc2c52d230 14550Author: Michel Dänzer <daenzer@vmware.com> 14551Date: Tue Oct 6 00:13:09 2009 +0200 14552 14553 Don't call radeon_cs_flush_indirect() without KMS. 14554 14555 Fixes http://bugs.freedesktop.org/show_bug.cgi?id=24316 . 14556 14557commit 457646d734558672402b6bcd15cadb66741edbed 14558Author: Jerome Glisse <jglisse@redhat.com> 14559Date: Mon Oct 5 15:52:42 2009 +0200 14560 14561 kms: assume accel is working is info ioctl return -EINVAL 14562 14563 -EINVAL should only happen if the info ioctl doesn't support 14564 the get accel query. This patch assume that on such case accel 14565 is working. 14566 14567commit 9d596562496863d65850306d2126d8df98464de4 14568Author: Dave Airlie <airlied@linux.ie> 14569Date: Mon Oct 5 18:35:20 2009 +1000 14570 14571 kms: only touch gamma stuff on 1.7 14572 14573 for correct colormap operation you probably want 1.7 14574 14575commit 876f43d4682df4d47d3e158f75e0ca6040f41c70 14576Author: Alex Deucher <alexdeucher@gmail.com> 14577Date: Mon Oct 5 02:23:03 2009 -0400 14578 14579 R1/2xx EXA: Minimise number of draw primitives used for Composite operations. 14580 14581commit 86cafb8affea448bdf58300044e755201b822d2a 14582Author: Dave Airlie <airlied@redhat.com> 14583Date: Mon Oct 5 14:17:25 2009 +1000 14584 14585 kms: don't use scratch pixmaps when copying fbcon. 14586 14587 scratch pixmaps seem to interact badly with mixed pixmaps, it appears 14588 some state may be getting left around in the privates somewhere, since 14589 scratch pixmap headers don't get destroyed. 14590 14591commit f8471512ea9f1d38140dfe98a0f832e9f935f51b 14592Author: Csillag Kristof <csillag.kristof@united-consult.hu> 14593Date: Sun Oct 4 18:18:56 2009 -0400 14594 14595 radeon: add support for Custom EDID 14596 14597 Allows you to specify an edid per output from a file 14598 to override what is detected by DDC. Useful for 14599 problematic monitors or KVM switches that block 14600 DDC. Specifying an EDID that is not compatible with 14601 your monitor could damage your monitor so use with 14602 caution. 14603 14604 agd5f: cache the custom edid at startup so we don't 14605 have to read it from file every time the output is 14606 queried. 14607 14608commit 60d9685abddccec17c1a9a5ec48cbe9c92543e0f 14609Author: Dave Airlie <airlied@redhat.com> 14610Date: Mon Oct 5 12:41:30 2009 +1000 14611 14612 kms: setup colormap and gamma correctly. 14613 14614 We need to program the gamma at modeset. no need to use 14615 old cmap interface at all. 14616 14617commit b0b10e548fd69b3ff67a0f9b71456b89c61cb4d6 14618Author: Alex Deucher <alexdeucher@gmail.com> 14619Date: Sun Oct 4 17:07:32 2009 -0400 14620 14621 R6/7xx EXA: Add support for BGRA picture formats. 14622 14623 Also add mysteriously missing case statements for ABGR destination formats. 14624 14625commit 8404c37b2ff95aba31db97676f106ec2f30fda7f 14626Author: Dave Airlie <airlied@linux.ie> 14627Date: Sun Oct 4 18:18:51 2009 +1000 14628 14629 radeon: fix build without kms 14630 14631commit 9460ea864b12ec1fbd11c5d9a20bb5a4279d9d3d 14632Author: Michel Dänzer <daenzer@vmware.com> 14633Date: Sat Oct 3 16:33:33 2009 +0200 14634 14635 Fix KMS on big endian machines. 14636 14637 Requires at least xserver 1.7 to work properly. 14638 14639 Also make sure the front buffer is and stays tiled if colour tiling is enabled. 14640 14641commit aee6b6f2c120baf477b4845ddc1a94637c31db2f 14642Author: Michel Dänzer <daenzer@vmware.com> 14643Date: Sat Oct 3 16:33:33 2009 +0200 14644 14645 EXA: Bail earlier from PrepareCopy hook on trivially unsupportable conditions. 14646 14647commit 30031b169def5caadb394225919a6079a0957142 14648Author: Michel Dänzer <daenzer@vmware.com> 14649Date: Sat Oct 3 16:33:33 2009 +0200 14650 14651 R3/5xx EXA: Add support for BGRA picture formats. 14652 14653 Also add mysteriously missing case statements for ABGR destination formats. 14654 14655commit 4b4ce36081ca151c24e028c54b59986f41731a73 14656Author: Michel Dänzer <daenzer@vmware.com> 14657Date: Sat Oct 3 16:33:32 2009 +0200 14658 14659 R3/5xx EXA: Minimise number of draw primitives used for Composite operations. 14660 14661 This should reduce the kernel CS checker overhead, if nothing else. 14662 14663 I'll leave porting this to other chipset families to others who can test it. 14664 14665commit eade1e5be159c9f2965d611925596d33cab11d6d 14666Author: Michel Dänzer <daenzer@vmware.com> 14667Date: Sat Oct 3 16:33:32 2009 +0200 14668 14669 EXA: Don't always flush when switching between 2D/3D engines with KMS. 14670 14671 This seems to work fine here now and help 2D performance quite a bit, let's see 14672 if it breaks anybody else's setup... 14673 14674commit 926e414fc835ace141c066830d11d8ce32dbb06c 14675Author: Michel Dänzer <daenzer@vmware.com> 14676Date: Sat Oct 3 16:33:32 2009 +0200 14677 14678 KMS: Double-buffer textured video source image upload. 14679 14680 In order to avoid stalling on previous frame. 14681 14682 OTOH without KMS we can't do this but have to wait for the previous frame to 14683 finish rendering. 14684 14685commit 6a3b75fae4147244212298ff55c7ab36c3d80d30 14686Author: Michel Dänzer <daenzer@vmware.com> 14687Date: Sat Oct 3 16:33:31 2009 +0200 14688 14689 KMS: Don't hide HW cursor when it's updated. 14690 14691 Inspired by the intel driver; might avoid/reduce HW cursor flicker in some 14692 cases. 14693 14694commit c4ba15b4871c70cced7f1761e24be692cdb28d2b 14695Author: Michel Dänzer <daenzer@vmware.com> 14696Date: Sat Oct 3 16:33:31 2009 +0200 14697 14698 Fix some compiler warnings. 14699 14700commit cc45856a18dd3e6f7e44d9eb507b31419da70977 14701Author: Kenneth Graunke <kenny@whitecape.org> 14702Date: Fri Oct 2 14:31:36 2009 -0400 14703 14704 radeon: Remove gatos message from driver startup 14705 14706 Remove a message on driver startup which directs people to the old GATOS 14707 website; said site contains highly out of date software which is no longer 14708 necessary. 14709 14710commit d302481493ff61eb7f8648b7afe7407b6b653cf7 14711Author: Jerome Glisse <jglisse@redhat.com> 14712Date: Wed Sep 16 15:18:40 2009 +0200 14713 14714 radeon/kms: fallback to shadowfb if kernel report acceleration is off 14715 14716 This will fallback to shadowfb is GPU accel failed somewhere during 14717 KMS kernel init. 14718 14719commit 7968e1fb89f6b59d1654df48249bf4b81990c008 14720Author: Alex Deucher <alexdeucher@gmail.com> 14721Date: Fri Sep 25 19:05:38 2009 -0400 14722 14723 radeon: fix segfault in MMIO path in RADEONInit3DEngine 14724 14725 fixes fdo bug 24158. 14726 14727 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14728 14729commit 0380a9cf98d34e88dd81370ade5525680ec89c02 14730Author: Alex Deucher <alexdeucher@gmail.com> 14731Date: Fri Sep 25 16:46:43 2009 -0400 14732 14733 kms/r600: add support for vline relocs 14734 14735 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14736 14737commit 3efecebb10de7f7bacf9f8c57ae20fd508097294 14738Author: Alex Deucher <alexdeucher@gmail.com> 14739Date: Fri Sep 25 14:44:41 2009 -0400 14740 14741 radeon: fix vline handling for kms 14742 14743 drm crtc ids do not correspond to actual hw crtcs, 14744 as such the vline stuff was never enabled for Xv. 14745 14746 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14747 14748commit 9733dcde0a21b7503aa20254724f2910b541b990 14749Author: Dave Airlie <airlied@linux.ie> 14750Date: Sat Sep 26 06:15:58 2009 +1000 14751 14752 r600: fix build 14753 14754commit ff18595eace42ddfc931f4f948cb5adf59ebcf52 14755Author: Dave Airlie <airlied@linux.ie> 14756Date: Sat Sep 26 06:04:51 2009 +1000 14757 14758 r600/xv: fix dst bo write domain 14759 14760commit 718a611f3cc7f5eea95cd73a9dcc7913c922fcbc 14761Author: Dave Airlie <airlied@linux.ie> 14762Date: Sat Sep 26 06:03:51 2009 +1000 14763 14764 radeon: fix zaphod 14765 14766commit 86c367be517e992aabc1bcfea2a2dde1a1ae520c 14767Author: Alex Deucher <alexdeucher@gmail.com> 14768Date: Fri Sep 25 13:03:44 2009 -0400 14769 14770 r600: fix bo accounting for Xv 14771 14772 fixes Xv after changes in: 14773 8f80e37eed3ec028718b4e71bbb9b598847fd94e 14774 14775 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14776 14777commit 8f80e37eed3ec028718b4e71bbb9b598847fd94e 14778Author: Dave Airlie <airlied@linux.ie> 14779Date: Fri Sep 25 11:37:38 2009 +1000 14780 14781 r600: space check wasn't correct. 14782 14783 the r6xx space check code was incorrect for the DFS case, since 14784 the dst bo was in GTT but the hardcoded cp_start function 14785 assumed the bos were where it thought. 14786 14787 Ripped out assumptions and replaced with code more like other radeons. 14788 14789commit 90669f6cb5ada9067d6fccc7f54c69b367862f42 14790Author: Alex Deucher <alexdeucher@gmail.com> 14791Date: Thu Sep 24 10:40:40 2009 -0400 14792 14793 radeon: remove old pre-randr rotation cruft 14794 14795 leftover from the randr 1.2 conversion. 14796 14797 Signed-off-by: Alex Deucher <alexdeucher@gmail.com> 14798 14799commit 97a4e747bfac14f34646c55ddf639e8fe22f2f55 14800Author: Michael Olbrich <m.olbrich@pengutronix.de> 14801Date: Sun Sep 20 14:19:41 2009 +0200 14802 14803 use AC_CHECK_HEADER instead of AC_CHECK_FILE 14804 14805 AC_CHECK_FILE is not possible when cross-compiling. Use AC_CHECK_HEADER / 14806 AC_PREPROC_IFELSE instead. 14807 14808 [ Michel Dänzer: Shuffled things around slightly to make it work on my setup ] 14809 14810 Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de> 14811 Signed-off-by: Michel Dänzer <daenzer@vmware.com> 14812 14813commit 579070ed9f0218a975e840331b6fe9fa8fd05789 14814Author: Alex Deucher <alexdeucher@gmail.com> 14815Date: Sat Sep 19 19:15:14 2009 -0400 14816 14817 kms: disable dga 14818 14819 Can be dangerous with kms: 14820 http://lists.freedesktop.org/archives/xorg-devel/2009-September/002113.html 14821 14822commit da7487f6ac7b32ed7866af53da2925aa52ed7aae 14823Author: Alex Deucher <alexdeucher@gmail.com> 14824Date: Sat Sep 19 13:08:02 2009 -0400 14825 14826 kms: init dga using xf86DiDGAInit 14827 14828 fixes cursor problems in games like UT2004 and Penumbra Overture 14829 14830 see fdo bug 24034 14831 14832commit 52279251fae9df99c569c16e2522bbd346d8ec38 14833Author: Dave Airlie <airlied@redhat.com> 14834Date: Fri Sep 18 14:29:32 2009 +1000 14835 14836 avivo: disable VGA rendering core when starting X. 14837 14838 this thing can be rendering to VRAM when we don't expect it. 14839 turn it off. 14840 14841 Signed-off-by: Dave Airlie <airlied@redhat.com> 14842 14843commit 2c46bafcb77c4125a27c18ad6ca2f6de5f143a2d 14844Author: Pauli Nieminen <suokkos@gmail.com> 14845Date: Thu Sep 17 15:13:51 2009 +0300 14846 14847 Fix configuration to disable KMS if not finding libdrm_radeon. 14848 14849 Most of KMS building was disabled if LIBDRM_RADEON was not detected. But at 14850 least configure was reporting KMS enabled even tough it was realy disabled. 14851 14852commit ee9f6d802988a4d7f7985687036db854f4a14fe5 14853Author: Pauli Nieminen <suokkos@gmail.com> 14854Date: Mon Sep 14 16:36:48 2009 +0300 14855 14856 radeon: Add missing libdrm_radeon cflags to compiler paramaters. 14857 14858 This fixes compilation if libdrm_radeon is installed to non-standard location like /opt. 14859 14860commit ac853ca0a950d6147cd6a0b4a1b089f5d2f080ea 14861Author: Alex Deucher <alexdeucher@gmail.com> 14862Date: Fri Sep 11 14:25:31 2009 -0400 14863 14864 atom: fix typo in asus quirks 14865 14866 Should be DVI-I, not DVI-D 14867 14868commit a88f60e7b789cd23a79e6e18a7e3fa008d15aa77 14869Author: Alex Deucher <alexdeucher@gmail.com> 14870Date: Thu Sep 10 13:42:31 2009 -0400 14871 14872 r600: add CS UTS and DFS hooks 14873 14874commit 0bb0ff0e55aaf8e64017d204e376a14f0a7384a3 14875Author: Dave Airlie <airlied@linux.ie> 14876Date: Thu Sep 10 11:48:00 2009 +1000 14877 14878 r6xx: fix EXA crash with gtkperf -a 14879 14880 unmapping was getting unbalanced. 14881 14882commit 10a58d54857484b6e89763ecf463f54dc290c702 14883Author: Dave Airlie <airlied@redhat.com> 14884Date: Wed Sep 9 14:24:49 2009 +1000 14885 14886 kms: add property support. 14887 14888 this adds support to the userspace DDX to get properties from the kernel 14889 and expose them over randr. Its most liberated from the Intel driver. 14890 14891commit fad9ee6cdae5ae45e6406e3071a84276f1b59a85 14892Author: Alex Deucher <alexdeucher@gmail.com> 14893Date: Tue Sep 8 17:50:50 2009 -0400 14894 14895 radeon: replace dga code with DiDGA 14896 14897commit 651fe5a47435c6a5cc1cdfb33c5c6601d692adcb 14898Author: Adam Jackson <ajax@redhat.com> 14899Date: Tue Sep 8 13:21:07 2009 +1000 14900 14901 radeon: fix dri2 struct zeroing 14902 14903commit 6990f2ac6478bf92929a4400ef84fb2142699204 14904Merge: 917f2d7c 853f4c3d 14905Author: Dave Airlie <airlied@redhat.com> 14906Date: Tue Sep 8 11:26:32 2009 +1000 14907 14908 Merge branch 'r6xx-cs' 14909 14910commit 853f4c3d1ea8f975ab2855f18d3ae336a4095091 14911Author: Dave Airlie <airlied@redhat.com> 14912Date: Tue Sep 8 11:25:39 2009 +1000 14913 14914 r600: more alignment fixups + vb map/unmap 14915 14916 I'm not so sure the vb map/unmap is a good idea, I think 14917 it pretty much locksteps the cpu/gpu, so we should really 14918 work out if we really need to flush this often, since 14919 mesa doesn't have to and we are just doing 3D ops. 14920 14921commit 917f2d7cd1815279710939edc27133ca6609460f 14922Author: Alex Deucher <alexdeucher@gmail.com> 14923Date: Sun Sep 6 10:53:29 2009 -0400 14924 14925 RV280: add agp quirk 14926 14927 deb bug 545040 14928 14929commit 78fcbf577ad6eba6399cc39f74b7ce5f9c8e265e 14930Author: Michel Dänzer <daenzer@vmware.com> 14931Date: Thu Sep 3 14:55:05 2009 +0200 14932 14933 EXA: Check for solid/gradient pictures the same way for all generations. 14934 14935 In particular, also catch them for >= R300. 14936 14937commit 794ae743c305331741e7cf45dd71c755fe325ed5 14938Author: Michel Dänzer <daenzer@vmware.com> 14939Date: Wed Sep 2 09:19:17 2009 +0200 14940 14941 EXA: Fix Composite restart dst/src switcheroo. 14942 14943 Fixes http://bugs.freedesktop.org/show_bug.cgi?id=22636 . 14944 14945 Now, where's that brown paper bag? :} 14946 14947commit c4ab50c5eafca3d04129a72453496eb8afb26b34 14948Author: Alex Deucher <alexdeucher@gmail.com> 14949Date: Tue Sep 1 13:43:00 2009 -0400 14950 14951 radeon: add some notes about DVO setup 14952 14953commit 22fbd6476f4b3513378e848affc4dd269e9d4498 14954Author: Dave Airlie <airlied@itt42.(none)> 14955Date: Tue Sep 1 15:22:57 2009 +1000 14956 14957 r100/r200: seriously WRAP is illegal for rectangular textures 14958 14959 unless we scale all the bos up to POT 14960 14961commit 8f4196e88855f10762254fca9e0a0988e7b5562f 14962Author: Alex Deucher <alexdeucher@gmail.com> 14963Date: Mon Aug 31 19:41:59 2009 -0400 14964 14965 r6xx/r7xx: various CS fixes from Dave 14966 14967commit a32384f9b9ed73c14484a380dfa14652ed478ad6 14968Author: Alex Deucher <alexdeucher@gmail.com> 14969Date: Mon Aug 31 13:26:05 2009 -0400 14970 14971 r1xx: fix textured video 14972 14973 Copy/paste bug from when kms support was added. 14974 14975commit 3d6bf0f8cf128ee67d448491be44d345547b75c5 14976Author: Pauli Nieminen <suokkos@gmail.com> 14977Date: Fri Aug 28 17:05:44 2009 +0300 14978 14979 radeon: Fix DRI2BufferPtr to be DRI2Buffer2Ptr for xserver 1.6. 14980 14981 Bump requirement for xserver in KMS mode to 1.6.2 for DRI2Buffer2Ptr. 14982 14983commit 6521dbf3884185ca5692d35d879d251890cc4f08 14984Author: Roland Scheidegger <sroland@tungstengraphics.com> 14985Date: Sat Aug 29 12:11:03 2009 +0200 14986 14987 r100/r200: fix section size mismatch for textured video 14988 14989 this fixes the driver complaining though reportedly it still does 14990 not work (on r100) 14991 14992commit 39dfac41eebbcdacb1ae0c17cea325ea2966c4a5 14993Author: Michel Dänzer <daenzer@vmware.com> 14994Date: Fri Aug 28 12:20:04 2009 +0200 14995 14996 EXA: RENDER repeat fix and more cleanup. 14997 14998 Apparently changing pPict->repeatType causes badness, see 14999 http://bugs.freedesktop.org/show_bug.cgi?id=23560 . 15000 15001 Now we just use a normalized repeatType value derived from pPict->repeat and 15002 pPict->repeatType everywhere. 15003 15004commit 7d36f13afa5da742aa6c989a2652e92c96e5f80e 15005Author: Michel Dänzer <daenzer@vmware.com> 15006Date: Thu Aug 27 09:52:11 2009 +0200 15007 15008 KMS: Remove explicit radeon_bo_wait calls before radeon_bo_map. 15009 15010 libdrm_radeon always takes care of this implicitly now. 15011 15012commit 066c90e8f5b93c80195dc43486ad3efa083a1f3a 15013Author: Michel Dänzer <daenzer@vmware.com> 15014Date: Thu Aug 27 09:52:11 2009 +0200 15015 15016 KMS: Don't use a blit for UploadToScreen if the pixmap BO will be idle anyway. 15017 15018 Direct CPU writes should be at least as fast in that case. 15019 15020commit 174b61bb786a841cebd354e23b4b0caa85b541aa 15021Author: Michel Dänzer <daenzer@vmware.com> 15022Date: Thu Aug 27 09:52:11 2009 +0200 15023 15024 KMS: DownloadFromScreen improvements. 15025 15026 * Drop superfluous RADEONDownloadFromScreenGTT function, EXA does the same 15027 thing when we return FALSE. 15028 * Take unflushed operations into account for determining which GEM domain the 15029 pixmap BO will end up in. 15030 * Only use a blit if it ends up in VRAM. 15031 15032commit 7623e169e1f7d5afbd4108de03f28098bca519db 15033Author: Michel Dänzer <daenzer@vmware.com> 15034Date: Thu Aug 27 09:52:11 2009 +0200 15035 15036 EXA: Allocate pixmap BOs in VRAM. 15037 15038 This is the intention with EXA in xserver Git, and IME it improves performance 15039 significantly even with older versions. 15040 15041commit b90b823790657707b5bf7249095833bef959b0e8 15042Author: Michel Dänzer <daenzer@vmware.com> 15043Date: Thu Aug 27 09:52:11 2009 +0200 15044 15045 DRI2: Set tiling for depth/stencil buffers with all DRI2 interface versions. 15046 15047commit 27d9664f8862d1948dd32021f8e9bcb94f23e4ef 15048Author: Michel Dänzer <daenzer@vmware.com> 15049Date: Thu Aug 27 09:52:10 2009 +0200 15050 15051 EXA: RENDER repeat cleanups. 15052 15053 We can't rely on the server setting pPict->repeatType to something sensible 15054 when pPict->repeat is FALSE, but we can do it ourselves. 15055 15056commit 9b5517fcf7f2397a1a814d396fc9dc24769a6d9f 15057Author: Michel Dänzer <daenzer@vmware.com> 15058Date: Thu Aug 27 09:52:10 2009 +0200 15059 15060 Throttle DRI2 buffer swaps / frontbuffer flushes. 15061 15062commit e87f0f50f31a59ca1f60d4582d4a57ed00854fb7 15063Author: Alex Deucher <alexdeucher@gmail.com> 15064Date: Wed Aug 26 02:13:38 2009 -0400 15065 15066 r6xx/r7xx: set EXA_HANDLES_PIXMAPS 15067 15068commit 3212c26b90c0f6f1a7248b4da3ed985a9c2e9381 15069Author: Alex Deucher <alexdeucher@gmail.com> 15070Date: Wed Aug 26 01:42:10 2009 -0400 15071 15072 r6xx/r7xx: more WIP 15073 15074commit 5a08e68cc254fb255e631b456e331c32456ef0e7 15075Author: Alex Deucher <alexdeucher@gmail.com> 15076Date: Tue Aug 25 19:24:41 2009 -0400 15077 15078 r6xx/r7xx: fix some define problems in Xv code 15079 15080commit 9aa214e125b7927d62b9fe124a851d0373c24d7e 15081Author: Alex Deucher <alexdeucher@gmail.com> 15082Date: Tue Aug 25 18:45:49 2009 -0400 15083 15084 r6xx/r7xx: fix reloc for vtx buffer 15085 15086commit bba51187055932ecd466f5f817428d6c773747b9 15087Author: Alex Deucher <alexdeucher@gmail.com> 15088Date: Tue Aug 25 18:37:15 2009 -0400 15089 15090 R6xx/r7xx: unmap vb bo when done 15091 15092commit 599adfc1f5e6d708be7ad30f4871de3046775727 15093Author: Alex Deucher <alexdeucher@gmail.com> 15094Date: Tue Aug 25 18:13:14 2009 -0400 15095 15096 r6xx/r7xx: fix flipped domains 15097 15098commit 2e83cca8d7efaf1a6836cfb9ea5893fd9d70175f 15099Author: Alex Deucher <alexdeucher@gmail.com> 15100Date: Tue Aug 25 18:05:43 2009 -0400 15101 15102 r6xx/r7xx: more cs exa wip 15103 15104commit 65852de027989c105246fa4e4eed432f29525a22 15105Author: Alex Deucher <alexdeucher@gmail.com> 15106Date: Tue Aug 25 17:29:42 2009 -0400 15107 15108 r6xx/r7xx EXA: WIP 15109 15110commit 69ec7a35e2a0a3d802ec093a6aab2d7ed2cc88be 15111Author: Alex Deucher <alexdeucher@gmail.com> 15112Date: Tue Aug 25 16:14:02 2009 -0400 15113 15114 r6xx/r7xx: first pass at kms accel support 15115 15116 Adapted from various patches from Dave and Jerome. 15117 15118commit 04692f80678ffea15ac33820381ce45865e697b8 15119Author: Dave Airlie <airlied@linux.ie> 15120Date: Wed Aug 26 05:31:56 2009 +1000 15121 15122 radeon: afaics these aren't radeon regs but r128 regs. 15123 15124 so remove them from radeon use. 15125 15126commit 9cf965bbc977f0523437c0ecf1d7363b17de2468 15127Author: Alex Deucher <alexdeucher@gmail.com> 15128Date: Tue Aug 25 12:47:34 2009 -0400 15129 15130 R6xx/r7xx: add begin/end batch macros 15131 15132commit b6368cc572c79bce9a9366242c727c13cab3f006 15133Author: Alex Deucher <alexdeucher@gmail.com> 15134Date: Tue Aug 25 12:14:33 2009 -0400 15135 15136 r6xx/r7xx: move more common state to default state setup 15137 15138commit fc74e1194c980d978667e02c60a29a761a694bde 15139Author: Alex Deucher <alexdeucher@gmail.com> 15140Date: Tue Aug 25 11:39:56 2009 -0400 15141 15142 RV740: disable small DFS transfers 15143 15144 Seems problematic on all rv740 chips, so until we 15145 can find a solution, disable them. 15146 15147commit 1c909ecd88e75c6e16af7aa068e32a7a60b0af37 15148Author: Jerome Glisse <jglisse@redhat.com> 15149Date: Tue Aug 25 11:21:41 2009 +0200 15150 15151 radeon: Fix X recycling with KMS 15152 15153 On X recycling CloseScreen drop master so grab it again 15154 in InitScreen and avoid to reallocate buffer & various 15155 others cs/bo utilities on X recycling. 15156 15157commit 39aaf9addc923852598ba6fb914459671b78e492 15158Author: Dave Airlie <airlied@linux.ie> 15159Date: Mon Aug 24 18:59:22 2009 +1000 15160 15161 r600: use exa fallback code for solid/copy codepaths. 15162 15163 Solid seem to try and handle 1bpp which really no. 15164 15165commit e4b2d57ddca8739ec0b452565713398d036be0d5 15166Author: Dave Airlie <airlied@redhat.com> 15167Date: Mon Aug 24 14:11:24 2009 +1000 15168 15169 r600: fix bad unit in texture setup 15170 15171commit bac224912c750dc1c85ff2d9b8526dad6c23b572 15172Author: Dave Airlie <airlied@redhat.com> 15173Date: Sat Aug 22 21:17:59 2009 +1000 15174 15175 radeon: don't spec any initial placement for pixmaps. 15176 15177 allow the first use to decide placement. 15178 15179commit 77f98717d825162da106c6898cdbcbdf5c984ae6 15180Author: Dave Airlie <airlied@redhat.com> 15181Date: Sat Aug 22 21:16:25 2009 +1000 15182 15183 exa/cs: add DFS from GTT optimisation 15184 15185 This uses the new libdrm busy interface, once I had this in place 15186 I added a error if this happened and it does on my desktop here, 15187 so may as well add the optimisation that used to be in my old KMS tree. 15188 15189 Signed-off-by: Dave Airlie <airlied@redhat.com> 15190 15191commit b1b77a4d6fb7404af9568644e1a8e050fdfa956e 15192Author: Alex Deucher <alexdeucher@gmail.com> 15193Date: Wed Aug 19 13:29:04 2009 -0400 15194 15195 rs600: add support for DisplayPriority HIGH 15196 15197 Still haven't implemented the actual watermark 15198 calculation, but this forces the display requests 15199 to urgent if the user specifies DisplayPriority HIGH. 15200 15201commit 72e0d1b2cb11a67b8e4be4c74913ee44dc051c5b 15202Author: Dave Airlie <airlied@redhat.com> 15203Date: Wed Aug 19 19:01:30 2009 +1000 15204 15205 radeon/kms: add initial colortiling support (disabled by default). 15206 15207 This requires an X server from git with createpixmap2 support fixed up in it. 15208 15209 On 1.6 and previous it won't do any tiling, if you are running git server 15210 please upgrade to latest git. 15211 15212 Option "AllowColorTiling" "true" to enable and do some testing 15213 15214commit 479a6daefe46f985c415b0d000b1b1b820f3924e 15215Author: Alex Deucher <alexdeucher@gmail.com> 15216Date: Tue Aug 18 02:53:29 2009 -0400 15217 15218 AVIVO: add support for DisplayPriority HIGH 15219 15220 If the DisplayPriority option is set to HIGH, 15221 force the diplay mem requests to urgent. 15222 15223commit 8fd6f69c84103b277de70181988e843da23c85b0 15224Author: Dave Airlie <airlied@redhat.com> 15225Date: Mon Aug 17 15:50:21 2009 +1000 15226 15227 radeon: move detected monitor type debugging. 15228 15229 We should print this after DAC detection to be useful for TV 15230 15231commit d25185f33d06a8162469e18688e95ee19a5d94d8 15232Author: Dave Airlie <airlied@redhat.com> 15233Date: Mon Aug 17 15:31:16 2009 +1000 15234 15235 atom: recall crtc source script after tv dpms on 15236 15237 For some reason we lost the DACB source when we dpms tv back on. 15238 15239commit a971f1cfbbda1e3d16fb41f4a07f3b43fd59bc0d 15240Author: Dave Airlie <airlied@redhat.com> 15241Date: Mon Aug 17 15:25:17 2009 +1000 15242 15243 atombios: store dac detect result into save register. 15244 15245 We end up writing back an empty save register over the dac 15246 detection results so the atombios table can no longer 15247 figure out the S-video vs composite so sets the wrong one up. 15248 15249 this gets tv-out on my r580 with a composite connector working. 15250 15251commit efef7c01ad38e078de2fa3f9e528e4ef7d05d00a 15252Author: Dave Airlie <airlied@redhat.com> 15253Date: Mon Aug 17 13:44:20 2009 +1000 15254 15255 atom/tvout: set the correct scaler up for tv out 15256 15257 If we need to use the second scaler actually set the second scaler 15258 up. 15259 15260commit 032a8a06315d12e66240e7ef74b5b165f6d0d1a9 15261Author: Dave Airlie <airlied@redhat.com> 15262Date: Mon Aug 17 12:18:23 2009 +1000 15263 15264 kms: make tv out match the connector name 15265 15266commit 36e51c4d63836863dd7a17cbf6e5a9b7c1bbd31a 15267Author: Dave Airlie <airlied@redhat.com> 15268Date: Mon Aug 17 12:17:31 2009 +1000 15269 15270 r100/r200: dont emit wrong clamp modes. 15271 15272 for rect textures you don't want to use wrap clamping which 15273 is the default. 15274 15275commit 595f7c3a60ca982ca2836a0174e1f36f9d1a4fec 15276Author: Dave Airlie <airlied@itt42.(none)> 15277Date: Mon Aug 17 09:38:09 2009 +1000 15278 15279 radeon/tv: fix typo in crtc register modifications 15280 15281commit 92baeafde8f0c1eed2b6769e96f8093e3d6d7478 15282Author: Alex Deucher <alexdeucher@gmail.com> 15283Date: Tue Aug 11 15:29:17 2009 -0400 15284 15285 RV280: add agp quirk 15286 15287 from Hartmut Niemann 15288 15289commit 04288fdb8b8c62fc2eb32fd467fdf05151ad121a 15290Author: Jerome Glisse <jglisse@redhat.com> 15291Date: Tue Aug 11 19:32:52 2009 +0200 15292 15293 radeon/r600: abstract vertex buffer for r600 & r700 hardware 15294 15295 Abstract vertex buffer, preliminary work before others change 15296 to bring cs support along IB support. 15297 15298commit 72daee2a18bfb809260d1bcbe80cc14afea3b0fa 15299Author: Adam Jackson <ajax@redhat.com> 15300Date: Mon Aug 10 13:59:49 2009 -0400 15301 15302 Fix misleading filenames in DRI setup failure messages 15303 15304commit cd99d9f0d715f1f74de4fe22f2fd30046f2c7568 15305Author: Michel Dänzer <daenzer@vmware.com> 15306Date: Fri Aug 7 11:56:50 2009 +0200 15307 15308 Properly let radeon_cs_flush_indirect() re-emit 2D state if necessary. 15309 15310 Also remove a superfluous assignment. 15311 15312commit 577ff3ce922e457cc32f80d4365cb1da81552e72 15313Author: Michel Dänzer <daenzer@vmware.com> 15314Date: Fri Aug 7 11:56:49 2009 +0200 15315 15316 Add support for EXA_MIXED_PIXMAPS in xserver master. 15317 15318commit e755fa56d03aa338d0c6345ed41e32aa3115ad4b 15319Author: Michel Dänzer <daenzer@vmware.com> 15320Date: Fri Aug 7 11:56:49 2009 +0200 15321 15322 Minor fixes for KMS EXA DownloadFromScreen hook. 15323 15324 Check we can handle the bpp, and remove superfluous flush. 15325 15326commit 4cebafae81c1738a1c330d52a2c5248869f06411 15327Author: Alex Deucher <alexdeucher@gmail.com> 15328Date: Fri Aug 7 11:56:40 2009 +0200 15329 15330 Add KMS EXA UploadToScreen hook. 15331 15332 Fixups by Michel Dänzer: Doesn't seem to be a win in terms of raw numbers, but 15333 should allow more pipelining and avoiding BO memory waste for small glyph 15334 pixmaps. 15335 15336commit 9243791322e36b9231e6a3f04024ad66325385e3 15337Author: Dave Airlie <airlied@redhat.com> 15338Date: Fri Aug 7 14:07:32 2009 +1000 15339 15340 r200: fixup scissors for DDX. 15341 15342 a) turn of R200_RE_CNTL - SCISSOR_ENABLE - this save us emitting R200_RE_TOP_LEFT, note scissor is still enabled. 15343 b) disable aux scissors. 15344 15345commit bd03977e320591ca55b1a2fbb32414c53cb3f72e 15346Author: Alex Deucher <alexdeucher@gmail.com> 15347Date: Tue Aug 4 15:21:44 2009 -0400 15348 15349 radeon: fix compile with !kms 15350 15351commit 4fc1e67e606daf9227e74518c54315b60cd38301 15352Author: Alex Deucher <alexdeucher@gmail.com> 15353Date: Tue Aug 4 15:17:39 2009 -0400 15354 15355 radeon: re-add asus connector quirks 15356 15357 fixes bug 19943 15358 15359commit 22074cf0e58fddba743924532625e6fca49b6bdc 15360Author: Jerome Glisse <jglisse@redhat.com> 15361Date: Tue Aug 4 21:09:25 2009 +0200 15362 15363 radeon/kms: add simple DownloadFromScreen implementation 15364 15365 What we want to do is add userspace object support to radeon 15366 kernel modesetting. Also this DFS is dumb and might endup doing 15367 blit from GTT to GTT. 15368 15369commit fce31b61a88522733863a9b4e9f1c935c439cb4e 15370Author: Alex Deucher <alexdeucher@gmail.com> 15371Date: Tue Aug 4 11:16:05 2009 -0400 15372 15373 RS880: enable accel 15374 15375commit 6b1b4b9214d58175727f7992fc42bc7244c42bf2 15376Author: Alex Deucher <alexdeucher@gmail.com> 15377Date: Mon Aug 3 15:55:04 2009 -0400 15378 15379 radeon: add some new r7xx pci ids 15380 15381commit f564460e94c9d0f1cf3ff4b8535481b2b8b4e9c1 15382Author: Alex Deucher <alexdeucher@gmail.com> 15383Date: Sat Aug 1 16:53:47 2009 -0400 15384 15385 radeon: use XAA in some cases 15386 15387 Use XAA in low memory situations or when the DRI is disabled. 15388 Using shadowfb might also be a viable option, maybe even a better option... 15389 fixes bug 21683 15390 15391commit c71b2f050e8996787eae463eddbfdb5864ffa65a 15392Author: Bryce Harrington <bryce@canonical.com> 15393Date: Sat Aug 1 16:50:43 2009 -0400 15394 15395 radeon: AGPMode quirk needed for SiS 15396 15397 fixes bug 23065 15398 15399commit e3659ed06fc5bb8817f1dbd7c2d6bc94c67b30f7 15400Author: Bryce Harrington <bryce@canonical.com> 15401Date: Sat Aug 1 16:48:48 2009 -0400 15402 15403 radeon: AGPMode quirk needed for IBM Thinkpad T40 with Mobility M7 LW 15404 15405 fixes bug 23064 15406 15407commit 2391531ed6b7c11ddd5ab91b2369821cc5f8b8a7 15408Author: Bryce Harrington <bryce@canonical.com> 15409Date: Sat Aug 1 16:41:41 2009 -0400 15410 15411 radeon: AGPMode quirk needed for HP Omnibook 6200 15412 15413 fixed bug 23063 15414 15415commit 474eda02257152ced52364f38cbad24c20aebbc0 15416Author: Alex Deucher <alexdeucher@gmail.com> 15417Date: Sat Aug 1 16:06:52 2009 -0400 15418 15419 radeon: reload bicubic Xv texture on VT switch 15420 15421 fixed bicubic filtering after VT switch or suspend/resume 15422 15423commit 447a2ce1b88aa2d6d5713e93174c4002617720f7 15424Author: Dave Airlie <airlied@redhat.com> 15425Date: Thu Jul 30 09:54:01 2009 +1000 15426 15427 radeon: fix r600 shadow fb mode with cursor allocation in wrong place 15428 15429 From RHEL QE testing, we could end up with the cursor at 0 since 15430 we think EXA is in use when really it isn't. the info->useEXA = FALSe 15431 might be unnecessary but better to be explicit 15432 15433 Signed-off-by: Dave Airlie <airlied@redhat.com> 15434 15435commit 2b9a5be3bb22fad1d52ec9eea7733c5e956250b7 15436Author: Alex Deucher <alexdeucher@gmail.com> 15437Date: Wed Jul 29 03:34:11 2009 -0400 15438 15439 radeon: fix typo in object header to connector type conversion 15440 15441 Should fix bug 19943 15442 15443commit 63c873cbd4d1d21d9f688028c0900c79fadc42c1 15444Author: Dave Airlie <airlied@redhat.com> 15445Date: Tue Jul 28 15:22:40 2009 +1000 15446 15447 ati: change to using ABI version check 15448 15449commit 21a621c297ac71c65c239ea960c38706e718b91c 15450Author: Dave Airlie <airlied@redhat.com> 15451Date: Tue Jul 28 13:32:28 2009 +1000 15452 15453 ati: update for resources/RAC API removal 15454 15455commit 5ad2519c2a7c0df389b2cd7cf7151c7e4b7252a8 15456Author: Michel Dänzer <daenzer@vmware.com> 15457Date: Mon Jul 27 14:22:02 2009 +0200 15458 15459 KMS: Explicitly specify VRAM and GTT domains when allocating BOs for pixmaps. 15460 15461 Without this, apparently they are initially allocated in system RAM (or 15462 possibly GTT), so any GPU rendering to them first triggers a copy of the whole 15463 uninitialized data to VRAM... 15464 15465 We may want to explore more sophisticated schemes in the future, but for now 15466 this seems to improve KMS 2D performance quite a bit. 15467 15468commit 57f2c83a22f27567506c555af431f89e6031204c 15469Author: Alex Deucher <alexdeucher@gmail.com> 15470Date: Sat Jul 25 14:19:38 2009 -0400 15471 15472 radeon: Set PCI/PCIE bus type properly per asic 15473 15474 - r1xx-rv350 chips have the old pci gart 15475 - rv380+ chips have newer pcie gart 15476 15477 Select the right kind regardless of whether the user selects 15478 PCI or PCIE. 15479 15480commit 2afc46fa74ce7730f766a3370d323c6b59694186 15481Author: Jerome Glisse <jglisse@redhat.com> 15482Date: Thu Jul 23 20:13:15 2009 +0200 15483 15484 radeon: fix KMS shadowfb for r6xx and newer hw. 15485 15486 We need to map front buffer and also to avoid EXA initialization 15487 when falling back to shadowfb on newer r6xx/r7xx when KMS is enabled. 15488 15489commit 328e4c816348b6f867df30b39856dbe78c8e0dcc 15490Author: Alex Deucher <alexdeucher@gmail.com> 15491Date: Thu Jul 23 13:28:36 2009 -0400 15492 15493 RS780: fill in MC access functions 15494 15495 Also, r6xx/r7xx don't have indirect MC space, so 15496 don't try to access it. 15497 15498commit f96d5b255425fbd02be2cad26edb590d474a5640 15499Author: Alex Deucher <alexdeucher@gmail.com> 15500Date: Thu Jul 23 13:25:04 2009 -0400 15501 15502 R6xx/R7xx: clip rendering to destination surface dimensions 15503 15504commit e372f845b0defaf2d2c9ef3cbbf7498e09d9372e 15505Author: Michel Dänzer <daenzer@vmware.com> 15506Date: Sun Jul 19 16:05:29 2009 +0200 15507 15508 Guard reference to CRTC active field. 15509 15510 Fixes build against older xserver. 15511 15512commit e38305aebdc95f80f5b4b3e5ba541ea67dc05f01 15513Author: Michel Dänzer <daenzer@vmware.com> 15514Date: Sat Jul 18 22:23:45 2009 +0200 15515 15516 Also drop DRM master in KMS CloseScreen. 15517 15518 The LeaveVT hook isn't always called when the server dies, e.g. when quitting 15519 from the GDM greeter. This may cause existing servers to fall over if the dying 15520 server process still exists when they try to re-acquire master and set a mode. 15521 15522 Also use drmSet/DropMaster() rather than ioctl() directly. 15523 15524commit a43c660a00147bfae5ca601f4720b2680b75211f 15525Author: Michel Dänzer <daenzer@vmware.com> 15526Date: Sat Jul 18 22:21:59 2009 +0200 15527 15528 Set CRTC active flag to TRUE after a successful KMS mode set. 15529 15530 Otherwise some things like changing the colour map won't work properly. 15531 15532commit f32069f42b3e4643f15148d0b96164def00dcc74 15533Author: Michel Dänzer <daenzer@vmware.com> 15534Date: Sat Jul 18 22:20:08 2009 +0200 15535 15536 Set the STALL bit in the CRTC_GUI_TRIG_VLINE registers. 15537 15538 As recommended by the register reference when using the WAIT_CRTC_VLINE bit in 15539 the WAIT_UNTIL register, as we are. 15540 15541commit 27bd9fc9b2c0eeff488c1f26f1355a7fa3655520 15542Author: Michel Dänzer <daenzer@vmware.com> 15543Date: Sat Jul 18 22:14:27 2009 +0200 15544 15545 Enable vsync for DRI2 region copies. 15546 15547commit 4b1b28865ffb17f58a70d955212fd51cb1fb55ab 15548Author: Alex Deucher <alexdeucher@gmail.com> 15549Date: Thu Jul 16 17:09:33 2009 -0400 15550 15551 R6xx/r7xx: reduce the amount of default state setup 15552 15553 Most of the default state setup was extraneous or duplicated 15554 in the accel ops. There's no need to reset all the consts 15555 everytime you reset the default state for example. This leads 15556 to a 10x reduction in default state. 15557 15558commit 7e6557b81026d8a4d6c837839e68c28b151f8a97 15559Author: Peter Hutterer <peter.hutterer@who-t.net> 15560Date: Thu Jul 16 11:33:57 2009 +1000 15561 15562 Update to xextproto 7.1 support. 15563 15564 DPMS header was split into dpms.h (client) and dpmsconst.h (server). Drivers 15565 need to include dpmsconst.h if xextproto 7.1 is available. 15566 15567 Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net> 15568 15569commit 95d431e8b260dd43cf6d93b90eecab1dc0ec7c26 15570Author: Dave Airlie <airlied@redhat.com> 15571Date: Wed Jul 15 12:07:01 2009 +1000 15572 15573 radeon: emit colorpitch relocs. 15574 15575 This causes relocs to be emitted for the colorpitch. 15576 15577 Shouldn't have no effect on current kernels, but will keep DDX 15578 going on kms kernels when tiling lands 15579 15580 Also contains a missing reloc space for r200 15581 15582commit 8312763977b22e108b7d003cbf3256e0165e4439 15583Author: Alex Deucher <alexdeucher@gmail.com> 15584Date: Tue Jul 14 18:13:27 2009 -0400 15585 15586 Add an r420 AGP quirk 15587 15588 fixes bug 22726 15589 15590commit 5473eeeb45468fd200fec49a8506281628a5e4b5 15591Author: Alex Deucher <alexdeucher@gmail.com> 15592Date: Tue Jul 14 14:05:13 2009 -0400 15593 15594 R3/4/5xx: only upload the bicubic texture once 15595 15596 Upload the bicubic texture once during textured video init 15597 rather than once per frame. Suggested by Michel Daenzer on 15598 IRC. 15599 15600commit 0485f27bc3d75cb6ab320e8164dbe6ea2713c78e 15601Author: Alex Deucher <alexdeucher@gmail.com> 15602Date: Tue Jul 14 13:11:12 2009 -0400 15603 15604 R3/4/5xx: fix bicubic Xv filtering with KMS changes 15605 15606 fixes bug 22730 15607 15608commit ca4a9efef987108266ae59afbb5142d4d2000c8e 15609Author: Alex Deucher <alexdeucher@gmail.com> 15610Date: Mon Jul 13 11:28:41 2009 -0400 15611 15612 Add configure option to disable kms support 15613 15614 Useful to building with mixed kms/non-kms setups 15615 15616commit 035e8d1d5593c12828bb079de4e663cf1b1f1674 15617Author: Edward O'Callaghan <eocallaghan@auroraux.org> 15618Date: Sun Jul 12 13:45:29 2009 +0200 15619 15620 Fix a warning 15621 15622 Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com> 15623 15624commit 76af48c43f829e7aebacc9f2a623823fa26ee22b 15625Author: Gaetan Nadon <memsize@videotron.ca> 15626Date: Wed Jul 8 09:58:56 2009 -0400 15627 15628 xf86-video-ati: use XORG_CHANGELOG macro to create ChangeLog. #22611 15629 15630 Build break: Makefile.am: command not found: git-log 15631 Adding the macro in configure.ac and use it in Makefile.am 15632 Refer to: https://bugs.freedesktop.org/show_bug.cgi?id=22611 15633 Tested: running autogen.sh, make and 'make dist' 15634 15635 Signed-off-by: Gaetan Nadon <memsize@videotron.ca> 15636 Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net> 15637 15638commit 43db263d301082e84e9bc304816bcbb206fe280e 15639Author: Alex Deucher <alexdeucher@gmail.com> 15640Date: Thu Jul 9 12:48:06 2009 -0400 15641 15642 RV280: another AGP quirk 15643 15644 fdo bug 12544 15645 15646commit 69b5e5496f10a9f566d2e563862c96cb41952eb6 15647Author: Alex Deucher <alexdeucher@gmail.com> 15648Date: Thu Jul 9 12:42:45 2009 -0400 15649 15650 RV280: Add an AGP quirk 15651 15652 lp bug 370205 via fdo bug 12544 15653 15654commit a6d6c900e60a3685ee9e93c34eb7f6d237c45fd8 15655Author: Alex Deucher <alexdeucher@gmail.com> 15656Date: Thu Jul 9 11:30:57 2009 -0400 15657 15658 RS740: fixup display latency setting as per rs690 15659 15660 Both chips program the same in this regard 15661 15662commit ea407570d39ace3162d372eda56bf791bfd80c24 15663Author: Dave Airlie <airlied@redhat.com> 15664Date: Thu Jul 9 10:46:23 2009 +1000 15665 15666 radeon: add shave support. 15667 15668 clean up output a lot 15669 15670commit 8c03c1fdb5ea35570064946557050c87ca30582a 15671Author: Alex Deucher <alexdeucher@gmail.com> 15672Date: Wed Jul 8 18:34:57 2009 -0400 15673 15674 R6xx/R7xx: fix hangs on x2 cards with PM options 15675 15676 Changing the PCIE lanes on x2 cards results in a hang, 15677 so for now, disable it. 15678 15679 Fixes fdo bug 22669 15680 15681commit 9645838c57f6b40837fdce23ce7f9faefb3d9966 15682Author: Michel Dänzer <daenzer@vmware.com> 15683Date: Wed Jul 8 19:18:02 2009 +0200 15684 15685 Further non-KMS fixes / cleanups. 15686 15687 Only compile tested. 15688 15689commit 2d86ec4ddb1fb83be95ed9cdcbabf4d568a6d7b6 15690Author: Alex Deucher <alexdeucher@gmail.com> 15691Date: Wed Jul 8 12:33:34 2009 -0400 15692 15693 Fix the build for the !XF86DRM_MODE case 15694 15695 compile tested only. 15696 15697commit a68afc2a80b8581cc3d67e9b5a95a10b0d4ead0e 15698Author: Michel Dänzer <daenzer@vmware.com> 15699Date: Wed Jul 8 09:15:28 2009 +0200 15700 15701 Fix RADEON_TRACE_FALL build. 15702 15703commit 9ca0aeaad01b7c682396b30af48668f287a59d6e 15704Author: Michel Dänzer <daenzer@vmware.com> 15705Date: Wed Jul 8 09:13:36 2009 +0200 15706 15707 Pass on CS relocation offset values. 15708 15709 Not actually used ATM, but just in case... 15710 15711commit f718dd794477fbdd5eeb4183f43ff25517c63034 15712Author: Michel Dänzer <daenzer@vmware.com> 15713Date: Wed Jul 8 09:11:18 2009 +0200 15714 15715 Don't set EXA UploadTo/DownloadFromScreen hooks when they'd always return FALSE. 15716 15717commit 42e76d0b24b76fbcc6b08a4ea5003dbec368645c 15718Author: Michel Dänzer <daenzer@vmware.com> 15719Date: Wed Jul 8 08:51:10 2009 +0200 15720 15721 Don't always flush CS when switching from 2D to 3D. 15722 15723 This works fine here and doubles x11perf -aa10text scores, but may cause 15724 problems on other setups according to Dave. Let's see... 15725 15726commit 94fe049c35187ddc3ec0472e9309c3a22879f415 15727Author: Michel Dänzer <daenzer@vmware.com> 15728Date: Wed Jul 8 08:36:22 2009 +0200 15729 15730 KMS CS fixes, take two. 15731 15732 Avoid flushing CS in the middle of (setting up state for) an operation, 15733 properly finish/restart the operation. 15734 15735commit 0519f15af2fe5e7e5a6c94e7203a96fe363c643e 15736Author: Michel Dänzer <daenzer@vmware.com> 15737Date: Tue Jul 7 10:04:09 2009 +0200 15738 15739 Fix DRI2 with current xserver Git. 15740 15741commit 7ad11fec2579807371360cb680e593f723ebaa1a 15742Author: Dave Airlie <airlied@redhat.com> 15743Date: Mon Jul 6 18:43:08 2009 +1000 15744 15745 radeon: add make dist headers 15746 15747commit e68fbf743ec88e0fa3a9f1ffe83daf251e250b21 15748Author: Dave Airlie <airlied@redhat.com> 15749Date: Mon Jul 6 18:32:32 2009 +1000 15750 15751 git: fix git-log to git log 15752 15753commit 5eeb1fd19c6797cc76ce308570dffe34cb82b24d 15754Author: Dave Airlie <airlied@linux.ie> 15755Date: Mon Jul 6 17:45:50 2009 +1000 15756 15757 radeon: oops fix compile without libdrm 15758 15759commit 98824f37fc74dc1fec2677515d0dc3cfc263f00a 15760Author: Dave Airlie <airlied@redhat.com> 15761Date: Mon Jul 6 15:29:53 2009 +1000 15762 15763 radeon: move state changes into flush function 15764 15765commit 503b51343267248b4c6994b88238dd5384528f03 15766Author: Dave Airlie <airlied@redhat.com> 15767Date: Mon Jul 6 15:20:13 2009 +1000 15768 15769 radeon: taken from kms-support + CS fixes commit. 15770 15771 re-emit 2D state and reset 3D state emit on flush 15772 15773commit d78d50591851d14543f0935d051a59ef29751bbc 15774Author: Dave Airlie <airlied@redhat.com> 15775Date: Mon Jul 6 14:45:07 2009 +1000 15776 15777 radeon: port to new space checking in libdrm interface 15778 15779 This uses a new libdrm interface which shares code with mesa. 15780 15781 It also fixes the bo to flush when full instead of never flushing. 15782 15783 It survives gtkperf -a here which the driver didn't before now 15784 15785commit e1200cb89218930d01330ba0114e013438655cce 15786Author: Dave Airlie <airlied@redhat.com> 15787Date: Mon Jul 6 15:10:29 2009 +1000 15788 15789 Revert "KMS CS fixes." 15790 15791 This reverts commit 11cf8a28eb46f48d101888552266921bee8b63fc. 15792 15793 I'll bring this back in a few minutes, I've got a big change introducing 15794 a new cs space accounting scheme which this messes up and I think 15795 probably conflicts with what I've done. 15796 15797 I promise i'll go over this and pick the bits that are still useful :) 15798 15799commit 11cf8a28eb46f48d101888552266921bee8b63fc 15800Author: Michel Dänzer <daenzer@vmware.com> 15801Date: Sun Jul 5 15:14:59 2009 +0200 15802 15803 KMS CS fixes. 15804 15805 * Flush CS before running out of space. 15806 * Don't flush CS every time when switching between 2D and 3D engine, but make 15807 sure to re-initialize 2D/3D state after flushing. 15808 15809 Doesn't seem quite perfect yet (some rendering seems to get dropped on the floor 15810 sometimes), but this should fix a lot of CS failures people have been seeing. 15811 15812commit 8d2f712eaf1e569fd92bbe2db5aceb43b7b367d1 15813Author: Dave Airlie <airlied@linux.ie> 15814Date: Fri Jul 3 10:40:59 2009 +1000 15815 15816 radeon: don't set up overlay under KMS until we figure out how. 15817 15818commit 2a622168606d98c404708b879e6417c8f7d7305c 15819Author: Michel Dänzer <daenzer@vmware.com> 15820Date: Thu Jul 2 15:03:53 2009 +0200 15821 15822 Test BO pointer before passing it to radeon_bo_unref(). 15823 15824commit 36a4ea5a9c7062f4c8aa7480bb7488ce2062390f 15825Author: Michel Dänzer <daenzer@vmware.com> 15826Date: Thu Jul 2 14:52:44 2009 +0200 15827 15828 Warning fixes. 15829 15830commit 65465d56c1bf943113518a5b2f57b3170213bab9 15831Author: Dave Airlie <airlied@linux.ie> 15832Date: Thu Jul 2 18:38:38 2009 +1000 15833 15834 radeon: fix up configure check better 15835 15836commit ad17108b72c9adf773a122b4c054c7b92e6a8b9d 15837Author: Dave Airlie <airlied@linux.ie> 15838Date: Thu Jul 2 17:25:57 2009 +1000 15839 15840 radeon: fix it properly. 15841 15842 I blame the flu 15843 15844commit e1b9cb9ea665eaa84e64a386299efa0819d55df4 15845Author: Dave Airlie <airlied@redhat.com> 15846Date: Thu Jul 2 17:18:47 2009 +1000 15847 15848 radeon: fix compile without kms 15849 15850commit de427fac6912a625f7e4d8b982a1f7a1ba80dfbe 15851Author: Dave Airlie <airlied@redhat.com> 15852Date: Thu Jul 2 15:41:43 2009 +1000 15853 15854 radeon/dri2: fix dri2 - needs newest libdrm_radeon 15855 15856commit 2b671adba54ebff1756dd990123f4bade6e35d61 15857Author: Dave Airlie <airlied@redhat.com> 15858Date: Thu Jul 2 15:25:19 2009 +1000 15859 15860 radeon/dri2: call proper gem bo function 15861 15862commit 851b2b8cab883e4422745cca1827aa8ac9f193a3 15863Author: Dave Airlie <airlied@redhat.com> 15864Date: Thu Jul 2 15:17:13 2009 +1000 15865 15866 radeon/kms: add textured video BO size checks 15867 15868commit 9d092060236e8587697fdd57a7a01f3c0dddd680 15869Author: Dave Airlie <airlied@redhat.com> 15870Date: Thu Jul 2 15:17:03 2009 +1000 15871 15872 kms: drop DGA 15873 15874commit 207174ea964c561076bdd5297a2fb6f136b4b645 15875Author: Dave Airlie <airlied@redhat.com> 15876Date: Thu Jul 2 15:16:47 2009 +1000 15877 15878 radeon: fix CS imbalance 15879 15880commit 4d24a9d2c0be0c025266cc8cdf7e8d7bb065cdb1 15881Author: Dave Airlie <airlied@redhat.com> 15882Date: Thu Jul 2 14:55:45 2009 +1000 15883 15884 radeon: enable KMS build 15885 15886commit 9bcd3ae7966eb7a17b922bcd33e0d6f016245614 15887Author: Dave Airlie <airlied@redhat.com> 15888Date: Thu Jul 2 14:55:18 2009 +1000 15889 15890 radeon/kms: fixup 3d state emission for KMS 15891 15892 This fixes font rendering under KMS (should have looked at dmesg) 15893 15894commit ac1607bb23f0c7584648407a3ac3fb27142e186e 15895Author: Dave Airlie <airlied@redhat.com> 15896Date: Thu Jul 2 14:55:05 2009 +1000 15897 15898 kms: no need to reference fifo slots 15899 15900commit 1782ce28953184776c90eb1255208a3e0ad245f0 15901Author: Dave Airlie <airlied@redhat.com> 15902Date: Tue Jun 30 16:55:26 2009 +1000 15903 15904 radeon: add KMS support (still disabled) 15905 15906 This adds DRI2 + KMS + driver pixmaps support to the driver. 15907 15908 I've decided to just do a completely separate KMS driver file 15909 instead of hacking the crap out of radeon_driver.c. So now 15910 I do the KMS check in radeon_probe.c time and set the DDX 15911 pointed up to a completely different set at this stage. 15912 15913 This avoids a lot of if (kms) type crap in the code at 15914 the expense of making sure we make changes to both files 15915 if necessary. 15916 15917 This code is still disabled in configure.ac as I broke EXA composite 15918 rendering somehow in KMS mode 15919 15920commit ac4bd24a66c1bdda0293f770a3f891e2b88cc8ee 15921Author: Alex Deucher <alexdeucher@gmail.com> 15922Date: Tue Jun 30 16:00:06 2009 -0400 15923 15924 EXA: fix the build on servers without EXA_HANDLES_PIXMAPS 15925 15926commit 49a0ca19ee4109cdd39a43fbf4bedf3f0557d6ec 15927Author: Michel Dänzer <daenzer@vmware.com> 15928Date: Tue Jun 30 17:22:50 2009 +0200 15929 15930 Warning fixes. 15931 15932commit e932836691aeaec37794fdaed2dabb22710fd171 15933Author: Dave Airlie <airlied@redhat.com> 15934Date: Tue Jun 30 16:24:37 2009 +1000 15935 15936 radeon: initial preparation for kms patch. 15937 15938 This patch contains most of the changes to the EXA and texture video 15939 accel code. 15940 15941 It adds a few bits of pixmap support but doesn't actually do anything 15942 useful KMS yet. 15943 15944 Testing this should not have any regressions over what we have already, 15945 biggest worries are r6xx, I've fixed a textured video one, but no idea 15946 what other might lurk 15947 15948 It won't build against libdrm radeon yet either 15949 15950commit bb04b450ed00ca4b1aa44c33085567d47b33b547 15951Author: Fredrik Höglund <fredrik@kde.org> 15952Date: Mon Jun 29 22:38:26 2009 +0200 15953 15954 Check if the composite op is supported in R200CheckComposite. 15955 15956commit f39cafc5c4a93a16ac93756473ebb1f33c7881db 15957Author: Alex Deucher <alexdeucher@gmail.com> 15958Date: Sun Jun 28 20:42:49 2009 -0400 15959 15960 radeon: call hdmi-b connectors dvi in randr 15961 15962 Technically HDMI-B is a glorfied DL DVI, but this is confusing 15963 to users, so call it DVI when reporting randr outputs. This 15964 allows us to remove two quirks where the bios was actually 15965 correct. 15966 15967commit 407e676b04117c7ca3dcc0cc072f3552b03c64e5 15968Author: Roland Scheidegger <sroland@tungstengraphics.com> 15969Date: Sun Jun 28 16:54:32 2009 -0400 15970 15971 R6xx/R7xx: fix pixel centers 15972 15973 Make sure we are using GL mode (centers at 0.5) rather 15974 than D3D mode (centers at 0.0). This also fixes fdo 15975 bug 21963. 15976 15977commit 68001981f22173ff949720055dba89291f284474 15978Author: Alex Deucher <alexdeucher@gmail.com> 15979Date: Sun Jun 28 16:20:10 2009 -0400 15980 15981 Add some missing M96 pci ids 15982 15983 fixes fdo bug 22404 15984 15985commit 06d5d8d7fd92b58b2467c782e57a80a3520a2c90 15986Author: Adam Jackson <ajax@redhat.com> 15987Date: Thu Jun 25 13:51:36 2009 -0400 15988 15989 Add a (harmlessly) missing break statement. 15990 15991commit 7d1d968fdc8de1e0ab0be916a1f3a4a398af47aa 15992Author: Adam Jackson <ajax@redhat.com> 15993Date: Thu Jun 25 13:49:18 2009 -0400 15994 15995 Cosmetic cleanup to digital output detection 15996 15997 No functional change, but this logic would need extending anyway for 15998 DisplayID, so unify it now. 15999 16000commit 96581746eea2f4e29a04f9266ec9649f25480424 16001Author: Adam Jackson <ajax@redhat.com> 16002Date: Wed Jun 24 17:05:47 2009 -0400 16003 16004 atom: Fix phantom VGA connector when HDMI-B present. 16005 16006 HDMI-B is really a glorified DL-DVI connector, the analog pins work fine. 16007 16008commit 496adc4dc7cf53fc6acfecb1158f1491d5f36168 16009Author: Michel Dänzer <daenzer@vmware.com> 16010Date: Fri Jun 19 14:37:01 2009 +0200 16011 16012 Revert "Build fixes for -Werror that aren't ready for prime time yet." 16013 16014 This reverts commit ab20ea1530ec8a3a9fd4089775c4e157e50fe1bf. 16015 16016 Whoops, pushed wrong branch... 16017 16018commit ab20ea1530ec8a3a9fd4089775c4e157e50fe1bf 16019Author: Michel Dänzer <daenzer@vmware.com> 16020Date: Sun Jun 14 16:50:07 2009 +0200 16021 16022 Build fixes for -Werror that aren't ready for prime time yet. 16023 16024commit 77e3537d312175a25f0e21cc07c3a96f78c3b35a 16025Author: Markus Gapp <markus.gapp@gmx.net> 16026Date: Sat Jun 13 11:51:29 2009 -0400 16027 16028 Add quirk for asus hd3450 board 16029 16030 Fixes bug 22266 16031 16032commit e8e0bb1ba29761c94d473e42ca6ada161cfff557 16033Author: Alex Deucher <alexdeucher@gmail.com> 16034Date: Fri Jun 12 11:07:43 2009 -0400 16035 16036 Fix count in 3d state setup after 43374c7420e378918bec062f4cbd581f16adb6f0 16037 16038commit 808c90a24c48da7fa97e15e2f12be5bb8fd8cc96 16039Author: Alex Deucher <alexdeucher@gmail.com> 16040Date: Thu Jun 11 16:09:40 2009 -0400 16041 16042 Only enable frac fb divs on rs600/rs690/rs740 for now 16043 16044 fixes bug 21413 16045 16046commit d095bb3157faaee10b0130ea9b1032e0b533aaee 16047Author: Alex Deucher <alexdeucher@gmail.com> 16048Date: Thu Jun 11 12:34:15 2009 -0400 16049 16050 Re-enable frac feedback dividers on AVIVO hardware 16051 16052 Preferring a freq <= target frequency seems to fix the 16053 previous issues with frac fb divs. Appears to work fine 16054 on all hardware I've tested on. See fdo bug 22229 for 16055 details. Should fix bugs: 22229, 21553, 21413, and 16056 possibly other ones related to unstable images at certain 16057 resolutions. 16058 16059commit 55fbdbae83d1563b472f49d0436c9298e390be66 16060Author: Alex Deucher <alexdeucher@gmail.com> 16061Date: Thu Jun 11 12:31:53 2009 -0400 16062 16063 Add PLL flag to prefer frequencies <= the target freq 16064 16065 This appears to be needed when using fractional feedback 16066 dividers. Based on a patch from Tom Hirst. See fdo 16067 bug 22229 for more details. 16068 16069commit 43374c7420e378918bec062f4cbd581f16adb6f0 16070Author: Alex Deucher <alexdeucher@gmail.com> 16071Date: Tue Jun 9 14:39:54 2009 -0400 16072 16073 r1xx/r2xx: clip 3D rendering to destination buffer 16074 16075commit 74b0a38c75c1d6a6344015d32d84b43d9e80df1d 16076Author: Alex Deucher <alexdeucher@gmail.com> 16077Date: Tue Jun 9 14:26:21 2009 -0400 16078 16079 Consolidate indentical RADEONDoneSolid/RADEONDoneCopy functions 16080 16081commit f6dff8836729cc9d29264080c9cfa0e1df6182cf 16082Author: Alex Deucher <alexdeucher@gmail.com> 16083Date: Tue Jun 9 13:31:15 2009 -0400 16084 16085 r3xx/r4xx/r5xx: clip 3D rendering to destination buffer 16086 16087commit 48dbf09e1309c815ac8b733145a0310083a29933 16088Author: Alex Deucher <alexdeucher@gmail.com> 16089Date: Tue Jun 9 13:12:40 2009 -0400 16090 16091 r3xx/r4xx/r5xx: refactor exa prepare composite 16092 16093 - consolidate code duplicated between r3xx/r4xx and r5xx 16094 - make pixel shader setup easier to read and less 16095 redundant 16096 16097commit ec1a3ccdf61c2226617b571cc4a209a542e9d7e3 16098Author: Alex Deucher <alexdeucher@gmail.com> 16099Date: Mon Jun 8 15:31:57 2009 -0400 16100 16101 R6xx/r7xx: don't clamp tex coords in composite VS 16102 16103 Fixes broken repeat modes. see fdo bugs 21818, 22139 16104 16105commit 32c218c5633fb7ffeb19e717bf645c1b25dd4ef5 16106Author: Alan Coopersmith <alan.coopersmith@sun.com> 16107Date: Thu Jun 4 07:23:03 2009 -0700 16108 16109 Add link to patch submission instructions to README 16110 16111 Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com> 16112 16113commit 39f736d6cb65fc7c3fd769478023fc87c1af5979 16114Author: Alex Deucher <alexdeucher@gmail.com> 16115Date: Thu Jun 4 09:57:22 2009 -0400 16116 16117 atom: connector quirk 16118 16119 Some r6xx cards have a VGA and an HDMI port with 16120 a shared ddc line listed as an HDMI port with both 16121 analog and digital encoders. 16122 16123 Fixes bug 19943 16124 16125commit 7599dc40855c33a5fbd8e9bbc2b4cd62752fb7df 16126Author: Alex Deucher <alexdeucher@gmail.com> 16127Date: Fri May 29 16:29:50 2009 -0400 16128 16129 Add RV740 (HD4770) support 16130 16131 Verified by ernstp on IRC 16132 16133commit ee0424078a3ea26f58c2998c9e58e065fb63f1a0 16134Author: Alex Deucher <alexdeucher@gmail.com> 16135Date: Wed May 27 18:37:58 2009 -0400 16136 16137 radeon: remove dead code 16138 16139commit b34df233115c0d82d7bcf82e041afbc55981ce82 16140Author: Alex Deucher <alexdeucher@gmail.com> 16141Date: Tue May 26 17:47:01 2009 -0400 16142 16143 R3xx/R4xx: increase EXA render target limits 16144 16145 now that we are using 1/12 subpixel mode 16146 16147commit 1457bc089d1395d18cf582a4d52444005f013753 16148Author: Matthieu Herrb <matthieu.herrb@laas.fr> 16149Date: Tue May 26 17:21:07 2009 -0400 16150 16151 radeon: fix segfault if Option "NoAccel" is set 16152 16153commit 4079f97958cf1d5fd76a83aa61b93d6289bb9b1e 16154Author: Michel Dänzer <daenzer@vmware.com> 16155Date: Sat May 23 13:51:00 2009 +0200 16156 16157 Bump version to mark Git snapshot. 16158 16159 The fourth version component isn't visible in the log file. 16160 16161commit 88a9e98341d96e5e7f48b69aed597d1bada6313a 16162Author: Alex Deucher <alexdeucher@gmail.com> 16163Date: Tue May 19 17:11:34 2009 -0400 16164 16165 Pre-atom: Improve engine clock setup code 16166 16167 Also remove SetMemoryClk() for pre-atom cards for now 16168 as it requires quite a bit more asic specific work. 16169 To set the mclk we'll need to use the mem reset/dll tables 16170 in the bios. 16171 16172commit 1429133ffdbfe046249fcaaa4cbee432a1012e55 16173Author: Michel Dänzer <daenzer@vmware.com> 16174Date: Mon May 18 17:59:00 2009 +0200 16175 16176 EXA: Set new EXA_SUPPORTS_OFFSCREEN_OVERLAPS flag when we have a 2D engine. 16177 16178commit e9cca1e7dc521d3756bb2d97f1d703b9ad65cb47 16179Author: Alex Deucher <alexdeucher@gmail.com> 16180Date: Mon May 18 10:48:25 2009 -0400 16181 16182 r4xx: Add some missing pci ids 16183 16184 Noticed by Romain in bug 21798 16185 16186commit 76fffd51027820b3a6ea10de2be67d5fd4f3e843 16187Author: Alex Deucher <alexdeucher@gmail.com> 16188Date: Fri May 15 16:35:13 2009 -0400 16189 16190 R100: switch to normalized texture coordinates 16191 16192 for consistency with the other asics 16193 16194commit 1a1eef31e36f98b7a30e27e567c1b66e0e4cf7dc 16195Author: Alex Deucher <alexdeucher@gmail.com> 16196Date: Thu May 14 15:48:54 2009 -0400 16197 16198 radeon: try harder when checking if a card is posted 16199 16200 fixes debian bug 524280, possibly others. 16201 16202commit 22e39392297fa11003df90c175db3c449d8f9853 16203Author: Michel Dänzer <daenzer@vmware.com> 16204Date: Thu May 14 10:24:07 2009 +0200 16205 16206 Pre-R600 EXA: Fix coordinate limits off-by-one error. 16207 16208 Fixes http://bugs.freedesktop.org/show_bug.cgi?id=21598 . 16209 16210commit fa09b058c7a17689989e600ffd465856a058579d 16211Author: Alex Deucher <alexdeucher@gmail.com> 16212Date: Wed May 13 16:10:40 2009 -0400 16213 16214 R6xx/R7xx Xv: normalize texture coordinates in the vertex shader 16215 16216commit 026b6f820d6caea17d2a082193e850713d5770a8 16217Author: Alex Deucher <alexdeucher@gmail.com> 16218Date: Wed May 13 15:48:32 2009 -0400 16219 16220 R6xx/R7xx: do EXA transforms in the vertex shader 16221 16222commit cd89241396d1931b04cfbdd8d553be16dbf9c360 16223Author: Alex Deucher <alexdeucher@gmail.com> 16224Date: Tue May 12 17:30:02 2009 -0400 16225 16226 R3xx-R5xx: do EXA transforms in the vertex shader 16227 16228commit 15ee78d37f8b64c3e6d234e7ab37a31e82327b6b 16229Author: Alex Deucher <alexdeucher@gmail.com> 16230Date: Mon May 11 02:22:13 2009 -0400 16231 16232 RV770: add missing pci id 16233 16234 Fixes bug 21209 16235 16236commit 5ff51ea83f9fdd836e830c77e3561eaef4b83183 16237Author: Alex Deucher <alexdeucher@gmail.com> 16238Date: Fri May 8 13:19:26 2009 -0400 16239 16240 R7xx: add support for reading disabled bios images 16241 16242commit 76fdae9018330aa5af0d57339baff3ca4834e1b5 16243Author: Alex Deucher <alexdeucher@gmail.com> 16244Date: Thu May 7 02:59:32 2009 -0400 16245 16246 R3xx/R4xx tex vid: no need to check if IS_R300_3D 16247 16248 function is r3xx/r4xx specific already 16249 16250commit 23a8ada06a5b1ccc975e89f129c00382ecbf29df 16251Author: Alex Deucher <alexdeucher@gmail.com> 16252Date: Thu May 7 02:34:22 2009 -0400 16253 16254 R3xx/R4xx tex vid: increase the guardband limit for rendering with a tri 16255 16256 limit goes from 2880 to 4021 now that we've switched to 1/12 subpixel mode. 16257 16258 This allows us to render with a clipped tri at higher resolutions preventing 16259 diagonal tearing. 16260 16261commit 0c3541f7013c53fb8096ada63cc7b3577524aa4c 16262Author: Alex Deucher <alexdeucher@gmail.com> 16263Date: Wed May 6 12:24:35 2009 -0400 16264 16265 AVIVO: set default cursor mode in cursor init 16266 16267 Should fix bug 21557 16268 16269commit da58e351b6398fa18b1d9c1a57a6e49b285f610f 16270Author: Alex Deucher <alexdeucher@gmail.com> 16271Date: Mon May 4 16:03:53 2009 -0400 16272 16273 AVIVO: fix cursoe corruption bands for real 16274 16275 Don't have to leave both cursors enabled, just have to use 16276 the same mode for both cursors whether or not they are enabled. 16277 16278commit 2c8e130f73c680d4a7381b2ef37982b82c6ee478 16279Author: Alex Deucher <alexdeucher@gmail.com> 16280Date: Mon May 4 13:40:17 2009 -0400 16281 16282 radeon: switch to EXA by default 16283 16284commit 8155a62df6817358a12cc2d613c1470ca55996ef 16285Author: Alex Deucher <alexdeucher@gmail.com> 16286Date: Mon May 4 12:57:36 2009 -0400 16287 16288 AVIVO: move cusor offscreen when disabling 16289 16290 Adapted from Yang's patch. Setting size to zero is 16291 actually a 1x1 cursor. 16292 16293commit a1c64ea5224009779ccad66b0f84d861eae966ac 16294Author: Alex Deucher <alexdeucher@gmail.com> 16295Date: Sun May 3 19:42:44 2009 -0400 16296 16297 r3xx-r5xx: switch to 1/12 subpixel precision 16298 16299 - based on similar patch in Jerome's cs ddx tree 16300 - also fix clipping offsets 16301 - should eventually allow for 4k render targets 16302 - mesa driver uses 1/12 mode, this avoids changing the 16303 subpixel mode when switching between ddx and mesa 16304 16305commit 3880bc3c95764acccd62cfa438932458df95daf6 16306Author: Alex Deucher <alexdeucher@gmail.com> 16307Date: Sun May 3 19:01:28 2009 -0400 16308 16309 Remove some old debugging info 16310 16311commit 3362381aa3a97c1da564f8a479ed4e7b0d6368b8 16312Author: Alex Deucher <alexdeucher@gmail.com> 16313Date: Sun May 3 18:52:00 2009 -0400 16314 16315 r2xx-r4xx: fix typo in last i2c fix and clarify hw i2c pin sel 16316 16317commit e01b8dc827ce1913927e1a4fad05a99a2dd9e805 16318Author: Alex Deucher <alexdeucher@gmail.com> 16319Date: Sun May 3 13:09:39 2009 -0400 16320 16321 RS690: gpio/connector fixes 16322 16323 Fixes bug 21521 16324 16325commit 1f70c9f05df9017d87b37f887e1eccd6d0568a02 16326Author: Alex Deucher <alexdeucher@gmail.com> 16327Date: Thu Apr 30 19:18:13 2009 -0400 16328 16329 R3xx-R5xx: don't set TX_OFFSET_RS in RS_INST_COUNT 16330 16331 Isn't necessary and seems to cause problems for RS690 users 16332 16333commit de17d36c1f710305870e3c636cae3f742f1cb6d6 16334Author: Alex Deucher <alexdeucher@gmail.com> 16335Date: Thu Apr 30 19:13:09 2009 -0400 16336 16337 AVIVO: first pass at display watermark setup 16338 16339 This helps with flickering and blanking when 16340 there is contention for MC bandwidth. 16341 16342commit 7d9f643ae3d07e51e644a5979ca90bc2c102bc89 16343Author: Alex Deucher <alexdeucher@gmail.com> 16344Date: Wed Apr 29 12:38:15 2009 -0400 16345 16346 radeon: fix several dpms issues 16347 16348 - atom dpms was unblanking for standby/suspend 16349 - return if r600+ in radeon_crtc_modeset_ioctl() 16350 - remove seprate standby/suspend handling in legacy 16351 crtc dpms. we turn the outputs off, so turn the 16352 crtcs off too. 16353 - disable the crtcs in legacy crtc dpms 16354 - move radeon_crtc_modeset_ioctl() calls to radeon_crtc_dpms() 16355 so they will get called for both atom and legacy paths 16356 16357 Should fix bug 21321 16358 16359commit 35183ef528411eb0122cf48550d93f921d291d12 16360Author: Alex Deucher <alexdeucher@gmail.com> 16361Date: Wed Apr 29 01:55:10 2009 -0400 16362 16363 AVIVO: better fix for cursor flickering/corruption issues 16364 16365 Should prevent ghost cursor from being left on screen. 16366 16367commit 53e75cbd0ba8f39799d7b776cdc59d60aa2a6916 16368Author: Alex Deucher <alexdeucher@gmail.com> 16369Date: Wed Apr 29 01:05:31 2009 -0400 16370 16371 R6xx/r7xx: fix CURSOR_SWAPPING_* macros 16372 16373 r6xx/r7xx have different swapper regs 16374 16375commit 4e948e2aaa59267a9aa797396add2c6b7b4c1e7c 16376Author: Alex Deucher <alexdeucher@gmail.com> 16377Date: Tue Apr 28 19:28:25 2009 -0400 16378 16379 AVIVO: fix cursor corruption/flickering issues 16380 16381 When both crtcs are enabled, both cursors need to be enabled 16382 or you get corruption bands. 16383 16384 Thanks to Yang Zhao for figuring this out. 16385 16386commit 4d0beaf02b28d5a5dac4ff6053593a2f979f92ec 16387Author: Alex Deucher <alexdeucher@gmail.com> 16388Date: Tue Apr 28 12:33:24 2009 -0400 16389 16390 RV410: SE variants only have 1 quad pipe 16391 16392 Should fix EXA corruption with the 3D engine. 16393 Mesa and drm patches forthcoming. 16394 16395 Reported by Kano on IRC. 16396 16397commit abd565ae71d9844d5afcd658628ca7778ce26ab1 16398Author: Alex Deucher <alexdeucher@gmail.com> 16399Date: Mon Apr 27 14:21:38 2009 -0400 16400 16401 AVIVO: disable fractional feedback divs 16402 16403 Doesn't seem to work well in all cases. Needs more investigation. 16404 fixes bug 21413 16405 16406commit 93626b1277f57bd559cf71b263221209b02e0046 16407Author: Alex Deucher <alexdeucher@gmail.com> 16408Date: Mon Apr 27 03:35:24 2009 -0400 16409 16410 AVIVO: make sure cursor width isn't negative 16411 16412commit 32af0f097bccc4499d67e969d46ba35e7487c5e2 16413Author: Dave Airlie <airlied@linux.ie> 16414Date: Sat Apr 25 19:39:34 2009 +1000 16415 16416 Revert "r300/r500: set R clamp mode." 16417 16418 This reverts commit dab3d61619d91ae6a3da3e9b29b8258313653813. 16419 16420commit b2c8f57bccbd46164907049a26ad2dd209d7b108 16421Author: Dave Airlie <airlied@linux.ie> 16422Date: Sat Apr 25 19:39:19 2009 +1000 16423 16424 Revert "r3xx/r5xx: fix typo in last commit" 16425 16426 This reverts commit 1a7b982457ead5f21c6e62a409ca0dbf86005607. 16427 16428commit 253731dce6ab25122f2eda559233c098536b7143 16429Author: Dave Airlie <airlied@linux.ie> 16430Date: Sat Apr 25 19:39:05 2009 +1000 16431 16432 Revert "radeon: attempt to fix R clamping for repeat again" 16433 16434 This reverts commit 5998f262a52cb85b334fcc1fe24c7b2ae474ce93. 16435 16436commit 5998f262a52cb85b334fcc1fe24c7b2ae474ce93 16437Author: Dave Airlie <airlied@linux.ie> 16438Date: Fri Apr 24 17:30:57 2009 +1000 16439 16440 radeon: attempt to fix R clamping for repeat again 16441 16442commit 7ed27b926f28a1f088eace5034a7d9985f13752d 16443Author: Alex Deucher <alexdeucher@gmail.com> 16444Date: Thu Apr 23 14:06:06 2009 -0400 16445 16446 r2xx/r3xx/r4xx: further i2c fixups 16447 16448 - hw i2c engine has pin selection on r2xx/r2xx/r3xx chips 16449 - also switch hw i2c pin sel for external tmds 16450 16451commit 0da80f34b5aa5999d2030ffbc5187328fa4e4ae5 16452Author: Alex Deucher <alexdeucher@gmail.com> 16453Date: Wed Apr 22 18:42:36 2009 -0400 16454 16455 radeon: minor cleanup and warning fix 16456 16457commit 1a7b982457ead5f21c6e62a409ca0dbf86005607 16458Author: Alex Deucher <alexdeucher@gmail.com> 16459Date: Thu Apr 23 10:09:22 2009 -0400 16460 16461 r3xx/r5xx: fix typo in last commit 16462 16463 Noticed by osiris on IRC 16464 16465commit dab3d61619d91ae6a3da3e9b29b8258313653813 16466Author: Dave Airlie <airlied@redhat.com> 16467Date: Thu Apr 23 15:46:44 2009 +1000 16468 16469 r300/r500: set R clamp mode. 16470 16471 On rs690 without this set, we appear to end up fetching from well 16472 outside the texture sometimes into memory that isn't mapped. 16473 16474 This "shouldn't" have any effect. 16475 16476commit bbfd4625478e37d10fd58b23f46153f349370a52 16477Author: Alex Deucher <alexdeucher@gmail.com> 16478Date: Wed Apr 22 14:47:16 2009 -0400 16479 16480 radeon PM: update quirks 16481 16482commit e3c564114eea4d032f920fcebd4040f1934c21b0 16483Author: Alex Deucher <alexdeucher@gmail.com> 16484Date: Wed Apr 22 14:31:44 2009 -0400 16485 16486 AVIVO: fix for cursor corruption issue 16487 16488commit 24e4b73b4fbbb2c790e6120ede3caaa4e7e58359 16489Author: Alex Deucher <alexdeucher@gmail.com> 16490Date: Wed Apr 22 11:42:18 2009 -0400 16491 16492 radeon pll: add support for fractional feedback divs 16493 16494 Allows us to hit dot clocks much closer, especially on 16495 chips with non-27 Mhz reference clocks like most IGP chips. 16496 16497 This should fix most flickering and blanking problems with 16498 non-exact dot clocks. 16499 16500commit efa0825a86a8dc0f03ebb42c576ed26189e9d4bb 16501Author: Alex Deucher <alexdeucher@gmail.com> 16502Date: Tue Apr 21 13:51:11 2009 -0400 16503 16504 Tex vid: fix top clipping 16505 16506commit aa59b5b55e6ac64940fb3f0fa61f7b8fe36ba745 16507Author: Alex Deucher <alexdeucher@gmail.com> 16508Date: Tue Apr 21 12:06:24 2009 -0400 16509 16510 RS690: add quirk for acer board 16511 16512 thanks to Gino Badouri for testing 16513 16514commit 0e5164d968316bbb59b4484a7df087854738068f 16515Author: Alex Deucher <alexdeucher@gmail.com> 16516Date: Tue Apr 21 04:47:43 2009 -0400 16517 16518 Tex vid: rework part 2 16519 16520 - use pPriv->w/h directly for tex coords 16521 - take src x/y offset into account when calculating tex coords 16522 - when copying data, only optimize for y clipping. In order 16523 to deal with the x clipping optimization, the copy routines 16524 or tex coords would need to be fixed. This should fix clipping 16525 problems with the current code. 16526 16527commit 06a1b4169b506f69dd807380d9ce7210b3a3bc23 16528Author: Alex Deucher <alexdeucher@gmail.com> 16529Date: Tue Apr 21 01:21:51 2009 -0400 16530 16531 R4xx: avoid a PM related hang 16532 16533commit 228b9ac87668241b09980fca17e9d866891285e6 16534Author: Alex Deucher <alexdeucher@gmail.com> 16535Date: Tue Apr 21 01:02:34 2009 -0400 16536 16537 radeon PM: add some quirks 16538 16539commit b6febf7368d3df3f1d3fee6a87aa71c54e77cdcc 16540Author: Alex Deucher <alexdeucher@gmail.com> 16541Date: Mon Apr 20 18:04:18 2009 -0400 16542 16543 Tex vid: cleanup and unify common code 16544 16545commit d096d1189c8dc5e19f68cc605ccffe1a8fef048b 16546Author: Alex Deucher <alexdeucher@gmail.com> 16547Date: Mon Apr 20 13:05:18 2009 -0400 16548 16549 Endian fixes for object table 16550 16551commit 101ce28ec4af696549b7d6defebb5b4ad412176c 16552Author: Alex Deucher <alexdeucher@gmail.com> 16553Date: Sun Apr 19 12:13:18 2009 -0400 16554 16555 Remove old messages 16556 16557 R5xx/r6xx support isn't really experimental anymore 16558 16559commit db177c70ce88af19c8e05eb916a15f3e124876eb 16560Author: Alex Deucher <alexdeucher@gmail.com> 16561Date: Fri Apr 17 01:05:15 2009 -0400 16562 16563 Update Xv info in man page 16564 16565commit 00266177bed2dc8693df497ca3ec19f2dc4adc05 16566Author: Alex Deucher <alexdeucher@gmail.com> 16567Date: Thu Apr 16 20:33:28 2009 -0400 16568 16569 R3xx/R5xx: only apply Xv attributes if bicubic is disabled 16570 16571 Provides consistent output 16572 16573commit 85323a7f84381fef7fad20c7f7ec601637af9aa7 16574Author: Alex Deucher <alexdeucher@gmail.com> 16575Date: Thu Apr 16 11:10:51 2009 -0400 16576 16577 R3xx/R4xx: set tex caching for Y texture when doing planar rendering 16578 16579 Doesn't affect performance, but docs indicate its the right 16580 thing to do. 16581 16582commit 5ea5df22c038fc8f00984acc760e9d8c962bf902 16583Author: Alex Deucher <alexdeucher@gmail.com> 16584Date: Wed Apr 15 20:26:34 2009 -0400 16585 16586 Tex vid: remove remnants of XV_HWPLANAR 16587 16588 no longer needed as bicubic is the only thing that uses 16589 the old csc code. 16590 16591commit 9091b3f5f13dbea83ffd89679dac600e9f280bb2 16592Author: Alex Deucher <alexdeucher@gmail.com> 16593Date: Wed Apr 15 20:07:42 2009 -0400 16594 16595 R3xx/R4xx: fix up planar shader 16596 16597 We were overwriting the coord fetch address with the first 16598 tex fetch. Seemed to work however, luck I guess. Reorder 16599 the fetches to write to temp0 last. 16600 16601commit 14c13faeb9f9b7717a25fcc1ca97d46cc6ee0031 16602Author: Alex Deucher <alexdeucher@gmail.com> 16603Date: Wed Apr 15 19:53:12 2009 -0400 16604 16605 R5xx: add shader-based csc 16606 16607 - native planar support 16608 - Xv attributes 16609 16610commit 832efc7b90f5eb2da99512fcb902ab4838d2dcd1 16611Author: Alex Deucher <alexdeucher@gmail.com> 16612Date: Tue Apr 14 17:56:49 2009 -0400 16613 16614 R3xx/R4xx: Implement shader-based csc for packed formats 16615 16616commit 32625118c27041265d25811c00d25ab7e82fb340 16617Author: Alex Deucher <alexdeucher@gmail.com> 16618Date: Tue Apr 14 16:38:40 2009 -0400 16619 16620 tex vid: fix attribute setup typo for XV_COLORSPACE 16621 16622commit adf0912006b4f1597784dbfcc563d5c6d1c5667d 16623Author: Alex Deucher <alexdeucher@gmail.com> 16624Date: Tue Apr 14 16:16:33 2009 -0400 16625 16626 R6xx/R7xx: implement Xv attributes 16627 16628 - brightness, contrast, hue, etc. 16629 - TODO: implement gamma 16630 16631commit 8810fe92b5aed08888584c6914482586b59f71ab 16632Author: Alex Deucher <alexdeucher@gmail.com> 16633Date: Tue Apr 14 11:26:35 2009 -0400 16634 16635 R200: clean code, always use shader based csc 16636 16637 - consolidate common r2xx csc shader code 16638 - always use shader based csc for both packed and planar 16639 formats 16640 - always use native planar csc on r1xx 16641 16642commit 17685fefba68d188c7c0fe7a079180ec0722c046 16643Author: Alex Deucher <alexdeucher@gmail.com> 16644Date: Mon Apr 13 20:06:11 2009 -0400 16645 16646 Tex vid: general cleanup 16647 16648 - convert macros to more meaningful VTX_OUT_4 and VTX_OUT_6 16649 names to reflect that they actually do 16650 - fix indenting 16651 16652commit 093ab4c9a33b0b396b78c061c3321dc044bdccdc 16653Author: Alex Deucher <alexdeucher@gmail.com> 16654Date: Mon Apr 13 19:48:35 2009 -0400 16655 16656 R1xx: add support for native planar textured Xv 16657 16658commit ec0cb51df81c6c9a1de640d227fa9c9c33161083 16659Author: Alex Deucher <alexdeucher@gmail.com> 16660Date: Mon Apr 13 17:21:20 2009 -0400 16661 16662 R2xx tex vid: append verts for clip boxes 16663 16664 rather than sending a new draw packet for each rect 16665 16666commit fde075a30a8ee2c333aa1bbe8fbd177258b085ba 16667Author: Alex Deucher <alexdeucher@gmail.com> 16668Date: Mon Apr 13 17:13:51 2009 -0400 16669 16670 R1xx tex vid: append verts for clip boxes 16671 16672 rather than sending a new draw packet for each rect 16673 16674commit 12839fc17a2cca4ac14b9757bdaa63ba4679f96f 16675Author: Alex Deucher <alexdeucher@gmail.com> 16676Date: Mon Apr 13 17:04:31 2009 -0400 16677 16678 Tex vid: split by family 16679 16680commit a30737b337edb31528174b483c9094941a5d41bb 16681Author: Roland Scheidegger <sroland@tungstengraphics.com> 16682Date: Mon Apr 13 15:36:07 2009 -0400 16683 16684 r200/r300: implement brightness/contrast/hue/saturation/gamma controls for textured video 16685 16686 This implements 16687 contrast/brightness/hue/saturation controls for r200/r300 plus gamma (same 16688 gamma value for all channels used though separate values would be trivial) 16689 control for r300. 16690 Some issues left: 16691 - only r200/r300 16692 - still can't be combined with bicubic 16693 - controls will silently cease to work if the format used is packed and not 16694 planar (except for rv250) 16695 - gamma range is from 100 to 10000 corresponding to 0.1 and 10.0 like used in 16696 overlay. However, usable range is far smaller. Over 2.0 picture gets dark 16697 pretty quickly, and below 0.6 or so black seems to turn into purple (I've 16698 verified that even with gamma 1.0 black actually often seems to be RGB 1/0/1 so 16699 this explains this since that gets amplified by low gamma values - not sure if 16700 this is a rounding problem somewhere, bogus reference values or is somehow 16701 expected). 16702 - gamma adds a bit too many instructions for my taste (7) though the 16703 alternative (3 texture lookups + some swizzling instructions) doesn't seem any 16704 better. 16705 16706commit 9dd33cc8346073300312cd5f68de72d9bea1dd6b 16707Author: Dave Airlie <airlied@redhat.com> 16708Date: Fri Apr 17 10:30:21 2009 +1000 16709 16710 radeon: tweak ddc fix for all r300s. 16711 16712 Alex seems to think this might affect some other chips 16713 16714commit a846ec14dd65bb12f314f83ffb45c4c4be564a5e 16715Author: Dave Airlie <airlied@redhat.com> 16716Date: Fri Apr 17 10:24:15 2009 +1000 16717 16718 radeon: fix DDC on rv410 VGA ports for non-atom use. 16719 16720 The rv410 hw i2c block appears to hold down whatever i2c lines 16721 it is connected to when in reset, so when doing DDC change hw i2c to 16722 point at different i2c lines. 16723 16724commit efa1476fc0765b53298624198137dc5cc71d37a8 16725Author: Alex Deucher <alexdeucher@gmail.com> 16726Date: Thu Apr 16 16:43:02 2009 -0400 16727 16728 PM: fix reversed logic in pm block handler 16729 16730commit 2b95de17781959457a809c8fecc6bbb08336c83f 16731Author: Alex Deucher <alexdeucher@gmail.com> 16732Date: Thu Apr 16 12:52:41 2009 -0400 16733 16734 RS400/RS480: attempt to fix ClockGating on RS4xx chips 16735 16736commit c0215d2f2a27755fb04d20aaa4320a5e7da7ee04 16737Author: Alex Deucher <alexdeucher@gmail.com> 16738Date: Thu Apr 16 11:58:19 2009 -0400 16739 16740 RS690/RS740: fix connector enumeration in some cases 16741 16742 Should fix Masta-G's issue reported on IRC. 16743 16744commit 573339114660f06e5f257648ba413bab0d6083ce 16745Author: Alex Deucher <alexdeucher@gmail.com> 16746Date: Thu Apr 16 11:19:01 2009 -0400 16747 16748 R3xx/R4xx EXA: set explicit tex cache regions 16749 16750 When using a mask, set explicit cache partitions for 16751 each texture. Gives 1% performance improvement in 16752 x11perf with anti-aliased text. 16753 16754commit eea9800890b56bac9c07b7bd9c9e33fae2938af3 16755Author: Alex Deucher <alexdeucher@gmail.com> 16756Date: Thu Apr 16 10:41:28 2009 -0400 16757 16758 Revert "DCE3/3.2: cleanup dpms after previous changes" 16759 16760 This reverts commit f8c7d6a6162196a743f6885ecaf63ba50de1722a. 16761 16762 This is apparently still needed for some setups, however, I can't 16763 reproduce this locally anymore. 16764 16765commit 099cb27d760af61b967239d960bc77f71e828655 16766Author: Alex Deucher <alexdeucher@gmail.com> 16767Date: Wed Apr 15 16:23:01 2009 -0400 16768 16769 Change ForceLowPowerMode to 4 pcie lanes 16770 16771 This fixes corruption for some users 16772 16773commit 5abcea88d1f3307a7d08ae7757c4b535df1503bf 16774Author: Alex Deucher <alexdeucher@gmail.com> 16775Date: Wed Apr 15 15:52:04 2009 -0400 16776 16777 PM: don't touch clock gating regs unless ClockGating option is set 16778 16779 Fixes an M26 hang reported by tormod on IRC 16780 16781commit bae88f81e2a01c0d547132b5b91e717ff9d0497d 16782Author: Alex Deucher <alexdeucher@gmail.com> 16783Date: Wed Apr 15 14:38:35 2009 -0400 16784 16785 man page: fix missing break 16786 16787commit 4b3a3785f7b06205eeec6ca6b8913ecc2b94bd1f 16788Author: Alex Deucher <alexdeucher@gmail.com> 16789Date: Sat Apr 4 11:39:35 2009 -0400 16790 16791 Unify ForceLowPowerMode and DynamicPM Options 16792 16793 Force lower power mode and switch to a lower mode 16794 when idle. 16795 16796commit 63cb57c5eb0d6bc083f54eb100cb972b0e7bca69 16797Author: Alex Deucher <alexdeucher@gmail.com> 16798Date: Wed Apr 1 16:22:02 2009 -0400 16799 16800 Add DynamicPM Option 16801 16802 Dyanmically switch between power states. Switch to a low 16803 power state when the system is idle (DPMS off). 16804 16805commit a34a8b37afbea6ed4bf8ca42364195e174250c48 16806Author: Alex Deucher <alexdeucher@gmail.com> 16807Date: Wed Apr 1 15:23:17 2009 -0400 16808 16809 Set default low power PCIE lanes to 2 16810 16811 1 lane seems to cause occasional corruption when 16812 blitting to/from gart memory. 16813 16814commit 1f0dc778dc25f4f85fedd73c55c847cab2c79fc5 16815Author: Alex Deucher <alexdeucher@gmail.com> 16816Date: Wed Apr 1 15:12:27 2009 -0400 16817 16818 Add support for setting the number PCIE lanes 16819 16820commit 7e10b6222e8f44a3ecc6aaea55a7a7680d133bb3 16821Author: Alex Deucher <alexdeucher@gmail.com> 16822Date: Wed Apr 1 15:05:38 2009 -0400 16823 16824 Add ForceLowPowerMode Option 16825 16826 Force the chip to a low power mode at the expense 16827 of performance. 16828 16829commit 533d01a5933cd491bbc09cd463ea62475abf4bf2 16830Author: Alex Deucher <alexdeucher@gmail.com> 16831Date: Tue Mar 31 17:24:29 2009 -0400 16832 16833 radeon: Add functions to set sclk/mclk using atombios 16834 16835commit adb099409768e695b9928fa6aa5760f93dadd9af 16836Author: Alex Deucher <alexdeucher@gmail.com> 16837Date: Tue Mar 31 17:13:11 2009 -0400 16838 16839 radeon: Add functions to set sclk/mclk on r1xx-r4xx 16840 16841commit 68e2a959ccc3d1a5d0731f1b55fdf1b2412635b2 16842Author: Alex Deucher <alexdeucher@gmail.com> 16843Date: Tue Mar 31 17:02:45 2009 -0400 16844 16845 Add unified ClockGating Option 16846 16847 - replaces DynamicClocks Option as the name was misleading 16848 - unified interface for atom and com based bioses 16849 - fix up clock gating code for newer r3xx asics 16850 16851commit 336cee0e6d19068582b189b2c747d92cb2139d2f 16852Author: Dave Airlie <airlied@redhat.com> 16853Date: Mon Apr 13 18:11:48 2009 +1000 16854 16855 radeon: legacy bandwidth calcs for later r3xx/r4xx cards. 16856 16857 This is an attempt to rationalise the code using the register info 16858 files available to me here. 16859 16860 Further info is required: 16861 r350 check for stop_req > 15 then subtract 10: what other chips need it? 16862 get rs480 working instead of using magic. 16863 16864commit 367ba7af76e7f63f3186a6e8d9bbe21aa31fc6d1 16865Author: Alex Deucher <alexdeucher@gmail.com> 16866Date: Fri Apr 10 09:37:15 2009 -0400 16867 16868 radeon: fix tmds_pll randr property 16869 16870commit 15f1a2f31a7bc33f276a65283dbad0c106bee73e 16871Author: Alex Deucher <alexdeucher@gmail.com> 16872Date: Thu Apr 9 20:17:03 2009 -0400 16873 16874 radeon: remove old crtc size heuristics 16875 16876 IIRC, the old randr code used to use this to use for front buffer sizing, 16877 but it has since been changed. 16878 16879commit ce9077b2f5a5e749c0078dfb79d26ef8711a823a 16880Author: Alex Deucher <alexdeucher@gmail.com> 16881Date: Thu Apr 9 19:59:19 2009 -0400 16882 16883 R6xx/R7xx: cleanup macro 16884 16885commit 6532aeddacdeda3aa534264741c8648e1c449e76 16886Author: Michel Dänzer <daenzer@vmware.com> 16887Date: Thu Apr 9 09:41:56 2009 +0200 16888 16889 Clear the framebuffer before initializing modes. 16890 16891 Fixes garbage being visible shortly on server startup or when VT switching back 16892 to X. 16893 16894commit 0c19053975925fed637aaa41d440e0cab8f6b596 16895Author: Alex Deucher <alexdeucher@gmail.com> 16896Date: Wed Apr 8 13:34:23 2009 -0400 16897 16898 radeon: chip family typo 16899 16900commit 2888dd9ae9689b1cd72115dc0ceea1f5957299b0 16901Author: Alex Deucher <alexdeucher@gmail.com> 16902Date: Wed Apr 8 13:10:01 2009 -0400 16903 16904 R6xx/R7xx: set proper 3D client driver name for r600 16905 16906commit 61a083116a2c89e145aa19da951243536e07542e 16907Author: Alex Deucher <alexdeucher@gmail.com> 16908Date: Wed Apr 8 10:44:30 2009 -0400 16909 16910 Bump version post release 16911 16912commit 1b02b93895c31a0c9d641e47b46dce43b40edd97 16913Author: Alex Deucher <alexdeucher@gmail.com> 16914Date: Wed Apr 8 10:34:44 2009 -0400 16915 16916 Bump for release 16917 16918commit da74b94a5da8f0bcf4ae99b20e594d2f0f4fc6f6 16919Author: Alex Deucher <alexdeucher@gmail.com> 16920Date: Tue Apr 7 18:59:06 2009 -0400 16921 16922 ATOM: code cleanup 16923 16924 - remove cases that aren't possible (e.g., no TMDSA on DCE3 cards) 16925 - condense duplicate cases 16926 16927commit a24aa7250d09ae68ceaaba987314388b5e52096c 16928Author: Alex Deucher <alexdeucher@gmail.com> 16929Date: Tue Apr 7 18:52:08 2009 -0400 16930 16931 DCE3/3.2: further fixes 16932 16933 disable the dig transmitter and encoder before setup and enable 16934 16935commit 284e1943f6726ed08e796f99e49fbd90ac7fa8a4 16936Author: Alex Deucher <alexdeucher@gmail.com> 16937Date: Wed Apr 1 15:07:01 2009 -0400 16938 16939 Cast info->sclk to int 16940 16941commit b580e388e0e31843308c783c1ca859672dc6ca0e 16942Author: Alex Deucher <alexdeucher@gmail.com> 16943Date: Wed Apr 1 14:59:43 2009 -0400 16944 16945 Fix missing ')' in PCIE macro 16946 16947commit 8ec27be3c8043a6ddc816b0b1480dd19f8d73832 16948Author: Alex Deucher <alexdeucher@gmail.com> 16949Date: Tue Apr 7 11:33:44 2009 -0400 16950 16951 radeon: fix Xv vsync for multi-head 16952 16953commit f8c7d6a6162196a743f6885ecaf63ba50de1722a 16954Author: Alex Deucher <alexdeucher@gmail.com> 16955Date: Sat Apr 4 16:05:20 2009 -0400 16956 16957 DCE3/3.2: cleanup dpms after previous changes 16958 16959commit a707d355c3c6ff92252c5a060a1fc32d97547552 16960Author: Alex Deucher <alexdeucher@gmail.com> 16961Date: Sat Apr 4 15:20:01 2009 -0400 16962 16963 DCE3/3.2: fix up transmitter/encoder setup 16964 16965 Should fix bug 21050 16966 16967commit 215e12f9c0e8ac62c23af1add776ef88f9a0dc54 16968Author: Eduard Fuchs <edfuchs@uni-kassel.de> 16969Date: Thu Apr 2 12:36:59 2009 -0400 16970 16971 Enable byte swapping for r6xx/r7xx Hardware 16972 16973 Patch from Eduard Fuchs with some cleanup from me. 16974 16975 Tested at 32 bpp on MPC8641HPCN board (PowerPC) with 16976 HD2400 PCIe card 16977 16978commit 0ded9fab6bc3fecc1976ae8369b788659e4d7793 16979Author: Alex Deucher <alexdeucher@gmail.com> 16980Date: Thu Apr 2 12:09:00 2009 -0400 16981 16982 Add RV790 (HD 4890) support 16983 16984commit 0c6f20a9404c6a4210c5180bdca61c13fc186754 16985Author: Alex Deucher <alexdeucher@gmail.com> 16986Date: Tue Mar 31 15:51:11 2009 -0400 16987 16988 Add PCIE register accessors 16989 16990commit 229d4e20e5adcdf9207944c19ca721477936ff02 16991Author: Alex Deucher <alexdeucher@gmail.com> 16992Date: Tue Mar 31 15:39:42 2009 -0400 16993 16994 Unify mc_idle() 16995 16996commit f44b486848094b75479abef61c1aa075aede835c 16997Author: Alex Deucher <alexdeucher@gmail.com> 16998Date: Tue Mar 31 15:33:11 2009 -0400 16999 17000 COMBIS: Clarify case 0 for LCD DDC table 17001 17002commit e6f3c48899a92174a55f6fe7327412bc88039d87 17003Author: Alex Deucher <alexdeucher@gmail.com> 17004Date: Tue Mar 31 15:30:40 2009 -0400 17005 17006 COMBIOS: fix default sclk/mclk from bios 17007 17008commit 75bf5272936093bc024559e5744d97d8e79943b1 17009Author: Alex Deucher <alexdeucher@gmail.com> 17010Date: Tue Mar 31 15:17:37 2009 -0400 17011 17012 mac mini: 3rd time's the charm 17013 17014 fixes 10fbbac4bed59e12ad794ac873dd415d9ee4e340 17015 fixes bugs 20796 and 20979 17016 17017commit 10fbbac4bed59e12ad794ac873dd415d9ee4e340 17018Author: Alex Deucher <alexdeucher@gmail.com> 17019Date: Tue Mar 31 15:14:27 2009 -0400 17020 17021 Mac mini: fix connector setup for real 17022 17023 fixes f061308e7abcf93d1612681b935387fee23e0a24 17024 fixes bugs 20796 and 20979 17025 17026commit 8f2a03bfc86422669a46c2853050979f099a17a8 17027Author: Alex Deucher <alexdeucher@gmail.com> 17028Date: Mon Mar 30 11:30:34 2009 -0400 17029 17030 r1xx-r3xx TV: fix white noise 17031 17032 should fix bug 13872 17033 17034commit f061308e7abcf93d1612681b935387fee23e0a24 17035Author: Alex Deucher <alexdeucher@gmail.com> 17036Date: Fri Mar 27 01:14:12 2009 -0400 17037 17038 mac mini: fix connector setup regression 17039 17040 fixes bug 20796 17041 17042commit 47134ca77988f8c3b42d55ac4fcbb5dc182878d4 17043Author: Alex Deucher <alexdeucher@gmail.com> 17044Date: Thu Mar 26 17:18:47 2009 -0400 17045 17046 radeon: re-enable r/g/b intensity attributes for overlay 17047 17048 Noticed by sroland on IRC. 17049 17050commit e2b5e48e2652daf03fcae852d8368f29ea4abcd9 17051Author: Alex Deucher <alexdeucher@gmail.com> 17052Date: Thu Mar 26 13:42:25 2009 -0400 17053 17054 R6xx/R7xx EXA: rework composite pixel shader 17055 17056 - move to vram storage 17057 - move swizzle logic to tex setup 17058 17059commit 24dbd3f93c640f819a9f5029f5a49e41ab875bd4 17060Author: Alex Deucher <alexdeucher@gmail.com> 17061Date: Thu Mar 26 12:01:07 2009 -0400 17062 17063 R6xx/R7xx: clean up logic in EXA Comp PS setup 17064 17065commit 076e4e3ac1c81086aecb4e70dd30fd5d26ec7bc8 17066Author: Alex Deucher <alexdeucher@gmail.com> 17067Date: Thu Mar 26 11:36:26 2009 -0400 17068 17069 R6xx/R7xx: clean up bool const code 17070 17071 3 regs: 1 bit per bool, 32 bools per ps/vs/gs 17072 17073commit d2c3964fe04be42fe538f36439ed5ffca96e436a 17074Author: Roland Scheidegger <sroland@tungstengraphics.com> 17075Date: Wed Mar 18 01:55:12 2009 +0100 17076 17077 fix textured video allocation bug 17078 17079 size needs to be calculated after dstPitch adjustments, got already fixed 17080 for earlier than R600 chips by planar textured yuv patches, clean this up. 17081 17082commit 18e56eb179fde28477487c63e6f9ebf7579e2cd5 17083Author: Roland Scheidegger <sroland@tungstengraphics.com> 17084Date: Thu Mar 5 02:07:46 2009 +0100 17085 17086 don't convert planar yuv to packed for r200 17087 17088 uses 3 textures for planar yuv and does yuv->rgb conversion in the shader. 17089 Similar to r300 code, but might have precision issues - hardware alu should 17090 have enough precision but hardware consts are only 8bit and we'd want 17091 at least 11. 17092 This also enables textured video on rv250 (and also supports packed yuv 17093 on that chip by using basically the same shader with packed data). 17094 17095commit 58530bf4912800f9e09ebaea42a13cff8a80c19e 17096Author: Roland Scheidegger <sroland@tungstengraphics.com> 17097Date: Sat Feb 21 04:46:31 2009 +0100 17098 17099 don't convert planar yuv to packed for r300 17100 17101 uses 3 textures for planar yuv and does yuv->rgb conversion in the shader. 17102 small performance advantage, but manual texture cache setting is necessary 17103 otherwise it may be measurably slower (but probably not relevant) in some 17104 cases. 17105 Unlike some other drivers, using MADs instead of DP3s, since this requires 17106 less instructions due to no MOVs are required, the end result is the same 17107 though the constants need to be different. 17108 Use of this is user settable for now (XV_HWPLANAR attrib). 17109 17110commit 97e19d96ba65a3df2fa3bbf73cfcc01b6dc3e796 17111Author: Roland Scheidegger <sroland@tungstengraphics.com> 17112Date: Tue Dec 30 22:23:39 2008 +0100 17113 17114 clip fixes 17115 17116 This fixes some oddities observed when the video is only partly visible. 17117 Instead of recalculating the geometry of the video, always use the same. 17118 Also fixes a assignment present twice, and another issue (bring in line with 17119 what the overlay code does). 17120 17121commit 7aaf5f08e4147b75c4975e6d680d3f5a93424164 17122Author: Alex Deucher <alexdeucher@gmail.com> 17123Date: Mon Mar 23 12:55:02 2009 -0400 17124 17125 R4xx ATOM: fix crtc routing for CRT1 17126 17127 should fix bug 20814 17128 17129commit 2fe6b393d9dbb193d4f0a3351ad76618cda07e2c 17130Author: Alex Deucher <alexdeucher@gmail.com> 17131Date: Sun Mar 22 01:31:30 2009 -0400 17132 17133 radeon: add new chip ids 17134 17135commit 1ddee7cd6fd267b2fc86f21af27c5425eb0835a4 17136Author: Alex Deucher <alexdeucher@gmail.com> 17137Date: Fri Mar 20 13:44:00 2009 -0400 17138 17139 DCE3.2: use RMX for for non-native modes on DVI 17140 17141 DVI seems to have issues with low dotclocks, so 17142 use the scaler instead. 17143 17144 Fixes bug 20754 17145 17146commit 44081e8fd2bf974272d5447b45dd7b206cf939c1 17147Author: Alex Deucher <alexdeucher@gmail.com> 17148Date: Thu Mar 19 20:48:16 2009 -0400 17149 17150 R6xx/R7xx EXA: fix maxPitchBytes 17151 17152 should now allow accel up the hw max of 8192x8192 17153 17154commit e8e2367b66d9a4f5977d95f52b37cdbad75e409a 17155Author: Alex Deucher <alexdeucher@gmail.com> 17156Date: Thu Mar 19 20:09:32 2009 -0400 17157 17158 Add new pci ids 17159 17160commit 68e83f10adb23fc359c1da881f5d666cd08b4067 17161Author: Alex Deucher <alexdeucher@gmail.com> 17162Date: Wed Mar 18 20:58:19 2009 -0400 17163 17164 bump version post release 17165 17166commit a196aae19603c8651dc502f5cb0df4d55befe163 17167Author: Alex Deucher <alexdeucher@gmail.com> 17168Date: Wed Mar 18 20:52:12 2009 -0400 17169 17170 bump for release 17171 17172commit 537da7018b0fd2ad263fc5cdc55b3b85cf8e9205 17173Author: Alex Deucher <alexdeucher@gmail.com> 17174Date: Wed Mar 18 19:22:55 2009 -0400 17175 17176 ATOM: fix up tv-out support 17177 17178 - fix purple hue when using ATOMTvOut option 17179 - fix TV load detection on newer r6xx/r7xx chips 17180 - disable CV for now 17181 17182commit c0e2513ab128ddd5be0ed626d9e31777a98983ef 17183Author: Alex Deucher <alexdeucher@gmail.com> 17184Date: Wed Mar 18 14:08:30 2009 -0400 17185 17186 R6xx/R7xx AGP: disable gart data transfers 17187 17188 UTS/DFS/Xv 17189 17190 Doesn't seem to be reliable on AGP. 17191 17192 fixes bugs: 20436, 20236, several reports on ML and IRC 17193 17194commit e3e1f0c7b892b094b292081021a0e97932b8e079 17195Author: Fabio <fabio.ped@libero.it> 17196Date: Wed Mar 18 12:13:53 2009 -0400 17197 17198 man page update 17199 17200commit 6ae54d3ea023968887bf0cb0ec1be2115621d554 17201Author: Alex Deucher <alexdeucher@gmail.com> 17202Date: Wed Mar 18 11:50:22 2009 -0400 17203 17204 AVIVO: Fix dac load detection 17205 17206commit 04e36b794c216e4fa77c780fd989df1ef1c7f0d9 17207Author: Fabio <fabio.ped@libero.it> 17208Date: Tue Mar 17 11:29:26 2009 -0400 17209 17210 man page updates 17211 17212 from bug 20648 17213 17214commit a6855c370194b6df307ea33724fe17a85d67607e 17215Author: Alex Deucher <alexdeucher@gmail.com> 17216Date: Mon Mar 16 13:52:42 2009 -0400 17217 17218 radeon: add another AGP quirk 17219 17220 bug 20647 17221 17222commit a55ced5ee20c07e743c7c0978803fd10589c1531 17223Author: Alex Deucher <alexdeucher@gmail.com> 17224Date: Sun Mar 15 12:08:04 2009 -0400 17225 17226 Fix fallout from r6xx/r7xx EXA merge in IB handling 17227 17228 Noticed Christiaan van Dijk. 17229 Should fix bug 20510 17230 17231commit a065eeb9b301ab452cc8fbd3cac1fd9a3841c4e1 17232Author: Alex Deucher <alexdeucher@gmail.com> 17233Date: Sat Mar 14 18:40:28 2009 -0400 17234 17235 atom: fix up bios scartch register usage 17236 17237 the bios scratch registers keep 3 types of state for each output: 17238 1. dpms 17239 2. active 17240 3. connected 17241 17242 some of the command tables use the active flags. We used to set the active 17243 flags based on dpms state, but instead set them based on whether they are 17244 connected or not. 17245 17246commit 5c4ef924d082fd3b279f62fda7a97f563b009d38 17247Author: Alex Deucher <alexdeucher@gmail.com> 17248Date: Sat Mar 14 18:16:49 2009 -0400 17249 17250 radeon: fix typo in bios scratch reg setup 17251 17252commit 96fce5a56269509a935e8ee621a5e17cc9005ddc 17253Author: Alex Deucher <alexdeucher@gmail.com> 17254Date: Fri Mar 13 19:29:58 2009 -0400 17255 17256 Bump post-release 17257 17258commit 8da1b82a70de7b3ef38905bd77c58a20fd44ddd6 17259Author: Alex Deucher <alexdeucher@gmail.com> 17260Date: Fri Mar 13 19:23:19 2009 -0400 17261 17262 bump for release 17263 17264commit 28d9d699dcdff394aa561505f7f54f08db800828 17265Author: Alex Deucher <alexdeucher@gmail.com> 17266Date: Fri Mar 13 18:18:31 2009 -0400 17267 17268 radeon: man page updates 17269 17270commit 06963796d00b5cb6704412a7d4417fcefae4aa4d 17271Author: Adam Jackson <ajax@redhat.com> 17272Date: Fri Mar 13 18:14:43 2009 -0400 17273 17274 uniphy: start adding DisplayPort setup 17275 17276commit 1db81afe4f8db55e02a65cd22ad84c6bbb27e8ef 17277Author: Alex Deucher <alexdeucher@gmail.com> 17278Date: Fri Mar 13 17:51:56 2009 -0400 17279 17280 radeon: add support for 30 bit LUTs 17281 17282 All radeons have them. Thanks to Yang Zhao for figuring 17283 this out. 17284 17285commit 5bb326c64ad3bf5c93943280dcbcb25c8910d1aa 17286Author: Alex Deucher <alexdeucher@gmail.com> 17287Date: Fri Mar 13 14:10:58 2009 -0400 17288 17289 R6xx/R7xx: EXA VSync Option not supported yet 17290 17291commit 8c2d40894448329e721360811c124998a17ba2ba 17292Author: Alex Deucher <alexdeucher@gmail.com> 17293Date: Fri Mar 13 14:03:43 2009 -0400 17294 17295 R6xx/R7xx: fix up vline stuff along the lines of previous chips 17296 17297 Avoids an additional function call. 17298 17299commit 176c1b43fc3ee48e87c134446754360b3683c24b 17300Author: Alex Deucher <alexdeucher@gmail.com> 17301Date: Fri Mar 13 13:28:42 2009 -0400 17302 17303 R6xx/R7xx: return in RADEONWaitForIdleMMIO() if accel is off 17304 17305 Don't wait for engine idle if we haven't initialized the engine. 17306 Behavior should be consistent with versions of the driver prior 17307 to accel support being added. 17308 17309 Should fix bug 20645 17310 17311commit d538a4508248adf2debaf0cc870a2d1c48abd986 17312Author: Alex Deucher <alexdeucher@gmail.com> 17313Date: Fri Mar 13 13:22:03 2009 -0400 17314 17315 R6xx/r7xx: clarify accel messages 17316 17317 We only support EXA and and only with DRI. 17318 17319commit 1b7090a4c0d6a38701fbfb928fac8e15d40fb81f 17320Author: Adam Jackson <ajax@redhat.com> 17321Date: Fri Mar 13 17:41:59 2009 -0400 17322 17323 Add some more cases to the DVI single-link checks. 17324 17325commit 104e13da74de37477e3a31a5de96c5ebb8fe888b 17326Author: Cooper Yuan <cooperyuan@gmail.com> 17327Date: Fri Mar 13 11:47:43 2009 -0400 17328 17329 radeon: save bios scratch registers in Preinit() 17330 17331 The display detect routines can change these which causes 17332 havok with some bioses. 17333 17334commit f70d7a0fd9ac3659d93511ab362eb80a7ff49777 17335Author: Dave Airlie <airlied@redhat.com> 17336Date: Fri Mar 13 12:56:29 2009 +1000 17337 17338 r600: reload shaders into VRAM on resume 17339 17340 As VRAM gets zeroed out over s/r, we need to reload the 17341 shaders. 17342 17343commit 1fe0dcfe98dfe30499ba53c1f208c4ef10bef001 17344Author: Yang Zhao <yang@yangman.ca> 17345Date: Thu Mar 12 18:50:32 2009 -0400 17346 17347 R6xx/R7xx shader: Fix OFFSET_[XYZ] macro for TEX_DWORD2 to accept floats 17348 17349 Values for OFFSET_[XYZ] are 5-bits two's-complement fixed-point with 17350 one-bit after decimal point. Values in [-8.0, 7.5] are valid. Inputs 17351 that do not exactly land on 0.5 increments are rounded towards 0 to the 17352 nearest increment. 17353 17354commit f356ca7795531e0c080775efb598141eb7db44ba 17355Author: Alex Deucher <alexdeucher@gmail.com> 17356Date: Thu Mar 12 03:42:40 2009 -0400 17357 17358 RS600: fix page table size for rs600 as well 17359 17360commit ef69336584c45e35aabb797a06c4030e75d66507 17361Author: Dave Airlie <airlied@redhat.com> 17362Date: Thu Mar 12 14:54:18 2009 +1000 17363 17364 r600: fix sizing of PCI GART table for r600 17365 17366commit 945ccbbd4fa2b65ccdfb23716c178c95b036734d 17367Author: Alex Deucher <alexdeucher@gmail.com> 17368Date: Tue Mar 10 09:53:14 2009 -0400 17369 17370 radeon: clean more thoroughly in RADEONFreeRec() 17371 17372commit 03ec2099fdc859b23e5fa56ac8696ec045f1b110 17373Author: 오유연(Yu-yeon Oh) <yuyeon_oh@tmax.co.kr> 17374Date: Tue Mar 10 09:51:08 2009 -0400 17375 17376 radeon_driver.c small memory bug 17377 17378commit 60dac878cc8626b0821c4c7a6a6dfa12a5597cfa 17379Author: Alex Deucher <alexdeucher@gmail.com> 17380Date: Mon Mar 9 20:48:39 2009 -0400 17381 17382 R4xx: add R4xxATOM option 17383 17384 If set, enables the use of atombios for modesetting 17385 on r4xx cards. 17386 17387commit 758fcc33c232c32e7b66261fca51de161d1a4800 17388Author: Dave Airlie <airlied@redhat.com> 17389Date: Tue Mar 10 10:34:13 2009 +1000 17390 17391 r600: enable DRI by default 17392 17393 fallback to shadowfb if DRI fails pre init 17394 17395commit 7735e411954fd2cb6105c30ebd5a1b209974ce22 17396Author: Alex Deucher <alexdeucher@gmail.com> 17397Date: Mon Mar 9 17:49:25 2009 -0400 17398 17399 R6xx/r7xx: clarify EXA message 17400 17401commit 1d355798f3d421cabf2a3a0ba32a580483476ab0 17402Author: Alex Deucher <alexdeucher@gmail.com> 17403Date: Mon Mar 9 15:13:01 2009 -0400 17404 17405 AVIVO: fix panning 17406 17407 low 2 bits of X need to be 0. 17408 17409commit d1add18f3d238c755f9875b5e5c211e4af742482 17410Author: root <root@benzylpiperazine.boston.devel.redhat.com> 17411Date: Fri Mar 6 17:53:59 2009 -0500 17412 17413 output: Filter out dual-link modes from DP->DVI connections 17414 17415 There's not enough pins on a DP->DVI connector for this to possibly work. 17416 17417commit 04481bf450650d1ea5aadabc9213b899b0fe8119 17418Author: root <root@benzylpiperazine.boston.devel.redhat.com> 17419Date: Fri Mar 6 17:32:15 2009 -0500 17420 17421 atom: Enable DisplayPort source to DVI/HDMI sink 17422 17423 DP source to DP sink still doesn't work yet, of course. 17424 17425commit 0ba73e040b94590867f8b1071a26da2526a3c375 17426Author: Adam Jackson <ajax@redhat.com> 17427Date: Fri Mar 6 15:38:00 2009 -0500 17428 17429 output: Allow for multiple DisplayPort outputs. 17430 17431commit 8f7c17a6b77ca368f067c2acb803cc8f9ef19351 17432Author: Tormod Volden <debian.tormod@gmail.com> 17433Date: Fri Mar 6 09:42:39 2009 -0500 17434 17435 RV350: Add AGPMode quirk for Thinkpad 17436 17437 fixes bug 15096 (and Ubuntu bug #195051) 17438 17439commit 5dc4b69fbdd0e8ca48cc00321f193260a8f10bd0 17440Author: Alex Deucher <alexdeucher@gmail.com> 17441Date: Thu Mar 5 12:26:04 2009 -0500 17442 17443 radeon: add a few more default common modes for lvds 17444 17445commit 3b0fc22ad9e5c0f120a74c4f3d8e48c270f0ff29 17446Author: Alex Deucher <alexdeucher@gmail.com> 17447Date: Wed Mar 4 19:04:34 2009 -0500 17448 17449 R6xx/R7xx: wire up DMAForXv option like older asics 17450 17451 Select between SW and HW-assisted uploads 17452 17453commit 52f06ace04ad8141effc45fb6a0107a05bb46a73 17454Author: Alex Deucher <alexdeucher@gmail.com> 17455Date: Wed Mar 4 18:51:44 2009 -0500 17456 17457 R6xx/R7xx: disable XV_BICUBIC attribute 17458 17459 It's not implemented yet. 17460 17461commit b1fd883b59b85fed8782e035890098908902f4ce 17462Author: Alex Deucher <alexdeucher@gmail.com> 17463Date: Wed Mar 4 11:17:16 2009 -0500 17464 17465 RN50: fix up cloning on servers 17466 17467 Since the rn50 only has one crtc, but often has multiple outputs, 17468 occasionally one of the outputs doesn't get a crtc assigned eventhough 17469 the outputs may have common modes. If we see an output with modes, but 17470 no crtc, assign it so all the attached outputs light up. 17471 17472 fixes bugs 19457 and similar ones 17473 17474commit 592bcac52f113a95923a8f1cb8427e7552d5670b 17475Author: Alex Deucher <alexdeucher@gmail.com> 17476Date: Wed Mar 4 10:53:20 2009 -0500 17477 17478 ATOM: don't use fixed ref div for LVDS 17479 17480 The bios doesn't provide one. If there are problems 17481 with certain panels, we may need to adjust the PLL flags 17482 in certain cases. 17483 17484 Should fix bug 20458 17485 17486commit e88b0eac8cab8b54b769fd7b4ad3b3b65a90de09 17487Author: Alex Deucher <alexdeucher@gmail.com> 17488Date: Wed Mar 4 02:52:36 2009 -0500 17489 17490 radeon: just add some common modes for LVDS 17491 17492 avoids needing to hack around with the edid 17493 17494commit 2bb319fd80d20a3d4e5cc9416891bc7e4c3e431e 17495Author: Alex Deucher <alexdeucher@gmail.com> 17496Date: Wed Mar 4 02:50:45 2009 -0500 17497 17498 Revert "radeon: adjust LVDS so that default modes get added" 17499 17500 This reverts commit 1a2b16561d19ec9c027c562902f5fc086c856994. 17501 17502commit 1a2b16561d19ec9c027c562902f5fc086c856994 17503Author: Alex Deucher <alexdeucher@gmail.com> 17504Date: Tue Mar 3 20:21:17 2009 -0500 17505 17506 radeon: adjust LVDS so that default modes get added 17507 17508 we can scale, so add the default modes. for panels 17509 with an EDID, set the continous freq bit. for panels without 17510 and EDID, add a FAKE edid with the continous freq bit set. 17511 17512 Based on similar code in the Intel driver. 17513 17514commit 71117970df36cbe689ef15e9a6cca24439b4cd62 17515Author: Alex Deucher <alexdeucher@gmail.com> 17516Date: Tue Mar 3 19:40:30 2009 -0500 17517 17518 AVIVO: add aspect scaling mode 17519 17520 No luck yet for aspect on pre-avivo chips 17521 17522commit d586a2c6f821c821a4a7708a3382acb63187534f 17523Author: Alex Deucher <alexdeucher@gmail.com> 17524Date: Tue Mar 3 12:17:30 2009 -0500 17525 17526 Rotation: don't rotate if acceleration is not active 17527 17528commit fdc64948119e2218921b4dc0a492522442668875 17529Author: Alex Deucher <alexdeucher@gmail.com> 17530Date: Mon Mar 2 16:33:46 2009 -0500 17531 17532 R6xx/r7xx: remove some unneeded code I missed in the last commit 17533 17534commit 4ad1c4decfee653dbbc1ea2ca4270487be622382 17535Author: Alex Deucher <alexdeucher@gmail.com> 17536Date: Mon Mar 2 15:38:35 2009 -0500 17537 17538 R6xx/R7xx EXA: cleanup composite texture setup 17539 17540commit b7164ac4ad55e5d0fc474df8ae762b469b91ba30 17541Author: Alex Deucher <alexdeucher@gmail.com> 17542Date: Mon Mar 2 04:08:09 2009 -0500 17543 17544 R6xx/R7xx EXA: combine composite mask/non-mask VS 17545 17546 Also fix set_bool_const() 17547 the CF bool consts are not contiguous by shader type 17548 There are 96 boolean constants (32 each for PS, VS, GS) and 17549 they are ordered as follows: 17550 ps, vs, gs ... ps, vs, gs 17551 17552commit fa98f424de739be2c6005b740a74bbf1ee968a8b 17553Author: Alex Deucher <alexdeucher@gmail.com> 17554Date: Mon Mar 2 02:28:57 2009 -0500 17555 17556 R6xx/R7xx Xv: combine packed and planar shaders 17557 17558 use a bool const to select the tex fetch routine 17559 17560commit ccde35c3eda3fff0de29eb8c6fdc392629724a34 17561Author: Christian Koenig <deathsimple@vodafone.de> 17562Date: Sun Mar 1 23:38:37 2009 -0500 17563 17564 R6xx/R7xx: move shaders to r600_shader.c and fixup Xv PS 17565 17566 patches from Christian Koenig with some adjustments from me 17567 17568commit a8e631c1b1c9b46602aeca66f8e7e68154d0bfc8 17569Author: Alex Deucher <alexdeucher@gmail.com> 17570Date: Sun Mar 1 22:30:31 2009 -0500 17571 17572 R6xx/R7xx: code cleanups 17573 17574commit 528061d51bb4248d6fabec9579dead32a730467a 17575Author: Alex Deucher <alexdeucher@gmail.com> 17576Date: Sun Mar 1 18:28:13 2009 -0500 17577 17578 R6xx/R7xx: write vertexes directly to the IB 17579 17580 Reduces the vertex buffer setup overhead 17581 17582commit 96a0bafa16dbb67a3d10fa6fe45be13930495ddc 17583Author: Alex Deucher <alexdeucher@gmail.com> 17584Date: Sun Mar 1 17:58:36 2009 -0500 17585 17586 R6xx/R7xx: switch emit functions to macros 17587 17588 This improves performance due to decreased function call 17589 overhead. 17590 17591commit cbd579c41a16bddcbb8793eac2a41f2e958bef4d 17592Author: Alex Deucher <alexdeucher@gmail.com> 17593Date: Sun Mar 1 13:16:20 2009 -0500 17594 17595 Print a message when we have a shared DDC line 17596 17597commit a7f465f73363fce409870f62173d518b1bc02ae6 17598Author: Alex Deucher <alexdeucher@gmail.com> 17599Date: Sun Mar 1 12:39:31 2009 -0500 17600 17601 R300: Add AGP quirk 17602 17603 bug 7770 17604 17605commit 017bc7f34de2e40919f0e711205b78ee4e3643f6 17606Author: Tormod Volden <debian.tormod@gmail.com> 17607Date: Sat Feb 28 10:33:35 2009 -0500 17608 17609 M9+: Add AGP quirk for Sony Vaio 17610 17611commit 49b57767d0d2c041517b0764c2ed2d2ba5a7092c 17612Author: Bryce Harrington <bryce@bryceharrington.org> 17613Date: Fri Feb 27 20:31:03 2009 -0500 17614 17615 Quirk for RV280 on 82865G/PE/P DRAM Controller/Host-Hub 17616 17617 Solves issue where display suffers from line corruption. 17618 (See https://edge.launchpad.net/ubuntu/+bug/300304) 17619 17620commit 150abd4123664eab65ea43585dbdb9b3151c46af 17621Author: Tormod Volden <lists.tormod@gmail.com> 17622Date: Fri Feb 27 11:23:32 2009 -0500 17623 17624 Janitor: cosmetic clean-up of AGPMode quirk table 17625 17626commit c7d344f0ad255761d5ae30da59cf0a8b532dfc68 17627Author: Michel Dänzer <daenzer@vmware.com> 17628Date: Fri Feb 27 17:05:01 2009 +0100 17629 17630 EXA: Make sure Prepare/FinishAccess hooks can handle EXA_PREPARE_AUX* indices. 17631 17632 And signal this to the EXA core using the new EXA_SUPPORTS_PREPARE_AUX flag. 17633 17634commit f4d57105fc869e9318f447a1514f1b284a0a4ec0 17635Author: Michel Dänzer <daenzer@vmware.com> 17636Date: Fri Feb 27 09:23:24 2009 +0100 17637 17638 Revert "EXA: Adapt to EXA changes in xserver Git." 17639 17640 This reverts commit de358736dc696559ba99c71cf5b2a97508201630. 17641 17642commit 3285a4af17b19f2d1c24eebfb6aa51cddcf3ca50 17643Author: Alex Deucher <alexdeucher@gmail.com> 17644Date: Thu Feb 26 17:31:13 2009 -0500 17645 17646 DCE3.2: fix up Save()/Restore() 17647 17648commit 4fc978f3d83ef6370c9715e7ca5350640ccbfe10 17649Author: Alex Deucher <alexdeucher@gmail.com> 17650Date: Thu Feb 26 12:10:08 2009 -0500 17651 17652 R6xx/R7xx: use shadowfb if DRI is disabled 17653 17654commit 000756e052a291230e5c95e48b69a5aa9c4fab0e 17655Merge: 22d7746b 8373f439 17656Author: Alex Deucher <alexdeucher@gmail.com> 17657Date: Thu Feb 26 11:44:13 2009 -0500 17658 17659 Merge branch 'r6xx-r7xx-support' of git+ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati 17660 17661commit 22d7746bfc1bcbb3269afac88971254088ea9ece 17662Author: Alex Deucher <alexdeucher@gmail.com> 17663Date: Thu Feb 26 11:00:46 2009 -0500 17664 17665 RS600: enable the DRI by default 17666 17667commit 8373f4399b03961f2c928a9275d47e9f41bd92bb 17668Author: Alex Deucher <alexdeucher@gmail.com> 17669Date: Wed Feb 25 09:16:38 2009 -0500 17670 17671 R6xx/R7xx EXA: same surface and same coords equals nop 17672 17673 should fix bug 20305 17674 17675commit c74727015453ff3c3d6d06b812ebca9eb19a9767 17676Author: Alex Deucher <alexdeucher@gmail.com> 17677Date: Tue Feb 24 11:08:46 2009 -0500 17678 17679 R6xx/R7xx EXA: init copy_area to NULL 17680 17681commit 95ce13572dc2d9f5dd6cf55c23411e275c0aadf1 17682Author: Alex Deucher <alexdeucher@gmail.com> 17683Date: Tue Feb 24 10:51:13 2009 -0500 17684 17685 R6xx/R7xx EXA: Optimize temp surface for overlapping copies 17686 17687 - allocate temp surface in PrepareCopy() 17688 - fall back to old OverlapCopy() path if we are not able 17689 to allocate a temp surface 17690 17691commit 1a7db3fc2a0277d724d60d028064d8ef75019c28 17692Author: Mark van Doesburg <mark.vandoesburg@hetnet.nl> 17693Date: Tue Feb 24 10:44:19 2009 -0500 17694 17695 R6xx/R7xx EXA: use a temp surface for overlapping copy 17696 17697commit 957c2ba3459af152ad16a3f6dd4631044f9df6f9 17698Author: Alex Deucher <alexdeucher@gmail.com> 17699Date: Tue Feb 24 09:43:58 2009 -0500 17700 17701 RS600: fix MC addr mask 17702 17703 Noticed by osiris on IRC 17704 17705commit de358736dc696559ba99c71cf5b2a97508201630 17706Author: Michel Dänzer <daenzer@vmware.com> 17707Date: Tue Feb 24 08:30:23 2009 +0100 17708 17709 EXA: Adapt to EXA changes in xserver Git. 17710 17711 The PrepareAccess hook can now get six different indices. 17712 17713 Also remove superfluous instances of #include "exa.h". 17714 17715 Note that you may need to run autoheader manually to get config.h.in updated. 17716 17717commit 752b0218af36a45b333a12f66dfdce8c1e662aa8 17718Author: Michel Dänzer <daenzer@vmware.com> 17719Date: Tue Feb 24 08:13:06 2009 +0100 17720 17721 EXA: Pass pScrn and info into RadeonCompositeTile. 17722 17723 Saves some cycles looking them up. 17724 17725commit e2e508bc08d993fc108297e9cd2fa0ce183282bc 17726Author: Michel Dänzer <daenzer@vmware.com> 17727Date: Tue Feb 24 08:12:42 2009 +0100 17728 17729 Only call RADEONWaitForVLine if it might actually do anything useful. 17730 17731 Otherwise it and RADEONBiggerCrtcArea show up pretty high in x11perf -aa*text 17732 profiles. 17733 17734commit 5ab8699c97c050424cf813b9a60550aba32446a2 17735Author: Michel Dänzer <daenzer@vmware.com> 17736Date: Tue Feb 24 08:12:23 2009 +0100 17737 17738 Fix compile warning when building without EXA. 17739 17740commit 132e61741185f4a1114e66a34bdbf21e08e4d5d8 17741Author: Alex Deucher <alexdeucher@gmail.com> 17742Date: Mon Feb 23 12:13:41 2009 -0500 17743 17744 radeon: re-enable load detection output attribute for TV/CV 17745 17746commit 918d511f6f26d2fdeb63d6e32cd348707b4a2140 17747Author: Alex Deucher <alexdeucher@gmail.com> 17748Date: Mon Feb 23 10:22:19 2009 -0500 17749 17750 RS600: fix up MC setup 17751 17752 Please test if you have an RS600 17753 17754commit 27f8ca2cce65be2bcb3375231886d5444d251808 17755Author: Alex Deucher <alexdeucher@gmail.com> 17756Date: Thu Feb 19 21:18:07 2009 -0500 17757 17758 R6xx/R7xx: add wait for idle MMIO path 17759 17760commit 1f783117f79f95f1744681ccc7d80dc0e7b53057 17761Author: Alex Deucher <alexdeucher@gmail.com> 17762Date: Thu Feb 19 21:07:06 2009 -0500 17763 17764 RBBM_GUICNTL is pre-r6xx only 17765 17766commit f3933ed200dbf61b05a114f7697f2bb68d92abe7 17767Author: Alex Deucher <alexdeucher@gmail.com> 17768Date: Thu Feb 19 20:12:33 2009 -0500 17769 17770 Don't write new HDP location until we've written the new FB location 17771 17772commit 7f6801abb4bbd79830999d99ccdbd672a9984df6 17773Author: Alex Deucher <alexdeucher@gmail.com> 17774Date: Thu Feb 19 19:02:52 2009 -0500 17775 17776 radeon: one more 32 -> 64 just to be safe 17777 17778 missed this in cadae2a8d9979b2d61eae7d1f7a9ab8f5eb28e83 17779 17780commit ca22390c8a5256bbc8eeff775d5eebf0e892d890 17781Author: Alex Deucher <alexdeucher@gmail.com> 17782Date: Thu Feb 19 18:58:40 2009 -0500 17783 17784 R6xx/R7xx: fix up a few more paths 17785 17786commit cadae2a8d9979b2d61eae7d1f7a9ab8f5eb28e83 17787Author: Alex Deucher <alexdeucher@gmail.com> 17788Date: Thu Feb 19 18:50:49 2009 -0500 17789 17790 Fix MC setup on systems with more than 512 MB of VRAM 17791 17792 many thanks to app4des on IRC for helping me sort this out. 17793 17794commit 42492c70c481c88911337eefb97ba9d030adf7b6 17795Author: Alex Deucher <alexdeucher@gmail.com> 17796Date: Thu Feb 19 10:20:28 2009 -0500 17797 17798 R6xx/R7xx: wait for MC idle when changing the MC 17799 17800 should fix bug 19984 17801 17802commit e6475282486f4895bc68f6b093ecbb1aa6d25f72 17803Author: Alex Deucher <alexdeucher@gmail.com> 17804Date: Wed Feb 18 20:21:13 2009 -0500 17805 17806 R6xx/R7xx Xv: fix some missing bits from last commit 17807 17808commit adff8906c9899dde7711382577a63f4a726437ca 17809Author: Alex Deucher <alexdeucher@gmail.com> 17810Date: Wed Feb 18 20:01:57 2009 -0500 17811 17812 R6xx/R7xx EXA/Xv: properly deal with running out of vertex buffer space 17813 17814 As noted by mhopf, if VGT_MAX/MIN_INDX, etc. regs change, you need to re-emit 17815 CB blocks to avoid a hang. So, just set the VGT_MAX_INDX to a reasonably large value 17816 in the default state and don't touch them when drawing. When we run out of VB space, 17817 just draw the current buffer, grab a new one, and continue. 17818 17819commit cf85d4a1d43a3209c7ca9307aede2c2c243f7130 17820Author: Alex Deucher <alexdeucher@gmail.com> 17821Date: Wed Feb 18 19:02:56 2009 -0500 17822 17823 R6xx/R7xx: reset 3D state after VT switch 17824 17825commit 6545f0a2967414fa518a78440060b8b69c0146ee 17826Author: Alex Deucher <alexdeucher@gmail.com> 17827Date: Wed Feb 18 18:15:18 2009 -0500 17828 17829 R6xx/R7xx: fixup accel paths 17830 17831commit 61311b1de019c5a2344928a47347a966c229b08a 17832Author: Alex Deucher <alexdeucher@gmail.com> 17833Date: Wed Feb 18 17:27:05 2009 -0500 17834 17835 Bump version post release 17836 17837commit 0d3d1f1f7b450dabd3ad7d2df26ad1c72709b29b 17838Author: Alex Deucher <alexdeucher@gmail.com> 17839Date: Wed Feb 18 17:24:57 2009 -0500 17840 17841 bump for release 17842 17843commit 76a32f8391317513538dafee49cbb61c0d756356 17844Author: Alex Deucher <alexdeucher@gmail.com> 17845Date: Wed Feb 18 16:31:44 2009 -0500 17846 17847 R6xx: Connector quirk for asus board 17848 17849 bug 19943 17850 17851commit 97b8482dba4e99088b59d3ab1396be98a70e823e 17852Author: Alex Deucher <alexdeucher@gmail.com> 17853Date: Wed Feb 18 13:27:15 2009 -0500 17854 17855 Set default RMX type to FULL on LVDS 17856 17857commit 7d22b1799b34010d34c3600d6cb02ffe839a0780 17858Author: Alex Deucher <alexdeucher@gmail.com> 17859Date: Wed Feb 18 12:55:40 2009 -0500 17860 17861 DCE3.2+: allow output cloning 17862 17863commit 1a237a40958c006c56b80850bd77b2ac6c17e030 17864Author: Alex Deucher <alexdeucher@gmail.com> 17865Date: Wed Feb 18 11:46:20 2009 -0500 17866 17867 ATOM: reset crtc initialized flag on CloseScreen() 17868 17869 additional fix needed for bug 16781 17870 17871commit 9a108f0a0b7203458673ce6221e747a166d39617 17872Author: Alex Deucher <alexdeucher@gmail.com> 17873Date: Tue Feb 17 19:52:27 2009 -0500 17874 17875 ATOM: don't unblank uninitialized crtcs 17876 17877 If the crtc timing isn't setup, you might get stuck in a loop 17878 in the BlankCRTC command table 17879 17880 fixes bug 16781 17881 17882commit 3dff20e276615e8b77177689a4a5f8d91b3e8eac 17883Author: Alex Deucher <alexdeucher@gmail.com> 17884Date: Tue Feb 17 13:14:38 2009 -0500 17885 17886 R6xx/R7xx EXA: switch to surface sync packet 17887 17888commit 31bd6d28dd35021e0e706d4f094989deb856c26a 17889Author: Alex Deucher <alexdeucher@gmail.com> 17890Date: Tue Feb 17 11:24:02 2009 -0500 17891 17892 Fix crtc routing on pre-DCE3.2 systems 17893 17894 This should fix peterz and bug 20074 17895 17896commit 3c9e00c7f2ed494976713e8f77b729e0d76409db 17897Author: Alex Deucher <alexdeucher@gmail.com> 17898Date: Mon Feb 16 16:18:42 2009 -0500 17899 17900 RV350: add another AGP quirk 17901 17902commit 036a17bec5dabbb91640b907811da060dac154d6 17903Author: Alex Deucher <alexdeucher@gmail.com> 17904Date: Mon Feb 16 13:10:45 2009 -0500 17905 17906 RV280: add another agp quirk 17907 17908commit 00ac9d2f8d58725d0d3e6b1bf7f728d9ae7e6109 17909Author: David Miller <davem@davemloft.net> 17910Date: Sun Feb 15 18:26:14 2009 +1000 17911 17912 GART: Save/restore GART table consistently. 17913 17914 Always save the GART table with the RADEON_SURFACE_CNTL register 17915 zero'd out to make sure we always use the same endinanness. 17916 17917 fixed up as per Michel's suggestion for endianness. 17918 17919 Signed-off-by: David S. Miller <davem@davemloft.net> 17920 Signed-off-by: Dave Airlie <airlied@linux.ie> 17921 17922commit 2de90c8349409eb892edc57fd24b61b679eb7719 17923Author: David Miller <davem@davemloft.net> 17924Date: Sun Feb 15 18:24:01 2009 +1000 17925 17926 DRI: Fix page size used in RADEONDRIGetPciAperTableSize(). 17927 17928 The ATI GART has a fixed size of 4096 bytes. So using 17929 the system page size here is wrong. 17930 17931 Signed-off-by: David S. Miller <davem@davemloft.net> 17932 Signed-off-by: Dave Airlie <airlied@linux.ie> 17933 17934commit 7cde00f49649e25fd5816927c7a5e28b608fabcd 17935Author: Alex Deucher <alexdeucher@gmail.com> 17936Date: Thu Feb 12 17:39:44 2009 -0500 17937 17938 R6xx/R7xx: switch to drm for wait for idle 17939 17940 THIS REQUIRES AN UPDATED DRM 17941 17942commit e22cd4011b9be437ba89bff568e7fb82b4907d99 17943Author: Yang Zhao <yang@yangman.ca> 17944Date: Thu Feb 12 14:46:53 2009 -0500 17945 17946 R6xx/R7xx EXA: Further optimizations to overlapping copy 17947 17948 Diagonal overlapping copies can be reduced to either horizontal- or 17949 vertical-only offset, and the one with fewer copies is picked. 17950 17951commit da08b760bcf3d04d775c4440fafec10657bb1863 17952Author: Alex Deucher <alexdeucher@gmail.com> 17953Date: Thu Feb 12 13:53:11 2009 -0500 17954 17955 R6xx/R7xx EXA: handle running out of vertex space in the copy path 17956 17957commit e3be312b0b73982c24f1f5d9cf76d7caafae0853 17958Author: Alex Deucher <alexdeucher@gmail.com> 17959Date: Thu Feb 12 13:48:36 2009 -0500 17960 17961 R6xx/R7xx EXA: properly handle non repeat cases in the texture setup 17962 17963commit 3ff1eb7e50fef22adb6280dd7de77c3ecafdb451 17964Author: Michel Dänzer <daenzer@vmware.com> 17965Date: Thu Feb 12 19:24:41 2009 +0100 17966 17967 EXA: If making a pixmap offscreen fails, return ~0ULL as texture offset. 17968 17969 This allows AIGLX to fall back to the non-zero-copy code rather than messing up 17970 the 3D driver. 17971 17972commit c19aa4fc8da7ac4745624098b146bcc42c0436dd 17973Author: Michel Dänzer <daenzer@vmware.com> 17974Date: Thu Feb 12 19:22:12 2009 +0100 17975 17976 EXA: The source tiling code can't handle RepeatReflect yet. 17977 17978commit adc35636aa1e936ec74cba8931ccffe2b5c11656 17979Author: Alex Deucher <alexdeucher@gmail.com> 17980Date: Thu Feb 12 13:10:53 2009 -0500 17981 17982 R1xx/R2xx EXA: fix non repeat texture setup 17983 17984commit c6e1bf75c399807716b99486ab3301aeba60fa79 17985Author: Alex Deucher <alexdeucher@gmail.com> 17986Date: Thu Feb 12 12:49:24 2009 -0500 17987 17988 R3xx-R5xx EXA: fix texture setup for non-repeat case 17989 17990 fixes bug 19923 17991 17992commit 0804dde9012cf33f614a1c2ee6d5d0def040a9af 17993Author: Alex Deucher <alexdeucher@gmail.com> 17994Date: Thu Feb 12 12:02:57 2009 -0500 17995 17996 AVIVO: fix dualhead/rotation for real 17997 17998commit 3e5ac32f747bef70903e9cdf41652a827eece707 17999Author: Alex Deucher <alexdeucher@gmail.com> 18000Date: Wed Feb 11 17:40:20 2009 -0500 18001 18002 Revert "R6xx/R7xx: handle running out of vertex buffer space" 18003 18004 This reverts commit 4fd7228de7ad88edf825dbc3039df877795a9479. 18005 18006 Causes lockups. 18007 18008commit bd141aa73a77f68301715fb3b5664e2082327a80 18009Author: Alex Deucher <alexdeucher@gmail.com> 18010Date: Wed Feb 11 15:31:47 2009 -0500 18011 18012 R6xx/R7xx Xv: switch packed over to Yang's new shader code 18013 18014commit cb4a1ceaa32d6847b146a31291772e1183972ee7 18015Author: Yang Zhao <yang@yangman.ca> 18016Date: Wed Feb 11 15:18:00 2009 -0500 18017 18018 R6xx/R7xx Xv: Planar - Properly scale Y'CbCr values before converting to RGB 18019 18020 According to MPEG-2 spec, Y' and Cb/Cr values are scaled to [16, 235] 18021 and [16, 240], respectively, when packed into bytes. Properly take care 18022 of the reverse scaling before translating to RGB. 18023 18024 Conversion matrix has been simplified to remove 3rd column, as the fitting 18025 to [-0.5, 0.5] can be done with scaling. 18026 18027 Redundant MOV instructions were also removed, and now only 3 GPRs are required. 18028 18029commit 4fd7228de7ad88edf825dbc3039df877795a9479 18030Author: Alex Deucher <alexdeucher@gmail.com> 18031Date: Wed Feb 11 14:46:02 2009 -0500 18032 18033 R6xx/R7xx: handle running out of vertex buffer space 18034 18035 draw current VB and start a new one. 18036 18037commit ec60ef094762901ede4df5cde55c3f162e8c667b 18038Author: Alex Deucher <alexdeucher@gmail.com> 18039Date: Wed Feb 11 13:05:27 2009 -0500 18040 18041 R6xx/R7xx: be more verbose about what function ran out of VB space 18042 18043commit e7dedbc355970407eddb38370b58a8c96d204c1d 18044Author: Alex Deucher <alexdeucher@gmail.com> 18045Date: Wed Feb 11 12:45:14 2009 -0500 18046 18047 R6xx/R7xx: Move engine idle to sync functions 18048 18049commit 4a759a907ecd571460c7ff64cd6288380359b04b 18050Author: Alex Deucher <alexdeucher@gmail.com> 18051Date: Wed Feb 11 12:29:50 2009 -0500 18052 18053 adjust alignment 18054 18055commit d98359b20e1eb8eae893581ab199c299c8352dc4 18056Author: Alex Deucher <alexdeucher@gmail.com> 18057Date: Wed Feb 11 12:03:56 2009 -0500 18058 18059 Revert "Radeon EXA: wait for the engine to be idle before sw access" 18060 18061 This reverts commit 0c4694c58f2ec152b19023626bc7ced891d9da65. 18062 18063 This patch just adds latency and thus isn't a real solution 18064 18065commit d6c50b221cbab6e726948e2310c1def9fc38da64 18066Author: Alex Deucher <alexdeucher@gmail.com> 18067Date: Wed Feb 11 11:05:02 2009 -0500 18068 18069 R6xx/R7xx Xv: add support for packed uploads 18070 18071commit 8e437e996cc3f2c424c342701f4aa6bcf72ad08e 18072Author: Alex Deucher <alexdeucher@gmail.com> 18073Date: Wed Feb 11 10:53:50 2009 -0500 18074 18075 R6xx/R7xx Xv: Add native support for packed formats 18076 18077commit 0c4694c58f2ec152b19023626bc7ced891d9da65 18078Author: Alex Deucher <alexdeucher@gmail.com> 18079Date: Tue Feb 10 17:08:37 2009 -0500 18080 18081 Radeon EXA: wait for the engine to be idle before sw access 18082 18083 Really we need proper fencing with timestamps but this should 18084 help in the short term. 18085 18086 This should help with minor corruption issues and possibly 18087 lockups. 18088 18089commit 13ab61d90bb58112bef4ce7267ef127935315f80 18090Author: Fabio <fabio.ped@libero.it> 18091Date: Tue Feb 10 10:50:30 2009 -0500 18092 18093 man page updates 18094 18095 bug 20030 18096 18097commit 615f2f202cee5f0dd2c6b29329f843bb1c9c70d5 18098Author: Alex Deucher <alexdeucher@gmail.com> 18099Date: Tue Feb 10 10:17:47 2009 -0500 18100 18101 Bump version post RC release 18102 18103commit 1b4afc1c9f8458f9ab3434418f4ccf959f532ac1 18104Author: Alex Deucher <alexdeucher@gmail.com> 18105Date: Mon Feb 9 19:50:37 2009 -0500 18106 18107 R6xx/R7xx: Add checks to make sure we don't overrun VB space 18108 18109commit ff6efe513e0c3148812d6a9578d2265522417053 18110Author: Alex Deucher <alexdeucher@gmail.com> 18111Date: Mon Feb 9 17:43:24 2009 -0500 18112 18113 ATOM: warning fixes 18114 18115commit 87ad9006f3dafb45e131c9a3d6df713321f39998 18116Author: Alex Deucher <alexdeucher@gmail.com> 18117Date: Mon Feb 9 17:40:41 2009 -0500 18118 18119 RV350: add AGP quirk 18120 18121 bug 19981 18122 18123commit 610b8b65fd7aa97f9c03b2ff6ae607aa959eab0b 18124Author: Alex Deucher <alexdeucher@gmail.com> 18125Date: Mon Feb 9 16:33:20 2009 -0500 18126 18127 Bump for rc release 18128 18129commit e85b7a1c5948b5123eddf9dfbb9e002d74fbe9fb 18130Author: Alex Deucher <alexdeucher@gmail.com> 18131Date: Mon Feb 9 14:13:16 2009 -0500 18132 18133 R6xx/R7xx Xv: fix cache flush buffer size for planar 18134 18135commit 2a893bac1faffd28dce6d9a9693108196c71e6d3 18136Author: Alex Deucher <alexdeucher@gmail.com> 18137Date: Mon Feb 9 13:18:17 2009 -0500 18138 18139 R6xx/R7xx Xv: add accelerated uploads for planar formats 18140 18141commit 231aee18a73805be2f6c962e94a8345dd89fd0df 18142Author: Alex Deucher <alexdeucher@gmail.com> 18143Date: Mon Feb 9 13:02:27 2009 -0500 18144 18145 R6xx/R7xx Xv: implement native shader for planar formats 18146 18147commit 6c76bfe8105e3cf4e7e6ea1bfe1235be2079110f 18148Author: Alex Deucher <alexdeucher@gmail.com> 18149Date: Mon Feb 9 09:47:22 2009 -0500 18150 18151 R6xx/R7xx UTS: move actual upload to separate function 18152 18153 So it can be shared with Xv 18154 18155commit 04859362ce64ed4ae48305a83c92059899c1db4b 18156Author: Alex Deucher <alexdeucher@gmail.com> 18157Date: Sun Feb 8 20:13:06 2009 -0500 18158 18159 Add some missing r6xx/r7xx pci ids 18160 18161commit 132e4c575dc4675f4995e45f08c53c26bffd999a 18162Author: Yang Zhao <yang@yangman.ca> 18163Date: Sat Feb 7 13:54:51 2009 -0500 18164 18165 R6xx/R7xx EXA: Optimize overlapping copy 18166 18167 Overlapping copy is now done in chunks proportional to the 18168 non-overlapping area. 18169 18170 Diagonal overlaps are also handled properly. 18171 18172commit 8e9ef8ff581892cbe1b7ea56d48b9a1abd70179d 18173Author: Pierre Ossman <pierre@ossman.eu> 18174Date: Sat Feb 7 18:57:47 2009 +0100 18175 18176 Xv vsync support on r6xx/r7xx cards. 18177 18178commit 2222f0fd700f100b2e91fac2babe7d1b53f56c3e 18179Author: Pierre Ossman <pierre@ossman.eu> 18180Date: Sat Feb 7 18:56:42 2009 +0100 18181 18182 Fix bad range adjustment in VLINE code. 18183 18184commit 8ce6c024e31f6a3f5ae6c882738b1e64ae2944a1 18185Author: Alex Deucher <alexdeucher@gmail.com> 18186Date: Sat Feb 7 10:32:01 2009 -0500 18187 18188 R6xx/R7xx Xv: fix typos in cache flushing commands 18189 18190commit bba5d806cc46328ad9e4c521bfeeb2a59f96862b 18191Author: Alex Deucher <alexdeucher@gmail.com> 18192Date: Fri Feb 6 19:44:17 2009 -0500 18193 18194 AVIVO: better fix for rotation 18195 18196 This should handle initial rotation as well. 18197 18198commit 28cf3492f431f325cb08d7c6aade43d8886df3f6 18199Author: Alex Deucher <alexdeucher@gmail.com> 18200Date: Fri Feb 6 19:20:18 2009 -0500 18201 18202 AVIVO: fix rotation 18203 18204 When rotation is active, virtualX and virtualY change. 18205 18206commit 1b45936ae614244aa49b1a5d3c7fc39773c4f9b6 18207Author: Alex Deucher <alexdeucher@gmail.com> 18208Date: Fri Feb 6 18:29:55 2009 -0500 18209 18210 R6xx/R7xx EXA: Fix typo in DFS 18211 18212 noticed by pzad in IRC 18213 18214commit c06d89e16d5b2553142e8641e66080e1770c1563 18215Author: Alex Deucher <alexdeucher@gmail.com> 18216Date: Fri Feb 6 18:07:22 2009 -0500 18217 18218 R6xx/R7xx EXA: fallback on overlapping blits for now 18219 18220 Leave this disabled until we get a proper solution. 18221 18222commit 1d5fc3febf3470b94c423a1eda5e0683856909df 18223Author: Alex Deucher <alexdeucher@gmail.com> 18224Date: Fri Feb 6 14:43:30 2009 -0500 18225 18226 Revert "r6xx/r7xx EXA: Optimize overlapping copy" 18227 18228 This reverts commit 0dfadc1843e0d14b9cc1ee19a72f4fd60a2c495b. 18229 18230 This doesn't always work properly. 18231 18232commit def317e22d072405cd95ddb19d17eacd784ffd9e 18233Author: Alex Deucher <alexdeucher@gmail.com> 18234Date: Fri Feb 6 11:04:59 2009 -0500 18235 18236 R6xx/R7xx EXA: add accelerated UTS/DFS hooks 18237 18238 I'm not sure how much of a win these are. I need to 18239 do some benchmarking. 18240 18241commit 0dfadc1843e0d14b9cc1ee19a72f4fd60a2c495b 18242Author: Yang Zhao <yang@yangman.ca> 18243Date: Fri Feb 6 10:29:39 2009 -0500 18244 18245 r6xx/r7xx EXA: Optimize overlapping copy 18246 18247 When source and destination blocks are only offset horizontally, it 18248 appears to be unnecessary to perform careful, segment-by-segment copy. 18249 The code path that does this is taken out completely. 18250 18251 For the case where offset is only vertical, copying is now done by 18252 height of the non-overlapping area each time, instead of always 18253 line-by-line. 18254 18255commit 3d17bd199423e92d201c20f047d5e699942af976 18256Author: Alex Deucher <alexdeucher@gmail.com> 18257Date: Thu Feb 5 19:25:03 2009 -0500 18258 18259 Revert "R6xx/R7xx EXA: improve overlapping copy performance" 18260 18261 This reverts commit b24827c9d211e8a35da53b665385a7733d19910d. 18262 18263 This seems to cause corruption in some cases. 18264 18265commit bf74055f543e7f0664741620fb1fe827ebc12711 18266Author: Alex Deucher <alexdeucher@gmail.com> 18267Date: Thu Feb 5 17:31:43 2009 -0500 18268 18269 r6xx/r7xx EXA: fix corruption when doing sw access 18270 18271 need to wait until the engine is idle. Ideally we wait 18272 on a timestamp shadowed in memory, but polling the 18273 GRBM_STATUS reg will do for now. 18274 18275commit b24827c9d211e8a35da53b665385a7733d19910d 18276Author: Alex Deucher <alexdeucher@gmail.com> 18277Date: Thu Feb 5 16:21:37 2009 -0500 18278 18279 R6xx/R7xx EXA: improve overlapping copy performance 18280 18281 send vertices for each line of the copy, but only draw once 18282 18283commit 729fe756f809a41494dc161252d82313956e4a4b 18284Author: Alex Deucher <alexdeucher@gmail.com> 18285Date: Thu Feb 5 16:08:35 2009 -0500 18286 18287 r6xx/r7xx EXA: cleanup overlapping copy 18288 18289commit 16e01a5796a8e8b86ad25d3aa45b9e7044dc72d1 18290Author: Alex Deucher <alexdeucher@gmail.com> 18291Date: Tue Feb 3 19:46:10 2009 -0500 18292 18293 Fix encoder accounting 18294 18295 Should fix bug 19924 18296 18297commit d7bf7b9c17add31dc9a115b56a563c7f9bf2cdb6 18298Author: Alex Deucher <alexdeucher@gmail.com> 18299Date: Tue Feb 3 18:28:38 2009 -0500 18300 18301 EXA: fix and re-enable Solid() on R7xx 18302 18303 missing last bit in alu clause in solid PS 18304 18305commit 9abb09d5f43320bd27b336612d39d264fe93549f 18306Author: Alex Deucher <alexdeucher@gmail.com> 18307Date: Tue Feb 3 11:17:13 2009 -0500 18308 18309 Fix bad rv710 pci id 18310 18311commit e5b916770946a9eebcb4bd1e6f698220db8c718a 18312Author: Alex Deucher <alexdeucher@gmail.com> 18313Date: Tue Feb 3 10:48:41 2009 -0500 18314 18315 Allow rotation on r6xx/r7xx 18316 18317commit d1f071c7f1dad6babfbcfcc2cb2b722a4987f372 18318Author: Alex Deucher <alexdeucher@gmail.com> 18319Date: Tue Feb 3 10:44:10 2009 -0500 18320 18321 Initial R6xx/R7xx EXA and textured video support 18322 18323commit c88c3ef6f3db266c1aacba5297b8dfc8b66bf00e 18324Author: Wolke Liu <wolke.liu@amd.com> 18325Date: Mon Feb 2 17:01:34 2009 -0500 18326 18327 AVIVO: Save/restore vga pll registers 18328 18329 This fixes some VT switch issues on some chips 18330 18331commit 6fac3cefd1f46161c1e276ba40e72da2823aa9f6 18332Author: Alex Deucher <alexdeucher@gmail.com> 18333Date: Sat Jan 31 14:21:02 2009 -0500 18334 18335 Return NULL for encoder if no active device is assigned 18336 18337 fixes bug 19855 18338 18339commit db178c852d00e9b85513d750b5f8f7495f9e3360 18340Author: Alan Coopersmith <alan.coopersmith@sun.com> 18341Date: Fri Jan 30 20:37:41 2009 -0800 18342 18343 Add README with pointers to mailing list, bugzilla & git repos 18344 18345commit fa8e5a4fc236f8f15f462cb0d6164b194a65a118 18346Author: Michel Dänzer <daenzer@vmware.com> 18347Date: Fri Jan 30 16:16:26 2009 +0100 18348 18349 EXA: Accelerate Composite of RepeatPad/Reflect pictures when possible. 18350 18351commit 3a6e958c59b4b44fe620c8abe760ed1abf0007b1 18352Author: airlied <airlied@roo.(none)> 18353Date: Fri Jan 30 07:56:14 2009 +1000 18354 18355 rs780: include RS780 in the InitMemory to leave alone 18356 18357 This stops the MC_FB_LOCATION getting reassigned to 0, which 18358 is really bad thing to happen. 18359 18360 I've had reported memory corruption on these cards so hopefully this 18361 fixes it. 18362 18363commit bd8021d46e9066e4cd116c03a7b7adcfe2557aff 18364Merge: 4f88dd83 20d5dd38 18365Author: Alex Deucher <alexdeucher@gmail.com> 18366Date: Wed Jan 28 14:33:14 2009 -0500 18367 18368 Merge branch 'atom-rework' and fix conflicts 18369 18370commit 20d5dd387da555e895e2b73fb53e2b026dd91003 18371Author: Alex Deucher <alexdeucher@gmail.com> 18372Date: Wed Jan 28 14:12:26 2009 -0500 18373 18374 Move encoder specific data to encoder dev_priv 18375 18376commit 4f88dd83f290fb38e41256a7f1804008c0f28139 18377Merge: 1c5678fc 2ca970c7 18378Author: Alex Deucher <alexdeucher@gmail.com> 18379Date: Wed Jan 28 13:13:13 2009 -0500 18380 18381 Merge branch 'atom-tvout' of ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati 18382 18383 Conflicts: 18384 18385 src/radeon_atombios.c 18386 18387commit 1a62360e8260eaf5f98fa6be3f89f64c8cae7d32 18388Author: Alex Deucher <alexdeucher@gmail.com> 18389Date: Mon Jan 26 10:11:28 2009 -0500 18390 18391 Encoders not assigned yet, use supported devices 18392 18393 need to move this to encoder setup 18394 18395commit 44d9ceacb1fa6e1859d5d8bd1c55405aea24ad85 18396Author: Alex Deucher <alexdeucher@gmail.com> 18397Date: Mon Jan 26 03:47:32 2009 -0500 18398 18399 Fix legacy output setup 18400 18401 don't have assigned encoders at this point. Need to 18402 eventually move this stuff to encoder dev_priv. 18403 18404commit 0c24cbcea56c8f36374aca0f8ba849adfbf0ecfe 18405Author: Alex Deucher <alexdeucher@gmail.com> 18406Date: Mon Jan 26 03:24:19 2009 -0500 18407 18408 ATOM: print useful output info for DPMS events 18409 18410commit 2f92b885457c5cee2098ebbf8c6fe5a5eb98ec48 18411Author: Alex Deucher <alexdeucher@gmail.com> 18412Date: Mon Jan 26 03:12:44 2009 -0500 18413 18414 Warning fixes 18415 18416commit 4cbf6189f3a7945a1141b0451e4f7881ceaaadcd 18417Author: Alex Deucher <alexdeucher@gmail.com> 18418Date: Mon Jan 26 02:59:51 2009 -0500 18419 18420 Move legacy output setup functions to legacy_output.c 18421 18422commit 93ed767255b60fbdf5b416b9bd06c366036a5141 18423Author: Alex Deucher <alexdeucher@gmail.com> 18424Date: Mon Jan 26 02:40:41 2009 -0500 18425 18426 Fix off by one when printing encoder name 18427 18428 - also cleanup some debugging output 18429 18430commit b515eee4b1cafae78ed39cd91c18b4f5e671b20b 18431Author: Alex Deucher <alexdeucher@gmail.com> 18432Date: Mon Jan 26 01:50:18 2009 -0500 18433 18434 Additional output cleanup 18435 18436commit 1c5678fc452dbc93135e9ece50a4cff6583f6078 18437Author: Michel Dänzer <daenzer@vmware.com> 18438Date: Sat Jan 24 11:10:00 2009 +0100 18439 18440 Drop memcpy fallbacks from EXA UploadToScreen and DownloadFromScreen hooks. 18441 18442 While in theory it's possible for the PrepareAccess hook to fail on big endian 18443 platforms, in practice it isn't at this point because there's no other users of 18444 the surface registers. 18445 18446commit 5cbd6082b9f8d7009ee9ef5e036673f99e1bfb05 18447Author: Michel Dänzer <daenzer@vmware.com> 18448Date: Sat Jan 24 11:07:12 2009 +0100 18449 18450 Don't transform EXA Composite mask coordinates when there's no mask. 18451 18452commit 7e54c4b8acafead8de5d0dd6fdc1b619f832dd6f 18453Author: Thomas Jaeger <ThJaeger@gmail.com> 18454Date: Fri Jan 23 20:03:27 2009 -0500 18455 18456 Fall back to software for unsupported repeat modes 18457 18458commit 289b4a377cfff0c2a8ec938951cfbae1e270570e 18459Author: Alex Deucher <alexdeucher@gmail.com> 18460Date: Fri Jan 23 20:00:43 2009 -0500 18461 18462 Remove OutputType and other cruft 18463 18464commit 5f98508effbdc1fd62d0df15798722c147d88da1 18465Author: Maciej Cencora <m.cencora@gmail.com> 18466Date: Fri Jan 23 19:33:25 2009 -0500 18467 18468 Make sure gb_num_pipes is initialized when DRI is disabled 18469 18470 Found with valgrind 18471 18472commit 0faab02cc732414af3cfb2014d3ddc1139aff063 18473Author: Alex Deucher <alexdeucher@gmail.com> 18474Date: Fri Jan 23 17:35:01 2009 -0500 18475 18476 Remove some unused cruft 18477 18478commit 025082160c53c01a44f4d681cc63bf26ac4561d5 18479Author: Alex Deucher <alexdeucher@gmail.com> 18480Date: Fri Jan 23 17:31:18 2009 -0500 18481 18482 track encoder state 18483 18484commit 82f12e5a40c1fbcb91910a0f8b725c34fff02aae 18485Author: Alex Deucher <alexdeucher@gmail.com> 18486Date: Fri Jan 23 16:21:33 2009 -0500 18487 18488 Remove TMDSType, DACType, LVDSType from output rec 18489 18490commit 7b42b57d748f77ff6ef6f8ade7cbc983a98d7204 18491Author: Alex Deucher <alexdeucher@gmail.com> 18492Date: Wed Jan 21 16:28:00 2009 -0500 18493 18494 few more logic pasto's bits I missed 18495 18496commit fef1c12fbe4ceb4676258850d1fa5cac994f7c21 18497Author: Alex Deucher <alexdeucher@gmail.com> 18498Date: Wed Jan 21 15:34:01 2009 -0500 18499 18500 Fix compilation with RADEON_TRACE_FALL set 18501 18502 noticed by curaga on phoronix forum 18503 18504commit 1305376b009d17d65b4935896dc35cb7530f7c52 18505Author: Alex Deucher <alexdeucher@gmail.com> 18506Date: Wed Jan 21 15:20:47 2009 -0500 18507 18508 Move active_device setup to detect() 18509 18510 Also add encoder debugging printouts 18511 18512commit 9e0cc0ca5d7bc68756ae92ee3b5b2392366be71a 18513Author: Alex Deucher <alexdeucher@gmail.com> 18514Date: Wed Jan 21 14:47:08 2009 -0500 18515 18516 Fix logic cut and paste error 18517 18518commit 12010ae9296635ffa5f42ce0cd14482aaa290cf7 18519Author: Alex Deucher <alexdeucher@gmail.com> 18520Date: Wed Jan 21 13:52:11 2009 -0500 18521 18522 More legacy rework 18523 18524commit 3f8bbede8bf639050f773780888b89cbdcb425b8 18525Author: Alex Deucher <alexdeucher@gmail.com> 18526Date: Wed Jan 21 13:45:11 2009 -0500 18527 18528 fix legacy crtc routing and add some debugging info 18529 18530commit 80e0162ea2c1f793964731268251aba2d3243127 18531Author: Alex Deucher <alexdeucher@gmail.com> 18532Date: Wed Jan 21 13:28:14 2009 -0500 18533 18534 ATOM: fix encoder init 18535 18536commit aea27d9caabb0e54e868a21e6e547733a4ca2709 18537Author: Alex Deucher <alexdeucher@gmail.com> 18538Date: Wed Jan 21 12:33:49 2009 -0500 18539 18540 Switch legacy output code to use new encoder objects 18541 18542commit 9b44a4b57b00b150d041a5d82cb2df42159c062b 18543Author: Alex Deucher <alexdeucher@gmail.com> 18544Date: Wed Jan 21 11:32:24 2009 -0500 18545 18546 ATOM: more output cleanup 18547 18548commit 9e283fa9de34407fa6f0d5a8006a611b104f4bf2 18549Author: Alex Deucher <alexdeucher@gmail.com> 18550Date: Wed Jan 21 03:16:20 2009 -0500 18551 18552 Fixup encoder setup on pre-ATOM chips 18553 18554commit 5850abbdcd22b70bdbccd3642f8ba1a27e38e7cd 18555Author: Alex Deucher <alexdeucher@gmail.com> 18556Date: Wed Jan 21 02:28:47 2009 -0500 18557 18558 clean up encoder setup 18559 18560 - make sure not to duplicate encoders 18561 - use single encoder array 18562 18563commit 152d1d4c179c262be4d0a5618400aa13b4820342 18564Author: Alex Deucher <alexdeucher@gmail.com> 18565Date: Tue Jan 20 00:58:18 2009 -0500 18566 18567 First pass at converting legacy code to encoder objects 18568 18569commit 7cc7ba8bf9ce2f0eebfd9b48aa648c416c333ff4 18570Author: Alex Deucher <alexdeucher@gmail.com> 18571Date: Mon Jan 19 18:42:12 2009 -0500 18572 18573 ATOM: round 1 of output rework 18574 18575commit 25f3878858e6efb486888e5271b60d4ba08c73c6 18576Author: Alex Deucher <alexdeucher@gmail.com> 18577Date: Mon Jan 19 15:54:42 2009 -0500 18578 18579 start to re-org outputs 18580 18581commit cbeb310d35e5ea25ba7a45a2c555468391cdf29e 18582Author: Alex Deucher <alexdeucher@gmail.com> 18583Date: Fri Jan 16 19:53:49 2009 -0500 18584 18585 ATOM: switch to define for external tmds 18586 18587commit d0d157abced43f822885fa26f14a4924766594cd 18588Author: Alex Deucher <alexdeucher@gmail.com> 18589Date: Fri Jan 16 19:51:07 2009 -0500 18590 18591 ATOM: combine DAC setup functions 18592 18593 plus a few other clean-ups 18594 18595commit f6f969338f8187529c7ae9b2adc980b71a3f5837 18596Author: Alex Deucher <alexdeucher@gmail.com> 18597Date: Fri Jan 16 18:10:34 2009 -0500 18598 18599 DCE30: LVTMA requires DIG2 encoder 18600 18601 Fixes bug 19544 18602 18603commit 549ae63aecc2624a69b41ce4db75883d03d59c23 18604Author: Alex Deucher <alexdeucher@gmail.com> 18605Date: Thu Jan 15 19:26:40 2009 -0500 18606 18607 RV280 Add another AGP quirk 18608 18609commit 0e69068a8ed82e3fb7349d8224933451955e5f16 18610Author: Tormod Volden <debian.tormod@gmail.com> 18611Date: Tue Jan 13 20:31:17 2009 -0500 18612 18613 [PATCH] Add yet another AGP quirk for RV280 18614 18615commit b437da34cb4f7c637674fb432ad81b9bf32244a3 18616Author: Christiaan van Dijk <dvbmail@xs4all.nl> 18617Date: Mon Jan 12 11:54:15 2009 -0500 18618 18619 R3xx/R4xx: Maximize the use of clipped triangles for Xv rendering 18620 18621 While trying to get tear free video on my RS690 I found some problems in the code for rendering the 18622 video as a clipped triangle. The size limitations are quite strict and scaling the video output window 18623 creates all kinds of weird effects. This patch maximizes the use of this function and resolves the 18624 scaling issues. There could be a problem with zero target height/width, not sure if this is handled 18625 elsewhere. 18626 18627commit 3d106b990fa8daba96e590c57b164640a2faae7f 18628Author: Alan Coopersmith <alan.coopersmith@sun.com> 18629Date: Fri Jan 9 16:26:04 2009 -0800 18630 18631 Remove xorgconfig & xorgcfg from See Also list in man page 18632 18633commit 14d29a223028c183ba11e717a3737285d0ac8de1 18634Author: Alex Deucher <alexdeucher@gmail.com> 18635Date: Wed Jan 7 17:04:42 2009 -0500 18636 18637 RV280: add another AGP quirk 18638 18639commit 1795706a6bb13272a58e8fa1bd5ad2603d553dde 18640Author: Nicos Gollan <gtdev@spearhead.de> 18641Date: Wed Jan 7 14:25:24 2009 -0500 18642 18643 [PATCH] Fixed enumerations in radeon-output.c 18644 18645commit f69ad5aa17bb907105c1657803628d52b5ff3f11 18646Author: Alex Deucher <alexdeucher@gmail.com> 18647Date: Wed Jan 7 11:33:53 2009 -0500 18648 18649 Bump version post release 18650 18651commit a0dd5d7ee3f038a9bfe051db8dbfac4934a81581 18652Author: Alex Deucher <alexdeucher@gmail.com> 18653Date: Tue Jan 6 19:08:28 2009 -0500 18654 18655 Bump version for release 18656 18657commit c3fb8bb279959512a4ced644a64cc660c5cd97bd 18658Author: Alex Deucher <alexdeucher@gmail.com> 18659Date: Tue Jan 6 17:41:22 2009 -0500 18660 18661 ATOM: rework encoder/transmitter setup 18662 18663commit 9bb6b8ceaf772e1ce63610812a12260ee3860c27 18664Author: Alex Deucher <alexdeucher@gmail.com> 18665Date: Tue Jan 6 16:25:26 2009 -0500 18666 18667 ATOM: refactor output dpms 18668 18669commit 467995d960bdefb1911eabfc2a047cf816013101 18670Author: Alex Deucher <alexdeucher@gmail.com> 18671Date: Tue Jan 6 15:18:41 2009 -0500 18672 18673 ATOM: Adjust PLL setup for recent atom changes 18674 18675commit ee6310174c44809c761dbdb423e56e61392a0e10 18676Author: Alex Deucher <alexdeucher@gmail.com> 18677Date: Tue Jan 6 14:59:22 2009 -0500 18678 18679 ATOM: handle cases where TMDS uses linkb 18680 18681 fixes bug 18564 18682 18683commit 690999b682c0f954e27d9f1b3a3d8db4ef865c21 18684Author: Alex Deucher <alexdeucher@gmail.com> 18685Date: Tue Jan 6 14:55:49 2009 -0500 18686 18687 ATOM: rework object table parsing 18688 18689commit e23baa2fa8b25b68cd473aec7a3d8cf3ee7b0ea8 18690Author: Alex Deucher <alexdeucher@gmail.com> 18691Date: Tue Jan 6 14:38:06 2009 -0500 18692 18693 R100: Fix typo in e55598cc1b35d398c9eb2c3946df71456775f621 18694 18695 Noticed by Andreas Stenglein 18696 18697commit d5cb24075928690377fd2dfaa32abfebff9990df 18698Author: Alex Deucher <alexdeucher@gmail.com> 18699Date: Sun Jan 4 21:04:52 2009 -0500 18700 18701 ATOM: add format rev 2 for digitial command tables 18702 18703 should fix bug 18911 18704 18705commit 1813456af3def38c6f53103054976f3092491746 18706Author: Alex Deucher <alexdeucher@gmail.com> 18707Date: Sun Jan 4 21:01:05 2009 -0500 18708 18709 ATOM: fail on more tables with unhandled revisions 18710 18711commit 90ce00a3dbf37833a71db234372695460c8f0342 18712Author: Alex Deucher <alexdeucher@gmail.com> 18713Date: Sun Jan 4 19:13:53 2009 -0500 18714 18715 AGP: Print both device and subsytem pci ids for agp 18716 18717 helpful for adding quirks with libpciaccess 18718 18719commit 006d736d890a6b2f1d7eb8ee20ccf467a939516f 18720Author: Alex Deucher <alexdeucher@gmail.com> 18721Date: Sun Jan 4 19:08:37 2009 -0500 18722 18723 Another AGP quirk 18724 18725 debian bug 510208 18726 18727commit c39ca66a7e216b7492fc5e3680ac535831cfb9a1 18728Author: Alex Deucher <alexdeucher@gmail.com> 18729Date: Tue Dec 30 19:05:58 2008 -0500 18730 18731 R2xx: add another AGP quirk 18732 18733 from http://www.phoronix.com/forums/showthread.php?t=14626 18734 18735commit 877be1b268f275b8ad83d5e245a0f0d3b1a05091 18736Author: Fabio <fabio.ped@libero.it> 18737Date: Mon Dec 29 18:11:10 2008 -0500 18738 18739 man page updates 18740 18741 bug 19329 18742 18743commit 837f7e8a48ea35d21de24b620830a2c6db854861 18744Author: Dave Airlie <airlied@linux.ie> 18745Date: Tue Dec 30 06:50:54 2008 +1000 18746 18747 atombios: endian fix in r600 object table handling 18748 18749 This fixes the r600 object table on r600. 18750 18751commit 45f8b911659dba7f43b9b70bde4b9c3be51b0b5b 18752Author: Alex Deucher <alexdeucher@gmail.com> 18753Date: Mon Dec 29 10:25:33 2008 -0500 18754 18755 RS600: remove gpio quirk for ddc 18756 18757 Doesn't seem necessary on RS600. I'll try and find out more. 18758 18759commit e913990030c536427bfecdb7b3d1e18f1beb90d2 18760Author: Alex Deucher <alexdeucher@gmail.com> 18761Date: Mon Dec 29 10:22:20 2008 -0500 18762 18763 Revert "radeon: maybe special handling only needed for DVI port not HDMI one." 18764 18765 This reverts commit c877450a75d3fd4a2f200af197c691669966f6f4. 18766 18767 This breaks RS690 setups. I think for now we just apply this quirk to 18768 RS690/RS740. I'll follow up with the bios folks about RS600. 18769 18770commit c877450a75d3fd4a2f200af197c691669966f6f4 18771Author: Dave Airlie <airlied@linux.ie> 18772Date: Mon Dec 29 09:05:45 2008 +1000 18773 18774 radeon: maybe special handling only needed for DVI port not HDMI one. 18775 18776 fixes ciaranm's rs600 board on radeon DDC 18777 18778commit 2a9bc6be3246ec230b181391221e3a46beb94a23 18779Author: Dave Airlie <airlied@linux.ie> 18780Date: Mon Dec 29 09:05:16 2008 +1000 18781 18782 radeon: a-bit motherboard has no DVI output 18783 18784 thanks to ciaranm for finding this. 18785 18786commit eccf3fed2be54f5c02fefef32e39c4084acbbe22 18787Author: Alex Deucher <alexdeucher@gmail.com> 18788Date: Fri Dec 26 19:51:58 2008 -0500 18789 18790 Make sure to include radeon_drm.h in the tarball 18791 18792commit e55598cc1b35d398c9eb2c3946df71456775f621 18793Author: Alex Deucher <alexdeucher@gmail.com> 18794Date: Fri Dec 26 19:31:55 2008 -0500 18795 18796 R2xx EXA: limit composite to 2047 like r1xx 18797 18798 See bug 19269 18799 18800commit 631123d144d088d4f77a0599c34deaa31d551d71 18801Author: Dave Airlie <airlied@redhat.com> 18802Date: Tue Dec 23 10:08:46 2008 +1000 18803 18804 radeon: update to 6.9.0.91 18805 18806 do a prerelease for end of year. 18807 18808commit 047f7603174f6047090f4fafbe6488d697eb78bc 18809Author: Dave Airlie <airlied@linux.ie> 18810Date: Tue Dec 23 08:16:25 2008 +1000 18811 18812 radeon: pass distcheck 18813 18814commit c0c33dab44e6966b1702d4e8cfba3537fc6e2d5c 18815Author: Patrick Haller <patrick.haller@haller-consult.com> 18816Date: Mon Dec 22 03:06:23 2008 -0500 18817 18818 Fix off by one in EXA composite limit checking 18819 18820 Patch from Patrick, with some updates from me: 18821 - fix r200 limits 18822 - note about r100 limits 18823 18824commit d01a609a2b07da4ca0f182e79459432584ec61f6 18825Author: Alex Deucher <alexdeucher@gmail.com> 18826Date: Mon Dec 22 02:18:43 2008 -0500 18827 18828 R3xx-R5xx: better fix for xv primitive tearing issues 18829 18830 R5xx: always use single clipped triangle 18831 R3xx/R4xx: use single clipped triangle up to guardband limit, 18832 then use quad. 18833 18834commit 4d009d14c9c72f6e1bc8100ec8d9b8037ee4e514 18835Merge: d52882d8 4e96278b 18836Author: Dave Airlie <airlied@redhat.com> 18837Date: Mon Dec 22 16:00:01 2008 +1000 18838 18839 Merge branch 'atom-tvout' 18840 18841commit 4e96278b581e296c1203d97a6d7aa3bff3977222 18842Author: Dave Airlie <airlied@redhat.com> 18843Date: Mon Dec 22 15:58:35 2008 +1000 18844 18845 atombios/tv: add an option to enable atom tv-out for users. 18846 18847 This code is still experimental but we will allow users to enable it for experimental reasons 18848 18849commit d52882d8188830dd52fa112dadcf5ea7f3e5fd5e 18850Author: Dave Airlie <airlied@redhat.com> 18851Date: Mon Dec 22 09:27:47 2008 +1000 18852 18853 radeon: setup 3D engine even when no DRI. 18854 18855 This should fix some missing font issues in EXA without DRI cases. 18856 18857commit 2346fd7cbd90dcdce2b361d374a53da064fa6a12 18858Author: Dave Airlie <airlied@redhat.com> 18859Date: Thu Dec 18 16:35:13 2008 +1000 18860 18861 radeon: add all new pci ids for rv730/rv710 families 18862 18863commit 2ca970c7aa1f3f212f03bf3287d6b60eeafd68ae 18864Author: Alex Deucher <alexdeucher@gmail.com> 18865Date: Wed Dec 17 13:07:19 2008 -0500 18866 18867 Allow arbitrary tv-out modes 18868 18869commit 5da147eec70abb7c9125a8940089fd3ff8bdea95 18870Author: Alex Deucher <alexdeucher@gmail.com> 18871Date: Wed Dec 17 05:37:58 2008 -0500 18872 18873 missed one in last commit 18874 18875commit 6bfbb930c549e67f68969f7f0483380417f755b5 18876Author: Alex Deucher <alexdeucher@gmail.com> 18877Date: Wed Dec 17 05:36:09 2008 -0500 18878 18879 Make sure we hit the right bios reg 18880 18881commit 6ed0e68060fd2775e2d4027ff1533729e87848ba 18882Author: Alex Deucher <alexdeucher@gmail.com> 18883Date: Wed Dec 17 04:54:50 2008 -0500 18884 18885 properly handle EnableYUV 18886 18887 A bit of a hack. we really need to restructure the 18888 bios scratch reg handling. 18889 18890commit f21e673aced1bb5138bbb80e6d9724530eb95ebf 18891Author: Alex Deucher <alexdeucher@gmail.com> 18892Date: Wed Dec 17 03:17:45 2008 -0500 18893 18894 Fix colors on tv-out 18895 18896 Call EnableYUV to set the transform and enable/disable it. 18897 should also fix green tint with tv connected at boot issues. 18898 18899commit b2b43905a5385a8bb0b59b8e50952863d8dacb59 18900Author: Alex Deucher <alexdeucher@gmail.com> 18901Date: Wed Dec 17 00:59:52 2008 -0500 18902 18903 Pre-avivo: fix FP setup 18904 18905 - make sure to clear various shadow timing bits 18906 - crtc1 select bit was set wrong. 18907 - should fix bugs like 19100 18908 18909commit 3754cc23abac99880eda87d06aa42e16081b6c8a 18910Author: Alex Deucher <alexdeucher@gmail.com> 18911Date: Wed Dec 17 00:56:19 2008 -0500 18912 18913 R2xx: switch EXA composite to rects rather quads 18914 18915 For conistency with Xv path and to reduce potential diagonal tearing 18916 18917commit be12cf8fcc7a5a3c3ee2c630474e2f9d9368f4d4 18918Author: Dave Airlie <airlied@redhat.com> 18919Date: Wed Dec 17 15:28:25 2008 +1000 18920 18921 radeon: r500 tv-out force scaler values to nice set that looks correct 18922 18923commit 01f7a50d76f706a456de7a4eb8323939d905c82b 18924Author: Dave Airlie <airlied@redhat.com> 18925Date: Wed Dec 17 13:43:21 2008 +1000 18926 18927 r500: re-enable TV out 18928 18929commit 45b9ff5e250d69de0df41f68b33524f7e2ce13b7 18930Author: Dave Airlie <airlied@redhat.com> 18931Date: Wed Dec 17 13:42:10 2008 +1000 18932 18933 radeon: r500 PAL timings are slightly incorrect 18934 18935 this stablises the PAL image at least 18936 18937commit 435d996949185201a97493e131f9920a32863acd 18938Author: Dave Airlie <airlied@redhat.com> 18939Date: Wed Dec 17 11:50:48 2008 +1000 18940 18941 radeon: tv timings have crev/frev wrong way around for some reason. 18942 18943commit b9549e89cbdf8fff4aa7bd9f3adfb7d5732508cc 18944Author: Dave Airlie <airlied@redhat.com> 18945Date: Wed Dec 17 11:40:53 2008 +1000 18946 18947 atombios: re-enable TV dpms 18948 18949commit db85ff2e0c52a030546c0cc8ed0b851418f3e045 18950Author: Dave Airlie <airlied@redhat.com> 18951Date: Wed Dec 17 11:37:17 2008 +1000 18952 18953 atom/r600: re-enable TV outputs 18954 18955commit ff7b52c9a48f2e417acf6bec3b72546ad9b33471 18956Author: Dave Airlie <airlied@redhat.com> 18957Date: Wed Dec 17 11:36:38 2008 +1000 18958 18959 radeon: use latest atombios constants for TV encoders in set_crtc_source 18960 18961commit 1e0c7d3f0bf6be67b976fdea8fa2df0b4b16435f 18962Author: Dave Airlie <airlied@redhat.com> 18963Date: Wed Dec 17 11:35:06 2008 +1000 18964 18965 atom: spc3 only requires ucEncoderMode for DIG outputs 18966 18967commit 105cfd0b8696347ef71ad9daa941abb308bd6c2c 18968Author: Dave Airlie <airlied@redhat.com> 18969Date: Wed Dec 17 11:31:15 2008 +1000 18970 18971 atom/tv: enable scaler can set the TV output scaler up. 18972 18973 Add the tv standard to the scaler enable for TV use. 18974 18975commit 2a768abbd2f18abe203571213260bccc225f70fd 18976Author: Dave Airlie <airlied@redhat.com> 18977Date: Wed Dec 17 11:29:07 2008 +1000 18978 18979 atom/tv: fixup the tv out default if bios doesn't specify 18980 18981commit b7fdd5da08bfec705691363b8890699c794dce53 18982Author: Dave Airlie <airlied@redhat.com> 18983Date: Wed Dec 17 11:24:56 2008 +1000 18984 18985 atombios: fixup parsing of TV tables on r500 and r600 cards. 18986 18987 The TV tables changed between for later cards. 18988 18989commit e3b08cb59716b525b3063b184eab70e32b9d66f0 18990Author: Dave Airlie <airlied@redhat.com> 18991Date: Tue Dec 16 15:43:16 2008 +1000 18992 18993 atombios: correct fix for previous issue. 18994 18995 I committed an old patch. 18996 18997commit 1431bc6900b0e0d7f486b5b3ba281873472b3fdc 18998Author: Dave Airlie <airlied@redhat.com> 18999Date: Tue Dec 16 15:31:34 2008 +1000 19000 19001 atombios: fix issue with Clearing of MM_INDEX 0. 19002 19003 While playing with tv-out I discovered this issue with clearing 19004 the MM_INDEX register. 19005 19006commit f72367aff2d09161d30066a00673ba291b292a36 19007Author: Dave Airlie <airlied@redhat.com> 19008Date: Tue Dec 16 12:32:24 2008 +1000 19009 19010 radeon: add initial rv730 pciids. 19011 19012 Lots more to come, this just adds a few to start. 19013 19014commit 4d2429f3ec4d7d3e1512b268c7ff5cbff8d186b4 19015Author: Dave Airlie <airlied@redhat.com> 19016Date: Tue Dec 16 12:19:00 2008 +1000 19017 19018 DCE32: add support for DCE3.2 digital outputs. 19019 19020 This adds support for the digital outputs on the RV730. 19021 19022commit 4d3792fd1b9f733107582436540f483d8993199f 19023Author: Dave Airlie <airlied@redhat.com> 19024Date: Fri Dec 12 10:34:57 2008 +1000 19025 19026 radeon: fixes from Alex for some output engines 19027 (cherry picked from commit 1e8ac6ea7b35ccbc3649c96ead60f69158f66ebc) 19028 19029commit f892154dbd0d8b6dd8abdf2196ab0c7c18327412 19030Author: Alex Deucher <alexdeucher@gmail.com> 19031Date: Mon Dec 15 19:41:32 2008 -0500 19032 19033 R3xx-R5xx: switch back to quad rendering for Xv 19034 19035 The large clipped triangle trick hits the guardband limits on 19036 large displays. This gets textured video working again 19037 until we sort out how to deal with the situation better. 19038 19039 See bug 19046 19040 19041commit 62344e460f5d0cac81f94c535cda77d3d76528a9 19042Author: Alex Deucher <alexdeucher@gmail.com> 19043Date: Mon Dec 15 11:30:42 2008 -0500 19044 19045 Remove remnants of unused DDCMode option 19046 19047commit f1efee6a11dc9436276fb12cfb90aa79ada1192b 19048Author: Alex Deucher <alexdeucher@gmail.com> 19049Date: Mon Dec 15 11:27:17 2008 -0500 19050 19051 Remove mention of MergedFB from server log 19052 19053 noticed by Fabio, bug 19098 19054 19055commit 94bffac6968057cdedc792f8540fe44cb6ecfd7c 19056Author: Fabio <fabio.ped@libero.it> 19057Date: Mon Dec 15 11:22:56 2008 -0500 19058 19059 man page updates 19060 19061 from bug 19098 with some minor fixes from me 19062 19063commit c7899905ec824071f9cabbd1e271c4d76010b733 19064Author: Fabio <fabio.ped@libero.it> 19065Date: Mon Dec 15 10:58:23 2008 -0500 19066 19067 Remove duplicate ATI in some device names 19068 19069 from bug 19098 19070 19071commit 89a8fd5337c77abf1f3166e69ed061d85b5c20f0 19072Author: Dave Airlie <airlied@redhat.com> 19073Date: Mon Dec 15 14:41:10 2008 +1000 19074 19075 r600: don't idle engine 19076 (cherry picked from commit b77729448e487d37707ce028d8edc59352d1cc2d) 19077 19078commit 4ce255a9db8027f2471ba50e57724d3e4d5b16ba 19079Author: Fabio <fabio.ped@libero.it> 19080Date: Fri Dec 12 09:53:33 2008 -0500 19081 19082 Fix formatting in man page 19083 19084 fixes bug 19039 19085 19086commit 0e99017cb9a6d46b891705df73678c2705439532 19087Author: Joseph Adams <joeyadams3.14159@gmail.com> 19088Date: Thu Dec 11 00:57:03 2008 -0500 19089 19090 Add more eMac modes, fix crash in last patch 19091 19092commit fb46c30d316ab3ec54c54f3aec91d5164070a423 19093Author: Alex Deucher <alexdeucher@gmail.com> 19094Date: Wed Dec 10 17:16:56 2008 -0500 19095 19096 Add support for Apple eMacs 19097 19098 Thanks to Joseph Adams for helping me sort this out. 19099 Verified on eMac G4/1.0 with radeon 7500 19100 Later eMacs with radeon 9200 or 9600 chips may have 19101 different ddc setups. Need to verify. 19102 19103commit c0bcea9150ef215fa614733cef9a5b71a55a33bd 19104Author: Alex Deucher <alexdeucher@gmail.com> 19105Date: Wed Dec 10 11:40:47 2008 -0500 19106 19107 Add another AGP quirk 19108 19109 fixes bug 12544 19110 19111commit 607b02d6abf78147902a5944bd006897ad95e982 19112Author: Alex Deucher <alexdeucher@gmail.com> 19113Date: Wed Dec 10 03:54:35 2008 -0500 19114 19115 Add randr 1.3 panning support 19116 19117commit 24007acbf36bcfd09cb5e74ce4a8d918d393cf34 19118Author: Benjamin Close <Benjamin.Close@clearchain.com> 19119Date: Wed Dec 10 02:17:56 2008 -0500 19120 19121 [PATCH] xf86misc extension has been removed, update driver to reflect it 19122 19123commit 69e445f7f2ebd299070e44457b434bbc28b7e05c 19124Author: Dave Airlie <airlied@linux.ie> 19125Date: Tue Dec 9 18:59:11 2008 +1000 19126 19127 radeon: memset the pll setting function 19128 19129 Thanks to Peter Zijlstra for making me debug this and providing 19130 feedback. 19131 19132commit c72b732a2f5597cd72bd4ebe207799ae07f2681e 19133Author: Adrian Friedli <adi@koalatux.ch> 19134Date: Mon Dec 8 10:42:46 2008 -0500 19135 19136 Add another AGP quirk 19137 19138commit 979ad04d901676cb41aea5cf47a8a2e9c19064dd 19139Author: Alex Deucher <alexdeucher@gmail.com> 19140Date: Fri Dec 5 17:53:57 2008 -0500 19141 19142 Add missing .TP in man page 19143 19144commit b89d434011787fb71593395f462d142375783e35 19145Author: Pierre Ossman <pierre@ossman.eu> 19146Date: Fri Dec 5 17:52:02 2008 -0500 19147 19148 Fix scissor setup for Xv 19149 19150 We need to adjust the scissoring for each clip box, so move the setup 19151 to inside the while loop. 19152 19153commit 83c1115333cbe763f86422fa5aabf4ab21951e68 19154Author: Alex Deucher <alexdeucher@gmail.com> 19155Date: Fri Dec 5 11:52:18 2008 -0500 19156 19157 R200: Remove scissor setup from init3d() 19158 19159 We aren't using them, so no need to set them up. 19160 19161commit 7593d5399cdb2f42ecbcb5193580385b286a9db9 19162Author: Alex Deucher <alexdeucher@gmail.com> 19163Date: Fri Dec 5 11:18:15 2008 -0500 19164 19165 set coherent mode consistently (enabled for now) 19166 19167 It's set to enabled in the output setup, but then was set to 19168 disabled when teh output properties were setup. Set it up 19169 consistently. 19170 19171commit ad2579f8898251105a6b36b745afd1ce1dab103e 19172Author: Dave Airlie <airlied@linux.ie> 19173Date: Fri Dec 5 20:03:53 2008 +1000 19174 19175 radeon: damn you bios, HDMI cannot have DACs. 19176 19177commit bae30856f2cb3884387ef347327af6ff00b75854 19178Author: Alex Deucher <alexdeucher@gmail.com> 19179Date: Fri Dec 5 08:15:59 2008 +0100 19180 19181 Make VSync for EXA and Xv configurable 19182 19183commit 33638d9e388b330e2f4eb4debd05ba09924cf176 19184Author: Pierre Ossman <pierre@ossman.eu> 19185Date: Thu Dec 4 23:29:31 2008 +0100 19186 19187 Optimise RADEONWaitForVLine 19188 19189 Only avoid the vlines we are rendering to, instead of the entire 19190 screen. This way we don't stall the card for longer than we 19191 absolutely have to. 19192 19193 EXA calls fixed by Alex Deucher. 19194 19195commit d1690f5cc096e2f735c8b407c370a1c1cd7a8afc 19196Author: Pierre Ossman <pierre@ossman.eu> 19197Date: Wed Dec 3 21:49:37 2008 +0100 19198 19199 Improve tearing avoidance for Xvideo in two steps 19200 19201 - Fix up VLINE handling to trigger whenever scanout is outside the 19202 visible area. 19203 19204 - Render the video as a scissored triangle as R300+ cannot render a 19205 quad in a single pass. 19206 19207commit 4d98acbca2e630056bf56cdcd0e23007fded2ced 19208Author: Alex Deucher <alexdeucher@gmail.com> 19209Date: Wed Jul 16 03:14:06 2008 -0400 19210 19211 First pass at tear-free accel 19212 19213 if the dest pixmap is the front buffer, stall the pipe 19214 until the vline is outside the active area. 19215 For EXA, pick crtc based on the larger mode area; 19216 ideally we'd have one pixmap per crtc. 19217 For Xv, use dst window area to determine crtc. 19218 19219commit 82d67b70d0fb747cf7b1eb0b64244ae56cb84572 19220Author: Pierre Ossman <pierre@ossman.eu> 19221Date: Thu Dec 4 15:38:13 2008 -0500 19222 19223 Switch r200 Xv to use rect lists rather than quads to avoid tearing 19224 19225 Patch from Pierre with fixes from me. 19226 19227commit 9eccd0ad0827a1b8ec9abb2cebf5768d204a5ce9 19228Author: Alex Deucher <alexdeucher@gmail.com> 19229Date: Thu Dec 4 12:28:12 2008 -0500 19230 19231 Fix up US setup after r3xx bicubic merge 19232 19233commit e13fba853ba19e6b0f081b9b3d9fa76c38a0f82b 19234Merge: d2963370 79bbdd98 19235Author: Alex Deucher <alexdeucher@gmail.com> 19236Date: Thu Dec 4 12:25:29 2008 -0500 19237 19238 Merge branch 'bicubic' of git://git.infradead.org/users/drzeus/xf86-video-ati 19239 19240commit d29633708b5451f5541e88371d831ae03019a7f5 19241Author: Dave Airlie <airlied@redhat.com> 19242Date: Thu Dec 4 09:08:48 2008 +1000 19243 19244 radeon: apply atom quirks to object table 19245 19246commit 941ac0d3ab839dc5aaaef2f3b7f5c7085b2894c5 19247Author: Alex Deucher <alexdeucher@gmail.com> 19248Date: Wed Dec 3 17:31:25 2008 -0500 19249 19250 Make sure 3D state is emited when switching to 3D IB 19251 19252 - fixes bug 18864 19253 19254commit 79bbdd984c925e37f5b3db2605339f1a21377fcf 19255Author: Pierre Ossman <pierre@ossman.eu> 19256Date: Wed Dec 3 20:33:36 2008 +0100 19257 19258 Change the XV_BICUBIC attribute to a tristate, where the third state 19259 automatically turns the filter on and off as appropriate. 19260 19261commit bbdf821c179242266fd86edf9e688a6be9cf985f 19262Author: Pierre Ossman <pierre@ossman.eu> 19263Date: Wed Dec 3 19:42:50 2008 +0100 19264 19265 Make the R300 bicubic shader program a bit easier to understand by 19266 having comments that describe the operation, not how it has to be 19267 implemented in the hardware. 19268 19269commit 8f4b22932c91f1d51fcbf1d1a99d852fcdaafdcc 19270Author: Pierre Ossman <pierre@ossman.eu> 19271Date: Wed Dec 3 19:31:31 2008 +0100 19272 19273 Now that we have a bicubic code path that messes up the pixel shader, 19274 we need to restore it in the non-bicubic code. 19275 19276commit 0851daa6e7db77e440188d774ef8b3236146bcb9 19277Author: Pierre Ossman <pierre@ossman.eu> 19278Date: Wed Dec 3 19:17:36 2008 +0100 19279 19280 Fix macros so that they are safe to use with expressions as parameters. 19281 19282commit e04bcb0066781f61e97787856261e3380036d304 19283Author: Pierre Ossman <pierre@ossman.eu> 19284Date: Wed Dec 3 19:16:03 2008 +0100 19285 19286 Fix node setup on R300 bicubic pixel shader. It was backwards and used 19287 the wrong base for sizes. 19288 19289commit 927aee181f3637bda1b761caf93fbc0652e1207b 19290Author: Alex Deucher <alexdeucher@gmail.com> 19291Date: Wed Dec 3 01:39:38 2008 -0500 19292 19293 Only flush IB when direct rendering is enabled 19294 19295commit ca74ebf8813b7fab8d08465ef9600f001c948e94 19296Author: Alex Deucher <alexdeucher@gmail.com> 19297Date: Tue Dec 2 22:26:21 2008 -0500 19298 19299 move 3d state init after RADEON_SWITCH_TO_3D() 19300 19301commit fc079c5267baf431bbecee7744e484783d393152 19302Author: Dave Airlie <airlied@redhat.com> 19303Date: Tue Dec 2 22:11:57 2008 -0500 19304 19305 Don't mix 2D and 3D in the same IB 19306 19307 fix from radeon-gem-cs 19308 19309commit 678693aabd0b4c1812ab44bf5abb4c8394b254bb 19310Author: Dave Airlie <airlied@redhat.com> 19311Date: Tue Dec 2 22:02:21 2008 -0500 19312 19313 Stability fixes from radeon-gem-cs 19314 19315commit f231e4d9b2c683dc16f42cf09484fae926a892c1 19316Author: Dave Airlie <airlied@redhat.com> 19317Date: Wed Dec 3 11:39:04 2008 +1000 19318 19319 radeon: fix HDMI having DACs in some BIOSes 19320 19321commit fe73d9a7dfe8ec5c8f1a8dc08e14b4e138aa9276 19322Author: Bryce Harrington <bryce@canonical.com> 19323Date: Tue Dec 2 19:47:45 2008 -0500 19324 19325 Add another AGP quirk 19326 19327commit 6e0c48958c6a50a639368e369ce72d813256aee4 19328Author: Alex Deucher <alexdeucher@gmail.com> 19329Date: Mon Dec 1 18:50:40 2008 -0500 19330 19331 Add quirk for r200 card with the primary dac wired to both ports 19332 19333 refactor load detection setup as a byproduct 19334 should fix bug 18719 19335 19336commit fa496d7b0397d9be57db90d0860928e9ced73cca 19337Author: Pierre Ossman <pierre@ossman.eu> 19338Date: Mon Dec 1 08:29:29 2008 +0100 19339 19340 Fix comments for R500 fragment shader to reflect the code. 19341 19342commit 0d95fca114194664d9ab612c90d3222912cf41ff 19343Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19344Date: Sat Aug 30 12:38:43 2008 -0700 19345 19346 Fix indentation on IS_R300_3D Xv code. 19347 19348commit 36a62b8c47def3f23a0b735875f59d23ed3c095a 19349Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19350Date: Sat Aug 30 11:15:20 2008 -0700 19351 19352 r3xx: Various Xv fixes. 19353 19354commit 01ed5462fad56f5b7bf867041d00443bdf81c653 19355Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19356Date: Thu Aug 7 18:08:54 2008 -0700 19357 19358 Switch to Mesa-style 24-bit float packing. 19359 Seems like Mesa's got it right, so we should follow suit. 19360 19361commit 85bbe889045aa434f71a3eb0ba8153b7d2cc5023 19362Author: Maciej Cencora <m.cencora@gmail.com> 19363Date: Thu Aug 7 07:06:46 2008 -0700 19364 19365 Enable bicubic Xv on r3xx+. Still some buggies in the actual display, but looks pretty good. 19366 19367commit bf6263cff50bb85c19ff6dfebb739333fa8aca18 19368Author: Maciej Cencora <m.cencora@gmail.com> 19369Date: Thu Aug 7 07:00:36 2008 -0700 19370 19371 Add r3xx-specific bicubic filtering code. FP, RS, mostly, everything else is shared with r5xx. 19372 19373commit da021c36bbdf3bca31ee50ebe01cdb9495c09b36 19374Author: Dave Airlie <airlied@linux.ie> 19375Date: Sun Nov 30 18:44:24 2008 +1000 19376 19377 radeon_drm.h: remove kernel defines 19378 19379 remove these we should update from an installed kernel copy not from the 19380 original kernel copy. 19381 19382commit 783cdb7374941bb1d2b63eea375fbf2f1b808cc3 19383Author: Dave Airlie <airlied@redhat.com> 19384Date: Thu Nov 27 15:29:06 2008 +1000 19385 19386 r100/r200: EXA misrenders dst == a8 + dst alpha use. 19387 19388 Fedora BZ#469556 19389 19390 I've played with various dst and texture formats and rewritten large 19391 parts of this to no avail, so I'm punting for now. 19392 19393commit 2d7cfc421713286edd5121b56999246f9ae70286 19394Author: Eygene Ryabinkin <rea-fbsd@codelabs.ru> 19395Date: Tue Nov 25 13:25:33 2008 -0500 19396 19397 Catch unsubstituted macro for PCI region base access 19398 19399 also fix up some warnings. 19400 19401commit af7690b9c300d37c3a0c51e27e0ad2ca009224a2 19402Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 19403Date: Tue Nov 25 02:57:30 2008 -0500 19404 19405 [PATCH] radeon: Fix PCI usage of 32-bit driver on 64-bit platform 19406 19407 The radeon driver is storing PCI addresses in unsigned long's which 19408 won't work well on 32-bit platforms with 64-bit physical address space 19409 such as PowerPC 4xx. This fixes it by using unsigned long long instead. 19410 19411commit 065938617c0feab17f4274a5350de02a692ba065 19412Author: Ramon van der Stelt <rvanderstelt@xs4all.nl> 19413Date: Tue Nov 25 02:23:46 2008 -0500 19414 19415 Interlaced mode fixes 19416 19417 see bug 12626 19418 19419commit 36a7dc6ea1e1929e986ab1159497c71521cb2f10 19420Author: Bryce Harrington <bryce@canonical.com> 19421Date: Tue Nov 25 01:04:37 2008 -0500 19422 19423 Additional AGP quirks 19424 19425 bug 18693 19426 19427commit 2acb8e4ed220fccb21b22a53c5142ab3a35d32ef 19428Author: Alex Deucher <alexdeucher@gmail.com> 19429Date: Tue Nov 25 15:07:09 2008 +1000 19430 19431 radeon: fixup shared DDC lines for some rv610 cards. 19432 19433 Some cards share DDC between a HDMI and VGA output, and expect the 19434 driver to use load detect or EDID to figure it out. 19435 19436 airlied- shipped in RHEL5 19437 19438 Signed-off-by: Dave Airlie <airlied@redhat.com> 19439 19440commit b01e35168de40d192fd7a9ce6884b9c7419afbd4 19441Author: Alex Deucher <alexdeucher@gmail.com> 19442Date: Mon Nov 24 21:06:42 2008 -0500 19443 19444 Fix up posted logic 19445 19446 noticed by benh on IRC 19447 19448commit d7a03e93372b17d642b42f5490464616ae7c99c1 19449Author: Alex Deucher <alexdeucher@gmail.com> 19450Date: Mon Nov 24 15:52:05 2008 -0500 19451 19452 Get hardcoded edid from the bios for servers that support it 19453 19454 should fix bug 11300 19455 19456commit 3858e31fa9d77d52a5cabb02f53f06385eab40f7 19457Author: Alex Deucher <alexdeucher@gmail.com> 19458Date: Mon Nov 24 14:39:41 2008 -0500 19459 19460 Don't support rotation is accel is not available 19461 19462 Fixes bug 18068 19463 19464commit 3d27876d91ef69d71f17b6cf8c07c26bf12857b4 19465Author: Dave Airlie <airlied@redhat.com> 19466Date: Mon Nov 10 15:47:13 2008 +1000 19467 19468 radeon: always align dstPitch to 64 bytes for textured video 19469 19470 bz# 18454 19471 19472commit 6c8b352264540ccc4622a7c0b704ce452dcd707d 19473Author: Dave Airlie <airlied@redhat.com> 19474Date: Wed Oct 29 15:53:57 2008 +1000 19475 19476 pciid: add HD3300 - 790GX chipset. 19477 19478 RH BZ (#466706) 19479 19480commit 902eaf768142c6c7dcc487e10775027b84cd1f9a 19481Author: Alex Deucher <alexdeucher@gmail.com> 19482Date: Thu Nov 6 15:46:43 2008 -0500 19483 19484 Check for LVDS on all IGP chips 19485 19486 - fixes bug 18395 19487 19488commit 44abaa0ea8e5e3bbe833fa73e6d011618bb6d7e7 19489Author: Adam Jackson <ajax@redhat.com> 19490Date: Mon Nov 3 15:09:28 2008 -0500 19491 19492 Fix HDMI output setup on DIG1/DIG2. 19493 19494 The only reliable way to detect an HDMI sink is to check for the 19495 appropriate EDID extension block, so this will only work right on 19496 servers with xf86DoEEDID(). RANDR 1.2 will call this internally for 19497 us if it exists, but pre-1.6 servers are out of luck. 19498 19499commit 937b7ac2a259cf504a19dcf62a58b1db1afb8eb9 19500Author: Bryce Harrington <bryce@canonical.com> 19501Date: Wed Oct 29 01:31:50 2008 -0400 19502 19503 Add AGP quirk table 19504 19505 - set known good default agp modes for problematic chip/bridge combos 19506 - should fix bug 17360, others 19507 19508commit 67bc2ae7f227a2afedc8e699cf0458fa7960f301 19509Author: Julien Cristau <jcristau@debian.org> 19510Date: Sun Oct 26 16:05:04 2008 +0100 19511 19512 Bug#7148: Fix manpage formatting 19513 19514commit 98fdd78a1229584b922f816f331a45bfc178c961 19515Author: Dave Airlie <airlied@linux.ie> 19516Date: Tue Oct 21 18:36:22 2008 +1000 19517 19518 atombios: oops typo 19519 19520commit 763dff6c117a10b6a57ba1854c5b798359789e44 19521Author: Dave Airlie <airlied@linux.ie> 19522Date: Tue Oct 21 18:35:44 2008 +1000 19523 19524 atombios: set missing pixel clock 19525 19526commit 435cf7da68186f2601c4b888296117d4f652c625 19527Author: Alex Deucher <alexdeucher@gmail.com> 19528Date: Wed Oct 15 10:52:14 2008 -0400 19529 19530 Add LCD dithering quirk for macbook pro 19531 19532 fixes bug 17897. 19533 19534commit c0e6cb6d0eeef8f2ea60d840e1cd668fa92cd7f9 19535Author: Alex Deucher <alexdeucher@gmail.com> 19536Date: Fri Oct 10 11:09:27 2008 -0400 19537 19538 Fix grey level dithering setup 19539 19540commit 9eb2b6cd28224f0cbbabf3df451696ae1a845b9c 19541Author: Dave Airlie <airlied@redhat.com> 19542Date: Fri Oct 10 16:18:05 2008 +1000 19543 19544 radeon: hopefully fix plls for 30" monitors. 19545 19546 Only do the HIGH trick on the DCE3 cards. Tested on rv610/630/770. 19547 19548commit 0975e007ec7933acf9cd8ad2f281313ef8d32ac9 19549Author: Dave Airlie <airlied@redhat.com> 19550Date: Thu Oct 9 17:16:28 2008 +1100 19551 19552 radeon: rename radeon_memory to radeon_legacy_memory. 19553 19554 This is a setup patch for the kms/memory management code. 19555 19556commit d744dc013cf5ca3141c0fdf55e36d713fe9f53ba 19557Author: Adam Jackson <ajax@redhat.com> 19558Date: Wed Oct 8 15:50:25 2008 -0400 19559 19560 Fix connector table mapping for lvtma/kaleidoscope/uniphy. 19561 19562 Fixes DPMS on LVDS. 19563 19564commit dc795ba66a8c0eb2c3fdb86bd42dfc17e0aa3b6e 19565Author: Alex Deucher <alexdeucher@gmail.com> 19566Date: Tue Oct 7 12:32:47 2008 -0400 19567 19568 Fixups from last DFP/LCD refactor 19569 19570 - fix DFP1 -> LCD1 19571 - use _INDEX rather than _SUPPORT defs 19572 19573commit 4a9e8d9fe24278b56aef51677fd35289d01584a7 19574Author: Alex Deucher <alexdeucher@gmail.com> 19575Date: Mon Oct 6 18:00:08 2008 -0400 19576 19577 set grey levels correctly for temporal dithering 19578 19579commit 5708624d1151877dd66a462d4a75fed6774604a7 19580Author: Alex Deucher <alexdeucher@gmail.com> 19581Date: Mon Oct 6 15:11:56 2008 -0400 19582 19583 Turn coherent mode off by default on DVI 19584 19585commit a5c5ce96279d01eb519bfb92b94c06a58acb7f07 19586Author: Alex Deucher <alexdeucher@gmail.com> 19587Date: Mon Oct 6 15:07:31 2008 -0400 19588 19589 Refactor atom LCD/DFP output setup 19590 19591 Simplify the code, properly handle dithering, coherent mode, encoding, etc. 19592 Should fix bug 17897. 19593 19594commit f9826a5694b7adb6920eb5bdf45d840d8fb14d53 19595Author: Alex Deucher <alexdeucher@gmail.com> 19596Date: Fri Oct 3 15:24:12 2008 -0400 19597 19598 Add support for DDC via atom commands for RV410 19599 19600 The atom calls use the hw i2c engine for DDC. For some 19601 reason, sw i2c doesn't seem to work on the VGA GPIO on 19602 RV410 chips, so we use atom in that case. 19603 19604 This fixes the longstanding VGA DDC problems on RV410/M26 19605 chips. 19606 19607commit 23c17c0121d43c2fd11bb5cfbaa4b02abac2b16d 19608Author: Alex Deucher <alexdeucher@gmail.com> 19609Date: Fri Oct 3 14:30:33 2008 -0400 19610 19611 man page updates 19612 19613 Update render accel and page flipping information. 19614 fixes bug 17887. 19615 19616commit 4ac6e499089aec3c79af5dbede5fe631b105ce3d 19617Author: Alex Deucher <alexdeucher@gmail.com> 19618Date: Wed Oct 1 18:53:15 2008 -0400 19619 19620 proper fix for initial rotation 19621 19622 The initial modeset was setting the MC regs, so just set them 19623 rather than the full modeset. 19624 19625commit 346228b20c69a965c9e7d67157f42c1d36a5b6c3 19626Author: Alex Deucher <alexdeucher@gmail.com> 19627Date: Wed Oct 1 14:38:27 2008 -0400 19628 19629 Workaround to make initial rotation work 19630 19631 Acceleration needs to be initialized before setting the mode 19632 for initial rotation to work. Changing the order in RADEONScreenInit() 19633 doesn't work because RADEONDRIKernelInit() hangs in the ioctl to 19634 initialize the CP if the mode is not set prior (even if the heads 19635 are turned off after setting the mode). The workaround is to set the 19636 modes twice, once before accel and once after. 19637 19638 This needs to be looked into further. 19639 19640commit c359c2a31caf9f75b9fc6b6bcbc3e9dc1fe404ba 19641Author: Michel Dänzer <michel@tungstengraphics.com> 19642Date: Tue Sep 30 11:04:54 2008 +0200 19643 19644 Pass base offset into RADEONDisplayVideo() explicitly. 19645 19646 This prevents the base offset from flipping back and forth if double buffering 19647 is enabled and the two buffers start on different sides of a 4 MB boundary. 19648 19649 Should fix http://bugs.freedesktop.org/show_bug.cgi?id=17254 ; please don't 19650 reopen that bug report without attaching a log file captured after reproducing 19651 the problem with this fix. 19652 19653commit ba21d5ae7fbbf5828c44c41655aff8b182453e2e 19654Author: Michel Dänzer <michel@tungstengraphics.com> 19655Date: Tue Sep 30 11:03:04 2008 +0200 19656 19657 Update GEN_INT_CNTL register value after calling DRM modeset ioctl. 19658 19659 Fixes sync-to-vblank hangs after mode switch with DRM vblank-rework. 19660 19661commit 3110646daf05fa09faaa70066cc8c3db1bb07cec 19662Author: Michel Dänzer <michel@tungstengraphics.com> 19663Date: Sun Sep 14 11:46:36 2008 +0200 19664 19665 Fix -Werror build. 19666 19667commit d82f2938f69402c641a1c8362fdc513419b27659 19668Author: Alex Deucher <alexdeucher@gmail.com> 19669Date: Fri Sep 26 13:51:24 2008 -0400 19670 19671 atombios updates from upstream 19672 19673commit d0d58b39e49c5ce00bc8bd12f721ad8c7f554b91 19674Author: Alex Deucher <alexdeucher@gmail.com> 19675Date: Wed Sep 24 11:54:34 2008 -0400 19676 19677 radeon: Further pll adjustments 19678 19679commit d100a6af8f828eb94f8ba6e8a96c24389b5cf46f 19680Author: Alex Deucher <alexdeucher@gmail.com> 19681Date: Fri Sep 19 14:04:59 2008 -0400 19682 19683 cleanup macbook quirk 19684 19685commit 1218e38c75351a20f7414ba6fcb3545689339442 19686Author: Alex Deucher <alexdeucher@gmail.com> 19687Date: Fri Sep 19 13:52:07 2008 -0400 19688 19689 more macbook fixes 19690 19691commit 83e1f132965deadedda7a622184eb400cd4c120c 19692Author: Alex Deucher <alexdeucher@gmail.com> 19693Date: Fri Sep 19 09:36:59 2008 -0400 19694 19695 Fix copy/paste error in the last commit 19696 19697 DVI-D -> DVI-I 19698 19699commit 4af825f7b13cba8c3b204e72a07cab5009331240 19700Author: Alex Deucher <alexdeucher@gmail.com> 19701Date: Fri Sep 19 00:29:20 2008 -0400 19702 19703 radeon: Add quirk for macbook 19704 19705 - should fix bug 17625 19706 19707commit 38aa450fce6a91b35c191fd07112490a62cc29c6 19708Author: Alex Deucher <alexdeucher@gmail.com> 19709Date: Tue Aug 26 21:50:23 2008 -0400 19710 19711 fix some fallout from the common allocator 19712 19713 should fix bug 17317 19714 19715commit a3cc1d7a421456186024c5c069e403d374a0f0b9 19716Author: Dave Airlie <airlied@linux.ie> 19717Date: Tue Aug 26 08:03:20 2008 +1000 19718 19719 radeon: fix powerpc build 19720 19721commit 6cebfe257f7ddad855ee743e4eb899bd6fac7f46 19722Author: Alex Deucher <alexdeucher@gmail.com> 19723Date: Fri Jul 11 19:32:06 2008 -0400 19724 19725 Switch EXA path back to static cursor allocation 19726 19727 pre-AVIVO cards have address limits for the cursor offset 19728 19729commit 4dff54a3c8d7c9f2d6ec50354ff0b92f1b7fcbdf 19730Author: Alex Deucher <alexdeucher@gmail.com> 19731Date: Fri Jul 11 17:28:03 2008 -0400 19732 19733 Switch cursors over to generic allocator 19734 19735commit 5b1978a4796bcc31ac2f01d303dc8f8f44323025 19736Author: Alex Deucher <alexdeucher@gmail.com> 19737Date: Mon Aug 25 10:05:28 2008 -0400 19738 19739 Bicubic fixes from the last cherry-pick 19740 19741commit 7daba77ded1c718e93ae8c372a39a6e85228d513 19742Author: Alex Deucher <alexdeucher@gmail.com> 19743Date: Fri Jul 11 02:02:38 2008 -0400 19744 19745 Convert randr, Xv to a common allocator 19746 19747commit 610fe1a937da78f4ac813ac919c158dde8f42442 19748Author: Alex Deucher <alexdeucher@gmail.com> 19749Date: Thu Jul 10 22:47:35 2008 -0400 19750 19751 Switch to using radeon_drm.h from the drm 19752 19753 modelled after Matthias' similar rhd change 19754 19755commit a6561f2ec673b38907f7181235386f32e60c32ba 19756Author: Alex Deucher <alexdeucher@gmail.com> 19757Date: Thu Jul 10 22:31:13 2008 -0400 19758 19759 Move DRI to a separate struct 19760 19761 based on Matthias' similar work in rhd 19762 19763commit 75ef8dc214715d3c5c50996b293933842903ba65 19764Author: Egbert Eich <eich@freedesktop.org> 19765Date: Thu Jul 10 21:49:55 2008 -0400 19766 19767 Cleanups from rhd port 19768 19769 - remove unused vars 19770 - remove static exa render vars 19771 19772commit 71ad140fa11f3a504c38d6bddf40e3a3c0a20e60 19773Author: Alex Deucher <alexdeucher@gmail.com> 19774Date: Thu Jul 10 21:24:16 2008 -0400 19775 19776 Move accel state to a separate struct 19777 19778commit 5b2e095c31b88d8495a4f86e6cb46b49fa4acd65 19779Author: Alex Deucher <alexdeucher@gmail.com> 19780Date: Thu Jul 10 20:07:44 2008 -0400 19781 19782 Move CP into a separate struct 19783 19784commit 2145309230e36aee7758bd244deb1e72ada0c065 19785Author: Alex Deucher <alexdeucher@gmail.com> 19786Date: Mon Aug 25 08:47:59 2008 -0400 19787 19788 match textured video macro names with other accel code 19789 19790 OUT_VIDEO_REG() -> OUT_ACCEL_REG() 19791 etc. 19792 19793commit 01daef0f095fbbaee701d5fe97f3dd7838b5f915 19794Author: Alex Deucher <alexdeucher@gmail.com> 19795Date: Mon Aug 25 08:26:16 2008 -0400 19796 19797 Additional cleanups and re-arragement following bicubic merge 19798 19799commit 0a51d08c24af040fe48690662b1a912acad51700 19800Author: Alex Deucher <alexdeucher@gmail.com> 19801Date: Mon Aug 25 06:43:14 2008 -0400 19802 19803 Whitespace cleanup from bucubic merge 19804 19805commit c0170fef510b53a6ca2c6ea7a99119235229c929 19806Author: Maciej Cencora <m.cencora@gmail.com> 19807Date: Thu Aug 7 06:53:39 2008 -0700 19808 19809 Add needed FP registers, etc. for r3xx bicubic Xv. 19810 19811commit d9c38326cf70f57ab777ffdf9520b8cdea9d9cb6 19812Author: Dennis Kasprzyk <onestone@opencompositing.org> 19813Date: Tue Aug 5 18:45:38 2008 +0200 19814 19815 Fix typos. 19816 19817commit d38ceba62aa5cee76baa342ce7719a983a79f09e 19818Author: Dennis Kasprzyk <onestone@opencompositing.org> 19819Date: Tue Aug 5 16:21:57 2008 +0200 19820 19821 Fix bicubic fp calculation. 19822 19823commit 6f9c7fde75edc0399559f975db9a6c4dc22714ae 19824Author: Dennis Kasprzyk <onestone@opencompositing.org> 19825Date: Tue Aug 5 15:38:42 2008 +0200 19826 19827 Fixed bicubic fragment program comments. 19828 19829commit 48b09ca40ccb28b5584069316fd38786a78c1dd3 19830Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19831Date: Sat Aug 2 02:36:21 2008 -0700 19832 19833 Switch from 32-bit floats to 16-bit half-floats. 19834 Massive bandwidth savings, or so I'm told. Yay? 19835 19836commit ebbb7fb634fcadf28ff99c1df2c3db89fd56932d 19837Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19838Date: Sat Aug 2 01:45:43 2008 -0700 19839 19840 Change floats to uint32_t hex. 19841 Useful for moving to 16-bit half-floats. 19842 19843commit 3b46162d8a90be3524cb513d42b9ad3d0bea45f5 19844Author: Dennis Kasprzyk <onestone@opencompositing.org> 19845Date: Fri Aug 1 15:24:01 2008 +0200 19846 19847 Move some ALU instructions after the TEX instructions, so we can do something usefull while we are waiting for the texture values. 19848 19849commit 2ecdec4bafc97212dde4d6908ee4ccf618adc0e1 19850Author: Dennis Kasprzyk <onestone@opencompositing.org> 19851Date: Fri Aug 1 04:03:12 2008 +0200 19852 19853 Another uneeded SEM_AQUIRE. 19854 19855commit e93b5d1b80d6203f63543b7b678e2f1d9221b5b3 19856Author: Dennis Kasprzyk <onestone@opencompositing.org> 19857Date: Fri Aug 1 03:00:26 2008 +0200 19858 19859 Smarter usage of the texture semaphore. 19860 19861commit 286f7aa18f19bd21b263701adab38b736dbeda0f 19862Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19863Date: Thu Jul 31 17:01:44 2008 -0700 19864 19865 Enable bicubic filtering for all r5xx HW. 19866 19867commit 242aa4f630b4c60aefa3c12dc459a4d4d0b334a0 19868Author: Dennis Kasprzyk <onestone@opencompositing.org> 19869Date: Fri Aug 1 01:11:28 2008 +0200 19870 19871 Remove one constant. 19872 19873commit 413eacb0538977b0b3c92df074d40510f4539abc 19874Author: Dennis Kasprzyk <onestone@opencompositing.org> 19875Date: Fri Aug 1 00:52:09 2008 +0200 19876 19877 Heavy optimizations. 19878 19879commit a0c4a949cb49e5ac1e857aef08a8742b9f7b49da 19880Author: Dennis Kasprzyk <onestone@opencompositing.org> 19881Date: Thu Jul 31 21:19:18 2008 +0200 19882 19883 Remove uneeded negations. 19884 19885commit abb2b2e757d92591ab9277824b9d9746ba98f875 19886Author: Dennis Kasprzyk <onestone@opencompositing.org> 19887Date: Thu Jul 31 20:31:03 2008 +0200 19888 19889 Implement LPR in one instruction. 19890 19891commit c370b74bec13194573348583c38adf710b880e79 19892Author: Dennis Kasprzyk <onestone@opencompositing.org> 19893Date: Thu Jul 31 19:50:49 2008 +0200 19894 19895 Set helper texture filter correctly. 19896 19897commit 5e85d5a5d0c50b29086ec0c219c8b52d25dbc2e9 19898Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19899Date: Wed Jul 30 19:45:15 2008 -0700 19900 19901 Update bicubic tables. 19902 Now including the 2048x1 texture as well. 19903 19904commit 17e5e9573e59c3d82d51c261b9c5005f6aec7d43 19905Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19906Date: Wed Jul 30 11:57:25 2008 -0700 19907 19908 Force R580-only for bicubic. 19909 Initial reading of docs suggest RV560 and RV570 can't handle it, 19910 but they're welcome to test. 19911 19912commit 83b52473d0e7102265918b07be053fcda17a14b4 19913Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19914Date: Wed Jul 30 01:03:57 2008 -0700 19915 19916 Bump bicubic cutoff to R580. 19917 RV530 chipsets can't handle fullscreen bicubic... 19918 19919commit 69a4998d1286bcdd7bfe874cd5628bc1cc232bae 19920Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19921Date: Wed Jul 30 01:03:46 2008 -0700 19922 19923 Oops, made a mistake with vertices. 19924 19925commit a4ec30a677906ec2ff9824c7ddca586655f6d1a8 19926Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19927Date: Wed Jul 30 00:54:32 2008 -0700 19928 19929 Merge upstream changes to vertices, and also add Xv attributes for textured video, including bicubic filtering. 19930 19931commit f3b81c7582aed307fa44e134ee161cd8a3158657 19932Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19933Date: Mon Jul 28 19:50:10 2008 -0700 19934 19935 Fix constants. 19936 19937commit 7dbb7023ba023ec1a38be63af9c9f49e40222b7b 19938Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19939Date: Mon Jul 28 17:30:41 2008 -0700 19940 19941 Finally got the fragment program fully working for bicubic filtering on r5xx. 19942 19943commit 0e4dd73b9ebc6f608eeff945b4d463a00c02e07c 19944Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19945Date: Mon Jul 28 12:16:50 2008 -0700 19946 19947 More bicubic FP buggies. 19948 19949commit a87647e0c27e0950f4d0d8203a1242a994ad3419 19950Author: Dennis Kasprzyk <onestone@compiz-fusion.org> 19951Date: Sun Jul 27 10:43:01 2008 -0700 19952 19953 Fix texture size, texture filter, vertex offsets, etc. 19954 19955commit 232aa3e943fef4c4037b255c3b64a0aaff90ab5c 19956Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19957Date: Sat Jul 26 10:56:56 2008 -0700 19958 19959 Make vertices emit properly. 19960 *bangs head against wall* 19961 19962commit 8c84f67b93d926095633830aa8d95930a48b1c7b 19963Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19964Date: Sun Jul 20 14:25:44 2008 -0700 19965 19966 Try to get tex coords from the VAP to RS properly when bicubic filtering is enabled. 19967 I'm soo bad at this... :c 19968 19969commit a4a7d5f5967c51c394229de5eccaec44cfec8f50 19970Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19971Date: Sat Jul 19 13:02:02 2008 -0700 19972 19973 Upload pixel shader to card for r5xx. This was ridiculous. Also it doesn't work yet... 19974 19975commit 20c1db2d7c110ab5c1117a57b169baa1ab070518 19976Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19977Date: Fri Jul 18 13:48:08 2008 -0700 19978 19979 Upload bicubic filter to card. This was a LOT easier than I had feared, to be honest. 19980 19981commit a760e628134c6d7d42ec3c98118b6e5f6fcd3e7f 19982Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19983Date: Fri Jul 18 12:21:20 2008 -0700 19984 19985 Allocate memory for the bicubic filter texture. 19986 19987commit b6c9e2bb5365de82315c6814f915e57b0c4fa444 19988Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19989Date: Fri Jul 18 11:24:14 2008 -0700 19990 19991 Fixed typos in the bicubic texture tables. Whoops, looks like I'm still asleep. 19992 19993commit ca51f4f37e1dbf53bf7ffc0e8f612e9609e11209 19994Author: Corbin Simpson <MostAwesomeDude@gmail.com> 19995Date: Fri Jul 18 11:06:34 2008 -0700 19996 19997 Add bicubic texture table, as well as the script used to (re)generate it. To regenerate, just run "python bicubic_table.py > bicubic_table.h". 19998 19999commit 1cf7a5494fa94e8d9f30f9b2905dfbe6d4faa445 20000Author: Bryce Harrington <bryce@bryceharrington.org> 20001Date: Wed Aug 20 09:46:59 2008 -0400 20002 20003 radeon: Fix pasto in connector table setup for vga powerbooks 20004 20005 fixes bug 17214 20006 20007commit a55e85f742d1334bf88e4681e553f025d2de38df 20008Author: Michel Dänzer <michel@tungstengraphics.com> 20009Date: Thu Aug 14 15:21:51 2008 -0400 20010 20011 Make sure video offerlay offsets don't exceed the hardware limit of 128 MB. 20012 20013 Always set the overlay base address such that the buffer offsets are as small 20014 as possible. This could still break in theory if the buffers were more than 20015 128 MB apart, but in reality this can't happen ATM because we always allocate 20016 a single memory area for all buffers. 20017 20018 Fixes http://bugs.freedesktop.org/show_bug.cgi?id=16845 . 20019 20020commit 92ee21df344a989778e37369c7beb3904a00ead6 20021Author: Alex Deucher <alexdeucher@gmail.com> 20022Date: Thu Aug 14 14:49:45 2008 -0400 20023 20024 PLL adjustments 20025 20026 Seems higher dotclocks prefer a higher FB div. 20027 Someone with a lot of should try and find out where 20028 the div sweet spots are for various dotclock ranges. 20029 fixes bug 17125 20030 20031commit 0d5e0347af4322713075193154b8a348de4a0b52 20032Author: Alex Deucher <alexdeucher@gmail.com> 20033Date: Wed Aug 13 14:17:34 2008 -0400 20034 20035 Remove reset of 3D scissor registers when using the CP in the ddx 20036 20037 They should only affect 3D and init3d() should take care of that case 20038 noticed by libv on IRC. 20039 20040commit 7e456e9c427d48151b166bcac08c3e8a8b180f01 20041Author: Dave Airlie <airlied@linux.ie> 20042Date: Tue Aug 12 08:18:30 2008 +1000 20043 20044 radeon: add 0x9441 for hd 4870 x2 20045 20046 thanks to Michael Larabel for testing 20047 20048commit 4dbdeea7c9316575fba26b41fd347452e42cdcf2 20049Author: Alex Deucher <alexdeucher@gmail.com> 20050Date: Fri Aug 8 17:39:48 2008 -0400 20051 20052 Further cleanup and unification of i2c code 20053 20054 - unify the ddc and i2c code 20055 - add gpio mask support for legacy chips 20056 - remove the magic gpio dance for ancient monitors 20057 (if you have an ancient monitor that ddc stops 20058 working on let me know). This should speed up DDC 20059 on legacy chips. 20060 20061 -- radeon sw gpio i2c -- 20062 4 sets of gpio registers for clk and data and corresponding bit masks 20063 mask - locks the gpio for use by sw 20064 get - reads the value off the gpio pad 20065 put - sets the gpio direction to output 20066 a - "other stuff" On legacy chips you clear them if you want 20067 to use a gpio for i2c. In some cases they are used for the 20068 output value when the gpio in the output state. 20069 20070commit 268c848130ec1770bb645a74197b6aca7fc95abc 20071Author: Alex Deucher <alexdeucher@gmail.com> 20072Date: Fri Aug 8 15:50:07 2008 -0400 20073 20074 Fix VT switching on M6 chips 20075 20076 Some M6 chips have a faulty MEM_SIZE register that in 20077 some cases reports 0 on 8 MB cards. On EnterVT we check 20078 the MEM_SIZE reg as a check to see if the card is posted or 20079 not. Since this reg returns 0, the driver attempts to post 20080 the card which can lead to a hang. Switch this to check if 20081 either crtc is active as is done in the bios init code. 20082 fixes bug 13994 20083 20084commit 33f88f7fc90d9d93fdcbba9ad59dd70a6596bc3f 20085Author: Alex Deucher <alexdeucher@gmail.com> 20086Date: Thu Aug 7 17:57:07 2008 -0400 20087 20088 RS4xx DDC fixes take 3 20089 20090commit df0d1ef53100f0a19c5b5fdc349f5186c8d9bd87 20091Author: Alex Deucher <alexdeucher@gmail.com> 20092Date: Wed Aug 6 19:13:56 2008 -0400 20093 20094 RS4xx: Fix up ddc gpio 20095 20096 - I mixed up the regs previously 20097 20098commit 18429390440a829fb24ed3afd99ccf8278138496 20099Author: Alex Deucher <alexdeucher@gmail.com> 20100Date: Wed Aug 6 00:35:15 2008 -0400 20101 20102 Remove un-needed dac check for single crtc cards 20103 20104 Connector tables should be setup properly already and this 20105 check breaks systems with 2 DACs and 1 crtc like the es1000. 20106 20107commit 001c535687d8588873037ee5363d0a709f44b418 20108Author: Alex Deucher <alexdeucher@gmail.com> 20109Date: Tue Aug 5 17:14:48 2008 -0400 20110 20111 IGP: fix typo in IGP quirk handling 20112 20113commit eb65ddf70d182b6457e1ef5ebb820456039e8f6d 20114Author: Alex Deucher <alexdeucher@gmail.com> 20115Date: Tue Aug 5 15:01:33 2008 -0400 20116 20117 IGP: attempt to get external TMDS working 20118 20119 RS4xx chips have a tmds init table in the mobile info table 20120 20121commit 8b8990917809b9a35c6e9c1b9e3b12ff81c6dbb3 20122Author: Alex Deucher <alexdeucher@gmail.com> 20123Date: Tue Aug 5 12:16:06 2008 -0400 20124 20125 IGP: Updated quirk handling 20126 20127 - Hopefully finally fix DDC on RS4xx chips 20128 - RS3xx/RS4xx do not have internal TMDS 20129 - general quirk handling cleanup 20130 - we don't currently support mm gpio for external TMDS 20131 so return false. 20132 20133commit cb0deba5412a575d36f2f99377120b123506c946 20134Author: Calvin Fong <hoiwai930@gmail.com> 20135Date: Mon Aug 4 02:04:15 2008 -0400 20136 20137 IGP: add support for NTSC tv-out on legacy IGP chips 20138 20139 Patch from Calvin with some cleanups from me. 20140 No support for PAL yet. 20141 20142commit 942b18aca91819fa65d853cd15ffd1cd720cbd68 20143Author: Dave Airlie <airlied@redhat.com> 20144Date: Mon Aug 4 14:38:21 2008 +1000 20145 20146 radeon: make r600 use i2c table lookup for ddc. 20147 20148 This may fix DDC on rv770 cards. 20149 20150commit 1f3eee3682f3598a303c9c3accfbe01b245cacf9 20151Author: Alex Deucher <alexdeucher@gmail.com> 20152Date: Tue Jul 29 20:29:32 2008 -0400 20153 20154 Change prim types for exa and textured video to help avoid tearing 20155 20156 - r1xx - switch from tri fan to rect list 20157 - r2xx/r3xx/r4xx/r5xx - switch from tri fan to quad list 20158 20159commit 5b5441f8cc119db0d1e03dd35bd06015a26270dd 20160Author: Brice Goglin <bgoglin@debian.org> 20161Date: Tue Jul 29 23:31:14 2008 +0200 20162 20163 Increase default GART size to 32MB on >=R300 20164 20165commit 0a505297f09aefb1b4432176a263bfdf6f256f77 20166Author: Alex Deucher <alexdeucher@gmail.com> 20167Date: Mon Jul 28 17:12:41 2008 -0400 20168 20169 Fix error in driver connector table for powerbook w/ vga 20170 20171commit d5799ac53c3e1b2ca1da97780b171a44060c3aad 20172Author: Alex Deucher <alexdeucher@gmail.com> 20173Date: Mon Jul 28 11:09:10 2008 -0400 20174 20175 Add quirk for oem x300 card 20176 20177 - debian bug 492457 20178 20179commit 64dbadddcf6d069c0472f37afeab89d3e31e937d 20180Author: Owain Gordon Ainsworth <oga@openbsd.org> 20181Date: Mon Jul 28 10:40:58 2008 +1000 20182 20183 radeon: don't call ioctl unless DRI is initialised 20184 20185commit 1c5858484da4fb1c9bc3ac3b4d7a97863ab99730 20186Author: Alex Deucher <alexdeucher@gmail.com> 20187Date: Mon Jul 21 23:47:45 2008 -0400 20188 20189 First pass at InitDispBandwidth() for AVIVO chips 20190 20191 - support for LB allocation 20192 - MC priority bumps for display1/2 on RV515 variants and RS690 20193 If you are having display underflow problems (flickering on sides of 20194 screen in high res modes, etc.) on RV515 or RS690 boards, try setting: 20195 Option "DisplayPriority" "HIGH" in your config. 20196 - still no support for full display watermark programming yet 20197 20198 Something similar might be useful in rhd as well. 20199 20200commit b0378bb145c8a915c943bef7d17f2cdecfccc891 20201Author: Alex Deucher <alexdeucher@gmail.com> 20202Date: Mon Jul 21 13:47:09 2008 -0400 20203 20204 Interlaced mode fixups for AVIVO chips 20205 20206commit c18fad622a3c4f9572051120d83af68b625b5686 20207Author: Alex Deucher <alexdeucher@gmail.com> 20208Date: Mon Jul 21 10:36:48 2008 -0400 20209 20210 Clear display priority bits before resetting them 20211 20212commit dc231ff8e063313d2bcf5acccad67a9f8a7e3314 20213Author: Alex Deucher <alexdeucher@gmail.com> 20214Date: Mon Jul 21 10:30:41 2008 -0400 20215 20216 Clean up legacy display watermark setup 20217 20218 - makes crtc1 and crtc2 watermark setup independant. 20219 - fixes the case where only crtc2 is active 20220 20221commit f9034214f070fe3054cd9cefd7034252de234f38 20222Author: Michel Dänzer <michel@tungstengraphics.com> 20223Date: Mon Jul 21 09:09:02 2008 +0200 20224 20225 Call DRM modeset ioctl after the IRQ has been initialized. 20226 20227 This lets the DRM know it can safely disable the vblank interrupts. 20228 20229commit ecb6347a3d7071890600c98c2addef3a8ca260ee 20230Author: Alex Deucher <alexdeucher@gmail.com> 20231Date: Sat Jul 19 11:34:16 2008 -0400 20232 20233 Add oem quirk for external tmds setup on Dell Inspiron 8600 20234 20235 Noticed by fnord42 on IRC. 20236 20237commit df53d12a06fad41f00cff849458cb358ab5e2098 20238Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 20239Date: Thu Jul 17 14:37:05 2008 +1000 20240 20241 Fix console switch on R500 20242 20243 This patch fixes the console switch for me on R5xx. 20244 20245 There are two aspects to it: 20246 20247 - Fix the ordering of avivo_restore() to better match what's 20248 happening in the driver & ATOM, properly locking/unlocking and 20249 only enabling the CRTCs after everything has been properly 20250 programmed. 20251 20252 - Don't ASIC_INIT if the card has any CRTC enabled. This is the 20253 best I came up with for avoiding spurrious ASIC_INIT on cards that 20254 -are- POSTed but don't have the BIOS coming from c0000 on x86. The 20255 problem with spurrious ASIC_INIT is that we do it before we do 20256 RADEONSave(), so that screws up the console switch. 20257 20258 Note that I think we also should save/restore the palette, I don't think 20259 we do. right now, it's a minor issue for me because I fixed offb to be 20260 able to set it on AVIVO's but it might still have to be done in the long 20261 run. 20262 20263 Tested with a VGA analog setup on DACA and a DVI setup on TMDS 0. I 20264 haven't tested any other combo but that should get us going. 20265 20266 Cheers, 20267 Ben. 20268 20269 Signed-off-by: Dave Airlie <airlied@redhat.com> 20270 20271commit e8c07270e1fc2ee455320370692aaf90abac2773 20272Author: Alex Deucher <alexdeucher@gmail.com> 20273Date: Tue Jul 15 17:50:32 2008 -0400 20274 20275 ATOM: fix for r4xx cards with no VRAM_Usage tables 20276 20277 fixes bug 15954 20278 20279commit da41e71efd41907d9347a902720bce60b32550e5 20280Author: Alex Deucher <alexdeucher@gmail.com> 20281Date: Mon Jul 14 13:21:38 2008 -0400 20282 20283 ATOM: Add support for UpdateCRTC_DoubleBufferRegisters cmd table 20284 20285 locks/unlocks the crtc/grph/mode regs before updating the crtc 20286 20287commit a5e0cf13dc7ace6cf0e44e18b73b9a9266e740ab 20288Author: Julien Cristau <jcristau@debian.org> 20289Date: Sun Jul 6 12:17:28 2008 +0200 20290 20291 Link with -lpciaccess and -ldrm if needed 20292 20293 This makes sure the driver ends up with a DT_NEEDED reference to 20294 the libraries it's using. 20295 20296commit f8da849fd89322a54cc4d722767da755358eab70 20297Author: Michel Dänzer <michel@tungstengraphics.com> 20298Date: Sun Jul 13 11:31:03 2008 +0200 20299 20300 Only declare local variable when it's used. 20301 20302commit a6db4dc65aec560f3b1f37ff73ec6b0bd8207295 20303Author: Dave Airlie <airlied@linux.ie> 20304Date: Sat Jul 12 18:47:07 2008 +1000 20305 20306 atombios: fix typo in mode conversion for panel 20307 20308commit 19a0d5879851eff23a3e7d5cdea55bd784051115 20309Author: Dave Airlie <airlied@linux.ie> 20310Date: Sat Jul 12 17:04:25 2008 +1000 20311 20312 atombios: use macro to get offset into table 20313 20314commit a9746114369d516072d841ec356ec3ba3d0be71a 20315Author: Dave Airlie <airlied@linux.ie> 20316Date: Sat Jul 12 10:46:36 2008 +1000 20317 20318 pciid: add radeon HD3850. 20319 20320 pci id from legume on #radeon 20321 20322commit ab14f725676e4d4e45278c64b03fe2d328a3e439 20323Author: Alex Deucher <alexdeucher@gmail.com> 20324Date: Fri Jul 11 19:05:00 2008 -0400 20325 20326 R300: NUM_FPU adjustments for VAP_CNTL 20327 20328commit 810c28cc2660b73e03e4d27be97988fc5321a06f 20329Author: Alex Deucher <alexdeucher@gmail.com> 20330Date: Fri Jul 11 15:31:57 2008 -0400 20331 20332 Make sure cursor offsets are properly aligned when using EXA 20333 20334commit 7e67d0163579a44f104e8f354a929ac9b33e4c21 20335Author: Alex Deucher <alexdeucher@gmail.com> 20336Date: Fri Jul 11 14:30:17 2008 -0400 20337 20338 Fix cursor with multi-head and rotation 20339 20340commit 9086d008fb8c3cde0633b90ce19ffbf4eded388d 20341Author: Alex Deucher <alexdeucher@gmail.com> 20342Date: Tue Jul 8 09:57:04 2008 -0400 20343 20344 Add quirk for Falcon NW laptop 20345 20346commit 61f82ace0210251beb0bcc492218a75a193e1deb 20347Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 20348Date: Tue Jul 8 21:58:43 2008 +1000 20349 20350 atombios: add support for other endians. 20351 20352 This is a cleaned up (in as much as atombios can be..) of benh's patch. 20353 20354 airlied - removed benh's debugging for now, it might need to be put back later.. 20355 20356commit b4d1a47b5c6eafda5e274398eebe2701b030f22e 20357Author: Dave Airlie <airlied@linux.ie> 20358Date: Tue Jul 8 11:09:01 2008 +1000 20359 20360 rv770 initial support 20361 20362 adds pci ids and one register from AMD code 20363 20364commit 7ae4cec8cc8c90aee5dc4fa7abcce22321d4f4eb 20365Author: Roland Scheidegger <sroland@tungstengraphics.com> 20366Date: Mon Jul 7 14:39:47 2008 -0400 20367 20368 clamp tex coords (r100/r200) for textured video 20369 20370 fixes bug 14904 20371 20372commit 879f9a37856642b337e3c0233e38fd5443338eef 20373Author: Alex Deucher <alexdeucher@gmail.com> 20374Date: Fri Jul 4 13:54:49 2008 -0400 20375 20376 Revert "rotate: fix rotation in conf file." 20377 20378 This reverts commit 22c6ca0d5d226c725ba448e5db63fdba9e32332c. 20379 20380 There seems to be an ordering issue in that causes a nasty hang 20381 if the DRI is enabled. This needs more investigation. 20382 20383commit 22c6ca0d5d226c725ba448e5db63fdba9e32332c 20384Author: Dave Airlie <airlied@redhat.com> 20385Date: Thu Jul 3 20:14:59 2008 +1000 20386 20387 rotate: fix rotation in conf file. 20388 20389 need to set modes after init accel architecture 20390 20391 reported-by: Christoph Bartoschek <bartoschek@or.uni-bonn.de> 20392 20393commit ce1a3edd52b647744cfa4433301befb437d211bb 20394Author: Dave Airlie <airlied@redhat.com> 20395Date: Thu Jul 3 20:05:54 2008 +1000 20396 20397 radeon: drop all use of CPMode. 20398 20399 We never test the other codepath and I don't think I've ever recommended it 20400 for anyone. 20401 20402commit c037b4ce8769ad840a257e22b1e4ad58b8ed96fa 20403Author: Wolke Liu <wolke.lui@amd.com> 20404Date: Tue Jul 1 10:45:39 2008 -0400 20405 20406 Add pci id for FireMV 2400 20407 20408commit 72992668ca96717bf69bcea95a5deeaf7c57e826 20409Author: Alex Deucher <alex@botchco.com> 20410Date: Mon Jun 30 14:09:09 2008 -0400 20411 20412 RS4xx: enable the DRI by default on all variants 20413 20414 Some RS4xx chips had the DRI disabled by default. 20415 With the recent drm and ddx changes these chips have been 20416 reported to work now with the DRI. 20417 20418commit 0378c8ed88c829f09b5c6d51314325eaf2685fdb 20419Author: Michel Dänzer <michel@tungstengraphics.com> 20420Date: Sat Jun 28 14:06:45 2008 +0200 20421 20422 Restore versioning of interface for Mesa DRI driver. 20423 20424 This doesn't have anything to do with the package version. 20425 20426commit cfaa23d925e3c062cf87ea844566ac11ea02d69e 20427Author: Alex Deucher <alex@botchco.com> 20428Date: Fri Jun 27 20:36:54 2008 -0400 20429 20430 RADEON: IGP VGA fixes take 2 20431 20432 It seems only RS300/350/400/480 are missing the primary dac 20433 See bug 15708 20434 20435commit 31c27ffcb3c4c5334cf726ecd4e293a678b2a1ea 20436Author: Alex Deucher <alex@botchco.com> 20437Date: Fri Jun 27 20:29:04 2008 -0400 20438 20439 Revert "IGP: attempt to fix VGA on IGP chips" 20440 20441 This reverts commit e78e8a21b4040cd7f1983c241c860d9209398396. 20442 20443commit af79a77e6d8206b18bba97cf4e0770921dc439a7 20444Author: Brice Goglin <Brice.Goglin@ens-lyon.org> 20445Date: Fri Jun 27 17:09:13 2008 +0200 20446 20447 use PACKAGE_VERSION_* for radeon version 20448 20449 now radeon shows verion 6.9.0 instead of 4.3.0. 20450 20451 RADEON_VERSION_MAJOR_TILED is dropped from ddxDriverMajorVersion 20452 since RADEON_VERSION_MAJOR is bigger now. 20453 20454commit c83fbdfa076c107012b7dfbbfbbb2feede00542b 20455Author: Alex Deucher <alex@botchco.com> 20456Date: Thu Jun 26 19:48:45 2008 -0400 20457 20458 Bump for 6.9.0 release 20459 20460commit 32e1d1daf4f68ea37624afaa2bc1ea7742e1163b 20461Author: Brad Smith <brad@comstyle.com> 20462Date: Thu Jun 26 10:22:03 2008 -0400 20463 20464 RADEON: sys/endian.h needs sys/types.h on BSDs 20465 20466 See bug 16512 20467 20468commit dca522355a9039eca6efaba3b36397b246800f94 20469Author: Brad Smith <brad@comstyle.com> 20470Date: Thu Jun 26 10:13:59 2008 -0400 20471 20472 RADEON: fix copy/paste error in accel code 20473 20474commit f4292e110105910d0c19bf2db28e2682b27af2c7 20475Author: Dave Airlie <airlied@redhat.com> 20476Date: Fri Jun 20 15:31:43 2008 +1000 20477 20478 Revert "Revert "ATOM: disable TV-out for now"" 20479 20480 Oops I really didn't mean to do this, I was testing something and it slipped 20481 past. 20482 20483 This reverts commit dd18caa4b2efc430eaae0c4362b65641f9bef440. 20484 20485commit bb1cfcbfbc1ace1f257fc3bf6401501f1a6da827 20486Author: Alex Deucher <alex@botch2.com> 20487Date: Wed Jun 25 15:50:16 2008 -0400 20488 20489 RADEON: adjustments to Jerome's last commit 20490 20491 - Flush caches and wait for idle after drawing 20492 - Make sure 3D is idle too (after composite or textured video) 20493 20494commit 07be302f6dba61dd15918963fae66502230c74b7 20495Merge: 8c9b8de0 52459745 20496Author: Jerome Glisse <glisse@freedesktop.org> 20497Date: Wed Jun 25 10:30:45 2008 +0200 20498 20499 Merge branch 'master' of git://anongit.freedesktop.org/xorg/driver/xf86-video-ati 20500 20501commit 8c9b8de0373797cba6f0a27e6b6f461e1070fef9 20502Author: Jerome Glisse <glisse@freedesktop.org> 20503Date: Wed Jun 25 10:28:26 2008 +0200 20504 20505 radeon: flush & wait for 2d & dma idle after 2d blit 20506 20507 This should help to avoid 2d & 3d engine to step on each 20508 other dma transaction. 20509 20510commit 52459745ec05de88adbc087e9566fe6d97ef424b 20511Author: Dave Airlie <airlied@linux.ie> 20512Date: Wed Jun 25 17:09:11 2008 +1000 20513 20514 ati: hopefully fix byteswap mess for those other OSes. 20515 20516commit 67a6ac0001bc9d062aa426384a11a41fa7a1c09a 20517Author: Alex Deucher <alex@botch2.com> 20518Date: Tue Jun 24 21:06:37 2008 -0400 20519 20520 bump for rc release 20521 20522commit bd68507d2d66e03d8bcde5f6e7ea9b2dbfe8b8a0 20523Author: Alex Deucher <alex@botch2.com> 20524Date: Tue Jun 24 20:59:58 2008 -0400 20525 20526 RADEON: warning fix 20527 20528commit 8e534d69c9b19fc085f7c5ca9e18f5ea04f6fc12 20529Author: Alex Deucher <alex@botch2.com> 20530Date: Tue Jun 24 20:08:35 2008 -0400 20531 20532 RADEON: cleanups 20533 20534 - fix some warnings 20535 - RS400 and RS480 are separate families now (update default tmds and dac2 20536 tables) 20537 20538commit e78e8a21b4040cd7f1983c241c860d9209398396 20539Author: Alex Deucher <alex@botch2.com> 20540Date: Tue Jun 24 19:53:28 2008 -0400 20541 20542 IGP: attempt to fix VGA on IGP chips 20543 20544 VGA has never worked on some IGP chips. While the chip only has 20545 one DAC, it appears to use a mix of Primary DAC and TVDAC controls. 20546 20547 See bug 15708 20548 20549commit faa4b4b8dbe9b8a452cfa60d53874350bb04e0cb 20550Author: Jiří Paleček <jpalecek@web.de> 20551Date: Tue Jun 24 19:20:50 2008 -0400 20552 20553 RADEON: remove extraneous line from new pll code 20554 20555commit 72feaa37ea07620f5f2ead438dbc72a1c8883cd3 20556Author: Jiří Paleček <jpalecek@web.de> 20557Date: Mon Jun 23 15:53:58 2008 -0400 20558 20559 RADEON: PLL tweaks 20560 20561 Patch from Jiří Paleček (see debian bug 465864) with some tweaks 20562 by me. 20563 20564 - abort rather than programming bad dividers if no pll dividers can be found 20565 - improve the pll selection algorithm 20566 - in general, prefer lower ref dividers 20567 20568 I've tested this patch on a wide variety of chips (r1xx-r6xx) and clocks. 20569 20570commit 9c2f909ea437a63a408d2398ecabe0b378dbb982 20571Author: Alex Deucher <alex@botch2.com> 20572Date: Mon Jun 23 10:38:15 2008 -0400 20573 20574 RADEON: adjust randr crtc/output prepare/commit ordering 20575 20576 This fixes some occasional mode change problems with multiple heads active. 20577 It seems radeons generally like to turn on the whole output/crtc setup 20578 in one shot. 20579 20580commit aea9bf75cf0774afd3e65fcf3fd3851f5fb21ca3 20581Author: Alex Deucher <alex@botch2.com> 20582Date: Sat Jun 21 10:57:05 2008 -0400 20583 20584 0x1002 0x5657 is actually RV410 20585 20586 See bug 14289 20587 20588 com_bios_scratch.diff 20589 20590commit 38ce8a984f96056b7345bcc06505ba27e0e6d5b4 20591Author: Dave Airlie <airlied@redhat.com> 20592Date: Fri Jun 20 14:16:22 2008 +1000 20593 20594 legacy: use entity MMIO for dpms as this can crossover between zaphod infos 20595 20596commit 1a7d9bc53512b0a5240176c949e6abf1ae2fb1fd 20597Author: Dave Airlie <airlied@redhat.com> 20598Date: Fri Jun 20 14:14:21 2008 +1000 20599 20600 atombios: use MMIO from the entity not the info 20601 20602commit ef624b88903b1a87ef5b6388e18291f75776b93d 20603Author: Alex Deucher <alex@botch2.com> 20604Date: Thu Jun 19 18:20:52 2008 -0400 20605 20606 RADEON: fix read past the end of an array 20607 20608commit cfe814a481d8cf2005d738a0ca9782f1ed4177f5 20609Author: Dave Airlie <airlied@linux.ie> 20610Date: Fri Jun 20 07:51:27 2008 +1000 20611 20612 r600: don't add fb offset here to make shadowfb work. 20613 20614 discovered on irc with wpwrak. 20615 20616commit 5b323a2cbbc412b6dd4e782438b64ee996558d05 20617Merge: dd18caa4 eed24155 20618Author: Dave Airlie <airlied@linux.ie> 20619Date: Fri Jun 20 07:49:16 2008 +1000 20620 20621 Merge branch 'master' of ssh://git.freedesktop.org/git/xorg/driver/xf86-video-ati 20622 20623commit eed241553748125e902c44dcc8cf8e0605be071b 20624Author: Alex Deucher <alex@botch2.com> 20625Date: Tue Jun 17 17:42:31 2008 -0400 20626 20627 radeon: X_ERROR -> X_WARNING for num gb pipes 20628 20629 This just means your drm is old, not fatal or anything 20630 20631commit dd18caa4b2efc430eaae0c4362b65641f9bef440 20632Author: Dave Airlie <airlied@linux.ie> 20633Date: Sat Jun 14 18:37:56 2008 +1000 20634 20635 Revert "ATOM: disable TV-out for now" 20636 20637 This reverts commit effa245914823371e052cd9aa1143a02350891e7. 20638 20639commit b416e97a1f16ef700ba1aaca317dee82b7a3cf64 20640Author: Alex Deucher <alex@botch2.com> 20641Date: Fri Jun 13 02:21:31 2008 -0400 20642 20643 RADEON: man page updates 20644 20645commit 88ce87d79069662de28ecb3c10c140919927d584 20646Author: Alex Deucher <alex@botchco.com> 20647Date: Thu Jun 12 15:10:47 2008 -0400 20648 20649 Bump for 6.9.0rc1 release 20650 20651commit ada41723fdc8cbeeda12bb4ae2d16e8d1ba215de 20652Author: Alex Deucher <alex@botch2.com> 20653Date: Wed Jun 11 14:29:36 2008 -0400 20654 20655 R3xx: use DSTCACHE_CTLSTAT rather than RB2D_DSTCACHE_CTLSTAT 20656 20657 According to the hw guys, you should use DSTCACHE_CTLSTAT to 20658 flush the 2D dst cache rather than RB2D_DSTCACHE_CTLSTAT. 20659 20660commit 3d70c398afa0831387af81559453b77c218a54f9 20661Author: Dave Airlie <airlied@redhat.com> 20662Date: Thu Jun 12 06:18:26 2008 +1000 20663 20664 atom: fix another endian define 20665 20666commit f1794e4c34d801371461afed37e7bb7d1fc79305 20667Author: Dave Airlie <airlied@redhat.com> 20668Date: Thu Jun 12 06:10:30 2008 +1000 20669 20670 atom: fixup includes in cd operations to get correct endian bits in atombios.h 20671 20672commit f29976d12cc2020cd80c90fa3fd4b25c031559ba 20673Author: Dave Airlie <airlied@ppcg5.localdomain> 20674Date: Wed Jun 11 19:33:02 2008 +1000 20675 20676 radeon: add lots of endian bits into the atombios parser 20677 20678commit 09eb220971b5d2bfd7d1ff6f552060967a133152 20679Author: Dave Airlie <airlied@ppcg5.localdomain> 20680Date: Wed Jun 11 19:21:22 2008 +1000 20681 20682 atombios: fixup atombios bitfields for endianness 20683 20684 reverse the bits for big endian 20685 20686commit b80a3e85bf9dbfc8d96a745876d75d38bea97c1a 20687Author: Alex Deucher <alex@botchco.com> 20688Date: Mon Jun 9 20:29:47 2008 -0400 20689 20690 R5xx: fix RADEONSetAgpBase() for R5xx chips 20691 20692 Is there any reason to still do this in the ddx? 20693 Maybe real old drms? 20694 20695commit ab7936708dfbee4d3676262f8010e001b73a4a38 20696Author: Alex Deucher <alex@botchco.com> 20697Date: Mon Jun 9 17:05:04 2008 -0400 20698 20699 RS4xx: RC410 chips are actually RS400 based 20700 20701commit a54ca3c699c103e2e3df87101dca929a3eee4377 20702Author: Dave Airlie <airlied@linux.ie> 20703Date: Thu Jun 5 18:31:53 2008 +1000 20704 20705 radeon: fix typo in memory calcs 20706 20707 noticed by ghepeu on irc. 20708 20709commit 6e4e6d2a8f29f92efc219dca24ea31d1f37d5a0f 20710Author: Alex Deucher <alex@botchco.com> 20711Date: Tue Jun 3 20:32:20 2008 -0400 20712 20713 RADEON: minor cleanups 20714 20715commit 9d3afbf5fa4110928a9f965df2733c79db92ea99 20716Author: Michel Dänzer <michel@tungstengraphics.com> 20717Date: Tue Jun 3 11:40:49 2008 +0200 20718 20719 Call DRM_IOCTL_MODESET_CTL ioctl to avoid problems with DRM post vblank-rework. 20720 20721commit effa245914823371e052cd9aa1143a02350891e7 20722Author: Alex Deucher <alex@botchco.com> 20723Date: Sun Jun 1 20:47:50 2008 -0400 20724 20725 ATOM: disable TV-out for now 20726 20727 It only sort of works in NTSC mode and isn't handled correctly 20728 in all cases yet. Until we sort out the remaining details leave it 20729 disabled. 20730 See bugs: 16186, 16178, 16185 20731 20732commit 8504c6b0e40477ee544ad7f5366d569bdc53d6f0 20733Author: Alex Deucher <alex@botch2.com> 20734Date: Fri May 30 11:19:03 2008 -0400 20735 20736 RADEON: update RADEONGetVRAMType() for newer chips 20737 20738commit 7cb695a329a9b543ad61af08c2d7d6eaf56de35e 20739Author: Alex Deucher <alex@botch2.com> 20740Date: Fri May 30 10:10:56 2008 -0400 20741 20742 RADEON: add Int10 option 20743 20744 On some radeons you can't read the bios without initializing int10. 20745 On ATOM-based secondary cards, intitializing int10 tends to hang the card. 20746 20747commit 714b2c63ec1c4a1410a6d521c03e9d1f90937c01 20748Author: Alex Deucher <alex@botch2.com> 20749Date: Thu May 29 02:31:00 2008 -0400 20750 20751 RADEON: don't enable legacy bios table based init yet 20752 20753 Needs more work and we need to figure out how best to decide 20754 when to use it. 20755 20756commit e20b08525a64888287ec4a369d8f7dbde95c655d 20757Author: Alex Deucher <alex@botch2.com> 20758Date: Wed May 28 14:43:47 2008 -0400 20759 20760 ATOM: fail of we are not able to get a bios image 20761 20762commit d4fa17b426b0cafac3184b2ea9b675e2ff154ee9 20763Author: Alex Deucher <alex@botch2.com> 20764Date: Wed May 28 09:50:02 2008 -0400 20765 20766 R1xx: Gamma fixes for overlay 20767 20768commit c7e6a50f60a04bbbf1c1ddbf4738a5f75c94f950 20769Author: Alex Deucher <alex@botchco.com> 20770Date: Wed May 28 09:05:19 2008 -0400 20771 20772 RADEON: only attempt to read vbios from legacy ISA space if it's primary 20773 20774commit faea008806802ec0e045754ec1eca492ebae320e 20775Author: Alex Deucher <alex@botch2.com> 20776Date: Tue May 27 18:36:01 2008 -0400 20777 20778 R3/4/5xx: use get_param to get the num_gb_pipes from the drm 20779 20780commit 965a5dbcd9dc4bf1cdd7f2bbdec15e9733b2e090 20781Author: Alex Deucher <alex@botch2.com> 20782Date: Tue May 27 16:48:41 2008 -0400 20783 20784 RADEON: improve support for secondary cards 20785 20786 this should fix bugs 16115, 16035 20787 20788commit 5f951a5573f0c7572230c9aa4d3f75d67f91ed71 20789Author: Alex Deucher <alex@botch2.com> 20790Date: Wed May 21 00:35:14 2008 -0400 20791 20792 ATOM: remove duplicate code 20793 20794commit edce33e87fb79a540d8c285f205d4c3f2c2bc9f4 20795Author: Dave Airlie <airlied@redhat.com> 20796Date: Wed May 28 06:43:40 2008 +1000 20797 20798 radeon: fix surface access on avivo chips. 20799 20800 This should fix VT switch on vesafb 20801 20802commit 307bd65a25ee425d8359dd30572b002ce2338f91 20803Author: Michael Babej <mbabej@redhat.com> 20804Date: Tue May 27 19:13:49 2008 +1000 20805 20806 r600: fix vt switch issue 20807 20808 RH BZ 441492 20809 20810commit bf48c9eb8ec592515be9d1732d60283af715674d 20811Author: Dave Airlie <airlied@linux.ie> 20812Date: Sat May 24 11:51:20 2008 +1000 20813 20814 atombios: for LVDS set use ref div as per legacy 20815 20816commit 04500c8419b1aeaeac7968492b82e1d7cad1d05b 20817Author: Alan Coopersmith <alan.coopersmith@sun.com> 20818Date: Tue May 20 20:34:31 2008 -0700 20819 20820 Strip ^M's from src/AtomBios/includes/ObjectID.h 20821 20822commit 521a0488d1c414209c3534dc242000faa332a441 20823Author: Alan Coopersmith <alan.coopersmith@sun.com> 20824Date: Tue May 20 20:33:37 2008 -0700 20825 20826 On non-gcc compilers, use C99's __func__ instead of gcc's __FUNCTION__ 20827 20828commit b7c80d0c86646105d2bce5d4d59ba6c45aa7cafc 20829Author: Tobias Diedrich <ranma+freedesktop@tdiedrich.de> 20830Date: Tue May 20 19:16:54 2008 -0400 20831 20832 R200/R300: fix gamma setup for overlay 20833 20834 More pending. See bug 16001 20835 20836commit 69423482e2e94637142a9ba675589a1449a346a8 20837Author: Alex Deucher <alex@botch2.com> 20838Date: Tue May 20 18:57:13 2008 -0400 20839 20840 RADEON: cleanup connection detection and remove duplicate code 20841 20842commit 55e4469f59c82bb5762673de5f3f27d18b0bd9a3 20843Author: Alex Deucher <alex@botch2.com> 20844Date: Tue May 20 17:46:58 2008 -0400 20845 20846 RADEON: enable cloning on multi-crtc cards 20847 20848commit 12f185634071980041aaac6265d89708b181b215 20849Author: Alex Deucher <alex@botch2.com> 20850Date: Tue May 20 16:05:41 2008 -0400 20851 20852 ATOM: Ignore invalid connector entries 20853 20854commit 19e97f74e39fc2b35727708ac429de33d0b70162 20855Author: Egbert Eich <eich@freedesktop.org> 20856Date: Tue May 20 11:16:15 2008 +0200 20857 20858 Change RMX code to follow the programming algorithm suggested by ATI. 20859 20860 ATI provides the following algorithm to calculate the RMX scaling ratios 20861 in its programming specs: 20862 when RMX_AUTO_RATIO_HORZ_INC set to 1,Horizontal auto ratio result trucated, 20863 and then incremented by 1. 20864 Horz_Ratio = ( ((Active display width in characters (including overscan) + 1) 20865 / (Panel width in characters)) x 4096 + 1 ) 20866 else 20867 Horz_Ratio = ( ((Active display width in characters (including overscan)) 20868 / (Panel width in characters)) x 4096 + 1 ) 20869 20870 when RMX_AUTO_RATIO_VERT_INC set to 1, Vertical auto ratio result trucated, 20871 and then incremented by 1. 20872 Vert_Ration = ( ((Active display width in characters (including overscan) + 1) 20873 / (Panel width in characters)) x 4096 + 1) 20874 else 20875 Vert_Ration = ( ((Active display width in characters (including overscan)) 20876 / (Panel width in characters)) x 4096 + 1) 20877 20878 This patch implements this behavor. Additionally it avoids the use of floats. 20879 20880commit a4f3d0088ba763ed8eab1d331959b5ecde8262e8 20881Author: Alex Deucher <alex@botch2.com> 20882Date: Tue May 20 11:42:53 2008 -0400 20883 20884 R3/4/5xx: fixup logic from last commit 20885 20886 Spotted by otaylor and MrCooper 20887 20888commit 130e55738047f2a073bcc47e3e1400f7b694a81b 20889Author: Alex Deucher <alex@botch2.com> 20890Date: Tue May 20 11:32:42 2008 -0400 20891 20892 R3/4/5xx: fix EXA rotation 20893 20894 xrandr uses PictOpSrc for rotation which we were falling back on since 20895 render semanties require alpha=0 for REPEAT_NONE when there is no alpha 20896 channel and there is a transform. If the dst has no alpha channel we 20897 should be ok I think. 20898 20899 Otayler and I discussed this on IRC. the more general fix would be 20900 to clip the dst to the src and mask for bounded ops like in/add/over 20901 in the pixel exact transform case. 20902 20903commit cc9f510770700228e5d597c872e926a4e99bd950 20904Author: Corbin Simpson <MostAwesomeDude@gmail.com> 20905Date: Wed May 14 16:49:01 2008 -0400 20906 20907 RADEON: missing stdint.h includes 20908 20909commit 71fa57f871dba03260dba2180ce1dab44048ac1a 20910Author: Alex Deucher <alex@cube.(none)> 20911Date: Wed May 14 13:36:38 2008 -0400 20912 20913 Add RS600 support 20914 20915commit 582c1a1b2c7b1032e9f9f54ca36100c57f580c5c 20916Author: Alex Deucher <alex@cube.(none)> 20917Date: Tue May 13 20:50:25 2008 -0400 20918 20919 RS4xx: Split out RS400 and RS480 as separate families 20920 20921 RS400 (intel based IGP) and RS480 (AMD based IGP) have different 20922 MC setups and need to be handled differently 20923 20924commit 708e7c98f636738fbcc47a597bc94b309a4dc1c4 20925Author: Michel Dänzer <michel@tungstengraphics.com> 20926Date: Tue May 13 11:00:38 2008 +0200 20927 20928 Reinstate CARD* types that are part of external interfaces. 20929 20930 The compiler pointed this out... 20931 20932commit 908b7b940e3ef296836bc94680ebb8ab67650e31 20933Author: Matt Turner <mattst88@gmail.com> 20934Date: Mon May 12 12:06:33 2008 -0400 20935 20936 Replace CARD{8,16,32} with uint{8,16,32}_t 20937 20938 As has been done with xf86-video-intel, replace all CARD* datatypes with 20939 uint*_t datatypes available from stdint.h. 20940 20941commit 87e66ce76430890ab4939ffcd42f72b9288eb598 20942Author: Avi Rozen <avi.rozen@gmail.com> 20943Date: Mon May 12 08:33:38 2008 -0400 20944 20945 RADEON: fix lockup on start 20946 20947 see debian bug 480312 20948 20949commit 94bf8f01bd43cb103fffecfe04d04a214f892baf 20950Author: Dave Airlie <airlied@redhat.com> 20951Date: Mon May 12 20:02:51 2008 +1000 20952 20953 radeon: fix set_cursor_colours remove fb offset 20954 20955commit 1d0f1d31e2ed1d91ee87cb3e02ce48c8c07aa418 20956Author: Dave Airlie <airlied@redhat.com> 20957Date: Mon May 12 15:44:35 2008 +1000 20958 20959 radeon: rs485 vs rs485m.. mobile? non-mobile? bi? 20960 20961 So it appears we have in the wild two chipsets with the same PCI ID (0x1002:0x5974) 20962 that are mobile and non-mobile. 20963 20964 the RH bug references is a desktop system. 20965 The Dell Vostro 1100 also has this chipset with LVDS bits. 20966 20967commit 25e0c3945a51ae8c14b3a847ec75a256e1397f24 20968Author: Alex Deucher <alexdeucher@gmail.com> 20969Date: Mon May 12 10:28:33 2008 +1000 20970 20971 radeon: add initial support for cloning outputs. (single-CRTC cards only) 20972 20973 airlied - This code was originally written by Alex, and I've modified it to 20974 only run on single-CRTC cards for now until we can test it some more. 20975 20976commit 94405eb1c9e4e0ababc6aef03b753d6ed9eb7838 20977Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 20978Date: Fri May 9 05:28:44 2008 -0400 20979 20980 RADEON: fixup ifdef from last commit 20981 20982commit 18e20bc22a55ecfee9798c01079d7b24d19f0051 20983Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 20984Date: Thu May 8 08:45:51 2008 -0400 20985 20986 RADEON: disable MMIO paths for EXA composite/texvid on IGP/R5xx 20987 20988 The MMIO paths eventually lead to a hang on r5xx/IGP. I haven't 20989 been able to find out why yet. 20990 20991commit c3532268875fd24e6519bea2fb1b814d612bbdb4 20992Author: Dave Airlie <airlied@linux.ie> 20993Date: Wed May 7 02:37:18 2008 +1000 20994 20995 radeon: fix zaphod EXA with texture video 20996 20997commit ffc437f3606ab8ceba1ff152e4bb08988a58b54c 20998Author: Dave Airlie <airlied@linux.ie> 20999Date: Wed May 7 02:30:28 2008 +1000 21000 21001 avivo: fix zaphod cursor in theory 21002 21003commit e36ef14e3a1087e1fe41baa26ade2937f396001f 21004Author: Dave Airlie <airlied@linux.ie> 21005Date: Wed May 7 01:39:28 2008 +1000 21006 21007 radeon: fix textured-xv on zaphod 21008 21009commit 8fc19bee27c0f151d2ab3354f6ac0992b358436d 21010Author: Dave Airlie <airlied@linux.ie> 21011Date: Wed May 7 01:38:24 2008 +1000 21012 21013 radeon: zaphod: fix render accel for EXA 21014 21015commit fc41b9042a5220a8419cc7b69ca3850cae6b903c 21016Author: Dave Airlie <airlied@linux.ie> 21017Date: Wed May 7 01:32:23 2008 +1000 21018 21019 radeon: fix EXA pixmap offset on zaphod 21020 21021commit 4568cb820d567c6909a4be956d7e79b91232535e 21022Author: Dave Airlie <airlied@linux.ie> 21023Date: Wed May 7 01:19:39 2008 +1000 21024 21025 radeon: zaphod fixes for pciaccess not allowing multiple MMIO maps 21026 21027commit ca81fa086b21633a7fd926833fb6d1d4fa080646 21028Author: Dave Airlie <airlied@linux.ie> 21029Date: Wed May 7 01:12:01 2008 +1000 21030 21031 radeon: zaphod fix for cursor on second head 21032 21033 We don't need to add fbOffset here as the mmap we have of the framebuffer 21034 starts half way. 21035 21036commit 24b60c8965f6a0b3f0c2bb1e7236b4d6642c5918 21037Author: Julien Cristau <jcristau@debian.org> 21038Date: Fri May 2 15:30:45 2008 -0400 21039 21040 Add a test for __GLIBC__ to CD_Common_Types.h. 21041 21042 Atombios redefines the standard types but the definitions conflict 21043 with the ones from glibc (causes build failures on GNU/Hurd 21044 and GNU/kFreeBSD). 21045 21046commit f051359ac09c6b9416e39b9ca7d9dc0880aa1557 21047Author: thegraze <thegraze@googlemail.com> 21048Date: Fri May 2 14:02:16 2008 -0400 21049 21050 ATOM: add support for DragonFlyBSD 21051 21052commit 3d469cbc3225d890a895dac7cbc1ab7e08054b48 21053Author: Alex Deucher <alex@cube.(none)> 21054Date: Wed Apr 30 18:33:04 2008 -0400 21055 21056 RADEON: lock the cursors when updating 21057 21058 this should fix occasional corruption seen when updating 21059 the cursor. 21060 21061commit 445b71021843665ba32f37b2ce5c9d2857c07cc7 21062Author: Alex Deucher <alex@cube.(none)> 21063Date: Tue Apr 29 21:01:41 2008 -0400 21064 21065 RADEON: assorted fixes 21066 21067 - free rotate pixmaps on VT switch 21068 - save crtc/output status so we only turn on 21069 crtcs/outputs if they are off 21070 - show/hide cursors when changing modes 21071 21072commit 070cce5255a5c311f9d8b85ec54bd56655014933 21073Author: Stephan Wolf <stephan@letzte-bankreihe.de> 21074Date: Mon Apr 28 11:26:37 2008 -0400 21075 21076 R3xx+: further fix for IGP chips 21077 21078 see bug 15538 21079 21080commit 211e0041c7fc2df494b77428553943a2b526ee4e 21081Author: Alex Deucher <alex@cube.(none)> 21082Date: Sun Apr 27 21:08:00 2008 -0400 21083 21084 IGP: fix EXA composite corruption 21085 21086commit 656b06bdde129ca4fc370f5a2cf7311c9179b0ff 21087Author: Alex Deucher <alex@botch2.com> 21088Date: Sun Apr 27 20:20:49 2008 -0400 21089 21090 RADEON: remove duplicate register define 21091 21092 Also add more bit defs to wait_until register 21093 21094commit 8a9820a3aa49bc667f90ac291a27e4d7b4ae38b3 21095Author: Alex Deucher <alex@botch2.com> 21096Date: Sun Apr 27 19:02:22 2008 -0400 21097 21098 RADEON: decrease crtc/output verbosity 21099 21100commit c5d62fa0e8f52c3264ff9db3ff10cdf5a806bfc0 21101Author: Owen Taylor <otaylor@huygens.home.fishsoup.net> 21102Date: Thu Apr 17 13:14:53 2008 +0200 21103 21104 Emulate repeats by drawing in tiles 21105 21106 When we can't turn on hardware repeats, because the texture 21107 is non-power-of-two, or has padding at the ends of lines, 21108 try to draw the image in multiple tiles rather than falling 21109 back to software. (We can only do this when there is no 21110 transform.) 21111 21112commit eeb7b74bb6c813b0e3afa4b704f6ffb0d0aab92b 21113Author: Owen Taylor <otaylor@huygens.home.fishsoup.net> 21114Date: Thu Apr 17 13:14:25 2008 +0200 21115 21116 Turn on wrapping when repeating on R100 + R200 21117 21118 Actually enable repeats for R100 and R200. This corresponds 21119 to a R300 change made in the patch in: 21120 http://bugs.freedesktop.org/show_bug.cgi?id=15333 21121 21122commit e511f39dfef503006cf249b9f6934091eaade9b5 21123Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 21124Date: Thu Apr 17 05:04:34 2008 -0400 21125 21126 R300+: move more common code into init3d() 21127 21128 - pre-load r3xx tex instructions 21129 - setup RS instructions in init3d() 21130 21131commit 99435b7c18d931ea620044d0fdb4cc93dfcc6331 21132Author: Owen Taylor <otaylor@redhat.com> 21133Date: Thu Apr 17 02:46:11 2008 -0400 21134 21135 Radeon: Omit mask coordinates 21136 21137 Adapted from Owen's patch on bug 15546 21138 This fixes the slowness with aatext on r300 21139 and may speed up other chips marginally. 21140 21141commit 37614e1db9a595fbe8a21d7a045895e11d272db9 21142Author: Alex Deucher <alex@botch2.com> 21143Date: Tue Apr 15 09:48:16 2008 -0400 21144 21145 fix up some things from the last commit 21146 21147commit 1286fe5ce1c77453d57817b9b26b1bdb32ca7bc8 21148Author: Alex Deucher <alex@botch2.(none)> 21149Date: Mon Apr 14 20:02:14 2008 -0400 21150 21151 R300+: properly setup vap_cntl 21152 21153 this fixes tcl/pvs on RV515 among other things 21154 21155commit f72a4b805db26f10f69330b88459cbeae661189b 21156Author: Alex Deucher <alex@botch2.com> 21157Date: Mon Apr 14 14:10:40 2008 -0400 21158 21159 EXA: Don't wait for 3D idle after each Composite() 21160 21161 wait in CompositeDone() instead 21162 21163commit 4cd4acf1092aeb696b086a382a033aee471d2de9 21164Author: Alex Deucher <alex@botch2.com> 21165Date: Mon Apr 14 11:50:59 2008 -0400 21166 21167 R300: move more common code to init3d() 21168 21169commit 3c523c9a07402e17dff588fad842224c57e98223 21170Author: Alex Deucher <alex@botch2.com> 21171Date: Mon Apr 14 11:21:42 2008 -0400 21172 21173 R3xx+: 3D engine documentation and minor cleanups 21174 21175 - document the R300 exa/textured video code 21176 - minor cleanups of textured video code to clarify meaning 21177 21178commit ce025bbb2496d4de94b8d4ac450c64441b64ee04 21179Author: Alex Deucher <alex@botch2.com> 21180Date: Sat Apr 12 21:22:03 2008 -0400 21181 21182 R300+: consolidate some tcl/non-tcl paths 21183 21184 - Move more code to init3d() 21185 - MMIO textured video seems more reliable now on newer chips 21186 21187commit 11b54a319c7c9dd52e3fb13372697059dafe1cd3 21188Author: Alex Deucher <alex@botch2.com> 21189Date: Sat Apr 12 16:50:22 2008 -0400 21190 21191 R3xx+: fix XAA + textured video on non-TCL path 21192 21193commit dd15a2f5906725116b8cd9954243099055e88e37 21194Author: Alex Deucher <alex@botch2.com> 21195Date: Sat Apr 12 16:49:03 2008 -0400 21196 21197 R3xx+: more fixes to 2D/3D engine init 21198 21199commit f3e68d4b7afd2e23675bf6361c496814c9cb4b94 21200Author: Alex Deucher <alex@botch2.com> 21201Date: Fri Apr 11 10:59:07 2008 -0400 21202 21203 Fix exa glyph corruption on newer chips 21204 21205commit b59686d6427cbf8b35e36b020cbbc6a0c5149b22 21206Author: Alex Deucher <alex@botch2.com> 21207Date: Fri Apr 11 10:15:25 2008 -0400 21208 21209 R300+: pre-load vertex programs in init3D() 21210 21211commit acc5833a35ad6c29a57f659607afb27eebdc2ea5 21212Author: Alex Deucher <alex@botch2.com> 21213Date: Thu Apr 10 17:52:52 2008 -0400 21214 21215 R3xx+: consolidate more tcl code 21216 21217commit 6f8f75bd19ef1919c0291141675be2d0e29b3251 21218Author: Alex Deucher <alex@botch2.com> 21219Date: Thu Apr 10 17:08:50 2008 -0400 21220 21221 R3xx+: consolidate some common 3D code 21222 21223commit 4b9234e1c4f7c7f419cb4245d64f3f9756c98bb6 21224Author: Alex Deucher <alex@botch2.com> 21225Date: Thu Apr 10 16:58:22 2008 -0400 21226 21227 R3xx+: tcl wip 21228 21229commit 865c463e3afb4759758f569132be8bf1386da5cc 21230Author: Alex Deucher <alex@botch2.com> 21231Date: Thu Apr 10 16:51:04 2008 -0400 21232 21233 R300+: textured video tcl cleanup 21234 21235commit 79c8d4ca36a1c3e5fe759d4ccc379c36af8f1676 21236Author: Alex Deucher <alex@botch2.com> 21237Date: Thu Apr 10 16:28:18 2008 -0400 21238 21239 RADEON: cleanup 21240 21241commit c4821a287d29a65f3bcb7d60dc72ec13c0384008 21242Author: Alex Deucher <alex@botch2.com> 21243Date: Thu Apr 10 16:20:17 2008 -0400 21244 21245 Revert "R3xx/R5xx: move more VAP, etc. state setup into common init3d() function" 21246 21247 This reverts commit 305a3310963a5dd07b3495015b06aa8c7c4e6b02. 21248 21249 Conflicts: 21250 21251 src/radeon_commonfuncs.c 21252 src/radeon_exa_render.c 21253 src/radeon_textured_videofuncs.c 21254 21255commit 0032c80bf30bab189204e3e6929e18a19d753138 21256Author: Alex Deucher <alex@botch2.com> 21257Date: Thu Apr 10 14:35:00 2008 -0400 21258 21259 RADEON: store tcl status in driver rec 21260 21261commit 9e2ffe66d106abe34a670d2edc9905ed62c485e8 21262Author: Alex Deucher <alex@botch2.com> 21263Date: Thu Apr 10 14:24:04 2008 -0400 21264 21265 R3xx+: use the right register for engine flush 21266 21267commit e1a9f26c2d2cbca9ad159e723ec95b95be1ef349 21268Author: Alex Deucher <alex@botch2.com> 21269Date: Thu Apr 10 14:12:15 2008 -0400 21270 21271 R3xx+: minor textured video fixes 21272 21273 - set shader output swizzling correctly 21274 - flush the right cache register on r3xx+ 21275 21276commit d79040906cd25bd494feb5901f465bbd050aa923 21277Author: Alex Deucher <alex@botch2.com> 21278Date: Thu Apr 10 13:59:58 2008 -0400 21279 21280 R3xx+: EXA/textured video fixes 21281 21282 - get pipe config based on GB_PIPE_SELECT where applicable 21283 (adapted from a similar patch from Dave) 21284 - only flush the dst cache after submitting vertices, freeing 21285 the cache lines stalls the pipe 21286 - no need to wait for 3D idle after submitting vertices 21287 - fix PURGE_CACHE() and PURGE_ZCACHE() for r3xx+ 21288 - fix depth 16 with EXA composite 21289 21290commit 0a96173cc38e506728d4c3f2dd383ba56e856578 21291Author: Michel Dänzer <michel@tungstengraphics.com> 21292Date: Mon Apr 7 18:15:34 2008 +0200 21293 21294 Increase default CP timeout. 21295 21296 Helps avoid spurious timeouts causing problems, see 21297 http://bugs.freedesktop.org/show_bug.cgi?id=15203 . 21298 21299commit 255fbf465f5e7db2609a5a151bfa810249db52a0 21300Author: Owen W. Taylor <otaylor@fishsoup.net> 21301Date: Thu Apr 3 02:25:41 2008 -0400 21302 21303 Fix rendering of transformed sources for REPEAT_NONE with EXA on >= R300. 21304 21305 Use the border color when possible, otherwise fall back to software. 21306 21307commit bc0407e53237d7968808110bc0243076377acf6e 21308Author: Alex Deucher <alex@cube.(none)> 21309Date: Fri Apr 4 18:40:16 2008 -0400 21310 21311 ATOMBIOS: Add support for DynamicClocks option 21312 21313 This patch adds support for dynamic clock gating and static 21314 power management using the atom command tables. In some cases 21315 the bios may already set this up during post, so YMMV. 21316 21317 I was only able to test on desktop cards, so I haven't tested 21318 to see how much (if any) power this saves or how it affects the 21319 thermal footprint. 21320 21321commit 5f5e21bb50555c56bd371576074c28c929307ff1 21322Author: Alex Deucher <alex@cube.(none)> 21323Date: Fri Apr 4 14:29:45 2008 -0400 21324 21325 RADEON: warning fixes 21326 21327commit c8e9a973aaded24aad567a0e36d0c78a05d6b2fd 21328Author: Alex Deucher <alex@cube.(none)> 21329Date: Fri Apr 4 14:26:19 2008 -0400 21330 21331 RADEON: add some quirks 21332 21333commit 091963a635b79884afe77c026eabb48972fbe175 21334Author: Alex Deucher <alex@botch2.com> 21335Date: Thu Apr 3 22:35:16 2008 -0400 21336 21337 Minor cleanup 21338 21339commit 950e9860643c20acde0eca4e4ff26baacc1f2b69 21340Author: Alex Deucher <alex@botch2.com> 21341Date: Thu Apr 3 22:11:48 2008 -0400 21342 21343 Revert "RADEON: memmap rework 1" 21344 21345 This reverts commit dd8ee1b444f4b973a1e0fadca5f943f2162b5e94. 21346 21347 Conflicts: 21348 21349 src/radeon.h 21350 src/radeon_driver.c 21351 21352 This rework seems to have caused more trouble than it was worth. 21353 21354commit 88a1fe4a94c5d11aff22734b21c89890e4428cd5 21355Author: Alex Deucher <alex@botch2.com> 21356Date: Thu Apr 3 22:04:43 2008 -0400 21357 21358 Revert "RADEON: remove driver rec copies of mc info, use save rec directly" 21359 21360 This reverts commit be0858a84fbdf74c0b844f462933a221d48c707d. 21361 21362 Conflicts: 21363 21364 src/radeon_driver.c 21365 21366commit c40a7aa3989576a8144213e2f31b892d21df8686 21367Author: Owen W. Taylor <otaylor@fishsoup.net> 21368Date: Thu Apr 3 14:43:55 2008 -0400 21369 21370 R3xx/R5xx: Fix pitch and clamp mode for repeating textures 21371 21372 - We can always use TXPITCH on a R300 even when repeating, 21373 (previous check for pitch matching width was also wrong) 21374 - Fix clamp mode for repeating textures to be WRAP 21375 21376commit a8593482c1f2e0f2dbac06c2e5325ba8c83ed9ff 21377Author: Dave Airlie <airlied@redhat.com> 21378Date: Wed Apr 2 09:58:05 2008 +1000 21379 21380 atombios: fix the dual-head hopefully. 21381 21382 tested on r600 with DVI and VGA 21383 21384commit 61d883d116fab3e9b513432d65e705afc5bb39f1 21385Author: Dave Airlie <airlied@redhat.com> 21386Date: Wed Apr 2 09:57:38 2008 +1000 21387 21388 Revert "Revert "atombios: fixup the width/height to use the mode values not the scrn ones"" 21389 21390 This reverts commit fc9af578997b6f22ee8b17e83f37d98689291b0e. 21391 21392 I see your revert and raise you one... 21393 21394commit fc9af578997b6f22ee8b17e83f37d98689291b0e 21395Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 21396Date: Tue Apr 1 09:25:45 2008 -0400 21397 21398 Revert "atombios: fixup the width/height to use the mode values not the scrn ones" 21399 21400 This reverts commit c2b1c8b706a6c7c1fd0af80091958473133d54e7. 21401 21402 These registers hold surface size. Using the mode values 21403 breaks dualhead. 21404 21405commit 959509dd54de053f526b534e379a46934127231f 21406Author: Dave Airlie <airlied@redhat.com> 21407Date: Mon Mar 31 14:29:44 2008 +1000 21408 21409 radeon: use correct DDC interfaces so quirks get applied 21410 21411 Radeon seemed to mess up applying certain quirks, hopefully this will fix it. 21412 21413commit 18f5f1cd2f52afed89fc11ade0920f3dfea87306 21414Author: Dave Airlie <airlied@redhat.com> 21415Date: Mon Mar 31 14:11:49 2008 +1000 21416 21417 radeon: split quirks into separate function and new quirk for IBM RN50 21418 21419 Add a connector table quirk for the IBM RN50. 21420 21421commit c2b1c8b706a6c7c1fd0af80091958473133d54e7 21422Author: Dave Airlie <airlied@redhat.com> 21423Date: Sun Mar 30 11:44:14 2008 +1000 21424 21425 atombios: fixup the width/height to use the mode values not the scrn ones 21426 21427 this fixes it properly, legacy appears to be okay. 21428 21429commit c5edea3d8c9254d3a21e390b8309e39e4c9635db 21430Author: Dave Airlie <airlied@redhat.com> 21431Date: Sun Mar 30 11:11:22 2008 +1000 21432 21433 r500/r600: fix rotation to fill screen 21434 21435 I'm not 100% sure this is the correct fix (maybe we shouldn't be using scrn 21436 virtualX/Y)... this will fix it for now until I get more time. 21437 21438commit 9c62c820ba45ebc14d5f36f5d7885863800b6adb 21439Author: Michel Dänzer <michel@tungstengraphics.com> 21440Date: Fri Mar 28 12:37:29 2008 +0100 21441 21442 Include config.h, so FGL_LINUX can actually be defined when it's tested... 21443 21444commit a00d9260a85b94a522c442aee24bc5ea4dc31c5c 21445Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 21446Date: Thu Mar 27 20:03:13 2008 -0400 21447 21448 RADEON: fix lid issues on AVIVO chips for real this time :) 21449 21450commit f0e89c09074b2c7e641f73692bb39b0bf68eb49c 21451Author: Alex Deucher <alex@botch2.com> 21452Date: Thu Mar 27 19:15:18 2008 -0400 21453 21454 Revert "RADEON: attempt to fix lid issues" 21455 21456 This reverts commit 9b4473c1d830b88866dd22e8174a07195bd6fcf4. 21457 This doesn't help. 21458 21459commit 1442d396b938049b83f009a78ddabe2bf85641b6 21460Author: Dave Airlie <airlied@redhat.com> 21461Date: Thu Mar 27 14:02:51 2008 +1000 21462 21463 radeon: size bios to max of bar vs 64k. 21464 21465 reported by dwmw2: rhbz 438299 21466 21467commit de2f609ff0004ef8b74727bfebc2c74fb91205ea 21468Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 21469Date: Wed Mar 26 18:35:21 2008 -0400 21470 21471 AVIVO: no need to call PreinitXv() on AVIVO chips as they have no overlay 21472 21473commit 75884c257bc2bcfa5b498a77d4c403f09face036 21474Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 21475Date: Wed Mar 26 18:16:47 2008 -0400 21476 21477 XAA: update message about render so as to not confuse users 21478 21479commit 9b4473c1d830b88866dd22e8174a07195bd6fcf4 21480Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 21481Date: Wed Mar 26 18:01:29 2008 -0400 21482 21483 RADEON: attempt to fix lid issues 21484 21485 On some laptops the bios attempts to re-program the chip 21486 when a lid event comes in. This should hopefully prevent 21487 the bios from doing that. 21488 21489commit 8b144830fe9b4a0cee4745023de5e7d387070f60 21490Author: Alex Deucher <alex@samba.(none)> 21491Date: Tue Mar 25 01:15:05 2008 -0400 21492 21493 RV250: disable textured video due to HW bug 21494 21495 The YUV->RGB conversion in the texture engine is broken 21496 on RV250 so the colors come out wrong. 21497 21498commit 1789f11ab91633d3928f8b71988d51ff44bda9d1 21499Author: Alex Deucher <alex@samba.(none)> 21500Date: Mon Mar 24 19:03:30 2008 -0400 21501 21502 R3xx/R5xx: flush PVS state before enabling pvs-bypass 21503 21504commit 305a3310963a5dd07b3495015b06aa8c7c4e6b02 21505Author: Alex Deucher <alex@samba.(none)> 21506Date: Mon Mar 24 14:25:03 2008 -0400 21507 21508 R3xx/R5xx: move more VAP, etc. state setup into common init3d() function 21509 21510 Also some minor code cleanups 21511 21512commit 399b1d405e602c62d6deebea6d7e1f38886cd8e2 21513Author: Alex Deucher <alex@samba.(none)> 21514Date: Mon Mar 24 13:04:57 2008 -0400 21515 21516 R3xx/R5xx: use non VAP/TCP for textured video 21517 21518 Just extra state to emit. 21519 21520commit cd77ec18f32a7b36acb655c927bbfd7044019f97 21521Author: Dave Airlie <airlied@redhat.com> 21522Date: Mon Mar 24 18:42:21 2008 +1000 21523 21524 r300: don't bother with VAP/TCL for render. 21525 21526 We just send more data to the card to process per transaction, without getting 21527 any actual gains, as we already pre-compute the vertices without needing 21528 any clipping or transforms from the card. 21529 21530 Perhaps some stuff could be done on-card, but so far the code is a lot 21531 faster if we avoid sending this extra info. 21532 21533 pre: 150000 glyphs/sec 21534 post: 185000 glyphs/sec 21535 21536commit 301c6739b88676a0c78fc72194e993f894b8dc28 21537Author: Alex Deucher <alex@botch2.com> 21538Date: Sun Mar 23 11:14:02 2008 -0400 21539 21540 RS4xx: Revert back to previous fifo settings for now 21541 21542 Setup of these registers needs more investigation. 21543 21544commit 9bea60b3eb378de5e1d44cc02a2763f4feae7882 21545Author: Alex Deucher <alex@botch2.com> 21546Date: Sat Mar 22 11:46:15 2008 -0400 21547 21548 RS4xx: more work on disp/disp2 fifo setup 21549 21550commit 90f11c3986c28daa7b600b9662da145af325d264 21551Author: Alex Deucher <alex@botch2.com> 21552Date: Sat Mar 22 11:29:51 2008 -0400 21553 21554 RS4xx: missed this on the last commit. 21555 21556commit 6d5066a451017a2683addc9e2496987626795dda 21557Author: Alex Deucher <alex@samba.(none)> 21558Date: Fri Mar 21 16:21:54 2008 -0400 21559 21560 RS4xx: attempt to set up disp/disp2 fifos correctly 21561 21562 If you have an XPRESS chip, please test!!! 21563 21564commit fb1cffac05ae20c8365b25a2042b0ae961880faf 21565Author: Alex Deucher <alex@samba.(none)> 21566Date: Fri Mar 21 15:24:36 2008 -0400 21567 21568 RS4xx: attempt to fix TMDS/DVO support 21569 21570 XPRESS chips added a second set of FP control registers. 21571 I don't have the hw to test however. 21572 21573commit 5e3b21284482df9974c9a58f248f0100def2bb0c 21574Author: Alex Deucher <alex@samba.(none)> 21575Date: Wed Mar 19 19:15:05 2008 -0400 21576 21577 Disable the setting of HARDWARE_CURSOR_BIT_ORDER_MSBFIRST 21578 21579 See bug 11796 21580 21581commit 17cd42ed31814ba329a6a68edd0d75390a7da40e 21582Author: Matt Turner <mattst88@gmail.com> 21583Date: Wed Mar 19 18:17:10 2008 -0400 21584 21585 Enable BSR in Log2 functions 21586 21587 This patch edits RADEONLog2 and ATILog2 to use the x86 BSR instruction instead 21588 of looping through bits. It should provide a somewhat of a speed increase in 21589 this function on x86 and AMD64 architectures. 21590 21591 Note: the BSR instruction was added with the 80386 CPU and is therefore not 21592 compatible with earlier CPUs, though I highly doubt it's even possible to use a 21593 286 in conjunction with a Radeon. 21594 21595 The inline assembly also works with Intel's compiler (icc). 21596 21597commit c83827b4d2b6f03c54429e757a756eb99ff8be6b 21598Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br> 21599Date: Wed Mar 19 17:58:34 2008 -0400 21600 21601 [PATCH] Compile warning fixes. 21602 21603 Minor changes to avoid declarations mixed with code. 21604 Ansified functions with empty prototype to specify they don't 21605 receive arguments. 21606 Added some prototypes to radeon.h, and major reorder on radeon.h 21607 adding prototypes in alphabetical order and specifying to file that 21608 defines it. 21609 21610commit bed9754ad21d6c0a7f61067b04ba31c430a7cecb 21611Merge: 55e446b5 f71ac0e4 21612Author: Alex Deucher <alex@samba.(none)> 21613Date: Wed Mar 19 16:06:41 2008 -0400 21614 21615 Merge branch 'master' of ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati into r3xx-render 21616 21617commit 55e446b5bc091e6c7b3c2e9ae20b45130555c246 21618Author: Alex Deucher <alex@samba.(none)> 21619Date: Wed Mar 19 13:15:32 2008 -0400 21620 21621 R3xx/R5xx: Make sure to clamp the output of the FS 21622 21623commit b6aa4279cbe68cc8e4523795e9714fb798b62d98 21624Author: Alex Deucher <alex@samba.(none)> 21625Date: Wed Mar 19 12:45:01 2008 -0400 21626 21627 R5xx: bump textured video limits to 4096 21628 21629commit 4a445a3e8c4c5ecd9d4ef8daa26906c3ceaa94a1 21630Author: Alex Deucher <alex@samba.(none)> 21631Date: Wed Mar 19 12:31:51 2008 -0400 21632 21633 RADEON: add new macros to distinguish between R3xx and R5xx 3D 21634 21635commit 85d0c9e8d22ccc72bec87b3fd44da5d7609293e0 21636Author: Alex Deucher <alex@samba.(none)> 21637Date: Wed Mar 19 12:07:33 2008 -0400 21638 21639 RADEON: fixed textured video with XAA and tiling 21640 21641commit f5951db7b3522e0fe6af7f46a170c9c9a60a9bff 21642Author: Alex Deucher <alex@samba.(none)> 21643Date: Wed Mar 19 12:01:50 2008 -0400 21644 21645 RV515: fix textured video and EXA Composite 21646 21647 There seems to be an issue with the PVS setup on RV515, but 21648 bypassing it seems to work fine. 21649 21650commit 13573879fe56368ad06234712b677c23fabc56c6 21651Author: Dave Airlie <airlied@redhat.com> 21652Date: Wed Mar 19 15:06:47 2008 +1000 21653 21654 r500: make it work from startup. 21655 21656 I'm not sure why this worked or what is going wrong here, really the 21657 VAP internal architecture escapes me :) 21658 21659commit d331dd64d644a18ec99a2136cd0943b5edca1f03 21660Author: Alex Deucher <alex@samba.(none)> 21661Date: Tue Mar 18 19:44:26 2008 -0400 21662 21663 R3xx/R5xx: remove extra return after last commit 21664 21665commit bc34df7a9c35cdd38c49d5c22471f3f487a33d6e 21666Author: Alex Deucher <alex@samba.(none)> 21667Date: Tue Mar 18 19:39:47 2008 -0400 21668 21669 R3xx/R5xx: switch an ErrorF() to RADEONFALLBACK() 21670 21671commit 6f03f8fe0ecf4181dcf125049cf63bece0451fb2 21672Author: Alex Deucher <alex@samba.(none)> 21673Date: Tue Mar 18 19:36:05 2008 -0400 21674 21675 R3xx: we only use 2 temps, not 3 21676 21677commit 8bb71ab4a3eb4fb6ef7f709e87c8df387cb70ee3 21678Author: Tilman Sauerbeck <tilman@code-monkey.de> 21679Date: Tue Mar 18 14:36:08 2008 -0400 21680 21681 R3xx/R5xx: fix up a8-src-something_with_colors 21682 21683commit c362591d9b496df30668543158e4de44de742dc3 21684Author: Alex Deucher <alex@samba.(none)> 21685Date: Tue Mar 18 11:15:17 2008 -0400 21686 21687 R3xx/R5xx: remove some cruft 21688 21689commit 89fe6d2c7d7471e6088558130f6e49f46c31dd47 21690Author: Dave Airlie <airlied@linux.ie> 21691Date: Tue Mar 18 09:43:43 2008 -0400 21692 21693 R5xx: fix typ in r5xx render accel 21694 21695 This gets render working on r5xx 21696 21697commit 79b40ebcd8dedfc83e484c1024beeeaccc6124f3 21698Author: Alex Deucher <alex@samba.(none)> 21699Date: Tue Mar 18 02:46:49 2008 -0400 21700 21701 R5xx: first pass at render support (untested) 21702 21703commit 71292c8f193230255d1d980c2e996bb01d04fab6 21704Author: Alex Deucher <alex@samba.(none)> 21705Date: Tue Mar 18 00:45:37 2008 -0400 21706 21707 R5xx: bump tex/dst limits to 4096 21708 21709commit 30b52f8aa6a471455284f59b5b27252743892b13 21710Author: Alex Deucher <alex@samba.(none)> 21711Date: Mon Mar 17 23:20:10 2008 -0400 21712 21713 R3xx/R5xx: whitespace cleanup and cruft removal 21714 21715commit 9c9f1b538ed710c3066775fba0a8e936b63087b1 21716Author: Alex Deucher <alex@samba.(none)> 21717Date: Mon Mar 17 23:01:37 2008 -0400 21718 21719 R3xx: get masks working and cleanup 21720 21721 RS offset was wrong for mask texture 21722 21723commit ef94febd74f8ee63081b61e42f093a5a2b8fbf1e 21724Author: Alex Deucher <alex@samba.(none)> 21725Date: Mon Mar 17 22:27:19 2008 -0400 21726 21727 R3xx: minor adjustments 21728 21729commit f71ac0e40b9d950bcb3bba42a75d41f45b6ed1bf 21730Author: Alban Browaeys <prahal@yahoo.com> 21731Date: Mon Mar 17 20:48:48 2008 -0400 21732 21733 RADEON: Revert to old behavior when resetting the memmap on VT switch 21734 21735 Not sure why this needs to be done twice. Should fix bug 14980 21736 Probably needs more investigation. 21737 21738commit bedbbf196dc97ee5142e7dfae16fb6f317fca5a7 21739Author: Alex Deucher <alex@samba.(none)> 21740Date: Mon Mar 17 20:16:25 2008 -0400 21741 21742 R3xx: some progress 21743 21744commit af0e626c132de2dd9958fec657fcc85d4c0fe5e1 21745Author: Alex Deucher <alex@samba.(none)> 21746Date: Mon Mar 17 18:07:12 2008 -0400 21747 21748 R3xx: fix errant w 21749 21750commit 29ea5bfc0eb3194e2454fc3ee863df54f0300880 21751Author: Alex Deucher <alex@cube.(none)> 21752Date: Mon Mar 17 16:41:57 2008 -0400 21753 21754 RADEON: fix typo in RADEONAdjustMemMapRegisters() 21755 21756commit ab317e85c5ab1a249a510c34aeb3a908be1a66fc 21757Author: Alex Deucher <alex@cube.(none)> 21758Date: Mon Mar 17 15:28:09 2008 -0400 21759 21760 RADEON: make sure var is initialized properly in RADEONAdjustMemMapRegisters() 21761 21762commit 208d307227e15f37a6af5194398ed23266ff743a 21763Author: Dave Airlie <airlied@linux.ie> 21764Date: Sun Mar 16 19:39:23 2008 +1000 21765 21766 radeon: the 0x5974 appears to be a mobility chip... 21767 21768 After debugging with partymola on #radeon, adding this allowed his 21769 Dell Vostro 1000 to work properly 21770 21771commit 9bc7c2ec4048e1677547c1d60c51ccb954f7589a 21772Author: Alex Deucher <alex@samba.(none)> 21773Date: Fri Mar 14 20:12:22 2008 -0400 21774 21775 R3xx: odds and ends... 21776 21777 still not working. 21778 - swizzle US output for BGR formats 21779 - no need to write to temps in ALU ops, 21780 write to output only 21781 - flush the PVS before updating 21782 21783commit 96bea7906c4706fcd57a9cd8f1ce3feab6ac676d 21784Author: Alex Deucher <alex@samba.(none)> 21785Date: Fri Mar 14 15:59:36 2008 -0400 21786 21787 R3xx: theoretical support for component alpha 21788 21789 masks are still broken so... 21790 21791commit cffe3dcc8991cd7c457a9c1a9f41055aa9ea3436 21792Author: Alex Deucher <alex@samba.(none)> 21793Date: Fri Mar 14 14:37:43 2008 -0400 21794 21795 R3xx: VS WIP 21796 21797commit b73f52a50dfd6ff8d92f04d6b510c39582c6ac67 21798Author: Alex Deucher <alex@samba.(none)> 21799Date: Fri Mar 14 14:20:49 2008 -0400 21800 21801 R3xx/R5xx: enable VS for mask texture 21802 21803commit 569a14ca9be1e18fe9921edc816ac3dc32d6cca7 21804Author: Alex Deucher <alex@samba.(none)> 21805Date: Fri Mar 14 13:32:12 2008 -0400 21806 21807 R3xx/R5xx: Fix magic numbers in vertex shaders 21808 21809commit 4878997529601d62e257aa1c9112bd460561de73 21810Author: Alex Deucher <alex@samba.(none)> 21811Date: Thu Mar 13 21:23:40 2008 -0400 21812 21813 R3xx: make sure to set the FS code size correctly 21814 21815commit 22f46b88ef05afb6a6b6d70007ac4980a446430e 21816Author: Alex Deucher <alex@samba.(none)> 21817Date: Thu Mar 13 20:25:33 2008 -0400 21818 21819 R3xx: attempt to setup the rasterizer properly for mask texture 21820 21821 Not working yet 21822 21823commit 081fc9e892fa3d2e07b7db65b2e2719646255463 21824Author: Alex Deucher <alex@samba.(none)> 21825Date: Thu Mar 13 18:38:26 2008 -0400 21826 21827 R3xx: more mask work 21828 21829commit 2bf0236c03538ace3ce6d0e68f0829fc47d1385b 21830Author: Alex Deucher <alex@samba.(none)> 21831Date: Thu Mar 13 18:32:00 2008 -0400 21832 21833 R3xx: enable composite for non-mask cases 21834 21835commit 74286ba41302107d2fc626fee2181f7c4bc18164 21836Author: Alex Deucher <alex@samba.(none)> 21837Date: Thu Mar 13 18:25:32 2008 -0400 21838 21839 R3xx: add basic mask support 21840 21841commit a2bbe10d866567911b68f222b4758624bfe9bf84 21842Author: Alex Deucher <alex@samba.(none)> 21843Date: Thu Mar 13 18:16:53 2008 -0400 21844 21845 R300: setup source selects and output swizzling 21846 21847commit b9974ecce7d1932595226004858b08a7a6b188dc 21848Author: Alex Deucher <alex@samba.(none)> 21849Date: Thu Mar 13 17:35:38 2008 -0400 21850 21851 R3xx: set the texture id and add some register info 21852 21853commit 0ef700b7da5e554a0d0d166f3fde85ff45c9eb1f 21854Author: Alex Deucher <alex@samba.(none)> 21855Date: Thu Mar 13 17:02:25 2008 -0400 21856 21857 R3xx/R5xx: enable blending 21858 21859commit b35c09a597c93a1d9f06ef0091c96822b0653f98 21860Author: Dave Airlie <airlied@redhat.com> 21861Date: Thu Mar 13 18:42:29 2008 +1000 21862 21863 xv: fixup XAA on r500 textured video 21864 21865 the XAA area should never end up tiled. This may break with nooffscreen pixmaps 21866 21867commit d4446461c3630caff166456c351ace34f57cc119 21868Author: Matt Turner <mattst88@gmail.com> 21869Date: Tue Mar 11 21:20:53 2008 -0400 21870 21871 Properly fix uninitialized variables warnings 21872 21873 According to commit 9fd13e6773371c82b9799a5bda7c96ffa5cafe8c to 21874 xf86-video-intel by Kristian Høgsberg, there is a better way to fix the 21875 possibly initialized variables warnings. This patch will use Kristian's fix. 21876 21877commit 20adfd7390d9b1f100e0c4a14f175377b8335c82 21878Author: Alex Deucher <alex@cube.(none)> 21879Date: Tue Mar 11 20:09:35 2008 -0400 21880 21881 RADEON: enable output attributes that require a modeset immediately 21882 21883 This should fix bug 14915 21884 21885commit 53ba7f5771b0b53fb0d3bc29d64bdd3813756d10 21886Author: Alex Deucher <alex@cube.(none)> 21887Date: Tue Mar 11 19:12:40 2008 -0400 21888 21889 RADEON: fix vblank interrupts after VT switch or suspend/resume 21890 21891commit e946c097f0438afbea6f3dd37ee39d67d415708c 21892Author: Matt Turner <mattst88@gmail.com> 21893Date: Tue Mar 11 19:07:58 2008 -0400 21894 21895 [PATCH] Fix a few warnings 21896 21897commit 8e160508520c0a24ca90aad182f296461ca0d9b6 21898Author: Alex Deucher <alex@cube.(none)> 21899Date: Tue Mar 11 18:11:13 2008 -0400 21900 21901 DCE3: add support for PCIEPHY (untested) 21902 21903commit fbded88a2925f9f049936dad0736721e7b84a6ee 21904Author: Alex Deucher <alex@cube.(none)> 21905Date: Tue Mar 11 14:10:31 2008 -0400 21906 21907 ATOM: remove some cruft 21908 21909commit 3263f6e4a410281d620c288a92bb4521f7b6fc06 21910Author: Alex Deucher <alex@cube.(none)> 21911Date: Tue Mar 11 14:05:48 2008 -0400 21912 21913 DCE3: enable DPMS on DIG ports 21914 21915commit eb90e235b58c94f3d4d75394725ab2fe246a42ff 21916Author: Alex Deucher <alex@cube.(none)> 21917Date: Tue Mar 11 13:53:54 2008 -0400 21918 21919 DCE3: adjust PLL for DCE3 chips 21920 21921 this fixes stability issues on digital outputs and certain modes. 21922 21923commit 552615ccc5360baafb8bb41698c1ca27816fd4b2 21924Author: Alex Deucher <alex@cube.(none)> 21925Date: Tue Mar 11 13:38:29 2008 -0400 21926 21927 ATOMBIOS: enable load detection by default on both DACs 21928 21929 Load detection is reliable with atom, so enable it by default 21930 on both DACA and DACB, rather than just DACA. 21931 21932commit 78b10487cf222c96f8944ba25e2ea970506b3535 21933Author: Alex Deucher <alex@cube.(none)> 21934Date: Tue Mar 11 13:16:00 2008 -0400 21935 21936 DCE3: add output attribute to enable/disable coherent mode 21937 21938 Enabled by default. The TMDS transmitter can be programmed 21939 slightly differently depending on the chips in the panel. If you 21940 have problems with tmds on a particular panel, try disabling it. 21941 21942commit d20be31c46fbec623af4c3628a7c603ceacf500f 21943Author: Alex Deucher <alex@botch2.(none)> 21944Date: Mon Mar 10 21:05:43 2008 -0400 21945 21946 RV550: MC setup is like RV515 not RV530 21947 21948commit 38606b08b68842fbcc81c233009c1117269f3be9 21949Author: Matthieu Herrb <matthieu@bluenote.herrb.net> 21950Date: Sat Mar 8 23:22:59 2008 +0100 21951 21952 Makefile.am: nuke RCS Id 21953 21954commit 9d710ee1a44cf2f3a948fbdbe17ef09521cbe744 21955Author: Alex Deucher <alex@cube.(none)> 21956Date: Fri Mar 7 15:09:14 2008 -0500 21957 21958 AVIVO: clean up some unused variables 21959 21960commit c28c30c9f3d7bfebfd56a5c982c96f0090982054 21961Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 21962Date: Fri Mar 7 14:10:49 2008 -0500 21963 21964 RADEON: Fix crash in last commit 21965 21966commit c3a3635865e380c784a226c8ead069d4716d6b75 21967Author: Dave Airlie <airlied@redhat.com> 21968Date: Thu Mar 6 20:17:45 2008 -0500 21969 21970 RADEON: fix tiling/interlaced interaction with randr 1.2 21971 21972commit df1b94dc4eb1f35b636dbf2ec0ab1c2da9937c0d 21973Author: Alex Deucher <alex@botch2.(none)> 21974Date: Thu Mar 6 19:22:08 2008 -0500 21975 21976 DCE3: Ignore outputs with DIN connectors for now 21977 21978commit cb2dc19387c7b6494c47c76d683cf38a48700768 21979Author: Alex Deucher <alex@cube.(none)> 21980Date: Thu Mar 6 18:33:12 2008 -0500 21981 21982 AVIVO: fix typo from a previous commit 21983 21984 Leave tv dpms hook disabled or you may get bad interactions 21985 with the shared DAC 21986 21987commit 77355de48057e5e7e0d5b3f3cf5a7a92220a53b1 21988Author: Alex Deucher <alex@cube.(none)> 21989Date: Thu Mar 6 17:46:00 2008 -0500 21990 21991 AVIVO: don't add outputs for invalid connectors 21992 21993commit 600dbe080997a01ceaf6be86723189d518bc1281 21994Merge: 594743a9 5b7875d0 21995Author: Alex Deucher <alex@cube.(none)> 21996Date: Thu Mar 6 17:31:37 2008 -0500 21997 21998 Merge branch 'master' of ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati 21999 22000commit 594743a99811a8b0f391412892414fcd158eeb56 22001Author: Alex Deucher <alex@cube.(none)> 22002Date: Thu Mar 6 17:30:21 2008 -0500 22003 22004 AVIVO: fix up memsize detection for IGP chips 22005 22006commit 5b7875d0cbfbdbcd1515c4e942d30de298b49dff 22007Author: Doug Chapman <doug.chapman@hp.com> 22008Date: Thu Mar 6 14:31:06 2008 -0500 22009 22010 Bug #14826: Fix a bogus check around xf86SetOperatingState. 22011 22012commit 651fe23f4c650ed91843dec48db24e18e8b91219 22013Merge: 3de2dc88 766f464d 22014Author: Adam Jackson <ajax@redhat.com> 22015Date: Thu Mar 6 14:28:27 2008 -0500 22016 22017 Merge branch 'master' of git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-ati 22018 22019commit 41171c25cd235bafad26bcbabced16ead4b8c54b 22020Author: Alex Deucher <alex@cube.(none)> 22021Date: Thu Mar 6 14:05:18 2008 -0500 22022 22023 DCE3.0: add support for crtc memreq table 22024 22025commit 766f464dfdfccadef23e4232f2bce5db22195513 22026Author: Alex Deucher <alex@cube.(none)> 22027Date: Thu Mar 6 13:35:43 2008 -0500 22028 22029 RADEON: take 2 on proper pragma pack support for bsds 22030 22031 See bug 14594. Based on suggestion by Henry Zhao 22032 22033commit a842ce9ca6494e724a7828cead9b61c9ef02b6aa 22034Author: Alex Deucher <alex@cube.(none)> 22035Date: Thu Mar 6 12:32:18 2008 -0500 22036 22037 DCE3.0: Minor fixups 22038 22039commit 8a1ba374033591c725a78923aa30829e4de2a5ae 22040Author: Alex Deucher <alex@cube.(none)> 22041Date: Thu Mar 6 09:53:51 2008 -0500 22042 22043 RADEON: option to override TVDAC adj values from bios with driver defaults 22044 22045 If you have a washed out image on the tv dac, try this option. 22046 Option "DefaultTVDACAdj" "TRUE" 22047 22048commit 0ed48f8f651a28e189f9fee8c6b593da0178d21c 22049Author: Alex Deucher <alex@cube.(none)> 22050Date: Wed Mar 5 18:41:01 2008 -0500 22051 22052 AVIVO: Initial support for DCE 3.0 using atombios 22053 22054 DACs are working well, DIG support (DVI, HDMI, LVDS, etc.) 22055 still has some issues. 22056 22057commit 2901e99f1942842856cd39c1dcc8b22f3cf7d9e3 22058Author: Alex Deucher <alex@botch2.(none)> 22059Date: Wed Mar 5 10:40:06 2008 -0500 22060 22061 RADEON: fix fetching of dac2 adj values from newer bios tables 22062 22063commit 74eb981287d76836327830bd51272f605a07e0cc 22064Author: Alex Deucher <alex@botch2.(none)> 22065Date: Mon Mar 3 12:02:44 2008 -0500 22066 22067 ATOMBIOS: fix atombios parser support on *bsd 22068 22069 bsd requires a different pragma pack than Linux. 22070 See bug 14594. 22071 22072commit f7769ea86e265f347eb58c517ccb5ef8b35eec27 22073Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br> 22074Date: Sun Mar 2 14:49:21 2008 -0500 22075 22076 [PATCH] Ensure symbols used by other modules are visible. 22077 22078 The xf86-video-ati drivers are one of the cases where LoaderSymbol is 22079 widely used in some obscure ways. This patch fixes the problem, and 22080 allows compiling with -fvisibility=hidden. 22081 22082commit a4398ac3ad77216f2c8aa628425bef5f2912a0a9 22083Author: Alex Deucher <alex@cube.(none)> 22084Date: Sat Mar 1 18:52:26 2008 -0500 22085 22086 RS6xx: change isIGP checks to CHIP_FAMILY_RS690 22087 22088 these paths are only relevant on RS6xx chips 22089 22090commit 67d4d04836c05293b844bc505f303cfb04c0f8a4 22091Author: Alex Deucher <alex@cube.(none)> 22092Date: Sat Mar 1 18:33:18 2008 -0500 22093 22094 RADEON: use xf86SetDesiredModes() in screeninit and enterVT 22095 22096 this should restore the proper output state on VT switches 22097 22098commit be0858a84fbdf74c0b844f462933a221d48c707d 22099Author: Maciej Cencora <m.cencora@gmail.com> 22100Date: Sat Mar 1 18:11:58 2008 -0500 22101 22102 RADEON: remove driver rec copies of mc info, use save rec directly 22103 22104 info->mc_* were used and the immediately copied into info->ModeReg 22105 ones. Just use the ModeReg copies directly. 22106 22107commit dd8ee1b444f4b973a1e0fadca5f943f2162b5e94 22108Author: Alex Deucher <alex@cube.(none)> 22109Date: Sat Mar 1 16:23:51 2008 -0500 22110 22111 RADEON: memmap rework 1 22112 22113 Don't restore memmap regs on every mode switch. 22114 Just do memmap save/restore/setup on server start and VT switch. 22115 22116commit 1f6a23000001f3d1c21b5c04f94714a8caa7aa8b 22117Author: Alex Deucher <alex@cube.(none)> 22118Date: Sat Mar 1 15:53:42 2008 -0500 22119 22120 RADEON: only restore legacy dac regs on legacy radeons 22121 22122commit dee6cef8e62d0651c00319e03eea92940fd24aa4 22123Author: Alex Deucher <alex@botch2.(none)> 22124Date: Sat Mar 1 14:39:32 2008 -0500 22125 22126 RS4xx: enable exa render accel and textured video 22127 22128 RS6xx paths seem to work fine on RS4xx 22129 22130commit 129f737efe4e8d1a368e7db4b063bdcd9339cb09 22131Author: Alex Deucher <alex@cube.(none)> 22132Date: Sat Mar 1 14:32:30 2008 -0500 22133 22134 AVIVO: save/restore regs by block 22135 22136 Save/Restore the entire block for each output. 22137 This should fix VT switch problems. 22138 22139commit b069aadaa63a95d7a71b5cfbab83577b49501094 22140Author: Alex Deucher <alex@botch2.(none)> 22141Date: Fri Feb 29 22:36:02 2008 -0500 22142 22143 AVIVO: LVDS panels need dithering enabled 22144 22145 Fixes bug 14760 22146 22147commit fe87bdee815372b4b4d7d4c705e34681625b90f2 22148Author: Alex Deucher <alex@botch2.(none)> 22149Date: Fri Feb 29 13:10:13 2008 -0500 22150 22151 AVIVO: disable pageflipping on avivo chips until we have proper drm support 22152 22153commit fb3678c7f511d539a51cd090cb8b5041d7d2ba26 22154Author: Alex Deucher <alex@botch2.(none)> 22155Date: Fri Feb 29 13:01:21 2008 -0500 22156 22157 R5xx: fix register count when sending fragment program for textured video 22158 22159commit a66d37d1a896ec934989592457c2beff8e6f1639 22160Author: Alex Deucher <alex@botch2.(none)> 22161Date: Fri Feb 29 04:07:05 2008 -0500 22162 22163 fix off-by-one in last commit 22164 22165commit e56062960be0c8d3947861dd5e0691fce6516b99 22166Author: Alex Deucher <alex@cube.(none)> 22167Date: Thu Feb 28 19:16:39 2008 -0500 22168 22169 AVIVO: save/restore scaler regs 22170 22171commit ae1c39a9b3e666404d0931679c9078c2e125a8bc 22172Author: Alex Deucher <alex@cube.(none)> 22173Date: Thu Feb 28 18:53:55 2008 -0500 22174 22175 RS6xx: rework output parsing 22176 22177 Turns out it's not as complex as I originially thought. 22178 IGP chips just have non-standard GPIO entires for DDC. 22179 22180commit d8d6c9fe4ae7e1ab67dd041a251e901d97c29ed6 22181Author: Alex Deucher <alex@botch2.(none)> 22182Date: Thu Feb 28 17:01:14 2008 -0500 22183 22184 RS6xx: fix typos in previous commit 22185 22186 Noted by Maciej Cencora on IRC 22187 22188commit 46547ae8bdbc5c10f1fd028b95ec4c5c31a5b318 22189Author: Alex Deucher <alex@cube.(none)> 22190Date: Thu Feb 28 14:29:30 2008 -0500 22191 22192 AVIVO: disable dithering on DFPs 22193 22194 This should fix the color banding some people have noticed. 22195 Also save/restore DDIA regs on RS6xx 22196 22197commit 72a53d6f20ac29b3baddb7d8af04f19b76d2e04f 22198Author: Michel Dänzer <michel@tungstengraphics.com> 22199Date: Thu Feb 28 17:38:04 2008 +0100 22200 22201 Handle EXA coordinate limits more cleverly. 22202 22203 Generally set the 2D engine limits, and only enforce the 3D engine limits in the 22204 CheckComposite hook. This should still prevent useless migration of pixmaps the 22205 3D engine can't handle but allows for basic acceleration of bigger ones. 22206 22207 Fixes http://bugs.freedesktop.org/show_bug.cgi?id=14708 . 22208 22209commit 5249f450a2487475a95531603cc8668db2c21c33 22210Author: Michel Dänzer <michel@tungstengraphics.com> 22211Date: Thu Feb 28 12:23:58 2008 +0100 22212 22213 Fix 16 bit packed YUV XVideo playback on big endian systems with DRI disabled. 22214 22215 http://bugs.freedesktop.org/show_bug.cgi?id=14668 22216 22217commit e40d75fd8b2aece9dae8076fac822a4a83025fb2 22218Author: Alex Deucher <alex@samba.(none)> 22219Date: Wed Feb 27 22:53:10 2008 -0500 22220 22221 R500: fragment program clean up and magic number conversion 22222 22223commit 140dadba36b2191f0e18e41dd987785abd5f55d2 22224Author: Alex Deucher <alex@samba.(none)> 22225Date: Wed Feb 27 22:21:12 2008 -0500 22226 22227 R300: fix up magic numbers in fragment program 22228 22229commit e521476bb5e2dfabc93747e43eb911a8a101357e 22230Author: Alex Deucher <alex@botch2.(none)> 22231Date: Wed Feb 27 21:26:55 2008 -0500 22232 22233 R300/R400: bump up the clip limits for textured video 22234 22235 This allows up to 2560x2560 (hw limit) 22236 22237commit 10db46f11d7e1c055c9ad6034c65ad163dad17dc 22238Author: Alex Deucher <alex@samba.(none)> 22239Date: Wed Feb 27 15:28:50 2008 -0500 22240 22241 AVIVO: make sure we select the right LUT for each crtc 22242 22243commit ea944f38dcfd871b27345698afea1cb986ecb049 22244Author: Alex Deucher <alex@samba.(none)> 22245Date: Wed Feb 27 14:37:52 2008 -0500 22246 22247 R300+: update RADEONCP_REFRESH() to reflect new location of scissor regs 22248 22249commit b865faf95666e2172c3eec143f77fe9c524e4983 22250Author: Alex Deucher <alex@samba.(none)> 22251Date: Wed Feb 27 14:05:44 2008 -0500 22252 22253 R100/R200: move r100/r200 specific 3D setup into appropriate blocks 22254 22255 R3xx+ doesn't have these regs. 22256 22257commit ee485ee0fc550414cd39cdb8fd2769fe0bc996a8 22258Author: George Sapountzis <gsap7@yahoo.gr> 22259Date: Wed Feb 27 18:41:31 2008 +0200 22260 22261 radeon: pass distcheck 22262 22263commit ff110c781299fb729e93572ba3d630ca59e8127d 22264Author: George Sapountzis <gsap7@yahoo.gr> 22265Date: Wed Feb 27 18:41:17 2008 +0200 22266 22267 add warning for r128 22268 22269commit b4a344f36332a12a51c9e896d93e9cb0672d31ef 22270Author: George Sapountzis <gsap7@yahoo.gr> 22271Date: Wed Feb 27 18:40:50 2008 +0200 22272 22273 add warning for mach64 22274 22275commit 81ce4097c12ab64a52d0992f981f588f5a297b7b 22276Author: George Sapountzis <gsap7@yahoo.gr> 22277Date: Wed Feb 27 18:40:23 2008 +0200 22278 22279 drop r128 22280 22281commit 873ff2033f1d0b84c91c48e829aa945f41b490b7 22282Author: George Sapountzis <gsap7@yahoo.gr> 22283Date: Wed Feb 27 18:40:09 2008 +0200 22284 22285 drop r128 (build system) 22286 22287commit 2b8b2676e7b2fd2952f5307a934ef0a11006f838 22288Author: George Sapountzis <gsap7@yahoo.gr> 22289Date: Wed Feb 27 18:39:49 2008 +0200 22290 22291 drop mach64 22292 22293commit 05bbbec11a04e7f6321b92c9dbec4ad4c0d130ff 22294Author: George Sapountzis <gsap7@yahoo.gr> 22295Date: Wed Feb 27 18:39:32 2008 +0200 22296 22297 drop mach64 (build system) 22298 22299commit 68888189cf8d460ef6f8f2f1431a6ffe9fcd8134 22300Author: Alex Deucher <alex@samba.(none)> 22301Date: Wed Feb 27 03:09:28 2008 -0500 22302 22303 RADEON: fix textured video uploads 22304 22305commit 7cfad216a3f1486be194c972b9b6bcf070358586 22306Author: Alex Deucher <alex@samba.(none)> 22307Date: Wed Feb 27 02:13:10 2008 -0500 22308 22309 RADEON: fix clipping for textured video 22310 22311commit a2dca1d68d751def34ef3c6f836574173737bf76 22312Author: Alex Deucher <alex@samba.(none)> 22313Date: Tue Feb 26 16:38:29 2008 -0500 22314 22315 RADEON: Convert textured video to use pipelined uploads 22316 22317commit b4fa1ce9d2da04d94521a82d3c2e95f0fe985ccc 22318Author: Alex Deucher <alex@samba.(none)> 22319Date: Tue Feb 26 14:45:25 2008 -0500 22320 22321 R300/R500: fix up clipping 22322 22323 Based on patch from Peter Zubaj. 22324 22325commit 44e527a117ab0a363135ff066c7f7e0c12e3dc89 22326Author: Alex Deucher <alex@samba.(none)> 22327Date: Tue Feb 26 14:33:24 2008 -0500 22328 22329 R300: fix cordinate clamping in render code 22330 22331 Based on Peter's fix for textured video 22332 22333commit 00ec17ad53d7ad43f19c9b723794ac1b8ef86826 22334Author: Peter Zubaj <pzubaj at marticonet.sk> 22335Date: Tue Feb 26 14:26:14 2008 -0500 22336 22337 R300: Fix clamping and pitch for textured video 22338 22339commit 4207793f9b6c891cb17ba715223d2002668885e3 22340Author: Alex Deucher <alex@samba.(none)> 22341Date: Tue Feb 26 13:16:22 2008 -0500 22342 22343 R300/R500: set the number graphics pipes properly 22344 22345 This should fix the checkerboard issues on r300/r350 cards. 22346 22347commit 823d8bf49e4c0a8c839354179942c5a1944c178d 22348Author: Alex Deucher <alex@samba.(none)> 22349Date: Tue Feb 26 12:25:14 2008 -0500 22350 22351 R300/R500: clean up magic numbers in textured video 22352 22353commit 8ea75b268f11794f4a7e7bac52cb256490ed3fd2 22354Author: Dave Airlie <airlied@linux.ie> 22355Date: Tue Feb 26 16:29:19 2008 +1000 22356 22357 regs: fix spelling properly 22358 22359commit f2816064a6c2c4c35ccba74b9aa80547e25c012e 22360Author: Dave Airlie <airlied@linux.ie> 22361Date: Tue Feb 26 16:28:24 2008 +1000 22362 22363 regs: fix spelling mistake 22364 22365 pointed out by plaes on irc 22366 22367commit 9d2ca30b90607085578dde1f314db663bd5f82ec 22368Author: Alex Deucher <alex@samba.(none)> 22369Date: Mon Feb 25 17:34:00 2008 -0500 22370 22371 R300/R500: clean up magic numbers in render code 22372 22373commit 153ad6fcf704cbf9f811d9986cd4baf04e82c9d2 22374Author: Dave Airlie <airlied@linux.ie> 22375Date: Mon Feb 25 07:10:48 2008 +1000 22376 22377 fixup check for EXA composite pointed out by Alan Swanson 22378 22379commit 85e470e64f629de72e361c77770e2e29998d1bf4 22380Merge: 27ddb39b 1b84c76f 22381Author: Alex Deucher <alex@samba.(none)> 22382Date: Sun Feb 24 05:37:22 2008 -0500 22383 22384 Merge master and fix conflicts 22385 22386 Merge branch 'master' of ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati 22387 22388 Conflicts: 22389 22390 src/radeon_commonfuncs.c 22391 22392commit 27ddb39b12a0b54e099fd5274c4c91f08e2d2822 22393Author: Alex Deucher <alex@samba.(none)> 22394Date: Sun Feb 24 05:30:11 2008 -0500 22395 22396 R300: clean up magic numbers in RADEONInit3DEngine 22397 22398commit d4c20f33ad6a1f88615cd7e09ad3638896873f9e 22399Author: Alex Deucher <alex@samba.(none)> 22400Date: Sun Feb 24 04:46:10 2008 -0500 22401 22402 R300: replace magic numbers in cache flush 22403 22404commit e52f1c8d2647b81d891ec0728dd582941a76c83f 22405Author: Alex Deucher <alex@samba.(none)> 22406Date: Sun Feb 24 04:43:18 2008 -0500 22407 22408 R300: fill in some more 3D bitfields 22409 22410commit 1b84c76f27c8d24cb42beae26abf000721901c1c 22411Author: Dave Airlie <airlied@redhat.com> 22412Date: Sun Feb 24 19:20:36 2008 +1100 22413 22414 rs690: initial textured video support 22415 22416commit 68158124366db883a901e960fe5f05f8df5efa42 22417Author: Dave Airlie <airlied@redhat.com> 22418Date: Tue Feb 19 19:51:18 2008 +1100 22419 22420 rs690: initial rs690 rotate 22421 22422 suffers same problem as r500 with clipping 22423 22424commit e614bb6965588bf09dcb87f5e08e67120ec9847f 22425Author: Dave Airlie <airlied@redhat.com> 22426Date: Sun Feb 24 01:46:05 2008 -0500 22427 22428 r500: convert fragprog to use register values 22429 22430commit 6ce9ee47c75620b2e5d211c5d59d17271a6a7b19 22431Author: Dave Airlie <airlied@redhat.com> 22432Date: Sat Feb 23 22:49:34 2008 -0500 22433 22434 r500: add textured video Xv adapter support 22435 22436commit 9aaf8b33b22b6ba112869558ae54e021b9487ad2 22437Author: Dave Airlie <airlied@redhat.com> 22438Date: Sat Feb 23 22:16:25 2008 -0500 22439 22440 r500: initial rotate support - not fully working yet. 22441 22442 Just an example of how to setup and run the r500 3D engine for rotation. 22443 this rotates for me but I get some strange clipping on the bottom of my screen 22444 22445commit 05dc3e4fc19d056ce99a7b110665adab2ca1ea21 22446Author: Adam Jackson <ajax@redhat.com> 22447Date: Sat Feb 23 20:29:51 2008 -0500 22448 22449 Clarify R500 US setup. 22450 22451commit 080606ad528972623e1ed1124d8efe7705a73446 22452Author: Adam Jackson <ajax@redhat.com> 22453Date: Sat Feb 23 20:21:17 2008 -0500 22454 22455 Add the R500 US index/data pair. 22456 22457commit edfb3b6bbf0ee17ace8e6ba704a6f54e249fec63 22458Author: Alex Deucher <alex@samba.(none)> 22459Date: Sat Feb 23 19:59:33 2008 -0500 22460 22461 RADEON: no textured video yet on XPRESS chips 22462 22463 Still need to sort out the VAP and PVS stuff 22464 22465commit 4146bfe5d00e40a86d17826fac50d04b2469621d 22466Author: Alex Deucher <alex@samba.(none)> 22467Date: Sat Feb 23 19:21:52 2008 -0500 22468 22469 R500: fix typo in new r5xx regs 22470 22471commit d9be9f34b0d3313e7b22b2a8bb0a8924ad3116bf 22472Author: Alex Deucher <alex@samba.(none)> 22473Date: Sat Feb 23 19:06:30 2008 -0500 22474 22475 RADEON: add textured video support for r1xx-r4xx radeons 22476 22477 Based on the kdrive ati video code by Eric Anholt. 22478 R3xx/R4xx still have some clipping issues in certain situations 22479 22480commit 9dc4acad79196e9d5d94dd710773bfa83456d47f 22481Author: Alex Deucher <alex@cube.(none)> 22482Date: Sat Feb 23 18:29:00 2008 -0500 22483 22484 RS6xx: gpio entry for DDIA varies depending on the number of DFP ports 22485 22486commit ed0a93edf28155308e7ab9d8705581bb38455ea0 22487Author: Adam Jackson <ajax@redhat.com> 22488Date: Sat Feb 23 15:02:17 2008 -0500 22489 22490 Fix R500_US_CONFIG. 22491 22492commit bc2bd6f841b51aeed3b6b4a47dbe758c200bc5a6 22493Author: Adam Jackson <ajax@redhat.com> 22494Date: Sat Feb 23 14:34:18 2008 -0500 22495 22496 Add R500 unified shader register block. 22497 22498commit a7b5c3bb74fc4de5e38a75ac31656445ce823464 22499Author: Alex Deucher <alex@cube.(none)> 22500Date: Fri Feb 22 19:35:11 2008 -0500 22501 22502 RS6xx: fix DDC on DDIA output (usually HDMI port) 22503 22504commit 3327a681e21101cc6f6e162f4e29f9937b69ccc3 22505Author: Alex Deucher <alex@cube.(none)> 22506Date: Fri Feb 22 17:05:56 2008 -0500 22507 22508 ATOM: properly set up DDIA output on RS6xx boards 22509 22510commit 1d0e9ab8b9451101b1b91943546f6c5833c21b3f 22511Author: Michel Dänzer <michel@tungstengraphics.com> 22512Date: Wed Feb 20 10:21:49 2008 +0100 22513 22514 radeon: Fix typo flagged by gcc -Wall. 22515 22516commit b5bd442b60dbc72fe4c1e928ab864aeb0fd7a3cb 22517Author: Alex Deucher <alex@botch2.(none)> 22518Date: Tue Feb 19 20:47:40 2008 -0500 22519 22520 R100: fix render accel for transforms 22521 22522 Not sure why we had a separate broken path for r100 vertex 22523 submission. 22524 22525commit a0a73208a21546ac120fb9a463261836c9ea7b55 22526Author: Alex Deucher <alex@botch2.(none)> 22527Date: Tue Feb 19 20:11:19 2008 -0500 22528 22529 RADEON: restore clock gating and CP clock errata on VT switch 22530 22531 This may help people with hangs on resume 22532 22533commit b77e2aff7453a9f370beba37ca3c25b92b3f97ff 22534Author: Alex Deucher <alex@botch2.(none)> 22535Date: Tue Feb 19 19:55:41 2008 -0500 22536 22537 RADEON: fix DDC types 5 and 6 22538 22539commit af82172a82f2bdf96e571def659a1c70f92dfdbf 22540Author: Alex Deucher <alex@botch2.(none)> 22541Date: Tue Feb 19 19:39:35 2008 -0500 22542 22543 RADEON: update man page with supported chips 22544 22545commit 1302d1b019c8091986aaf9c86b25c1f36a037788 22546Author: Alex Deucher <alex@botch2.(none)> 22547Date: Tue Feb 19 00:34:31 2008 -0500 22548 22549 Bump for 6.8 release 22550 22551commit 830d29e8b280c11e798fb5bf5278bf24ecbac7bf 22552Author: Alex Deucher <alex@cube.(none)> 22553Date: Mon Feb 18 21:17:12 2008 -0500 22554 22555 ATOM: fix RS6xx connector table parsing and fix some warnings 22556 22557commit 03aa4cc6d6e8c715a1c1d677cc1845223505b358 22558Author: Roland Scheidegger <sroland@tungstengraphics.com> 22559Date: Mon Feb 18 20:19:58 2008 -0500 22560 22561 bring back to life planar-to-packed conversion for rs4xx 22562 22563 Here's a patch to bring back the code for converting planar yuv to packed yuv, 22564 if a RS400 family chip is used (though I've no idea if they really all fail 22565 with planar yuv). 22566 22567 fixes bug 12744 22568 22569commit 66b0b3f43bb714bc007169d5ba6dabe34cfc008b 22570Author: George Sapountzis <gsap7@yahoo.gr> 22571Date: Mon Feb 18 10:33:59 2008 +0200 22572 22573 r128/radeon: probably correct fix for non pci access code 22574 22575commit d055b9e800ae50d08cca9db75fc666ce1da9ab52 22576Author: Dave Airlie <airlied@linux.ie> 22577Date: Sat Feb 16 09:13:43 2008 +1000 22578 22579 r128/radeon: hopeful fix for non pci access code 22580 22581commit c773bc6a314327da29e21e4ebac6fa7f3e98a0a7 22582Author: Dave Airlie <airlied@linux.ie> 22583Date: Sat Feb 16 09:05:07 2008 +1000 22584 22585 r128/radeon: fix build without pciaccess 22586 22587commit 690a52da5248f47a587a878d05fce9784957970b 22588Author: Dave Airlie <airlied@linux.ie> 22589Date: Sat Feb 16 08:33:36 2008 +1000 22590 22591 mach64: fix non pciaccess build 22592 22593commit fc85188fd95bf78b7f965cdde3e22b644c74ff81 22594Author: George Sapountzis <gsap7@yahoo.gr> 22595Date: Fri Feb 15 18:36:10 2008 +0200 22596 22597 ati: convert to pci probe 22598 22599 add pciids for each subdriver, make no use of the match_data functionality. 22600 22601 thanks to Alex Deucher for reviewing and testing. 22602 22603commit 665bd7e2f61cac3e029bbad5024034e5136deec1 22604Author: George Sapountzis <gsap7@yahoo.gr> 22605Date: Fri Feb 15 18:34:56 2008 +0200 22606 22607 ati wrapper: translate Device lines 22608 22609commit a596c1618f72179a45289a50a1f9e89462ce9667 22610Author: George Sapountzis <gsap7@yahoo.gr> 22611Date: Fri Feb 15 18:34:31 2008 +0200 22612 22613 r128: do not compile in PciChipsets twice 22614 22615commit 99cd8ff9a7e15fc2b4e55f8bc020f584173a8c2d 22616Author: George Sapountzis <gsap7@yahoo.gr> 22617Date: Fri Feb 15 18:33:51 2008 +0200 22618 22619 ati: drop duplicate pci-id defines 22620 22621commit 146b01b51069dc227d0b125babb3f6957c9b9de2 22622Author: George Sapountzis <gsap7@yahoo.gr> 22623Date: Fri Feb 15 18:33:29 2008 +0200 22624 22625 mach64: clean probe a little 22626 22627 do not report I/O ports now, they are reported later. 22628 22629commit f47d461331a032f9bdcf6f63336e848778cec6cc 22630Author: George Sapountzis <gsap7@yahoo.gr> 22631Date: Fri Feb 15 18:32:56 2008 +0200 22632 22633 mach64: minor cosmetic, I2C type 22634 22635commit 2c66f2e812195167df9ca113044d46deece776ac 22636Author: Dave Airlie <airlied@clockmaker.usersys.redhat.com> 22637Date: Fri Feb 15 10:59:30 2008 +1000 22638 22639 make distcheck pass 22640 22641commit ed87f367ddab7366f84917b43b31d76df4ce1110 22642Author: Alex Deucher <alex@botch2.(none)> 22643Date: Wed Feb 13 12:53:46 2008 -0500 22644 22645 RADEON: disable LVDS if panel size is 0x0 22646 22647 if we can't get the panel size from the bios or edid 22648 or a user specified option, assume it's not connected. 22649 22650commit 422d7f441fdbb318d22d44db661ac9bd61387bd2 22651Author: Alex Deucher <alex@cube.(none)> 22652Date: Tue Feb 12 15:35:46 2008 -0500 22653 22654 R6xx: when both crtcs are in use make sure they are both enabled 22655 22656 sometimes setting a mode on one crtc can cuase a blank screen on the other. 22657 make sure they are both enabled if they should be. 22658 22659commit 860f5af75274cb236f536e1da09da6bd9a579530 22660Author: Alex Deucher <alex@cube.(none)> 22661Date: Tue Feb 12 14:46:49 2008 -0500 22662 22663 R6xx: fix up use of bios scratch regs to reflect the new offsets 22664 22665commit 32f2119b43a0faf6069d8cc0816f0d9f7914c07f 22666Author: Alex Deucher <alex@cube.(none)> 22667Date: Tue Feb 12 14:20:02 2008 -0500 22668 22669 R6xx: bios scratch regs moved. 22670 22671commit 8d64be6ebd7f50d4bcb587afeee8252c1367dc77 22672Author: Alex Deucher <alex@samba.(none)> 22673Date: Tue Feb 12 12:53:09 2008 -0500 22674 22675 RADEON: make sure we always set up the FP crtc shadow registers for crtc0 22676 22677 The behavior changed when I added rmx center mode support. In cases where 22678 crtc0 drives a DAC this can lead to a blank screen. 22679 22680commit 810d192ee046077a894e0fb5f2dfd6a7c0130766 22681Author: Alex Deucher <alex@cube.(none)> 22682Date: Mon Feb 11 19:23:01 2008 -0500 22683 22684 R6xx: add missing objects 22685 22686commit 85043439426e534e561259ce98bebdd8508b36a9 22687Author: Alex Deucher <alex@botch2.(none)> 22688Date: Mon Feb 11 16:36:58 2008 -0500 22689 22690 R6xx: make sure we set up the HDP base properly 22691 22692commit 9ab5d2ec7c583c74f364d7cfbb54bcd2cd8ae2f5 22693Author: Alex Deucher <alex@botch2.(none)> 22694Date: Mon Feb 11 15:26:51 2008 -0500 22695 22696 RADEON: always restore crtc1 before crtc0 when using both crtcs 22697 22698 In some rare cases restoring crtc0 first will result in a blank screen 22699 on crtc1. If you are having issues with a blank screen on crtc1 22700 that used to work on 6.6.3 or before, this should help. 22701 22702commit e33edca75bd9df0aa19a33e74c38a6d02610befd 22703Author: Alex Deucher <alex@botch2.(none)> 22704Date: Mon Feb 11 00:33:12 2008 -0500 22705 22706 RADEON: Implement proper (hopefully) BIOS/DRIVER interaction for ATOM bios 22707 22708 Tested on atom-based Desktop cards. It'd nice to get some testing 22709 on atom-based laptops. 22710 22711commit 6524e33435a786f7de0064cdd1b04c1120d21593 22712Author: Alex Deucher <alex@botch2.(none)> 22713Date: Sun Feb 10 18:52:52 2008 -0500 22714 22715 RADEON: Implement proper (hopefully) BIOS/DRIVER interaction for COM bios 22716 22717 Tested on my M10-based laptop. 22718 22719commit 8606c1bd175893c77e8c758246a1aed166c8be51 22720Author: George Sapountzis <gsap7@yahoo.gr> 22721Date: Fri Feb 8 19:07:03 2008 +0200 22722 22723 mach64: factor out BIOS panel info parsing 22724 22725commit 9f33218c80f5a6d6d9464aa3db8ae25a4759f351 22726Author: George Sapountzis <gsap7@yahoo.gr> 22727Date: Fri Feb 8 19:06:41 2008 +0200 22728 22729 mach64: minor cosmetic, LCD panel id 22730 22731commit 8cd5a465a03834b4b2f3a19f8d36fc9d2e18c6d4 22732Author: George Sapountzis <gsap7@yahoo.gr> 22733Date: Fri Feb 8 19:06:27 2008 +0200 22734 22735 mach64: minor cosmetic, DAC type 22736 22737commit 7f4db96123fdcba9874226352802d702c2a498bd 22738Author: Dave Airlie <airlied@redhat.com> 22739Date: Fri Feb 8 10:47:10 2008 +1000 22740 22741 add rv670 pciids 22742 22743commit e8899b9978291c62a65f468c92f340f65ad5479d 22744Author: Alex Deucher <alex@botch2.(none)> 22745Date: Thu Feb 7 19:27:38 2008 -0500 22746 22747 R6xx: fix ddc after my i2c rework 22748 22749 Seems r6xx does something different for its i2c table, 22750 revert to the old behavior for now. 22751 22752commit 435de6c4e46ff2bebd4cee58888a66b936cd3fdf 22753Author: Alex Deucher <alex@samba.(none)> 22754Date: Thu Feb 7 19:14:13 2008 -0500 22755 22756 RADEON: sync up with latest upstream versions 22757 22758 atombios.h 22759 ObjectID.h 22760 22761commit 692789a293970f70b88ccb6adcf0676d8b495ae2 22762Author: George Sapountzis <gsap7@yahoo.gr> 22763Date: Thu Feb 7 18:03:37 2008 +0200 22764 22765 mach64: factor out BIOS multimedia parsing 22766 22767commit 933328ffd6d1d872a18d3de8624c4df845a64588 22768Author: George Sapountzis <gsap7@yahoo.gr> 22769Date: Thu Feb 7 18:03:03 2008 +0200 22770 22771 mach64: complement hint for sparc 22772 22773commit 956c8c81f3ff434930a0cb17b027b2f8e4eeabb2 22774Author: George Sapountzis <gsap7@yahoo.gr> 22775Date: Thu Feb 7 18:02:31 2008 +0200 22776 22777 mach64: consolidate refclk #2 22778 22779commit dce4cc26a8e2bf53805ec63763243f3ff6b4a6d3 22780Author: George Sapountzis <gsap7@yahoo.gr> 22781Date: Thu Feb 7 18:02:17 2008 +0200 22782 22783 mach64: consolidate refclk #1 22784 22785commit f7ed807f0d82a7446ebc4acdd4e94df44a675f19 22786Author: George Sapountzis <gsap7@yahoo.gr> 22787Date: Thu Feb 7 18:01:59 2008 +0200 22788 22789 mach64: cosmetic 22790 22791commit cda1cd198f33c26ef1b51532a2126468369743b8 22792Author: George Sapountzis <gsap7@yahoo.gr> 22793Date: Thu Feb 7 18:01:33 2008 +0200 22794 22795 mach64: factor out BIOS clock parsing 22796 22797commit 73ff279469be9c7cbf9f533b85fcb553694ff413 22798Author: George Sapountzis <gsap7@yahoo.gr> 22799Date: Thu Feb 7 18:00:55 2008 +0200 22800 22801 mach64: BIOSBase is no longer used 22802 22803commit caea326cc6f1932bb299f451be013651a5749ea7 22804Author: Dave Airlie <airlied@linux.ie> 22805Date: Wed Feb 6 06:36:13 2008 +1000 22806 22807 r300: move fragprog setup code to prepare composite for now 22808 22809commit 470cd6a401c6a3e8fea981a8fe97c28be3cfb81d 22810Author: Dave Airlie <airlied@linux.ie> 22811Date: Wed Feb 6 06:04:13 2008 +1000 22812 22813 r300: remove r300 specific vertex emission 22814 22815 Set the vertex and fragment engine to expect the mask coords. 22816 22817commit f65e8dfac23adfa199026765fe3a1ea08cf4da67 22818Author: Alex Deucher <alex@cube.(none)> 22819Date: Sun Feb 3 00:09:59 2008 -0500 22820 22821 RADEON: rework i2c handling 22822 22823 Split out clk, data, and lock regs and masks. some cards use different 22824 regs and masks for each. For cards with ATOMBIOS, use the i2c bios 22825 table to grab the i2c data. 22826 22827commit a38a903debc0a50dbc73f59dc2741bbea76d2bd9 22828Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 22829Date: Fri Feb 1 02:59:07 2008 -0500 22830 22831 RADEON: don't restore LVDS_PLL_CNTL for now 22832 22833 seems to cause problems with resume for some users. 22834 this needs further investigation. 22835 see bug 12596 22836 22837commit 5d7bea2b62c86868baf1960665a40249a43bafc5 22838Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 22839Date: Fri Feb 1 02:53:22 2008 -0500 22840 22841 RADEON: remove redundant RADEONDisableDisplays() 22842 22843 use RADEONBlank() instead 22844 22845commit 73b437ce232c94c0067a0d2f70538b6e1e8c07a7 22846Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 22847Date: Fri Feb 1 02:47:06 2008 -0500 22848 22849 RADEON: remove unused "aspect" scaler option 22850 22851commit f1fb9e4daa29bc379f653f847254db1496b625fd 22852Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 22853Date: Fri Feb 1 02:39:18 2008 -0500 22854 22855 RADEON: Implement "center" mode for RMX on legacy radeons 22856 22857commit bcd590103e04bfdb4f12413beacebf344f07e88e 22858Author: Alex Deucher <alex@botch2.(none)> 22859Date: Tue Jan 29 12:12:54 2008 -0500 22860 22861 RADEON: update man page to reflect AGP 1x default revert 22862 22863commit 09348a83d06ba9d3129499c4daedd44a68771530 22864Author: Alex Deucher <alex@botch2.(none)> 22865Date: Tue Jan 29 12:09:24 2008 -0500 22866 22867 Seems the default is more reliable... we can't win. 22868 22869 Revert "radeon: Default to 1x again with non-v3 AGP cards." 22870 22871 This reverts commit b653e5a628bfa4dfb168e96f93f41eb910f409fb. 22872 22873commit 0c26806245381b925b73ff9f3b747e3dcf0ebd6f 22874Author: Alex Deucher <alex@botch2.(none)> 22875Date: Tue Jan 29 10:26:48 2008 -0500 22876 22877 RADEON: Add new RV380 pci id 22878 22879 bug 14289 22880 22881commit ce77ed78a877023da72dbe51609aef9a07e250b1 22882Author: Maciej Cencora <m.cencora@gmail.com> 22883Date: Mon Jan 28 19:02:56 2008 -0500 22884 22885 RS690: Implement MC idle check 22886 22887commit b7de4ff52cfbdd85ee65000613632e21b92af24e 22888Author: George Sapountzis <gsap7@yahoo.gr> 22889Date: Sat Jan 26 19:28:05 2008 +0200 22890 22891 mach64: add hint for sparc and minor cosmetic. 22892 22893commit 09d713aa0ed6367b4457420b3c2832fe8eca9b00 22894Author: Alex Deucher <alex@botch2.(none)> 22895Date: Wed Jan 23 18:05:41 2008 -0500 22896 22897 Revert "RADEON: adjust PAL mode hstart" 22898 22899 This reverts commit 719a9a376e34d99032af75e3f7b002670ccb816b. 22900 22901 This breaks TV out on some cards. 22902 22903commit 719a9a376e34d99032af75e3f7b002670ccb816b 22904Author: Alex Deucher <alex@botch2.(none)> 22905Date: Sun Jan 20 18:40:53 2008 -0500 22906 22907 RADEON: adjust PAL mode hstart 22908 22909 Most people seem to get a more aligned picture with this setting 22910 22911commit b2db8657fb888cff6d64c6dcb182caac389776ce 22912Author: Alex Deucher <alex@botch2.(none)> 22913Date: Sun Jan 20 18:33:22 2008 -0500 22914 22915 RADEON: re-work i2c for DDC 22916 22917 Unify the radeon/avivo paths and grab the data/clk masks from bios 22918 if available 22919 22920commit a43003b24022a833e604f41b2873c0350b34181c 22921Author: Alex Deucher <alex@botch2.(none)> 22922Date: Sat Jan 19 18:49:53 2008 -0500 22923 22924 RADEON: get dac2 adj values from the bios tables 22925 22926commit d4596c52ac9994be26e9ec2d7d57b3892c34abdb 22927Author: Alex Deucher <alex@botch2.(none)> 22928Date: Sat Jan 19 17:17:26 2008 -0500 22929 22930 RADEON: grab pll_in_min/pll_in_max from bios tables if available 22931 22932commit 9a0947c812d0d38d1bca6a91140ac210831a6cb4 22933Author: Alex Deucher <alex@botch2.(none)> 22934Date: Sat Jan 19 13:57:50 2008 -0500 22935 22936 ATOM: Use LVDS edid from bios if available 22937 22938commit 7238258c12def8ef273e5362f716d165f720c5a5 22939Author: Kusanagi Kouichi <slash@ma.neweb.ne.jp> 22940Date: Sat Jan 19 15:04:21 2008 +0100 22941 22942 radeon: Partial fix for XVideo RGB image distortions. 22943 22944commit 32be3cf9d6c34e60ff8c3d6cfe9f73f1869c50e4 22945Author: Brice Goglin <Brice.Goglin@ens-lyon.org> 22946Date: Fri Jan 18 14:42:14 2008 -0500 22947 22948 RADEON: print the name of the output when printing the EDID 22949 22950commit 4ba9430ee97dbce8f77db8de6ce9b753a75e453d 22951Author: George Sapountzis <gsap7@yahoo.gr> 22952Date: Fri Jan 18 18:18:18 2008 +0200 22953 22954 mach64: workaround for corruption at upper-left 22955 22956 commit possible workaround, it's reported multiple times ... 22957 22958commit 12c00111b68c9cf4872a424258c6f8b7247aac47 22959Author: George Sapountzis <gsap7@yahoo.gr> 22960Date: Fri Jan 18 18:17:40 2008 +0200 22961 22962 use stand-alone drivers for each chip family. 22963 22964 Do not load the ati wrapper when the user specifies the sub-driver name in 22965 the Driver line of xorg.conf. Also, for -configure cause the wrapper to fail 22966 and let each sub-driver speak for themselves. 22967 22968commit 24c7d134cd450f9e2cca85e4a2fc3253d250be04 22969Author: Alex Deucher <alex@botch2.(none)> 22970Date: Fri Jan 18 09:50:38 2008 -0500 22971 22972 RADEON: make sure EXA Composite is actually disabled on XPRESS chips. 22973 22974commit eaf425436008092abe81208321a2b3b6698a5d79 22975Author: Dave Airlie <airlied@redhat.com> 22976Date: Fri Jan 18 20:11:57 2008 +1000 22977 22978 fixup register 6594 save/restore 22979 22980commit 3de2dc88cf26ff5932f11cecdf975777b8aa2a4a 22981Author: Adam Jackson <ajax@redhat.com> 22982Date: Wed Jan 16 14:55:05 2008 -0500 22983 22984 Bump CRTC size limits on AVIVO chips so 30" displays work without tweaking. 22985 22986 Note that the CRTC size limits we're using right now are _not_ the 22987 hardware limits, they're just heuristics until we can resize the front 22988 buffer properly. 22989 22990commit 2a89a31481d71a56a9930073cf99d3ae7b4290e1 22991Author: Alex Deucher <alex@samba.(none)> 22992Date: Thu Jan 17 15:08:17 2008 -0500 22993 22994 RADEON: use radeon_output->Flags for tracking RMX rather than mode->Flags 22995 22996commit 495e3119250ffb48489debbaabe560d23753cc43 22997Author: Alex Deucher <alex@samba.(none)> 22998Date: Thu Jan 17 14:56:19 2008 -0500 22999 23000 AVIVO: Add support for RMX 23001 23002 Both centered and expansion modes are supported. Select 23003 using output attributes. 23004 23005commit 6bd510a211f25d52e74791e4a429cd2218ced541 23006Author: Alex Deucher <alex@samba.(none)> 23007Date: Wed Jan 16 18:09:49 2008 -0500 23008 23009 RADEON: add a message about render accel on newer cards 23010 23011commit 85bf3439fe2579aec48f5cd8d65a9d51b1ae8535 23012Author: Alex Deucher <alex@samba.(none)> 23013Date: Wed Jan 16 17:52:06 2008 -0500 23014 23015 R300: only init3d on r3xx 23016 23017commit dbb2ca471dfbff245b30c5055871dee0dc0e3d15 23018Author: Alex Deucher <alex@samba.(none)> 23019Date: Wed Jan 16 17:10:02 2008 -0500 23020 23021 R300: only enable render accel on non-IGP r3xx/r4xx chips for now 23022 23023commit 3c72b100bcfacee600644669b586e86cfd32754e 23024Author: Alex Deucher <alex@samba.(none)> 23025Date: Wed Jan 16 16:55:42 2008 -0500 23026 23027 R300: First pass at render accel 23028 23029 This first pass is pretty limited. All it currently supports 23030 is transforms for rotation. No blending yet. 23031 23032 Based on inital implementation from Wolke Liu with 23033 additional lock-up fixes by Dave Airlie. 23034 23035commit 2ba3562d2af911fdd90881049599e239d27260bc 23036Author: George Sapountzis <gsap7@yahoo.gr> 23037Date: Sat Jan 12 17:11:59 2008 +0200 23038 23039 ati wrapper: xf86PciInfo.h is enough 23040 23041commit c2caeb11a97dad5379d70881c5c0fd834a8c3d54 23042Author: George Sapountzis <gsap7@yahoo.gr> 23043Date: Sat Jan 12 16:18:34 2008 +0200 23044 23045 ati wrapper: add DriverRec's and use them 23046 23047commit 19e1b180fec6f83a474e125465bc60111c0f43e0 23048Author: George Sapountzis <gsap7@yahoo.gr> 23049Date: Sat Jan 12 16:18:07 2008 +0200 23050 23051 mach64: load for both "ati" and "mach64" as driver names 23052 23053 similar to r128/radeon 23054 23055commit 92f54400d5450b29b3a76d5ecc927cf0d73e156e 23056Author: George Sapountzis <gsap7@yahoo.gr> 23057Date: Sat Jan 12 16:17:47 2008 +0200 23058 23059 mach64: add version (need not match with ati) 23060 23061 similar to r128/radeon 23062 23063commit 311ec7b6c54a50a4b8a5a445f7283da2b0b2e0f5 23064Author: George Sapountzis <gsap7@yahoo.gr> 23065Date: Sat Jan 12 16:17:21 2008 +0200 23066 23067 atimisc: rename to mach64 23068 23069commit 5244e235262290eab8a3546f449295c12ed8f6c7 23070Author: Alex Deucher <alex@samba.(none)> 23071Date: Mon Jan 14 16:11:09 2008 -0500 23072 23073 ATOMBIOS: disable the scaler by default on avivo cards 23074 23075 Fixes bug 14001 23076 23077commit 729da30c80d6545b75c8faea44754634f477cc09 23078Merge: 000741e2 625a885a 23079Author: Alex Deucher <alex@samba.(none)> 23080Date: Mon Jan 14 10:05:01 2008 -0500 23081 23082 Merge branch 'master' of ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati 23083 23084commit 625a885a964488da2a35065eb597a17ee57b22a9 23085Author: George Sapountzis <gsap7@yahoo.gr> 23086Date: Wed Jan 9 17:39:56 2008 +0200 23087 23088 ati wrapper: use pci ids 23089 23090commit 48865200ca9f1daf87e52620b9f8175f88dd886f 23091Author: George Sapountzis <gsap7@yahoo.gr> 23092Date: Wed Jan 9 19:01:04 2008 +0200 23093 23094 radeon: remove stray include 23095 23096commit a5a6b873353611cb7e46e5e375f039957c7051a7 23097Author: George Sapountzis <gsap7@yahoo.gr> 23098Date: Wed Jan 9 13:03:32 2008 +0200 23099 23100 radeon: remove stray _X_EXPORT 23101 23102commit 80c9974b6cdc0147d777df6990b3f3aacd87969d 23103Author: George Sapountzis <gsap7@yahoo.gr> 23104Date: Tue Jan 8 18:15:47 2008 +0200 23105 23106 r128: pci-rework conversion 23107 23108 compile-tested only 23109 23110commit cd4b39403d74f314e1c2cfa4cf0e4fe542891dc3 23111Author: George Sapountzis <gsap7@yahoo.gr> 23112Date: Sat Mar 10 23:12:15 2007 +0200 23113 23114 Drop symbol lists from r128, radeon, theatre. 23115 23116 compile-tested only 23117 23118commit 63b4b8213fabc5a57f897b60b6eaa9f78d86e6ff 23119Author: George Sapountzis <gsap7@yahoo.gr> 23120Date: Wed Mar 21 20:24:22 2007 +0200 23121 23122 [mach64] Drop symbol lists. 23123 23124 egrep LOADER\|SymLists\|Symbols src/*.[hc] 23125 23126 not needed after server commit bca9364f3f4a2376edbcf57a34f704ce28be21ba, i.e. 23127 xserver 1.2 23128 23129 compile-tested only 23130 23131commit 000741e250e54122b0adc91694eb4bfa320a70fb 23132Author: Alex Deucher <alex@samba.(none)> 23133Date: Thu Jan 10 14:49:48 2008 -0500 23134 23135 RADEON: clean up output handling 23136 23137commit 10e7636c02478b8ffe183bb0c46229ca0d6584e1 23138Author: Kristian Høgsberg <krh@bitplanet.net> 23139Date: Wed Jan 9 12:47:39 2008 -0500 23140 23141 RADEON: fix crtc routing on r4xx cards when using atom to init DVO chip 23142 23143commit 3af671f5963810dbfd63abc9889b1d46b68f404c 23144Author: Alex Deucher <alex@botch2.(none)> 23145Date: Wed Jan 9 11:30:25 2008 -0500 23146 23147 RADEON: restore FP2 regs before external encoders 23148 23149 This may fix krh's dvi problem 23150 23151commit 2a54c6bb09ade2ec8f998dfc1624017029d47fa3 23152Author: Alex Deucher <alex@botch2.(none)> 23153Date: Tue Jan 8 18:43:54 2008 -0500 23154 23155 RADEON: Make default output actually work... 23156 23157commit fa3e2055225c27e25465fc46786da1b7574fd3cc 23158Author: Alex Deucher <alex@botch2.(none)> 23159Date: Mon Jan 7 01:13:09 2008 -0500 23160 23161 RADEON: add default outputs if no connected devices 23162 23163 If no connected devices found at server startup, default 23164 to something so the server comes up. LVDS on mobility chips, 23165 DAC or TMDS on others. 23166 23167commit d972cc9237eb90b49b11f8d2bdc5b61f628911dc 23168Author: Alex Deucher <alex@botch2.(none)> 23169Date: Sat Jan 5 17:19:06 2008 -0500 23170 23171 RADEON: Fix TVStandard option 23172 23173commit 45656b9d5a426053da2a094de39c2690c0c6f799 23174Author: Alex Deucher <alex@botch2.(none)> 23175Date: Sat Jan 5 12:00:55 2008 -0500 23176 23177 R128: Like powerpc, don't use VGA by default on sparc 23178 23179commit b8e8db4675d07e45782de0d7c67ee0fd85eaedb3 23180Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23181Date: Fri Jan 4 20:16:19 2008 -0500 23182 23183 RADEON: fix tvdac load detection at server start up 23184 23185commit 3ba7f393d0669df36848715799de8affc10a5534 23186Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23187Date: Fri Jan 4 20:00:48 2008 -0500 23188 23189 RADEON: further fixup for pScrn->pScreen issue 23190 23191 the previous fix seems to cause the driver to hang on 23192 some cards. 23193 23194commit a0de9c0844f9e066e0f02e8cd8045bdd278e6494 23195Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23196Date: Fri Jan 4 19:48:30 2008 -0500 23197 23198 RADEON: improve ntsc image centering 23199 23200commit 03d2f25801c8a8ec15030f06008df112d07c1a2d 23201Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23202Date: Fri Jan 4 18:19:54 2008 -0500 23203 23204 RADEON: fix compile 23205 23206commit 308848783ed9ae27aed7c7de6ee813d375ef495e 23207Author: Dave Airlie <airlied@linux.ie> 23208Date: Sat Jan 5 09:11:55 2008 +1000 23209 23210 i2c: a line mux of 0 is valid 23211 23212commit 94d5a432f72801f821d1c4ce952baba17552659d 23213Author: Dave Airlie <airlied@linux.ie> 23214Date: Fri Jan 4 17:16:54 2008 +1000 23215 23216 atombios: i2c lines are all done with sw so we can accept any of them. 23217 23218 hch reported this working on his laptop with 0x7e60 as EDID for LVDS 23219 23220commit 2e4473b63d65801ae8ac5a8391de232b2201d958 23221Author: Alex Deucher <alex@samba.(none)> 23222Date: Thu Jan 3 15:47:50 2008 -0500 23223 23224 RADEON: fix crash when setting rotation in the config file 23225 23226 xf86CrtcRotate() accesses pScrn->pScreen which is not set 23227 during ScreenInit(). This should also be fixed in the server. 23228 See bug 12129 23229 23230commit c652208861bffca94f06b7f67688ce220e050bfb 23231Author: Michel Dänzer <michel@tungstengraphics.com> 23232Date: Thu Jan 3 17:54:58 2008 +0100 23233 23234 radeon: Adapt manpage to reality wrt AGP options. 23235 23236commit ab451e4b7a5423d61b57cf0646599267d8504af4 23237Author: Michel Dänzer <michel@tungstengraphics.com> 23238Date: Thu Jan 3 17:52:39 2008 +0100 23239 23240 radeon: Miscellaneous warning fixes. 23241 23242commit 394c52273328e90518568b694ee79dc1a8dab651 23243Author: Dave Airlie <airlied@linux.ie> 23244Date: Thu Jan 3 18:56:16 2008 +1000 23245 23246 r500: tvout avoid doing dpms here it makes my tv mode go all crappy 23247 23248 need to investigate further 23249 23250commit 1c647279f021d01e110980727b7c7dd7efae1642 23251Author: Dave Airlie <airlied@linux.ie> 23252Date: Thu Jan 3 11:55:28 2008 +1000 23253 23254 r600: change devices list depending on connector 23255 23256commit f36db6e10d32a68b32d20ae4ad02cfc0bfd1c9c3 23257Author: Dave Airlie <airlied@linux.ie> 23258Date: Thu Jan 3 11:27:47 2008 +1000 23259 23260 r600: fixup crash on unknown output type 23261 23262 not sure this is 100% the correct approach 23263 23264commit 1accfdd590828e95e0d68a576c8ee05a06a86e43 23265Author: Alex Deucher <alex@samba.(none)> 23266Date: Wed Jan 2 19:48:28 2008 -0500 23267 23268 RADEON: various avivo tv-out related clean-ups 23269 23270commit ce34090c758ac91171cb6adb9b8a36e4adbf99cf 23271Merge: 2180f04b 30cab1db 23272Author: Alex Deucher <alex@samba.(none)> 23273Date: Wed Jan 2 16:41:36 2008 -0500 23274 23275 Merge branch 'master' of git+ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati 23276 23277commit 2180f04b6fb94a638f6274cb4455d5688b324dbc 23278Author: Alex Deucher <alex@samba.(none)> 23279Date: Wed Jan 2 16:41:06 2008 -0500 23280 23281 RADEON: preliminary support for ATOM RMX 23282 23283 not functional yet. 23284 23285commit dab4dc285154d40303aadaa849b85f8e251e578e 23286Author: Alex Deucher <alex@samba.(none)> 23287Date: Wed Jan 2 16:27:19 2008 -0500 23288 23289 RADEON: add support for ATOM component video table 23290 23291 Component output is still not working. 23292 23293commit 30cab1dbebb7bdb925f2fe9f91b84183312bbbfd 23294Author: Alex Deucher <alex@botch2.(none)> 23295Date: Wed Jan 2 15:41:01 2008 -0500 23296 23297 RADEON: Make sure all old IGP chips have HasSingleDac set 23298 23299 fix the csv file and re-gen the headers. 23300 23301commit e8e585651215b011e3ad07c59d0eab9107ccd8c6 23302Author: Dave Airlie <airlied@redhat.com> 23303Date: Wed Jan 2 10:14:46 2008 +1000 23304 23305 PLL/r600: tweak pll to pick first one found instead of keeping going 23306 23307commit 14aa4060ad27ecb3d40b2b17ee4cf7cc55a121dd 23308Author: Dave Airlie <airlied@redhat.com> 23309Date: Wed Jan 2 09:49:44 2008 +1000 23310 23311 r600: fix tv-out output naming 23312 23313commit f65374f5e15bfd391a1838a686cd87d3bab8043d 23314Author: Maciej Cencora <m.cencora@gmail.com> 23315Date: Mon Dec 31 09:44:34 2007 +1000 23316 23317 atombios: initial rs690 patches 23318 23319commit a674f683e6699c30664d9cd468a64de991c3fd7e 23320Author: Dave Airlie <airlied@linux.ie> 23321Date: Sun Dec 30 17:43:17 2007 +1000 23322 23323 atombios: enable TV-out detection code - tv out works for me with this 23324 23325commit aa7c28cbd943bb525698515d444cb5097880e364 23326Author: Dave Airlie <airlied@linux.ie> 23327Date: Sun Dec 30 17:40:37 2007 +1000 23328 23329 atombios: enable support for using tv timings 23330 23331 enable support for the atombios tv timings selection by programming the crtc 23332 with the tv timings if a tv is detected on the output 23333 23334commit 0bc3fd595a73e12a424571697d164a09a6a4c072 23335Author: Dave Airlie <airlied@linux.ie> 23336Date: Sun Dec 30 16:39:58 2007 +1000 23337 23338 atombios: add support for reading tv standard from atombios 23339 23340 fix typo in atombios header file 23341 23342commit bfa22d676a6f333503104041f62222f4de9bb7d8 23343Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23344Date: Wed Dec 26 12:29:47 2007 -0500 23345 23346 RADEON: fix typo that broke tv load detection 23347 23348commit 0c99554c6fab1192f5e8595469c21b5f33e1eb4f 23349Author: David Miller <davem@davemloft.net> 23350Date: Wed Dec 26 02:19:12 2007 -0500 23351 23352 [RADEON]: Add missing break in SCLK calculation. 23353 23354commit 6e0d5cc1c62fbfc1962fa0d6333f0c0a8c6327bd 23355Author: David Miller <davem@davemloft.net> 23356Date: Wed Dec 26 02:17:34 2007 -0500 23357 23358 [RADEON]: Like powerpc, don't use VGA by default on sparc. 23359 23360commit 2b1fae668ddabbc72e5fc31365302ea722174df1 23361Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23362Date: Wed Dec 26 02:04:06 2007 -0500 23363 23364 RADEON: fix PAL tv-out 23365 23366 Many thanks to Andrew Randrianasulu for providing me with 23367 pll reg dumps 23368 23369commit d736eb5732da573162c70712dc4e8b0114986702 23370Author: Alex Deucher <alex@botch2.(none)> 23371Date: Mon Dec 24 12:34:15 2007 -0500 23372 23373 RADEON: default "IgnoreLidStatus" to TRUE 23374 23375 Seems there are lots of busted ACPI lid status and people 23376 starting X with the lid closed. 23377 23378commit 653da558148cc601bc1f80253e92ef98c75ef37a 23379Author: Alex Deucher <alex@botch2.(none)> 23380Date: Mon Dec 24 01:11:56 2007 -0500 23381 23382 RADEON: restore crtc regs before VGA regs. 23383 23384 It seems some radeons don't restore text console properly if 23385 the crtc regs are restored after the VGA regs. 23386 Thanks to Sverre Froyen for helping me track this down 23387 23388commit ad3325f672a241449ca239c8ee3a24b6d7703d76 23389Author: Alex Deucher <alex@botch2.(none)> 23390Date: Sun Dec 23 17:18:42 2007 -0500 23391 23392 RADEON: Add "IgnoreLidStatus" option 23393 23394 Generally, users that start X with the laptop lid closed 23395 want to use one or more external monitors rather than the 23396 internal panel and do not want the internal panel to be on 23397 by default. Others, it seems, want to always have the 23398 internal panel on, regardless of the lid. I can't win. 23399 Enable this option to force the latter. 23400 23401commit 20eedf348a527e1e0a5450bc22d7564895034a66 23402Author: Alex Deucher <alex@botch2.(none)> 23403Date: Sun Dec 23 11:27:55 2007 -0500 23404 23405 RADEON: fix pll input setup on mac cards 23406 23407 the function was exiting before the complete setup was finished. 23408 23409commit 4f2e833e8ebaba3ad85ec5314fff8fa05395b679 23410Author: Arkadiusz Miskiewicz <arekm@maven.pl> 23411Date: Fri Dec 21 18:56:34 2007 -0500 23412 23413 configure.ac fixes 23414 23415commit 2b6e8e2b8f74e94560de89693ecbc7260536591e 23416Author: Arkadiusz Miskiewicz <arekm@maven.pl> 23417Date: Fri Dec 21 18:10:17 2007 -0500 23418 23419 RADEON: various cleanups 23420 23421commit 4c6f60e3b19ac55ab1255c79df03b1df5950864e 23422Author: Alex Deucher <alex@botch2.(none)> 23423Date: Fri Dec 21 17:33:04 2007 -0500 23424 23425 RADEON: clean up prototypes 23426 23427commit 8c761afdcb9baf1649b93449692fb9ab67bc2c80 23428Author: Alex Deucher <alex@botch2.(none)> 23429Date: Fri Dec 21 16:24:49 2007 -0500 23430 23431 RADEON: more re-org 23432 23433 move save/restore routines into legacy_crtc/output 23434 23435commit 910773d3a6c717f9d4762ea7b9ee6c3ae061781e 23436Author: Alex Deucher <alex@botch2.(none)> 23437Date: Fri Dec 21 15:53:15 2007 -0500 23438 23439 RADEON: more re-org. move XAA Mem init to radeon_accel.c 23440 23441commit 0631a23bd103f9b74e525da2c41304eab60c6f17 23442Author: Alex Deucher <alex@botch2.(none)> 23443Date: Fri Dec 21 15:40:18 2007 -0500 23444 23445 RADEON: fix rn50 reversion from last merge 23446 23447commit 5b917797a13c6caa80028d1842a284598e874288 23448Author: Alex Deucher <alex@botch2.(none)> 23449Date: Fri Dec 21 15:30:20 2007 -0500 23450 23451 RADEON: remove no longer used radeon_display.c 23452 23453commit bf14aa5f88fc3b4e69d71db5b23248b8bb2018d0 23454Author: Alex Deucher <alex@botch2.(none)> 23455Date: Fri Dec 21 15:29:47 2007 -0500 23456 23457 RADEON: re-org legacy crtc/output code into separate files 23458 23459commit be7f8fd338f5af8b632f16a83db41e15d00af469 23460Author: Dave Airlie <airlied@linux.ie> 23461Date: Fri Dec 21 10:03:49 2007 +1000 23462 23463 fix mode bandwidth configure check 23464 23465commit 3f9b597dedc45379b0bc0b631f3f924c403bca48 23466Author: Dave Airlie <airlied@linux.ie> 23467Date: Fri Dec 21 09:55:42 2007 +1000 23468 23469 fixup clip test include handling 23470 23471commit 9a5b501332c0a1f10af20845af48c9ddd2ce26a0 23472Author: Dave Airlie <airlied@linux.ie> 23473Date: Fri Dec 21 09:45:55 2007 +1000 23474 23475 set ddc line correctly post-merge 23476 23477commit 3c31b96afa20913ad947e68fe0c3a662e5eafbdd 23478Merge: eb99c3c5 f5e8c185 23479Author: Dave Airlie <airlied@linux.ie> 23480Date: Fri Dec 21 09:36:22 2007 +1000 23481 23482 Merge remote branch 'origin/atombios-support' 23483 23484 Conflicts: 23485 23486 src/radeon_display.c 23487 src/radeon_driver.c 23488 23489commit eb99c3c5c9a2249cb84920f0f225e525fc3a4144 23490Author: Alex Deucher <alex@botch2.(none)> 23491Date: Thu Dec 20 18:14:38 2007 -0500 23492 23493 Bump for RC release 23494 23495commit 8d49ff1da917b7f8240267953ef6ce4ff04daecb 23496Author: Alex Deucher <alex@botch2.(none)> 23497Date: Thu Dec 20 01:09:57 2007 -0500 23498 23499 RADEON: check for xf86_crtc_clip_video_helper() in configure.ac 23500 23501 use xf86_crtc_clip_video_helper() from the server if available. 23502 23503commit fb7a4e24f2da3561ef81371ca4013a4f13806e91 23504Author: Adam Jackson <ajax@redhat.com> 23505Date: Wed Dec 19 19:15:19 2007 -0500 23506 23507 Fix RN50 mode filtering. 23508 23509 The old code would attempt to limit the maximum pixel size of the screen 23510 by limiting the maximum PLL frequency. This ends up confusing the PLL 23511 computation code since sometimes your maximum freq can be lower than your 23512 minimum freq. More to the point it's just wrong, maximum PLL frequency 23513 isn't the same thing as maximum pixel clock, and even that isn't the same 23514 thing as maximum scanout pixels per second. 23515 23516 The correct thing to do is filter by the mode's effective memory bandwidth. 23517 23518commit c1b0b69cc50516c3b7e881b0eb46cb3cd2e9dce6 23519Merge: 0e663487 ce4fa1ce 23520Author: Alex Deucher <alex@botch2.(none)> 23521Date: Wed Dec 19 11:54:50 2007 -0500 23522 23523 Merge branch 'master' of ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati 23524 23525commit 0e6634870d1ab38ee8c83f6bda1ba60364997853 23526Author: Alex Deucher <alex@botch2.(none)> 23527Date: Wed Dec 19 11:54:27 2007 -0500 23528 23529 RADEON: skip empty connectors when creating outputs 23530 23531commit 6afbf718c151dc3c5c59bd3136b58a93a114d798 23532Author: Alex Deucher <alex@botch2.(none)> 23533Date: Wed Dec 19 11:48:38 2007 -0500 23534 23535 RADEON: add support for legacy radeons with DVI and no connector table 23536 23537commit f5e8c185001e62e744310667c2d1bd3fe6542a62 23538Author: Dave Airlie <airlied@clockmaker.usersys.redhat.com> 23539Date: Wed Dec 19 10:38:58 2007 +1000 23540 23541 more endian related fixage 23542 23543commit 98b247066d00db66abe91f518cd93b5c4da4cfb4 23544Author: Dave Airlie <airlied@clockmaker.usersys.redhat.com> 23545Date: Wed Dec 19 10:25:41 2007 +1000 23546 23547 fix big endian build since zaphod fixups 23548 23549commit ce4fa1cedec0cf56b9979dfaa12a8d3a7c643df4 23550Author: Arkadiusz Miskiewicz <arekm@maven.pl> 23551Date: Tue Dec 18 15:34:14 2007 -0500 23552 23553 RADEON: fix fd leak in lid detect code 23554 23555commit 20568f66f9a9a60a33bd9a69ccc14a891c656836 23556Author: Arkadiusz Miskiewicz <arekm@maven.pl> 23557Date: Tue Dec 18 15:32:10 2007 -0500 23558 23559 RADEON: more cleanups and warning fixes 23560 23561commit 1496194200adbcb044ec3977367a0908262e389c 23562Author: Arkadiusz Miskiewicz <arekm@maven.pl> 23563Date: Tue Dec 18 15:29:53 2007 -0500 23564 23565 RADEON: driver cleanups, warning fixes 23566 23567commit bd7206fa120495037e3fea0c920d0031d7715bf6 23568Author: Alex Deucher <alex@botch2.(none)> 23569Date: Tue Dec 18 03:03:11 2007 -0500 23570 23571 RADEON: fix another merge error 23572 23573 this broken legacy radeons 23574 23575commit 65a3ac7530e11bb7d818a988fd0cf1dde7688fa4 23576Author: Alex Deucher <alex@samba.(none)> 23577Date: Tue Dec 18 00:15:38 2007 -0500 23578 23579 RADEON: more PLL tweaks 23580 23581commit d93a0e10b8bc6e3797a3cf6c1e28ca413a7c38e4 23582Author: Alex Deucher <alex@samba.(none)> 23583Date: Mon Dec 17 20:32:45 2007 -0500 23584 23585 RADEON: post div tweaks for legacy radeon 23586 23587commit 03b8b49f6f502c45552b018fd8c44d366b2d576f 23588Author: Alex Deucher <alex@samba.(none)> 23589Date: Mon Dec 17 20:20:04 2007 -0500 23590 23591 RADEON: fix typo from merge 23592 23593commit 19b9d3708852b7efe2b05249c8359dadb924dd94 23594Merge: cf685f37 29706ca5 23595Author: Alex Deucher <alex@samba.(none)> 23596Date: Mon Dec 17 20:07:32 2007 -0500 23597 23598 Merge branch 'atombios-support' of git+ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support 23599 23600commit cf685f37ec874f0aacd09e7c4eb0402c6daec1b0 23601Merge: 2a134af0 44d07c4c 23602Author: Alex Deucher <alex@samba.(none)> 23603Date: Mon Dec 17 20:07:07 2007 -0500 23604 23605 Merge branch 'master' of git+ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support 23606 23607 merge master and fix conflicts 23608 23609commit 29706ca585ebd9b7b3521521a17016322e9ecccb 23610Author: Dave Airlie <airlied@linux.ie> 23611Date: Tue Dec 18 10:55:38 2007 +1000 23612 23613 fixup shadow setup on !r600 23614 23615commit 44d07c4ccce9acb5bd21a17acb082e91f7225764 23616Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23617Date: Mon Dec 17 18:56:12 2007 -0500 23618 23619 RADEON: typo from last commit 23620 23621commit 4da3782239921eb377216d4de4a9cc5bb55e0e8a 23622Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23623Date: Mon Dec 17 18:51:31 2007 -0500 23624 23625 RADEON: add output enable masks 23626 23627 add output enable masks for outputs that drive 23628 more than one connector. Make sure we don't turn off 23629 an output that's driving another connector. 23630 23631commit 5c5d2d19b2b032a06dd333b4ecc029aac342fb93 23632Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23633Date: Mon Dec 17 18:15:55 2007 -0500 23634 23635 RADEON: whitespace clean-ups 23636 23637commit 9f1d8220315c8894a17f2cc328025dc682b0c6e0 23638Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23639Date: Mon Dec 17 18:04:05 2007 -0500 23640 23641 RADEON: more PLL fixes 23642 23643 - reduce the calculation accuracy 23644 - certain LVDS panels seem to only like certain ref_divs 23645 - add pll flags to handle special cases 23646 - adjust the pll limits on legacy cards 23647 23648commit 2a134af01bc85de758ad83a85b23d1bba4a1a0f5 23649Author: Dave Airlie <airlied@redhat.com> 23650Date: Mon Dec 17 15:00:36 2007 +1000 23651 23652 r600: add shadow support to r600 driver to at least make 2d useable 23653 23654commit 614414611a9f246cbc74f579a79987fff97cf571 23655Author: Dave Airlie <airlied@redhat.com> 23656Date: Mon Dec 17 11:10:14 2007 +1000 23657 23658 radeon: cleanup pitch calculation and make r600 work again 23659 23660commit 79a375dbc7f323e2f551490a35f44ec36bed877c 23661Author: George Wu <geo@flood.OCF.Berkeley.EDU> 23662Date: Mon Dec 17 10:55:36 2007 +1000 23663 23664 r600: might as well fix VT for R600 23665 23666commit bc213ee723a45f2c396b4ed211a50f7642349973 23667Author: Alex Deucher <alex@samba.(none)> 23668Date: Sun Dec 16 14:54:00 2007 -0500 23669 23670 RADEON: fix sclock printout 23671 23672commit 4747c1f3cd4167b6a51d4864a297719ea48b9346 23673Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23674Date: Sun Dec 16 14:07:29 2007 -0500 23675 23676 RADEON: Make sure LVDS_EN bit is set when enabling LVDS 23677 23678commit a9817b2cf436a536dbc43ad77abc3bdcc53d346d 23679Author: Alex Deucher <alex@samba.(none)> 23680Date: Sat Dec 15 20:51:53 2007 -0500 23681 23682 RADEON: clean up units in PLL calculation 23683 23684commit b653e5a628bfa4dfb168e96f93f41eb910f409fb 23685Author: Michel Dänzer <michel@tungstengraphics.com> 23686Date: Sat Dec 15 00:50:10 2007 +0100 23687 23688 radeon: Default to 1x again with non-v3 AGP cards. 23689 23690 Seems more reliable in general than what was set up by firmware - fingers 23691 crossed... 23692 23693commit 6229825fa5d6715569098afbb21a40f7a2e7e6be 23694Author: Michel Dänzer <michel@tungstengraphics.com> 23695Date: Sat Dec 15 00:48:26 2007 +0100 23696 23697 radeon: Warning fixes. 23698 23699 The lid detection code probably wouldn't work on other non-x86 platforms 23700 though... 23701 23702commit 818ccf0fd4b5879171c5f20526d5a58638f8fde5 23703Author: Fredrik Höglund <fredrik@kde.org> 23704Date: Fri Dec 14 23:56:12 2007 +0100 23705 23706 RADEON: Fix the vertex coordinates for transformed pictures 23707 23708 This partially fixes transformed pictures on R100/R200 based 23709 cards. The texture still doesn't appear to be clamped correctly, 23710 but since that doesn't matter for rotations at perpendicular 23711 angles, I'm committing this now so randr rotation and reflection 23712 will work properly. 23713 23714commit 3cfbcf4cafbdfdb33411d16e51fb1f77cd0f93dd 23715Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23716Date: Fri Dec 14 17:11:00 2007 -0500 23717 23718 RADEON: Fix PLL set up on certain notebooks 23719 23720 Some LVDS panels require specific PLL dividers as 23721 specified in the bios tables. Make sure to use them 23722 if the output is LVDS. 23723 23724commit a84d446fd301d456bcea8f7abdc52e5a30776412 23725Author: Alex Deucher <alex@botch2.(none)> 23726Date: Fri Dec 14 02:17:14 2007 -0500 23727 23728 RADEON: select fb_div0 for LVDS on RV410 (x700) mobility 23729 23730 Fixes bug 8038 23731 I wonder if desktop RV410 need a similar fix. 23732 If your x700 laptop panel has problems after this let me know. 23733 23734commit b3eed3d87f76779b5a62a3115f99a31484dc38e0 23735Author: Alex Deucher <alex@samba.(none)> 23736Date: Fri Dec 14 00:20:10 2007 -0500 23737 23738 RADEON: fix typo in previous cursor fix 23739 23740commit 814c6c48aebba2e45ce257289b922cd7e92caf2a 23741Author: Alex Deucher <alex@samba.(none)> 23742Date: Thu Dec 13 18:45:09 2007 -0500 23743 23744 RADEON: rework PLL calculation 23745 23746 - Take into account the limits from the bios tables 23747 - Unify the PLL calculation between legacy and avivo chips 23748 23749commit f5ac34983411e4c4f41ab1817dce582830f398fd 23750Merge: f2b2e080 6ccf5b33 23751Author: Alex Deucher <alex@samba.(none)> 23752Date: Wed Dec 12 22:37:44 2007 -0500 23753 23754 Merge branch 'master' of git+ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support 23755 23756 merge and fix conflicts 23757 23758commit f2b2e0804183b52d9d3f56ad85b3552ece76c544 23759Author: Alex Deucher <alex@samba.(none)> 23760Date: Wed Dec 12 22:18:37 2007 -0500 23761 23762 RADEON: fix rotation on avivo chips 23763 23764 There are still some issues, but this is better than before. 23765 23766commit 6ccf5b33d27218ae1c45ab93c122438ed536d8ba 23767Author: Alex Deucher <alex@botch2.(none)> 23768Date: Wed Dec 12 20:12:06 2007 -0500 23769 23770 RADEON: only enable vblanks if we want them 23771 23772 should fix bug 13610 23773 23774commit 1668f2056f56370f1b5681c13f1e14904e301216 23775Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23776Date: Wed Dec 12 19:39:08 2007 -0500 23777 23778 RADEON: use /proc/acpi to determine lid status 23779 23780 Linux only 23781 23782commit 33a39947f7f79533cd90007a17d57b20126642c6 23783Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23784Date: Wed Dec 12 18:50:18 2007 -0500 23785 23786 RADEON: fix cursors when using rotation 23787 23788 allocate separate cursor buffers for each crtc 23789 23790commit 9e5efdecd12092031a4aebce58747cb4a6f48f28 23791Author: Arkadiusz Miskiewicz <arekm@maven.pl> 23792Date: Tue Dec 11 23:53:03 2007 -0500 23793 23794 sparse fixes and cleanups from arekm 23795 23796commit 372bf41818fdafc6a9d2914aee3a8e359f668f02 23797Author: Alex Deucher <alex@botch2.(none)> 23798Date: Tue Dec 11 14:04:58 2007 -0500 23799 23800 RADEON: handle HMDI properly (untested) and fix some merge leftovers 23801 23802commit 3c22ad977c25d5ca2811821fcac6bb8ecd79994a 23803Merge: c9a0cee9 f3d2ec3a 23804Author: Alex Deucher <alex@botch2.(none)> 23805Date: Tue Dec 11 13:11:15 2007 -0500 23806 23807 Merge branch 'master' of ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support 23808 23809 fix conflicts 23810 23811commit f3d2ec3a5ae61215c792018320158750e7aa937c 23812Author: Alex Deucher <alex@botch2.(none)> 23813Date: Tue Dec 11 11:57:27 2007 -0500 23814 23815 RADEON: rewrite PLL computation 23816 23817 Algorithm adapted from BeOS radeon driver with some tweaks by me. 23818 Actually calulate and use the reference divider rather than using the bios default. 23819 Also, always calculate the PLL, rather than falling back to bios dividers. 23820 This should fix bugs 12913, 13590, 13533, possibly others. 23821 23822commit c9a0cee97ca69e8fe1e1937c7670fa903214cded 23823Author: Dave Airlie <airlied@linux.ie> 23824Date: Tue Dec 11 06:03:46 2007 +1000 23825 23826 more zaphod fixes - some other work maybe needed 23827 23828commit 9b125312ab6edc585e4f5931a6a6de81e13b6acc 23829Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23830Date: Mon Dec 10 13:53:15 2007 -0500 23831 23832 RADEON: only update crtc values when RMX is active 23833 23834commit 3a161e1b5d80361e318ced8da5c19e797749d693 23835Author: Alex Deucher <alex@botch2.(none)> 23836Date: Mon Dec 10 00:57:26 2007 -0500 23837 23838 RADEON: bios PLL cleanup 23839 23840commit 731830297f2fc4a416882aacfb0d9b5f8ed32520 23841Author: Dave Airlie <airlied@linux.ie> 23842Date: Mon Dec 10 15:50:38 2007 +1000 23843 23844 fixup some warnings 23845 23846commit 2818e2b02ca90c9dfa50905b5311b2ae83ac3b0c 23847Author: Dave Airlie <airlied@linux.ie> 23848Date: Mon Dec 10 15:43:52 2007 +1000 23849 23850 add more to configure.ac for using out-of-tree mode src 23851 23852commit cc3c36100986f9d8060bc2d433373d4806f8e730 23853Author: Dave Airlie <airlied@linux.ie> 23854Date: Mon Dec 10 15:25:56 2007 +1000 23855 23856 add support for building against legacy servers similiar to Intel codepaths 23857 23858commit 9c278cb7fa7f18d13bde053fd75221cfba9da377 23859Merge: 6451ea2d cc167b9b 23860Author: Dave Airlie <airlied@linux.ie> 23861Date: Mon Dec 10 15:18:03 2007 +1000 23862 23863 Merge branch 'zaphod-lolz' of git://git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support 23864 23865 Conflicts: 23866 23867 src/radeon.h 23868 src/radeon_crtc.c 23869 src/radeon_driver.c 23870 src/radeon_output.c 23871 23872commit 6451ea2dcc4fac762442f699935864f4a8d445f7 23873Merge: 0d89556b 5896ca40 23874Author: Dave Airlie <airlied@linux.ie> 23875Date: Mon Dec 10 15:08:42 2007 +1000 23876 23877 Merge branch 'master' into atombios-support 23878 23879commit 5896ca4097d439f59f90f397939132c061c3c364 23880Author: LisaWu <liswu@ati.com> 23881Date: Fri Dec 7 09:45:05 2007 +0100 23882 23883 radeon: Use %u instead of %d for unsigned value. 23884 23885commit df44f8380268c27d3978c4e91d736f093322b8b8 23886Author: Michel Dänzer <michel@tungstengraphics.com> 23887Date: Fri Dec 7 09:41:47 2007 +0100 23888 23889 radeon: Use gettimeofday instead of xf86getsecs. 23890 23891commit cc167b9bb7f1c3b8579e51e7bc2fca2f8eba6bd1 23892Author: Dave Airlie <airlied@redhat.com> 23893Date: Fri Dec 7 15:41:36 2007 +1000 23894 23895 disable tiling for zaphod heads 23896 23897commit 2ce8d192533a8c795714c5a9fb308ec74db40287 23898Author: Dave Airlie <airlied@redhat.com> 23899Date: Fri Dec 7 15:35:21 2007 +1000 23900 23901 don't add fboffset to info->FB it already is mapped at the offset 23902 23903commit 0dcd926d3092100854b3e362d6659d4950508aeb 23904Author: Dave Airlie <airlied@redhat.com> 23905Date: Fri Dec 7 14:45:04 2007 +1000 23906 23907 radeon: bring back zaphod all is forgiven. 23908 23909 You've whined, you've cried, you've nagged, and you're guilt trippin has 23910 made me do it... It actually wasn't as hard as I thought it would be. 23911 23912 Still not perfect, couple of things to fix yet 23913 23914commit bb5ede557bf32a42eef158ff0fbcfe1c6ede098a 23915Author: Dave Airlie <airlied@redhat.com> 23916Date: Fri Dec 7 14:30:32 2007 +1000 23917 23918 radeon: move savedreg/modereg into entity instead of info 23919 23920commit 64ab1cdf343a9a69e7e9e64f0bba77c54a94e9d0 23921Author: James Cloos <cloos@jhcloos.com> 23922Date: Thu Dec 6 15:51:12 2007 -0500 23923 23924 Add missing PHONY line for automatic ChangeLog generation 23925 23926commit 0d89556bfa41a3acbd6afe85b062e3a21f2ca057 23927Author: Dave Airlie <airlied@ppcg5.localdomain> 23928Date: Thu Dec 6 19:23:06 2007 +1100 23929 23930 powerpc: build fixes from last merge 23931 23932commit dbe3d2608ecc9896db9c23b3a347b50748c51e13 23933Merge: 48e31cda 21ed4353 23934Author: Dave Airlie <airlied@redhat.com> 23935Date: Thu Dec 6 14:22:03 2007 +1000 23936 23937 Merge branch 'master' into atombios-support 23938 23939 Conflicts: 23940 23941 src/radeon_output.c 23942 23943commit 21ed435398e4a398dd8a0a5d7c1d4cc45e916332 23944Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23945Date: Tue Dec 4 17:08:58 2007 -0500 23946 23947 RADEON: add MacModel imac-g5-isight for iMac G5 iSight 23948 23949 Thanks to Étienne Bersac for helping to figure this out. 23950 23951commit 54bfd522405d9fdfb69d3a59e111ac3d63483dbb 23952Author: Étienne Bersac <bersace03@laposte.net> 23953Date: Tue Dec 4 14:22:42 2007 -0500 23954 23955 RADEON: fix typo 23956 23957commit 5022d006cfc06ca0395981526b2c2c94c6878567 23958Author: Michel Dänzer <michel@tungstengraphics.com> 23959Date: Sun Dec 2 17:27:33 2007 +0100 23960 23961 radeon: Further XVideo fixes. 23962 23963 * Make sure pitch constraints are always met for DMA upload blits. 23964 * RGB24 is not affected by endianness. 23965 23966commit 6ed55b70b23dfdc7b41103ea59c1df2bda5e41e6 23967Author: Kusanagi Kouichi <slash@ma.neweb.ne.jp> 23968Date: Sun Dec 2 17:18:46 2007 +0100 23969 23970 radeon: Fix crash with XVideo 24bit RGB images. 23971 23972 See https://bugs.freedesktop.org/show_bug.cgi?id=13274 . 23973 23974commit a697b590899bb7704ec4d7ae9a9c3cbbfcaef382 23975Author: Michel Dänzer <michel@tungstengraphics.com> 23976Date: Sun Dec 2 17:11:20 2007 +0100 23977 23978 Fix build against xserver master. 23979 23980 (DE)ALLOCATE_LOCAL are gone. 23981 23982commit 48e31cdaa0caa21573879af5b9267773fe89176a 23983Author: George Wu <geo@ocf.berkeley.edu> 23984Date: Sun Dec 2 15:25:09 2007 +1000 23985 23986 RADEON/R600: small code cleanup 23987 23988commit 00b4480aa2c5d7f751e34fc964f431b90b14c8d2 23989Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 23990Date: Sat Dec 1 14:18:40 2007 -0500 23991 23992 RADEON: add options for force TV out as detected and to set TV standard 23993 23994 Also fix a typo in internal tv-out parsing 23995 23996commit 1e029fef5fe264f2ced445b80bf6070abcb84b82 23997Author: Alex Deucher <alex@samba.(none)> 23998Date: Sat Dec 1 00:58:51 2007 -0500 23999 24000 RADEON: move GPIO lookup to a separate function 24001 24002commit dcbef1ba9dfcf35c28e058832a55adf00afb472e 24003Author: Alex Deucher <alex@samba.(none)> 24004Date: Sat Dec 1 00:35:25 2007 -0500 24005 24006 RADEON: fix typo in previous commit 24007 24008 check gpio table revision before connector table revision 24009 24010commit fdce0598a2228c48c84deae1d7bebb2d7b3e979b 24011Author: Alex Deucher <alex@samba.(none)> 24012Date: Sat Dec 1 00:15:34 2007 -0500 24013 24014 RADEON: convert atombios connector table parsing to use ATOM structs 24015 24016 convert and add hpd info 24017 24018commit e3d7de9cc956aec5f940ad6db09e826b3a69523a 24019Author: Alex Deucher <alex@botch2.(none)> 24020Date: Fri Nov 30 20:14:42 2007 -0500 24021 24022 RADEON: remove unused cruft from last atom import 24023 24024commit d5d83411e8a884154d671aad440524507cce313e 24025Author: Alex Deucher <alex@botch2.(none)> 24026Date: Fri Nov 30 20:11:42 2007 -0500 24027 24028 RADEON: save/restore avivo crtc cursor control 24029 24030 this should prevent the cursor from showing up on in text 24031 mode or vesafb etc. after running the driver. 24032 24033commit af0196f7bf0d1d5d211391149c18935d64ed2b06 24034Merge: d9858a2b 0175b799 24035Author: Alex Deucher <alex@botch2.(none)> 24036Date: Fri Nov 30 16:40:28 2007 -0500 24037 24038 Merge branch 'master' of ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support 24039 24040 merge master and fix conflicts 24041 24042commit d9858a2b3744b99003cfb9f31b743a2d31b322e9 24043Author: Dave Airlie <airlied@linux.ie> 24044Date: Sat Dec 1 06:49:59 2007 +1000 24045 24046 radeon: add in pll spread spectrum workaround 24047 24048commit 0175b79987ef4d7b0ce8238c3bdde989e504516a 24049Author: Alex Deucher <alex@botch2.(none)> 24050Date: Fri Nov 30 15:37:42 2007 -0500 24051 24052 RADEON: rework MacModel option 24053 24054 this brings in some previous research from Michel Dänzer, 24055 Sjoerd Simons, and myself. Hopefully, the driver will pick 24056 the correct MacModel in more cases. This also changes the 24057 default connector table for desktop Macs to dual DVI rather 24058 than DVI+VGA as that seems to be the case more often than not. 24059 External TMDS chips are handled separately now as well. 24060 Eventually we should add an option to allow the user to specify 24061 what external TMDS chip they need, but we don't have enough info 24062 yet, so we'll rely on OF to init the external chip in most cases 24063 for now. 24064 24065commit e1945f1f25a34310bd58ce128c8ff27ecc985618 24066Merge: b368b0f2 df7777bf 24067Author: Alex Deucher <alex@botch2.(none)> 24068Date: Fri Nov 30 14:30:55 2007 -0500 24069 24070 Merge branch 'atombios-support' of ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support 24071 24072commit b368b0f22cd1d7ef9b4c65d82929c76f3b82d573 24073Author: Alex Deucher <alex@botch2.(none)> 24074Date: Fri Nov 30 14:29:27 2007 -0500 24075 24076 RADEON: disable atom pll set for r4xx cards 24077 24078 the clocks do not get set correctly in all cases. this needs 24079 further investigation. 24080 24081commit 5af15739571c09260750bcfd3620e16fd7fec862 24082Author: Alex Deucher <alex@botch2.(none)> 24083Date: Fri Nov 30 14:24:30 2007 -0500 24084 24085 RADEON: small cleanup of pll code 24086 24087commit df7777bff40c1feabcc12d2148ad6ac5213efbb3 24088Author: George Wu <geo@ocf.berkeley.edu> 24089Date: Fri Nov 30 17:49:33 2007 +1000 24090 24091 Add LVTMA PWRSEQ registers to fix VT switching for LVDS 24092 24093commit 9840a0fd4fc8c980533fcd4a02c55cd0d5634b6d 24094Author: Alex Deucher <alex@botch2.(none)> 24095Date: Thu Nov 29 13:27:37 2007 -0500 24096 24097 RADEON: add MacModel "mini-internal" for minis with internal TMDS 24098 24099 Some macs (minis and powerbooks) use internal tmds, others use external tmds 24100 and not just for dual-link TMDS, it shows up with single-link as well. 24101 Unforunately, there doesn't seem to be any good way to figure it out. 24102 24103commit 017c939cf0a2b12fbdc1681cc70c28b23ae3b397 24104Author: Alex Deucher <alex@samba.(none)> 24105Date: Thu Nov 29 02:52:14 2007 -0500 24106 24107 RADEON: implement CLUT adjust support 24108 24109commit 9963b0fe01feb6dd0cb555b874a48f6fa3b255cb 24110Author: Alex Deucher <alex@samba.(none)> 24111Date: Thu Nov 29 00:46:23 2007 -0500 24112 24113 RADEON: fix cursor offset on avivo chips 24114 24115commit 6c56e3d7655b17e93e8823aefe34b05291104695 24116Author: Alex Deucher <alex@botch2.(none)> 24117Date: Tue Nov 27 15:27:36 2007 -0500 24118 24119 RADEON: switch r4xx to atombios load detection 24120 24121 works great 24122 24123commit 7561242e5b79bc2798ca3aace2b79e1a36949488 24124Author: Alex Deucher <alex@botch2.(none)> 24125Date: Tue Nov 27 14:50:36 2007 -0500 24126 24127 RADEON: re-org load detection for legacy chips 24128 24129commit febdcc2dccd42acbcd68ae630b7811cae5c58e8a 24130Author: Dave Airlie <airlied@linux.ie> 24131Date: Wed Nov 28 05:10:57 2007 +1000 24132 24133 legacy: fix fb/agp read/writes 24134 24135commit bb8545146959b748994be055d5b3de66ec66c8b2 24136Author: Alex Deucher <alex@samba.(none)> 24137Date: Mon Nov 26 17:34:51 2007 -0500 24138 24139 RADEON: first pass at TV/Component video 24140 24141 Untested and not likely to work just yet. 24142 24143commit e2bde646b864dca9056d9ecfe23a0d905647ea9a 24144Author: Alex Deucher <alex@samba.(none)> 24145Date: Mon Nov 26 14:35:57 2007 -0500 24146 24147 RADEON: move crtc output source selection into atombios_output.c 24148 24149 The function fits better as an output function and should now 24150 work with clones as well. 24151 24152commit 16c9be4107678a2a58d3418b7f1cc94d695ca8d6 24153Author: Alex Deucher <alex@samba.(none)> 24154Date: Mon Nov 26 14:20:54 2007 -0500 24155 24156 RADEON: add default connector table for avivo chips 24157 24158commit 6f080d00e6f4f84d5e0d6b4eff302bf42c230e81 24159Author: Arkadiusz Miskiewicz <arekm@maven.pl> 24160Date: Mon Nov 26 12:43:30 2007 -0500 24161 24162 RADEON: fix backlight control on some laptops 24163 24164 It seems the bios scratch regs are involved in backlight control 24165 on some laptops. This patch fixes the problematic laptops and doesn't 24166 seem to break the previous bios lid and output control fixes. 24167 24168commit dcf22aed87366f4625fb5042cb84fecccd9ceece 24169Author: Alex Deucher <alex@botch2.(none)> 24170Date: Mon Nov 26 11:10:03 2007 -0500 24171 24172 RADEON: only return status unknown for XPRESS chips 24173 24174 this seems to cause more issues than it attempted to fix 24175 so limit it to XPRESS chips for now. 24176 24177commit 206e280f02324641b4fe5a1986e26adf0e021fd4 24178Author: Alex Deucher <alex@botch2.(none)> 24179Date: Mon Nov 26 09:39:27 2007 -0500 24180 24181 RADEON: fix typo in man page 24182 24183commit 4e792db655dc92d2864db36b4d8f6714908de8e8 24184Author: Dave Airlie <airlied@redhat.com> 24185Date: Fri Nov 23 15:44:44 2007 +1000 24186 24187 r500: set default minimum pll freq 24188 24189commit a13b4c69c105c096dd05e6de2c5c154c9a8bcc71 24190Author: Dave Airlie <airlied@redhat.com> 24191Date: Fri Nov 23 15:25:06 2007 +1000 24192 24193 r5xx: cleanup pll code.. 24194 24195 Clean the PLL code to use the radeon pll structs. 24196 24197commit 5d792b5673bbf4784eb0ec059221e9b57232a122 24198Author: Dave Airlie <airlied@redhat.com> 24199Date: Fri Nov 23 15:03:13 2007 +1000 24200 24201 radeon: fix up memory mapping issues for vt switch 24202 24203commit 558a2ef266c1ca517c7fb464b0ccfef83238c913 24204Author: Dave Airlie <airlied@redhat.com> 24205Date: Fri Nov 23 14:39:32 2007 +1000 24206 24207 fix silly spaces 24208 24209commit 1bda4424a7031de437acfca9c41d4b3668e36051 24210Author: Dave Airlie <airlied@redhat.com> 24211Date: Fri Nov 23 14:39:19 2007 +1000 24212 24213 r600: add hi agp address for mc 24214 24215commit 133234c31a294f24a3968a576aad2bb8b89d0f6a 24216Author: Dave Airlie <airlied@redhat.com> 24217Date: Fri Nov 23 14:15:18 2007 +1000 24218 24219 atombios: use values from object header 24220 24221commit dbf6eae7e7a4bd1bc60fefdc7ab6276ed3f097c4 24222Author: Dave Airlie <airlied@redhat.com> 24223Date: Fri Nov 23 11:55:05 2007 +1000 24224 24225 atombios: add initial object header parsing for r600 cards 24226 24227commit 6b103915c11fc79d2efc43c44fc2a00c3bc64ede 24228Author: Dave Airlie <airlied@redhat.com> 24229Date: Fri Nov 23 09:24:20 2007 +1000 24230 24231 r500: make vt switch work again for me 24232 24233commit d24208276aad7669feeed527dced60c76d39eae6 24234Author: Dave Airlie <airlied@redhat.com> 24235Date: Fri Nov 23 08:50:25 2007 +1000 24236 24237 avivo: fix typo in register saving 24238 24239commit 197a62704742a4a19736c2637ac92d1dc5ab34ed 24240Author: Adam Jackson <ajackson@redhat.com> 24241Date: Thu Nov 22 20:26:23 2007 +1000 24242 24243 radeon: fix openoffice/render bug on r100 chips 24244 24245commit 64010fc4eae8359c01e430f64252931242c91435 24246Author: Dave Airlie <airlied@linux.ie> 24247Date: Thu Nov 22 20:25:31 2007 +1000 24248 24249 Revert "Disable RENDER acceleration by default on some RV200 chips." 24250 24251 This reverts commit 145da701bf4fb9c0ad9f95620b20914ae0126852. 24252 24253 pull in fix from ajax next commit 24254 24255commit 145da701bf4fb9c0ad9f95620b20914ae0126852 24256Author: Stefan Dirsch <sndirsch@suse.de> 24257Date: Thu Nov 22 08:38:09 2007 +0100 24258 24259 Disable RENDER acceleration by default on some RV200 chips. 24260 24261 Novell Bug #341699: Render acceleration is known to be broken 24262 on at least "Radeon 7500 QW" and "Radeon Mobility M7 LW". 24263 24264commit e810c3ae9908cd57e95b1b091cded87cbfc12fdc 24265Author: Roland Scheidegger <sroland@tungstengraphics.com> 24266Date: Thu Nov 22 02:37:55 2007 +0100 24267 24268 really do not set up surface regs for depth buf on r100-class igps (bug #13080) 24269 24270 fix the if condition testing for these chips... 24271 24272commit 0cc7e94849c1525750fabd04cf58f2dee88372e0 24273Author: Alex Deucher <alex@botch2.(none)> 24274Date: Wed Nov 21 17:06:17 2007 -0500 24275 24276 RADEON: reorder crtc/pll setup 24277 24278commit d56bde98efceaa8344e62f8e98db90c4bb642331 24279Author: Alex Deucher <alex@botch2.(none)> 24280Date: Wed Nov 21 17:03:39 2007 -0500 24281 24282 RADEON: fix crtc to output routing 24283 24284 Thanks to AMD for the information 24285 24286commit a12390c832abe423def60e39cd5a9118e5910339 24287Merge: d5317922 e74dca19 24288Author: Alex Deucher <alex@botch2.(none)> 24289Date: Wed Nov 21 02:24:48 2007 -0500 24290 24291 Merge branch 'atombios-support' of git://git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support 24292 24293commit d5317922f29a57b6c4127826a2fc126c5fd7c117 24294Author: Alex Deucher <alex@botch2.(none)> 24295Date: Wed Nov 21 02:23:37 2007 -0500 24296 24297 RADEON: attempt to fix crtc to output routing 24298 24299 The output routing seems to be based on the output ids from the bios 24300 connector tables and the connected status in the bios scratch regs. 24301 I don't fully understand this yet, but this seems to work 24302 for the most part, however changing modes can sometimes 24303 lead to a blanked head. This can be remedied by forcing 24304 a dpms off cycle. 24305 24306commit 81ce299bffd75540925b4c8234adf11226147165 24307Author: Alex Deucher <alex@botch2.(none)> 24308Date: Wed Nov 21 01:35:44 2007 -0500 24309 24310 RADEON: provide clearer debugging info 24311 24312commit e74dca19416b13f97db9d1fc06299b988057d6a4 24313Author: Dave Airlie <airlied@redhat.com> 24314Date: Wed Nov 21 16:24:25 2007 +1000 24315 24316 re-enable mobility chips 24317 24318commit 9c5b813dd6b3492cbc9833bc59792a5cec457e51 24319Author: Alex Deucher <alex@botch2.(none)> 24320Date: Wed Nov 21 01:22:42 2007 -0500 24321 24322 RADEON: major re-org and clean up of atom output control 24323 24324 - use radeon_output->devices to determine output 24325 - clean up and simplify dpms and mode set 24326 24327commit 7634cb6b96f938bc6615eb2c49ae75aaefd04cce 24328Author: Alex Deucher <alex@botch2.(none)> 24329Date: Wed Nov 21 00:10:14 2007 -0500 24330 24331 RADEON: make naming consistent and remove some cruft 24332 24333commit 908748343fc9a6cdc38af0fc028c63a82766da3f 24334Author: Alex Deucher <alex@botch2.(none)> 24335Date: Wed Nov 21 00:05:42 2007 -0500 24336 24337 RADEON: store devices ids from bios for each driver output 24338 24339commit 3975da2ea8cb628f7f66c3f26c5dfa181cd1c532 24340Merge: e283aa33 295ce277 24341Author: Alex Deucher <alex@botch2.(none)> 24342Date: Tue Nov 20 23:52:29 2007 -0500 24343 24344 Merge branch 'atombios-support' of git://git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support 24345 24346commit e283aa332adf0134243a4fa3d14263719cd8a3fd 24347Author: Alex Deucher <alex@botch2.(none)> 24348Date: Tue Nov 20 23:35:46 2007 -0500 24349 24350 RADEON: add LVDS atom support 24351 24352commit e4bc3e1e7bb45571367d41b5328ff2590810b0f9 24353Author: Alex Deucher <alex@botch2.(none)> 24354Date: Tue Nov 20 18:09:29 2007 -0500 24355 24356 RADEON: enable/disable the right TMDS controller 24357 24358commit 3e47683ffaa44a89cda9bcddf530643befb27efa 24359Author: Alex Deucher <alex@botch2.(none)> 24360Date: Tue Nov 20 18:01:15 2007 -0500 24361 24362 RADEON: fixup for bios tables with wrong connector types 24363 24364commit 7412952eb1d1e9857cdab8417f7305f676900827 24365Author: Alex Deucher <alex@botch2.(none)> 24366Date: Tue Nov 20 18:00:12 2007 -0500 24367 24368 RADEON: switch to using ATOM defines for bios device table 24369 24370commit 295ce277bb0a44b9539b3dba575e7aff279dc2d0 24371Author: Dave Airlie <airlied@redhat.com> 24372Date: Wed Nov 21 08:56:40 2007 +1000 24373 24374 add missing files for make dist 24375 24376 noted by ndim on #radeonhd 24377 24378commit 20083b0695987b25e442ecbdec24f3cb6f1ac2ae 24379Author: Dave Airlie <airlied@redhat.com> 24380Date: Wed Nov 21 08:53:44 2007 +1000 24381 24382 LVDS on r500/r600 needs some work disable mobile chips for now 24383 24384commit e4b8a4479ddea9b083b3a763dc0b9302e7b9a82a 24385Author: Dave Airlie <airlied@redhat.com> 24386Date: Wed Nov 21 08:06:12 2007 +1000 24387 24388 r600: add memory controller regs from AMD 24389 24390commit aa88da974b97ea1e9bbb47b3494543575c09d912 24391Author: Dave Airlie <airlied@redhat.com> 24392Date: Wed Nov 21 08:01:35 2007 +1000 24393 24394 rs690 is !> r600 24395 24396commit c8872603454e6a4ffed9fc7d9adc2c364a429608 24397Author: Dave Airlie <airlied@clockmaker.usersys.redhat.com> 24398Date: Tue Nov 20 22:33:39 2007 +1000 24399 24400 radeon: restructure pci ids to avoid effort later 24401 24402 This uses a single file with all the pciids and parameters for radeon 24403 family and setup in it. I don't run the perl script at build time to avoid 24404 a perl dependency on build but adding pci ids should be done via the csv file 24405 with openoffice or gnumeric if possible. 24406 24407commit a5b34c2f1f7d5346c4489cb30e07291d1217026c 24408Author: Dave Airlie <airlied@redhat.com> 24409Date: Tue Nov 20 16:37:29 2007 +1000 24410 24411 r600: use standard memory controller setup paths 24412 24413commit 4a523da5221d53f2efa49da2326500e9b0b9f14d 24414Author: Dave Airlie <airlied@redhat.com> 24415Date: Tue Nov 20 15:31:11 2007 +1000 24416 24417 r600: get r600 to work non-accelerated. 24418 24419 DDC still not working yet 24420 24421commit 80023441ba46882bc810ff3790c7148059f155f5 24422Author: Dave Airlie <airlied@redhat.com> 24423Date: Tue Nov 20 14:10:23 2007 +1000 24424 24425 r600: block r600 startup due to lack of memory controller info 24426 24427commit f6fbbacc17bf9b1073d3e993b225987fd9173182 24428Author: Dave Airlie <airlied@redhat.com> 24429Date: Tue Nov 20 13:41:55 2007 +1000 24430 24431 atombios: add warnings for r500 and r600 24432 24433commit 45a8b083c123b820c008f04ab857a64a8facec14 24434Author: Dave Airlie <airlied@redhat.com> 24435Date: Tue Nov 20 13:37:00 2007 +1000 24436 24437 atombios: add all r5xx and r6xx pci ids 24438 24439commit 5d023e2c3c2ab44ea57ffadc9607025d602c376c 24440Merge: 0d1e0c78 c8872603 24441Author: Dave Airlie <airlied@redhat.com> 24442Date: Tue Nov 20 13:02:43 2007 +1000 24443 24444 Merge branch 'master' into atombios-support 24445 24446 Conflicts: 24447 24448 src/radeon_chipset.h 24449 src/radeon_driver.c 24450 src/radeon_probe.c 24451 24452commit 0d1e0c7805b3d8e56ccb49465e6b144afb7bdc51 24453Author: Dave Airlie <airlied@redhat.com> 24454Date: Tue Nov 20 09:08:04 2007 +1000 24455 24456 r5xx: add 71c5 for macbook pro 24457 24458commit d5909b30595c103bb5f42cd1704330f944bba49c 24459Author: Dave Airlie <airlied@redhat.com> 24460Date: Tue Nov 20 08:15:58 2007 +1000 24461 24462 r5xx: cleanups after last merge 24463 24464commit fe2f7a09050fb7a345a1f52239f8f3c4f1053891 24465Merge: 744c8cb6 49055d8a 24466Author: Dave Airlie <airlied@redhat.com> 24467Date: Tue Nov 20 08:04:32 2007 +1000 24468 24469 Merge branch 'master' into agd-atom-merge 24470 24471 Conflicts: 24472 24473 src/radeon_cursor.c 24474 src/radeon_output.c 24475 24476commit 744c8cb6c293fcaa687566f52901644e699baace 24477Merge: e258fbe4 e530af79 24478Author: Dave Airlie <airlied@redhat.com> 24479Date: Tue Nov 20 07:56:33 2007 +1000 24480 24481 Merge branch 'agd-atom' of ssh://people.freedesktop.org/~agd5f/xf86-video-ati-atom into agd-atom 24482 24483commit e258fbe411d255a1044b61d7ff738aee3fb5b7f4 24484Author: Dave Airlie <airlied@redhat.com> 24485Date: Mon Nov 19 16:35:05 2007 +1000 24486 24487 makes 2-headed cursor work 24488 24489commit e530af79adf51b3e95a0eca676c915a34dcbf4a7 24490Merge: 69e197f2 52aba8d7 24491Author: Alex Deucher <alex@botch2.(none)> 24492Date: Mon Nov 19 00:59:30 2007 -0500 24493 24494 Merge branch 'agd-atom' of /home/alex/git/airlied/xf86-video-ati2 into agd-atom 24495 24496commit 69e197f2c8002aacf2587754c8d3bd63c88f85b1 24497Merge: 5e8940fa 862dcabf 24498Author: Alex Deucher <alex@botch2.(none)> 24499Date: Mon Nov 19 00:57:34 2007 -0500 24500 24501 Merge branch 'agd-atom' of /home/alex/git/airlied/xf86-video-ati2 into agd-atom 24502 24503commit fca47ad083449f4cf9063dd970cdcebea6a7f110 24504Author: Dave Airlie <airlied@redhat.com> 24505Date: Mon Nov 19 15:53:40 2007 +1000 24506 24507 add z3ro's pciids 24508 24509commit 5e8940fa6e33d09091aa4bcf04b0f9e79596e1b8 24510Author: Alex Deucher <alex@botch2.(none)> 24511Date: Mon Nov 19 00:52:38 2007 -0500 24512 24513 fix logic in connector table check for TVs and switch counter to symbolic names 24514 24515commit c19123fd9483758eb6b286c3dffcb6d79d5b1ee5 24516Author: Dave Airlie <airlied@redhat.com> 24517Date: Mon Nov 19 15:46:58 2007 +1000 24518 24519 add firegl card on ajaxs machine 24520 24521commit f02f340e466a415b4e01648ca1e323f4ce125885 24522Author: Alex Deucher <alex@botch2.(none)> 24523Date: Mon Nov 19 00:39:19 2007 -0500 24524 24525 Don't assign a gpio for TV 24526 24527commit 52aba8d73189ba959f19c0437499d5e7a8829827 24528Merge: 862dcabf 5e8940fa 24529Author: Dave Airlie <airlied@redhat.com> 24530Date: Mon Nov 19 15:30:46 2007 +1000 24531 24532 Merge branch 'agd-atom' of ssh://people.freedesktop.org/~agd5f/xf86-video-ati-atom into agd-atom 24533 24534commit 862dcabfe0c10751d815e5cdd7436c10c2c2db10 24535Author: Dave Airlie <airlied@redhat.com> 24536Date: Mon Nov 19 15:30:08 2007 +1000 24537 24538 r520: nail i2c enable/disable issue 24539 24540commit f3dd7f413b670eeb6b8639f6677d72050ad5fe04 24541Author: Alex Deucher <alex@botch2.(none)> 24542Date: Mon Nov 19 00:19:39 2007 -0500 24543 24544 Don't detect TV out for now 24545 24546commit 94de0e22d7229ca71e18e1e849d8545d9ca7bafe 24547Author: Dave Airlie <airlied@redhat.com> 24548Date: Mon Nov 19 14:47:33 2007 +1000 24549 24550 i2c: fix bus enable stuff 24551 24552commit 3f1fc7eef13ea02fa5119e9b51d499841b801f2d 24553Author: Alex Deucher <alex@botch2.(none)> 24554Date: Mon Nov 19 00:02:14 2007 -0500 24555 24556 CRTs/DFPs may share a DVI port, but TV and CV don't 24557 24558commit 8f84c5ad4c4af14612ea68fe6f24d0d527f00acc 24559Author: Alex Deucher <alex@botch2.(none)> 24560Date: Sun Nov 18 23:43:06 2007 -0500 24561 24562 fix typo in loop 24563 24564commit 384cd8f52c89d089c6559e2eedbae45641fcd14e 24565Merge: f3f0e4ec 234b6073 24566Author: Dave Airlie <airlied@redhat.com> 24567Date: Mon Nov 19 14:02:55 2007 +1000 24568 24569 Merge branch 'agd-atom' of ../xf86-video-ati into agd-atom 24570 24571commit f3f0e4ec92c935c89ddb2f4241fe4335a521b439 24572Author: Alex Deucher <alex@botch2.(none)> 24573Date: Sun Nov 18 23:14:01 2007 -0500 24574 24575 RADEON: unify connectortype handling 24576 24577commit 234b6073054ac7630e82781683e666b94b2f12de 24578Author: Dave Airlie <airlied@redhat.com> 24579Date: Mon Nov 19 14:02:09 2007 +1000 24580 24581 restore avivo memory map registers at correct places 24582 24583commit 459a30ba511fe2fa8051380a9741fcfd9bb401ef 24584Author: Dave Airlie <airlied@redhat.com> 24585Date: Mon Nov 19 13:44:38 2007 +1000 24586 24587 fix type for r520 agp code 24588 24589commit 760af92412ef0d5cc44e52e7cec11fd80c4aaaeb 24590Author: Alex Deucher <alex@botch2.(none)> 24591Date: Sun Nov 18 22:34:59 2007 -0500 24592 24593 RADEON: unify DDC line handling 24594 24595commit e73bf6290da20dd61798ace775999ce1cb550934 24596Author: Dave Airlie <airlied@redhat.com> 24597Date: Mon Nov 19 13:32:16 2007 +1000 24598 24599 add x1900xt support 24600 24601commit 2e37937bacd624d616b91c41006c113791ebe98d 24602Author: Alex Deucher <alex@botch2.(none)> 24603Date: Sun Nov 18 20:18:50 2007 -0500 24604 24605 RADEON: step one in output rework 24606 24607 re-organize the output type 24608 24609commit 679e7a2e0d1b213524b8109193483bc9840fb116 24610Author: Alex Deucher <alex@botch2.(none)> 24611Date: Sun Nov 18 17:56:51 2007 -0500 24612 24613 Few fixes from the last commit. 24614 24615 Update parser works fine on r4xx. 24616 24617commit 1cd7cc3e6758ab1012f3ced6e958a1517f45557f 24618Author: Alex Deucher <alex@botch2.(none)> 24619Date: Sun Nov 18 17:44:36 2007 -0500 24620 24621 WIP: new atom code comples. 24622 24623 I commented out the object parsing for the time being as 24624 using it will require some thought as to new output 24625 related data structures. 24626 24627commit b155fa872ee4ca5d801e942aee6e619cef104f35 24628Author: Alex Deucher <alex@botch2.(none)> 24629Date: Sat Nov 17 00:34:56 2007 -0500 24630 24631 WIP: more new ATOM integration work 24632 24633commit 67db114d97abed7a607467e5d67c7b4ffa2c347e 24634Merge: 7d06a879 ea153465 24635Author: Alex Deucher <alex@botch2.(none)> 24636Date: Fri Nov 16 14:29:53 2007 -0500 24637 24638 Merge branch 'agd-atom' of /home/alex/git/airlied/xf86-video-ati2 into agd-atom 24639 24640commit ea1534659de87d3d75eb20d808d039cff22cb537 24641Author: Dave Airlie <airlied@linux.ie> 24642Date: Fri Nov 16 18:46:02 2007 +1000 24643 24644 avivo: fixup some i2c stuff 24645 24646commit 7d06a8791839ce6b22e2449646832b79cebf1b21 24647Author: Alex Deucher <alex@botch2.(none)> 24648Date: Fri Nov 16 02:43:00 2007 -0500 24649 24650 WIP: sync up with the latest ATOM bios code in rhd 24651 24652 doesn't compile ATM 24653 24654commit 3614d80ceb9a7b3615b0baab3cf2dd34ed4ab464 24655Author: Dave Airlie <airlied@linux.ie> 24656Date: Fri Nov 16 17:22:39 2007 +1000 24657 24658 add missing hpd register 24659 24660commit 9a2715fda97ac0ebcb45650a416e0652aab575b8 24661Author: Dave Airlie <airlied@linux.ie> 24662Date: Fri Nov 16 17:22:25 2007 +1000 24663 24664 make i2c unlock/lock registers for gpios 24665 24666commit 20dc549fff9a4137c93ebed449d05e0c437b6bc1 24667Author: Dave Airlie <airlied@linux.ie> 24668Date: Fri Nov 16 17:01:33 2007 +1000 24669 24670 avivo i2c: consolidate the avivo i2c code 24671 24672commit 3e62730f79a13883a65a568bc821bc56055a4ab7 24673Author: Dave Airlie <airlied@linux.ie> 24674Date: Fri Nov 16 15:19:00 2007 +1000 24675 24676 atombios: fixup warnings in atombios files 24677 24678commit cca7af3c4910983f7f090792986fcbfa0dc97cfb 24679Author: Dave Airlie <airlied@linux.ie> 24680Date: Fri Nov 16 15:04:01 2007 +1000 24681 24682 remove avivo_reg.h 24683 24684commit d39eb2077c6b2fc094ccd952772528eb9428c587 24685Author: Dave Airlie <airlied@linux.ie> 24686Date: Fri Nov 16 15:00:50 2007 +1000 24687 24688 radeon: rename a large section of avivo regs to documented names 24689 24690commit b7774c28dde72a205a40be78003df72eabfb9b1f 24691Author: Dave Airlie <airlied@linux.ie> 24692Date: Fri Nov 16 14:48:36 2007 +1000 24693 24694 Add copyright headers 24695 24696commit 3cfcd2164b400bd0d1cb4ede8eeb01abba9d75c8 24697Merge: efac14e6 718bfd3b 24698Author: Alex Deucher <alex@botch2.(none)> 24699Date: Thu Nov 15 23:25:39 2007 -0500 24700 24701 Merge branch 'agd-atom' of /home/alex/git/airlied/xf86-video-ati2 into agd-atom 24702 24703commit efac14e669a0c6184f8848191eb49ffb21934ee1 24704Author: Dave Airlie <airlied@linux.ie> 24705Date: Thu Nov 15 23:17:25 2007 -0500 24706 24707 r5xx: fix typo for crtc offset 24708 24709commit e6db621c37ff615be286462f000d67a662c5c331 24710Author: Alex Deucher <alex@botch2.(none)> 24711Date: Thu Nov 15 23:15:56 2007 -0500 24712 24713 fix INMC() and OUTMC() on !AVIVO chips 24714 24715 WR_EN is bit 8 so don't use OUTREG8. 24716 24717commit 52ba3fdd1ce05983fabedff234cfaf4c60fba38d 24718Author: Alex Deucher <alex@botch2.(none)> 24719Date: Thu Nov 15 23:12:30 2007 -0500 24720 24721 atombios_dac_detect() takes care of primary vs tv dac itself 24722 24723commit 49055d8aff91ff12186feaf5343c8fd2f96bcba0 24724Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 24725Date: Thu Nov 15 22:56:09 2007 -0500 24726 24727 RADEON: set proper defaults for tv dac BGADJ/DACADJ 24728 24729 we should get these values from the bios tables, but for now use 24730 some reasonable defaults. This should fix the washed out color 24731 problems on bugs 1082 and 12844. 24732 24733commit 821acf38b716ab87c3d07263d6e4a139fe54803f 24734Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 24735Date: Thu Nov 15 22:28:42 2007 -0500 24736 24737 RADEON: Make sure we set the MT properly for connected status unknown 24738 24739commit a94123f33ec6584fbdfc4b9ecd543d1357de8814 24740Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 24741Date: Thu Nov 15 22:19:54 2007 -0500 24742 24743 Revert "Portability fix from netbsd" 24744 24745 This reverts commit c9264aa53bf1470ad9104d1e7c4a8ce13c49c270. 24746 This breaks damage support. See bug 13244 24747 24748commit 718bfd3b61879172eee819fdab7080d5d4c0a756 24749Author: Dave Airlie <airlied@linux.ie> 24750Date: Fri Nov 16 10:37:04 2007 +1000 24751 24752 r5xx: fix typo for crtc offset 24753 24754commit e9d721c31372db045550f9562534b28f16121bb9 24755Author: Roland Scheidegger <sroland@tungstengraphics.com> 24756Date: Tue Nov 13 23:42:42 2007 +0100 24757 24758 ignore sometime bogus agp_mode bit from chip (bug #13190) 24759 24760 bit is wrong on at least X700 cards with rialto pcie-agp bridge chip, 24761 should be safe to use just the bit from the bridge hopefully to make 24762 agp setup work on these cards and not adversely affect others. 24763 24764commit b865886d00907899297ae864358eb26d9980975d 24765Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 24766Date: Sun Nov 11 20:38:41 2007 -0500 24767 24768 bump for RC release 24769 24770commit 7aeb35e5ad1aed6e78a3d8565fbfbfe66232ab45 24771Author: Alex Deucher <alex@botch2.(none)> 24772Date: Sun Nov 11 19:29:30 2007 -0500 24773 24774 fix from last commit 24775 24776commit 2ea95900547165e86ad3f8a41ce3331a05bad60e 24777Author: Alex Deucher <alex@botch2.(none)> 24778Date: Sun Nov 11 18:43:03 2007 -0500 24779 24780 Add full parsing support for atom bios connector table 24781 24782commit 7ce730828c293f0810dfdc554df48dfd76e35c49 24783Author: Alex Deucher <alex@botch2.(none)> 24784Date: Sun Nov 11 12:37:01 2007 -0500 24785 24786 rework crtc output source setup 24787 24788commit d61b6c78aa7810a2f9b9e2d9d95aab4295de80ce 24789Author: Alex Deucher <alex@botch2.(none)> 24790Date: Sun Nov 11 11:58:17 2007 -0500 24791 24792 make sure i2c bus exists before using it 24793 24794commit 342e3e207efda42ba679731c30dfb9d5e9d5643f 24795Author: Alex Deucher <alex@botch2.(none)> 24796Date: Fri Nov 9 17:11:43 2007 -0500 24797 24798 combine outputs based on id 24799 24800commit 8078c299d5941460243944d55051547c1a4d3791 24801Author: Alex Deucher <alex@botch2.(none)> 24802Date: Fri Nov 9 16:35:08 2007 -0500 24803 24804 use atom to program plls on r4xx 24805 24806commit 5febe2c96642f61d006abe6e8081e69d5b95adc0 24807Author: Alex Deucher <alex@botch2.(none)> 24808Date: Fri Nov 9 16:24:56 2007 -0500 24809 24810 turn off vga control when using ext modes 24811 24812commit 83f170c32c08c74a9e3466ffa0e0a0606c74427b 24813Author: Alex Deucher <alex@botch2.(none)> 24814Date: Thu Nov 8 19:33:13 2007 -0500 24815 24816 add pci ids for X1550PRO 24817 24818commit 5cdcaba0f6e9de2d15cfcc109ab97d9fd423e3bf 24819Author: Alex Deucher <alex@botch2.(none)> 24820Date: Thu Nov 8 19:28:03 2007 -0500 24821 24822 make sure to assign gpio 24823 24824commit 2dcb852778301b9284a2b4906dcf64f95ed638b7 24825Author: Alex Deucher <alex@botch2.(none)> 24826Date: Thu Nov 8 18:39:23 2007 -0500 24827 24828 pull in another of Dave's fixes 24829 24830commit 96273016a0bbdfa4d3a4e6275a3b09eeeadaa534 24831Author: Alex Deucher <alex@botch2.(none)> 24832Date: Wed Nov 7 01:22:30 2007 -0500 24833 24834 fix and move crtc source set up atombios_output.c 24835 24836 it's really more output related. 24837 24838commit 5c495c81cc3bcd4a38d06954243ed3bdc85bdc07 24839Author: Alex Deucher <alex@botch2.(none)> 24840Date: Wed Nov 7 01:04:11 2007 -0500 24841 24842 add support for initing external tmds via ATOM 24843 24844commit 5c13d9355280e6de44ebbf8de7ea89a6b91c7388 24845Author: Alex Deucher <alex@botch2.(none)> 24846Date: Wed Nov 7 00:17:28 2007 -0500 24847 24848 add avivo output stuff 24849 24850commit 68e7f5c67e2e9d2162b469ce31f452f3f89756b5 24851Author: Alex Deucher <alex@botch2.(none)> 24852Date: Tue Nov 6 23:43:29 2007 -0500 24853 24854 more avivo updates 24855 24856commit 303562dfb57e13c027b2aa9289d54e547c829ff1 24857Author: Alex Deucher <alex@botch2.(none)> 24858Date: Tue Nov 6 23:06:46 2007 -0500 24859 24860 add additional connector types 24861 24862commit 0d3e0735f710cb7b9505e4330997aa332f73c102 24863Author: Alex Deucher <alex@botch2.(none)> 24864Date: Tue Nov 6 22:59:25 2007 -0500 24865 24866 First round of avivo support 24867 24868commit 0abfe3150ce3eed4db93ccc2975bd4622dfa54a7 24869Author: Alex Deucher <alex@botch2.(none)> 24870Date: Tue Nov 6 18:47:00 2007 -0500 24871 24872 Add atombios files 24873 24874commit 20f01950e42babc308b4470df6a3c6628c932003 24875Author: Alex Deucher <alex@botch2.(none)> 24876Date: Tue Nov 6 18:04:43 2007 -0500 24877 24878 for r4xx ATOM cards, just use ATOM for PLL 24879 24880 while crtc timing and pll seem to work fine, output setup 24881 and routing don't seem to work too reliably with atom. 24882 AMD claims ATOM was still pretty new with r4xx so 24883 it's probably better to stick with direct programming for 24884 some things. 24885 24886commit 37af2f6b52c8f217b205486742a6b46a864da8ed 24887Author: Michel Dänzer <michel@tungstengraphics.com> 24888Date: Mon Nov 5 10:58:25 2007 +0100 24889 24890 radeon: Use indirect register access for (un)hiding the cursor. 24891 24892 This avoids hangs that occur when touching these registers while the card is 24893 busy with some cards. 24894 24895commit 5659e2f1b08e44d0c4f21a403e91b4801ef78f62 24896Author: Michel Dänzer <michel@tungstengraphics.com> 24897Date: Mon Nov 5 10:32:04 2007 +0100 24898 24899 radeon: Fix build warning on powerpc. 24900 24901commit d41d73f5f9d8baee7137e94da2c7852ea0412180 24902Author: Michel Dänzer <michel@tungstengraphics.com> 24903Date: Mon Nov 5 10:30:23 2007 +0100 24904 24905 Revert "radeon: Reinstate sync in radeon_crtc_show/hide_cursor." 24906 24907 This reverts commit 0241cac643fa1c08a45ea44f5c670b290e760ad8. 24908 24909 Taking the DRI lock in paths called from the SIGIO handler isn't safe. For 24910 example, the SIGIO handler may be called while the lock is held by an AIGLX 24911 context, resulting in deadlock. See e.g. 24912 https://bugs.freedesktop.org/show_bug.cgi?id=13005 . 24913 24914commit 78a3eabff382e8ebe33df2039076fb083bcc361b 24915Author: Alex Deucher <alex@botch2.(none)> 24916Date: Sun Nov 4 14:11:26 2007 -0500 24917 24918 WIP: get ATOM crtc stuff working on r4xx 24919 24920commit 5e9ebd8e496b72b051053d637c63b2956b7861d3 24921Author: Alex Deucher <alex@botch2.(none)> 24922Date: Sat Nov 3 20:46:17 2007 -0400 24923 24924 Initial support for r4xx 24925 24926 - add r3xx/r4xx MC access macros and functions 24927 24928commit c106075ccb81ca2ee4894743e676fd37653c8dce 24929Author: Alex Deucher <alex@botch2.(none)> 24930Date: Sat Nov 3 18:29:20 2007 -0400 24931 24932 More of Dave's ATOM init code. 24933 24934commit e18f5d61806b445ad77d93e258fbce9422b52bb6 24935Author: Alex Deucher <alex@botch2.(none)> 24936Date: Sat Nov 3 18:20:55 2007 -0400 24937 24938 Initial integration of Atom code and some of Dave's code. 24939 24940 ATOM builds, but it's not hooked up yet. 24941 24942commit aaafc6a8f33d4736c37c43d3fd5e43085a2590fb 24943Author: Brice Goglin <bgoglin@debian.org> 24944Date: Sun Oct 28 21:10:32 2007 +0100 24945 24946 Fix typo in 'disposing cursor' 24947 24948 Bug #12984. 24949 24950commit a5c55c418319c1315ad73c305ef48b44fe1d9580 24951Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 24952Date: Sat Oct 27 19:46:11 2007 -0400 24953 24954 RADEON: disable FP DETECT for DVI. 24955 24956 These bits don't seem to be reliable or routed right 24957 in all cases. 24958 24959commit 49cf7cb3db36ce7734f7e314a040a240191d6477 24960Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 24961Date: Thu Oct 25 21:23:09 2007 -0400 24962 24963 RADEON: Fix IgnoreEDID option 24964 24965commit 70bbc0cfe699588f8ae4123efca0e1baffc91000 24966Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 24967Date: Thu Oct 25 20:56:11 2007 -0400 24968 24969 RADEON: clean up mode handling and improve reliability of DDC 24970 24971 - Reduce the number of times we do DDC and don't 24972 skip the magical GPIO dance when doing DDC. 24973 - If no DDC modes are found, add the screen modes 24974 24975commit 1a4f590c2ac071f0c348f38e9f26fa6093722453 24976Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 24977Date: Thu Oct 25 19:25:36 2007 -0400 24978 24979 RADEON: document DefaultTMDSPLL option 24980 24981commit 2d26e0d0bcf25e210dc7771e76638dcfafcc2356 24982Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 24983Date: Thu Oct 25 19:18:18 2007 -0400 24984 24985 RADEON: Add driver option to force TV DAC load detection on 24986 24987commit a9306b7986467c0794c91625ad2ece597b718462 24988Author: Alex Deucher <alex@botch2.(none)> 24989Date: Thu Oct 18 20:06:58 2007 -0400 24990 24991 RADEON: fix possible crash in radeon_crtc_mode_set() 24992 24993 when radeon_crtc_mode_set() is called during ScreenInit(), 24994 we don't have pScrn->pScreen yet. 24995 24996commit 115b4b65ef679a46d90b7fc8ac22ace37038b3fc 24997Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 24998Date: Wed Oct 17 20:25:47 2007 -0400 24999 25000 RADEON: Remove LVDSBiosNativeMode Option 25001 25002 Shouldn't be needed any longer as I've finally sorted out 25003 the LVDS issues due to the crtc setup. 25004 25005commit f2eb5c1cc69a4f7b0754ec6b2efde4fa1650ab6d 25006Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25007Date: Tue Oct 16 18:54:28 2007 -0400 25008 25009 RADEON: finish fixing LVDS/RMX 25010 25011 For panels, both the mode and crtc values need to be updated based 25012 on the native mode timing. This shoudl fix any remaining problems 25013 with RMX. 25014 25015commit 93561688087493fccb7ac387d728feec61212180 25016Author: Michel Dänzer <michel@tungstengraphics.com> 25017Date: Tue Oct 16 10:57:06 2007 +0200 25018 25019 radeon: Handle 64 bit PCI resource bases better. 25020 25021commit 08945fc8a14ba57e32350e85daa3354619e9e72e 25022Author: Alex Deucher <alex@botch2.(none)> 25023Date: Mon Oct 15 21:12:46 2007 -0400 25024 25025 RADEON: allow the user to set any tv standard 25026 25027 allow the user to set any tv standard rather than 25028 limiting it to what the bios says is supported. 25029 25030commit 571548b2b95ce289f26a77c11deb6266f0b18516 25031Author: Alex Deucher <alex@botch2.(none)> 25032Date: Mon Oct 15 20:33:56 2007 -0400 25033 25034 RADEON: use native timing for RMX 25035 25036 RMX should work again with native timings after the previous 25037 patch. 25038 25039commit 2f87bff293a343b40c1be096933a5ae126632468 25040Author: Alex Deucher <alex@botch2.(none)> 25041Date: Mon Oct 15 20:06:28 2007 -0400 25042 25043 RADEON: Fix subtle change in crtc reg init 25044 25045 At some point we changed how hsync_wid and vsync_wid were clipped. 25046 Previously we used the field size as a mask when building the 25047 register. This got changed to setting the value to the field 25048 size if it was greater. this probably explains number stange 25049 mode bugs. 25050 25051commit fd5bb7bb5e968127b87102320eccc1222f205e5a 25052Author: Michel Dänzer <michel@tungstengraphics.com> 25053Date: Sat Oct 13 18:56:43 2007 +0200 25054 25055 radeon: Improve detection of default value for Option "MacModel" on Linux. 25056 25057 * Detect all PowerBooks with dual link DVI (according to developer.apple.com) 25058 and both Mac Mini G4 models. Other PowerBooks and iBooks should be covered 25059 by the 'detected as' line in /proc/cpuinfo. 25060 * Give specific instructions for overriding and reporting incorrect detection. 25061 * Only perform detection when Option "MacModel" isn't present with a valid 25062 value. 25063 * Close /proc/cpuinfo file handle when done with it. 25064 * Coding style cleanups. 25065 25066commit fbf121add5584049627f07345fc502b2aefc88ce 25067Author: Eugen Dedu <Eugen.Dedu@pu-pm.univ-fcomte.fr> 25068Date: Sat Oct 13 16:51:58 2007 +0200 25069 25070 radeon: Derive default value of Option "MacModel" from /proc/cpuinfo on Linux. 25071 25072commit 5db3afaa1fdb69d382ac769ef40191a4b964d28e 25073Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25074Date: Thu Oct 11 20:25:20 2007 -0400 25075 25076 RADEON: return status unknown for flaky chip/connector combinations 25077 25078 This should at least get something on the screen. 25079 XPRESS chips, I'm looking in your general direction... 25080 25081commit 1148d332f52f3780897aae3fcd5d6a67687c42cd 25082Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25083Date: Thu Oct 11 19:09:45 2007 -0400 25084 25085 RADEON: fix potential crash in mode handling 25086 25087commit 905b19df33d6f912336b65895e97775ffec07728 25088Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25089Date: Thu Oct 11 18:55:08 2007 -0400 25090 25091 RADEON: small cleanup from last commit 25092 25093commit 7afd04c1e4ffa6e4e5ba08ae90ba002237dc282b 25094Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25095Date: Thu Oct 11 00:16:45 2007 -0400 25096 25097 RADEON: tell the bios not to muck with the hardware while the driver is active 25098 25099 by toggling the appropriate bios scratch regs you can tell 25100 the bios not the touch the hw while the driver is active. 25101 This should prevent the bios from scrambling the hardware 25102 when users open the lid or toggle bios hotkeys. 25103 25104 fixes bug 12567 25105 25106commit 1b231d28fdda5cdc44bb9d2075d4edfd8f17e21f 25107Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25108Date: Wed Oct 10 21:54:46 2007 -0400 25109 25110 RADEON: attempt to detect lid status on legacy bios 25111 25112 I'm not sure how reliable this is. The lid closed 25113 bit only seems to be set if the lid is closed at boot 25114 so LVDS will default to on if you restart X with the 25115 lid closed after boot. 25116 25117commit 47ef583b993607bc3a372dd125c6e5228c4ec7bf 25118Author: Alex Deucher <alex@botch2.(none)> 25119Date: Wed Oct 10 19:35:31 2007 -0400 25120 25121 RADEON: more XPRESS connector table hacks 25122 25123 Seems XPRESS desktop chips with DVI-D list a proprietary 25124 connector type rather than DVI-D. try and do the right thing. 25125 Also, XPRESS chips seem to include connector type 8 and ddc 25126 type 8 which are unknown at this time. 25127 25128commit c9264aa53bf1470ad9104d1e7c4a8ce13c49c270 25129Author: Joerg Sonnenberger <joerg@britannica.bec.de> 25130Date: Tue Oct 9 23:18:29 2007 -0400 25131 25132 Portability fix from netbsd 25133 25134commit 80eee856938756e1222526b6c39cee8b5252b409 25135Author: Matthieu Herb <> 25136Date: Tue Oct 9 16:17:50 2007 -0400 25137 25138 RADEON: fix console restore on netbsd 25139 25140 Include the mode restore bugfix from monolithic Xorg, that is derived 25141 from the version in xsrc which in turn was provided by Matthieu Herb 25142 over 3 years ago on the XFree86 lists. Suggested by various 25143 developers, hold-back due to the working state in xorg-server 1.1.1. 25144 Tracing down the exact change showed that the changed default color 25145 depth made this issue a lot more prominent again. Discussed with Eric 25146 Anholt. 25147 25148commit 547fef4c6382f8a4951c086ee531b804cba31075 25149Author: Lars Weiler <pylon@gentoo.org> 25150Date: Tue Oct 9 16:10:34 2007 -0400 25151 25152 include stdint.h to fix the build on some platforms 25153 25154 fixes bug 12733 25155 25156commit ddec5783103045084a66a34d71326cc6e4a3f598 25157Author: George Sapountzis <gsap7@yahoo.gr> 25158Date: Tue Oct 9 13:27:27 2007 +0300 25159 25160 theatre compiles with pci-rework. 25161 25162commit 051435610a66735fd455bbb526fa294fcfe8c0b6 25163Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25164Date: Sun Oct 7 19:39:47 2007 -0400 25165 25166 RADEON: still more LVDS fixes 25167 25168 Seems some laptops need the native mode from the bios for 25169 LVDS while others seem to prefer a CVT mode. Add an option 25170 to pick the preferred mode. The default it to use the bios 25171 table timing. 25172 25173commit 0b03a73b7dcb4aa192c42f2a4c842d324c358122 25174Author: Dave Airlie <airlied@linux.ie> 25175Date: Sat Oct 6 20:11:19 2007 +1000 25176 25177 radeon: add support for DDC type 6 25178 25179 thanks to arekm on xorg-devel for testing/feedback 25180 25181commit 83d7506a603889b6c18736cff9b2cd3296c95c84 25182Author: Alex Deucher <alex@botch2.(none)> 25183Date: Fri Oct 5 20:25:04 2007 -0400 25184 25185 Bump for RC release 25186 25187commit fe66cc4e02a52e1d014ae5de192c65e8abd7f58a 25188Author: Tormod Volden <bugzi06.fdo.tormod@xoxy.net> 25189Date: Fri Oct 5 20:10:55 2007 -0400 25190 25191 RADEON: rename to PCI_CHIP_RS485_5975 everywhere 25192 25193 RS482_5975 is actually RS485. see bug 12048 25194 25195commit 314e8fdbe5369e4c1786d176d03d0d2a046f25d5 25196Author: Tormod Volden <bugzi06.fdo.tormod@xoxy.net> 25197Date: Fri Oct 5 20:05:54 2007 -0400 25198 25199 RADEON: Disable DRI by default for RS482 cards 25200 25201 fixes bug 12048 25202 25203commit abbafe6db7c0d24efc61a05950d4214c00d3f043 25204Author: Michel Dänzer <michel@tungstengraphics.com> 25205Date: Thu Oct 4 11:33:09 2007 +0200 25206 25207 radeon: Fix compile warnings. 25208 25209commit 1110957ba703475847b8e72023c6edcbd1068aef 25210Author: Michel Dänzer <michel@tungstengraphics.com> 25211Date: Thu Oct 4 11:32:48 2007 +0200 25212 25213 radeon: Fix for pci-rework. 25214 25215 Surprisingly easy, thanks to George's pci-rework changes. 25216 25217commit 7addf41885ec5658f531624a9c24ea5bd7d22d19 25218Merge: 3a958ba1 78c6bd30 25219Author: Michel Dänzer <michel@tungstengraphics.com> 25220Date: Thu Oct 4 11:06:08 2007 +0200 25221 25222 Merge remote branch 'origin/pci-rework' 25223 25224commit 3a958ba136c3fae5a4ddd56373ac7cd47046f10e 25225Author: Alex Deucher <alex@botch2.(none)> 25226Date: Wed Oct 3 22:59:14 2007 -0400 25227 25228 RADEON: remove some cruft 25229 25230commit b6bda79f72df5e5bf9c6b71fa3298e765da506bd 25231Author: Alex Deucher <alex@botch2.(none)> 25232Date: Wed Oct 3 22:56:27 2007 -0400 25233 25234 RADEON: more fixes for single crtc chips 25235 25236commit 5f5c4e6ad61c45c24f1443b91b4bc5375efdebc0 25237Author: Alex Deucher <alex@botch2.(none)> 25238Date: Wed Oct 3 22:46:21 2007 -0400 25239 25240 RADEON: remove RADEONSaveMode() 25241 25242 It's only called by RADEONSave(), so just move its 25243 contents into RADEONSave(). this mirrors RADEONRestore(). 25244 25245commit bfede412b3a3cd11769a580b167c528734146096 25246Author: Alex Deucher <alex@botch2.(none)> 25247Date: Wed Oct 3 22:19:16 2007 -0400 25248 25249 RADEON: minor fixes for external TMDS 25250 25251commit 597dffce9bdc200003d0be880235258386a0bdd7 25252Author: Alex Deucher <alex@botch2.(none)> 25253Date: Wed Oct 3 21:43:08 2007 -0400 25254 25255 RADEON: final fix for RMX/LVDS 25256 25257 It seems that on quite a few laptops the LVDS H/V timing 25258 from the bios tables are way off or just doesn't work period. 25259 Either we are using it wrong or we need some additional checks 25260 when we parse it. Only the dot clock seems to really matter, 25261 so use cvt modes and update the dotclock. This seems to work 25262 correctly in all cases. 25263 25264commit cc0c2d8e61600652b1f9cb3dc49db2ef62b1e40d 25265Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25266Date: Wed Oct 3 00:59:37 2007 -0400 25267 25268 RADEON: Finally sort out LVDS modes 25269 25270 - The panel timing from the bios tables is for the crtc, 25271 not the native panel mode, so add cvt modes for the panel 25272 and fix up the crtc values in mode_fixup(). 25273 - rename and reconfigure RADEONValidateFPModes() to what it 25274 really does: add screen modes. I suppose for backwards 25275 compatibility we ought to add the screen modes to every 25276 reasonable output, but everyone should really use xrandr 25277 or an output monitor section to add custom modes. Hopefully 25278 this will go away at some point. 25279 25280commit 0ca184c3c35032df39ea7ce5d2d4aba1a97b6426 25281Author: Alex Deucher <alex@botch2.(none)> 25282Date: Tue Oct 2 19:38:18 2007 -0400 25283 25284 RADEON: fix RADEONInitDispBandwidth() on single crtc cards 25285 25286 Should fix bug 12490 25287 25288commit a5fb2a5d54cdd562a97881cdea1a4af35b483aab 25289Author: Dave Airlie <airlied@linux.ie> 25290Date: Mon Oct 1 08:48:27 2007 +1000 25291 25292 radeon: fix powerpc build 25293 25294commit 22519fde1e002f28d6036d448fcd18452d00f1bb 25295Author: Alex Deucher <alex@botch2.(none)> 25296Date: Sun Sep 30 13:11:20 2007 -0400 25297 25298 RADEON: add support for ext tmds table and ext tmds chip init 25299 25300 This probably won't work on all chips as the various gpio lines 25301 seem to need special magic to to actually talk to the i2c slave 25302 chips. 25303 25304commit dcc376e2d2a13329dd03f1bc4b471329757a6f5f 25305Author: Alex Deucher <alex@botch2.(none)> 25306Date: Wed Sep 26 23:27:38 2007 -0400 25307 25308 RADEON: remove cruft 25309 25310commit 99ceaefa18c6e07b55106cca0ea8996fa73667be 25311Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25312Date: Sun Sep 30 12:02:54 2007 -0400 25313 25314 RADEON: more work on LVDS mode validation and fixups 25315 25316 Hopefully this will fix up the weird mode issues with LVDS and 25317 native and RMX modes, but I can't seem to get quite the right 25318 combo to fix everyone. 25319 25320commit d808781d48adf01e80b5bb476bae2d2f599030f1 25321Author: Dave Airlie <airlied@redhat.com> 25322Date: Fri Sep 28 14:50:42 2007 +1000 25323 25324 rc410: disable DRI by default due to it not working 25325 25326commit 4d9f944e40c6b9ac7e80428dff49c6e83fda1d8f 25327Author: Michel Dänzer <michel@tungstengraphics.com> 25328Date: Thu Sep 27 17:50:15 2007 +0200 25329 25330 radeon: Fix build on little endian platforms. 25331 25332 This was accidentally broken when moving away from using pixmap data pointers 25333 directly. 25334 25335commit 5902e913010a2fb0c17efe01f9aade73f9646f16 25336Author: Michel Dänzer <michel@tungstengraphics.com> 25337Date: Thu Sep 27 12:53:34 2007 +0200 25338 25339 radeon: EXA pitch fixes. 25340 25341 The combined pitch/offset registers only support pitches up to 16320 bytes. 25342 25343 With EXA >= 2.3, set the maxPitchBytes field accordingly. With older versions, 25344 limit maxX such that the pitch of 32bpp pixmaps doesn't exceed the limit. 25345 25346 Also check the limit in RADEONGetOffsetPitch just in case. 25347 25348 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=12551 . 25349 25350commit ca728973bce504e03bf6228d18287906c35eaa3f 25351Author: Michel Dänzer <michel@tungstengraphics.com> 25352Date: Thu Sep 27 11:14:21 2007 +0200 25353 25354 radeon: Don't use pixmap data pointer directly with EXA. 25355 25356 Fixes segfault with current xserver master without AccelDFS. 25357 25358commit 21593d04d222b05dbba9abd31eaa3bfb91d999b6 25359Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25360Date: Mon Sep 24 23:01:20 2007 -0400 25361 25362 RADEON: more clean up of mode code 25363 25364 also print out LVDS mode info from bios for both 25365 legacy and atom bioses 25366 25367commit 22cccb99d762567757d3fd38795f71e943f1faf9 25368Author: Alex Deucher <alex@botch2.(none)> 25369Date: Mon Sep 24 00:10:01 2007 -0400 25370 25371 bump for RC release 25372 25373commit fad9e7b00de051f632000eb2304ac23a524f7c8e 25374Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25375Date: Sun Sep 23 23:43:26 2007 -0400 25376 25377 RADEON: Rework LVDS mode validation 25378 25379 The old validation code wasn't really well suited to randr. 25380 This fixes several issues: 25381 - missing display size for panels with edid 25382 - broken duplicate modes 25383 25384commit bf9674b9df351108e069d037ec10869a4154a881 25385Author: Alex Deucher <alex@botch2.(none)> 25386Date: Sun Sep 23 15:01:34 2007 -0400 25387 25388 RADEON: properly set default RMX types 25389 25390commit b27135bce8b41d69290613b440a338b0a7fe0200 25391Author: Alex Deucher <alex@botch2.(none)> 25392Date: Sun Sep 23 14:44:38 2007 -0400 25393 25394 RADEON: fix up LVDS handling for r3xx and newer 25395 25396commit d5cf268be2c9b3e320d7101f5213f7d5609b6308 25397Author: Alex Deucher <alex@botch2.(none)> 25398Date: Sat Sep 22 18:31:51 2007 -0400 25399 25400 RADEON: Fix crash when Xv window is outside of either crtc 25401 25402commit 8aca0ca6be7122e0ece20985a1862553a1b225c0 25403Author: Alex Deucher <alex@botch2.(none)> 25404Date: Sat Sep 22 17:56:50 2007 -0400 25405 25406 Revert "RADEON: fix crash when Xv window is outside of either crtc" 25407 25408 This reverts commit 9109e62e3be7f96b41b534ab517fdf1baf458806. 25409 25410 This breaks ABI. better fix to come. 25411 25412commit d6e1d9be1cd5786a24ed4bb8e45fa7e04df56622 25413Author: Alex Deucher <alex@botch2.(none)> 25414Date: Sat Sep 22 16:14:06 2007 -0400 25415 25416 RADEON: Don't make the entity as shareable 25417 25418 This should prevent people from mistakenly trying to run 25419 zaphod mode 25420 25421commit 81114af1cdddc0b10d076f2e38c7a00c1223cc48 25422Author: Alex Deucher <alex@botch2.(none)> 25423Date: Sat Sep 22 15:51:23 2007 -0400 25424 25425 RADEON: preliminary support for mac mini 25426 25427 Option "MacModel" "mini" 25428 may not be 100% correct yet 25429 25430commit 6c482e453bc8156886294d0c1b8f3f1b3dcf4b36 25431Author: Alex Deucher <alex@botch2.(none)> 25432Date: Sat Sep 22 15:11:20 2007 -0400 25433 25434 RADEON: Fix RMX on LVDS 25435 25436 LVDS + RMX doesn't seem to like having the crtc values tweaked. 25437 25438commit 9109e62e3be7f96b41b534ab517fdf1baf458806 25439Author: Michel Dänzer <daenzer@debian.org> 25440Date: Sat Sep 22 14:19:10 2007 -0400 25441 25442 RADEON: fix crash when Xv window is outside of either crtc 25443 25444 Should fix bug 12175 25445 25446commit d2ce4a5003ce1291ea2327b2c00a0b24408fe26c 25447Author: Alex Deucher <alex@botch2.(none)> 25448Date: Sat Sep 22 10:41:55 2007 -0400 25449 25450 RADEON: RMX updates 25451 25452 - add option to turn off RMX 25453 - turn off RMX by default on DVI 25454 - add infrastructure to support more RMX modes 25455 25456commit f95b9ab729376083bf0d12987ee260ec2aba721e 25457Author: Alex Deucher <alex@botch2.(none)> 25458Date: Fri Sep 21 00:12:37 2007 -0400 25459 25460 RADEON: remove more old cruft 25461 25462commit 5a6f74103f0ec0d451d0e2573442efe5922848af 25463Author: Maciej Cencora <m.cencora@gmail.com> 25464Date: Thu Sep 20 23:56:08 2007 -0400 25465 25466 RADEON: fix video in on RV380 (tested on X600 VIVO) 25467 25468commit c72a365386e19f9257db041d44b09ad499cc9f6a 25469Author: Alex Deucher <alex@botch2.(none)> 25470Date: Thu Sep 20 23:49:57 2007 -0400 25471 25472 RADEON: fix up dvo support (still no external chip init) 25473 25474commit 5e4d98470b6412a686883c554e7eb7badbe78c4d 25475Author: Alex Deucher <alex@botch2.(none)> 25476Date: Thu Sep 20 23:22:48 2007 -0400 25477 25478 RADEON: round 3 on the PLLs. should fix the LVDS issues 25479 25480commit c5e2a2f09af807006c7ea493a8e90ff77abe207c 25481Author: Alex Deucher <alex@botch2.(none)> 25482Date: Wed Sep 19 19:58:28 2007 -0400 25483 25484 bump for RC release 25485 25486commit 46ff78b9f010ca24178d4363761be00eb3ecb632 25487Author: Alex Deucher <alex@botch2.(none)> 25488Date: Wed Sep 19 19:49:58 2007 -0400 25489 25490 RADEON: add default connector setup for single crtc chips 25491 25492commit 509ca0cb0e1cde905b47db2bbac6f2a58523b279 25493Author: Alex Deucher <alex@botch2.(none)> 25494Date: Wed Sep 19 19:41:17 2007 -0400 25495 25496 RADEON: fix external TMDS parsing in legacy bios connector table 25497 25498commit 5a59547ef7e986c9613023d941c3354cd476faf1 25499Author: Alex Deucher <alex@botch2.(none)> 25500Date: Wed Sep 19 19:28:09 2007 -0400 25501 25502 Revert "RADEON: initial pass at external TMDS support" 25503 25504 This reverts commit 4000a710c93dd2d82891e4082bc7fa922ba9c5f4. 25505 25506 This needs to be reworked and needs more soak time. so revert 25507 for now. 25508 25509commit dc333884c540d536bffe51a5ebfaf1822af6a91c 25510Author: Alex Deucher <alex@botch2.(none)> 25511Date: Tue Sep 18 22:23:12 2007 -0400 25512 25513 RADEON: fix plls again 25514 25515 previous fix seemed to break other chips. Lets see how 25516 this goes. 25517 25518commit 271e541088f455d1cfedff87e88c9a4fdbdbe424 25519Author: Alex Deucher <alex@botch2.(none)> 25520Date: Tue Sep 18 22:12:31 2007 -0400 25521 25522 RADEON: Remove more mergedfb cruft 25523 25524commit 8ae69c496eba701e744cca0605f73242673f7b3f 25525Author: Alex Deucher <alex@botch2.(none)> 25526Date: Tue Sep 18 20:12:21 2007 -0400 25527 25528 RADEON: adjust pll restore 25529 25530 some chips seem to be pickier than others. fixes bug 12467 25531 25532commit 8a84dc8ee260716026fa661a6dced2000621efa2 25533Author: Michel Dänzer <michel@tungstengraphics.com> 25534Date: Tue Sep 18 11:53:39 2007 +0200 25535 25536 radeon: Remove declaration of unused variable. 25537 25538commit 2fbbd9c5f0ec27e72780e8b054bc59502f1e5651 25539Author: Alex Deucher <alex@botch2.(none)> 25540Date: Tue Sep 18 00:45:56 2007 -0400 25541 25542 RADEON: note that MergedFB support has been replaced with randr 1.2 25543 25544commit 81b708dbd35c29bf2d67e8c655af97b60e61498b 25545Author: Alex Deucher <alex@botch2.(none)> 25546Date: Mon Sep 17 23:54:03 2007 -0400 25547 25548 RADEON: turn off load detection on tv dac by default 25549 25550 it's just too unreliable and even when it is most people 25551 don't want it. enable it via output attribute or force 25552 the outputs on with xrandr 25553 25554commit c3ab53159d42157985a25e9125eaa5362273b73d 25555Author: Alex Deucher <alex@botch2.(none)> 25556Date: Mon Sep 17 23:28:51 2007 -0400 25557 25558 RADEON: XPRESS connector tables are crap 25559 25560 besides specifying the pimary DAC (which they don't have), 25561 most seem to specify DDC_CRT2 for DDC, which does not seem 25562 to ever work. Some have reported success with DDC_MONID so 25563 lets give that a try. 25564 25565commit 1fafa970fa78ddea29a8804b6cd0aea52e4fa2e1 25566Author: Alex Deucher <alex@botch2.(none)> 25567Date: Mon Sep 17 23:10:13 2007 -0400 25568 25569 RADEON: ignore edid digital bit on analog ports 25570 25571 Some monitors have this bit set even on analog ports 25572 25573commit 4000a710c93dd2d82891e4082bc7fa922ba9c5f4 25574Author: Alex Deucher <alex@botch2.(none)> 25575Date: Mon Sep 17 23:04:17 2007 -0400 25576 25577 RADEON: initial pass at external TMDS support 25578 25579 - Based on Dave Airlie's initial work and the dvo support 25580 in the intel driver. 25581 - Only sil164 is supported at the moment. 25582 - Once we get some testing and such, we ought to move the dvo 25583 drivers out of the drivers so they can be shared among all drivers 25584 - Doesn't seem to work on my card (r4xx ATOM) card ATM 25585 - Legacy bios table programming sequence not implemented yet 25586 25587commit 38515d402555eaa61c686d42973e59f659b07466 25588Author: Alex Deucher <alex@botch2.(none)> 25589Date: Sat Sep 15 11:39:47 2007 -0400 25590 25591 RADEON: don't restore bios scratch regs 25592 25593 Leave them as programmed by the bios 25594 fixes bug 12424 25595 25596commit 78c6bd305f2ea70629d1dda3c0c48c067451f5fb 25597Author: George Sapountzis <gsap7@yahoo.gr> 25598Date: Mon Sep 10 19:00:16 2007 +0300 25599 25600 [mach64] Revert to multiple mappings per BAR. 25601 25602 libpciaccess added subrange mappings and mtrr handling. 25603 25604 compile-tested only. 25605 25606commit 0241cac643fa1c08a45ea44f5c670b290e760ad8 25607Author: Michel Dänzer <michel@tungstengraphics.com> 25608Date: Sat Sep 8 00:22:40 2007 +0200 25609 25610 radeon: Reinstate sync in radeon_crtc_show/hide_cursor. 25611 25612 Make sure the DRI lock is held though, as these can be called asynchronously. 25613 25614 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=12245 . 25615 25616commit 49933e2f7d590811f2bc8c0d51a09f3b7f14845e 25617Author: Alex Deucher <alex@botch2.(none)> 25618Date: Sat Sep 1 12:23:07 2007 -0400 25619 25620 RADEON: automatically disable tiling if requested virtual desktop exceeds surface limits 25621 25622 Also, tweak default desktop sizes and add some informational messages 25623 25624commit d49a60bc939c9f3e9d57c23e44263f5bb52f09fb 25625Author: Sjoerd Simons <sjoerd@luon.net> 25626Date: Thu Aug 30 11:40:13 2007 +0200 25627 25628 radeon: Fix Option "MacModel". 25629 25630commit 2b860726fff058623e0e8e1b0aca092c246875ce 25631Author: Michel Dänzer <michel@tungstengraphics.com> 25632Date: Thu Aug 30 11:38:28 2007 +0200 25633 25634 radeon: Don't build currently unused RADEONSavePalette. 25635 25636commit 12187a6aa93049c002a4171344d03c713f7f3c5d 25637Author: Alex Deucher <alex@botch2.(none)> 25638Date: Wed Aug 29 23:11:30 2007 -0400 25639 25640 RADEON: Add quirk and connector tables for apple laptops 25641 25642 As far as I can tell there are three apple laptop variants: 25643 ibook - LVDS, TVDAC drives TV or VGA via dongle 25644 powerbook-duallink - LVDS, TV, External TMDS/Primary DAC 25645 powerbook - LVDS, TV, Internal TMDS/Primary DAC 25646 use Option "MacModel" "<string>" 25647 to enable the appropriate quirks where string is one of the above 25648 25649 We can't yet init the external TMDS directly, but if OF inits it, 25650 it should work. This should also fix bug 9955. 25651 25652 Please test! 25653 25654commit 61c1fdaa8553581944f78a11e6f9aa76163a468a 25655Author: Alex Deucher <alex@botch2.(none)> 25656Date: Tue Aug 28 23:47:19 2007 -0400 25657 25658 RADEON: add option to force tmds pll to default table 25659 25660 Also rework the tmds pll output attribute handling a bit 25661 25662commit 0d9087bc60bb95c770b899cfed29699c02bdac49 25663Author: Alex Deucher <alex@botch2.(none)> 25664Date: Tue Aug 28 23:08:20 2007 -0400 25665 25666 RADEON: remove some cruft forgotten in a previous commit 25667 25668commit 6ff0645ecfe65727e8ef5d5e6215b4e03078e1a5 25669Author: George Sapountzis <gsap7@yahoo.gr> 25670Date: Tue Aug 28 15:37:11 2007 +0300 25671 25672 [mach64] Convert to pci-rework, keeping source-code compatibility. 25673 25674 It still uses the old probe method though, this is due to the ati wrapper. 25675 25676commit 7b38d9a1209f87255e5bb0aefe46a363ce4fb6ef 25677Merge: 2e3d43af 673f7997 25678Author: George Sapountzis <gsap7@yahoo.gr> 25679Date: Tue Aug 28 16:01:12 2007 +0300 25680 25681 Merge branch 'master' into pci-rework 25682 25683 Conflicts: 25684 25685 src/ati.c 25686 25687commit 673f799729824f4439dd5f681f75dd5aab50947f 25688Author: Alex Deucher <alex@botch2.(none)> 25689Date: Tue Aug 28 00:42:30 2007 -0400 25690 25691 RADEON: Update tv attributes immediately 25692 25693commit ad6f7ad1b2ccae0bc0a416b9b0ca22709c9d5199 25694Author: Alex Deucher <alex@botch2.(none)> 25695Date: Tue Aug 28 00:08:41 2007 -0400 25696 25697 RADEON: remove the "default" tv_standard option 25698 25699commit 17e0f9e6cbfdb115034d327bd34d46339fd632b7 25700Author: Alex Deucher <alex@botch2.(none)> 25701Date: Mon Aug 27 23:59:03 2007 -0400 25702 25703 RADEON: enable load detection for tvdac if output count for tvdac < 2 25704 25705commit 42839fb5a8584196e7b18375bff6c426ed0347d9 25706Author: Alex Deucher <alex@botch2.(none)> 25707Date: Mon Aug 27 23:44:13 2007 -0400 25708 25709 RADEON: make load detection an output attribute for analog outputs 25710 25711 Since TV/VGA/DVI-I can share the TV DAC, we often get false detection 25712 of all inputs that share that DAC. Make load detection an output 25713 attribute. Enabled by default on primary dac and on cards where 25714 tv dac is (usually) dedicated to tv (non-IGP mobilities). 25715 25716commit a5a1055d64ab4fa16bfb03a412ae6c4fe69ff65d 25717Author: Alex Deucher <alex@botch2.(none)> 25718Date: Mon Aug 27 22:42:22 2007 -0400 25719 25720 RADEON: make tmds pll an output attribute 25721 25722 sometimes the bios tmds plls are busted for certain 25723 monitors. sometimes the dirver tables are. Let the user 25724 pick at run time. 25725 25726commit a12e4aa01bf1c5723c3c791ff9bdc26eef21d5ea 25727Author: Alex Deucher <alex@botch2.(none)> 25728Date: Sun Aug 26 18:51:29 2007 -0400 25729 25730 Bump for new release 25731 25732commit d43596e5f5d7c60f96b57bc3e743a9b40eb7109d 25733Author: Alex Deucher <alex@botch2.(none)> 25734Date: Sun Aug 26 18:07:50 2007 -0400 25735 25736 RADEON: Fix rotation. works now 25737 25738commit 47eb3327c258bb0cfd9a1d5677624b9988a39057 25739Author: Alex Deucher <alex@botch2.(none)> 25740Date: Sun Aug 26 15:43:22 2007 -0400 25741 25742 RADEON: minor tweak to tv out 25743 25744commit f36720377737210c985b196d9a988efdd767f1c7 25745Author: Alex Deucher <alex@samba.(none)> 25746Date: Sun Aug 26 14:13:06 2007 -0400 25747 25748 RADEON: fix depth 16 palette for real this time 25749 25750commit f2b13f1457bf860b075310d3962254be0ed7bea3 25751Author: Alex Deucher <alex@botch2.(none)> 25752Date: Sun Aug 26 13:27:19 2007 -0400 25753 25754 RADEON: Only update pixclks_cntl when updating tv routing 25755 25756 No need to re-set all of the pll2 stuff 25757 25758commit 5d044b9f74c7aa7e12f2822896fed881e2ca9d19 25759Author: Alex Deucher <alex@botch2.(none)> 25760Date: Sat Aug 25 21:03:08 2007 -0400 25761 25762 RADEON: fix crtc clipping for Xv 25763 25764commit 3fd2d22a02812d5f86cdc1c9503f48362b0c362b 25765Author: Alex Deucher <alex@botch2.(none)> 25766Date: Sat Aug 25 17:37:35 2007 -0400 25767 25768 RADEON: remove fbdev option 25769 25770 FBDev support is currently broken, and it not really compatible 25771 with randr 25772 25773commit 3469e1aa08792890fa6a5c72da52a1992a0b382c 25774Author: Alex Deucher <alex@botch2.(none)> 25775Date: Fri Aug 24 20:42:13 2007 -0400 25776 25777 RADEON: add extra green data in depth 16 25778 25779 Apparently some radeons need this? 25780 25781commit 71f650d1bc432514516f7ac64a5e8a54c5227881 25782Author: Michel Dänzer <michel@tungstengraphics.com> 25783Date: Fri Aug 24 09:21:39 2007 +0200 25784 25785 Require xorg-server >= 1.3 for RandR 1.2. 25786 25787commit d7230939f523610c57f92bdfc72966bdbc6f1070 25788Author: Michel Dänzer <michel@tungstengraphics.com> 25789Date: Fri Aug 24 09:21:14 2007 +0200 25790 25791 64 bit warning fixes. 25792 25793 For printf vs. CARD32, use %u or %x and and a cast to unsigned. 25794 25795commit 91c45fedfd155a153dcd2c3f3e30986bfbd44e6f 25796Author: Dave Airlie <airlied@nx6125b.(none)> 25797Date: Fri Aug 24 15:05:01 2007 +1000 25798 25799 radeon: don't disable dac if either tv or vga is using it 25800 25801 On my rs480 I had to vt switch to get hotplug VGA working due to the tv-out 25802 code turning off the dac when the vga code was actually using it. 25803 25804commit 056ca6bb5adf974290693b55de6cd6880d2132d1 25805Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25806Date: Thu Aug 23 20:19:47 2007 -0400 25807 25808 RADEON: NONE to None to match other port info 25809 25810commit 13fd53286bdda2c55683bdb5f63e7d345f6c63ef 25811Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25812Date: Thu Aug 23 20:17:51 2007 -0400 25813 25814 RADEON: set (hopefully) reasonable default max desktop sizes 25815 25816 Based on the amount of vram. We really need ttm... 25817 25818commit 53bad86ca48a9b6529c1f0989ee568d9d48841c6 25819Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25820Date: Thu Aug 23 19:24:59 2007 -0400 25821 25822 RADEON: clean up the logic in crtc_mode_set() 25823 25824commit 4712dedea225e9e07177aebda2ffc6290d1f53c7 25825Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25826Date: Thu Aug 23 18:51:34 2007 -0400 25827 25828 RADEON: Always assume LVDS is connected 25829 25830 Not all bioses seem to set the right scratch bits. If we have 25831 LVDS (via bios table or otherwise) assume it's connected. 25832 25833commit b0f170c5f736ecba1a5899d602c4173fe9b9b1fa 25834Author: Michel Dänzer <michel@tungstengraphics.com> 25835Date: Thu Aug 23 12:55:40 2007 +0200 25836 25837 radeon: Remove unnecessary #include <time.h>. 25838 25839commit ac54c0e4360099697755d14b1030def73d8235b0 25840Author: Michel Dänzer <michel@tungstengraphics.com> 25841Date: Wed Aug 22 14:33:59 2007 +0200 25842 25843 radeon: Warning fixes. 25844 25845commit de26e406f52b3b13f03eee2b8023924ec6406f0a 25846Author: Alon Ziv <alonz@nolaviz.org> 25847Date: Mon Jul 30 22:47:59 2007 +0300 25848 25849 radeon: Sane handling of timeouts in WaitForVerticalSync(2). 25850 25851 RADEONWaitForVerticalSync() and RADEONWaitForVerticalSync2() need to wait 25852 for a timeout specified in milliseconds; looping around usleep() causes 25853 the timeout to be unnecessarily long, as the OS may sleep longer than 25854 requested (on Linux the minimum actual sleep value may be several ms). 25855 25856 The new logic uses gettimeofday() in the loop to see when the (absolute) 25857 timeout has arrived. 25858 25859 Signed-off-by: Alon Ziv <alonz@nolaviz.org> 25860 25861commit c66e5de26ae93caa368213f3cce139aacec955d2 25862Author: Sascha Sommer <saschasommer@freenet.de> 25863Date: Thu Aug 23 12:11:51 2007 +0200 25864 25865 radeon: Round down RMX stretch ratios. 25866 25867 Fixes issues with RMX scaling, see 25868 https://bugs.freedesktop.org/show_bug.cgi?id=8983 . 25869 25870commit 633c1fff10a3be4c9f48c1995e330d60bf6abbb2 25871Author: Michel Dänzer <michel@tungstengraphics.com> 25872Date: Thu Aug 23 12:11:41 2007 +0200 25873 25874 radeon: Sync pages when enabling page flipping with EXA as well. 25875 25876 Exclude the DRI window(s) though to avoid scribbling over 3D rendering. 25877 25878commit 4f8010ce22043c0f8d60c0f49d270ce98c9d2466 25879Author: Michel Dänzer <michel@tungstengraphics.com> 25880Date: Thu Aug 23 12:11:12 2007 +0200 25881 25882 radeon: Don't synchronize DRI windows between pages when possible. 25883 25884commit 7b527054a7c81d1d1dbc79d41b9e53064dab68cb 25885Author: Michel Dänzer <michel@tungstengraphics.com> 25886Date: Thu Aug 23 12:10:33 2007 +0200 25887 25888 radeon: Restore memmap registers even if only AGP location changed. 25889 25890commit d7ba9f001c0ab645984526afd0e64d1c6a6d654a 25891Author: Michel Dänzer <michel@tungstengraphics.com> 25892Date: Thu Aug 23 11:39:54 2007 +0200 25893 25894 radeon: Change a test to info->IsIGP from several IGP families. 25895 25896commit 8c7c22e22b6076abc80e4e1aaa8d1f4cf2f3ed14 25897Author: Michel Dänzer <michel@tungstengraphics.com> 25898Date: Thu Aug 23 11:38:17 2007 +0200 25899 25900 radeon: Wait for pending overlay flip to finish before emitting new one. 25901 25902commit 92fa7cc00688d7bfc1fb72e645ac30c6d92669c6 25903Author: Michel Dänzer <michel@tungstengraphics.com> 25904Date: Thu Aug 23 11:38:16 2007 +0200 25905 25906 radeon: Don't call RADEONDRIRefreshArea when the damaged region is empty. 25907 25908commit 5cb20c2dc5eca9d7d7d78e9924ea1b90076e7253 25909Author: Dave Airlie <airlied@linux.ie> 25910Date: Thu Aug 23 18:46:41 2007 +1000 25911 25912 radeon: bug 11899 + debian 435040 - test for usefbdev before calling int10 25913 (cherry picked from commit 3334c247540b113f9c4a5ce1879d8648a45a959a) 25914 25915commit 9d6261c024c697a211da4b6300420153eb72a264 25916Author: Roland Bär <roland@verifysoft.de> 25917Date: Thu Aug 23 18:33:34 2007 +1000 25918 25919 radeon: bug 11860 fix some mem leaks 25920 25921commit 7bc1f862bc5f992f213143fbafef52459ba7db4a 25922Author: Roland Bär <roland@verifysoft.de> 25923Date: Thu Aug 23 18:37:35 2007 +1000 25924 25925 radeon: bug 11861 - dead code removal in radeon_video.c 25926 (cherry picked from commit 8e3a6f83016cd8c4cfd43ceee4cbf0a8dc018b2a) 25927 25928commit 9d38c8aa1a7d6fb1af41ee8abdb4a95f94843538 25929Author: Dave Airlie <airlied@clockmaker.usersys.redhat.com> 25930Date: Thu Aug 23 20:10:24 2007 +1000 25931 25932 radeon: cleanup some warnings 25933 25934commit e4c8969b48a6c8dcc4e7f9852479d24a0204fc0d 25935Author: Dave Airlie <airlied@redhat.com> 25936Date: Thu Aug 23 19:56:21 2007 +1000 25937 25938 updated release numbering for randr 1.2 25939 25940commit 81a8093f2ffdce59d4e8a44f65f3d5c771fd6425 25941Merge: c08e6ec9 53a67e31 25942Author: Dave Airlie <airlied@redhat.com> 25943Date: Thu Aug 23 19:51:47 2007 +1000 25944 25945 Merge remote branch 'origin/randr-1.2' 25946 25947commit c08e6ec9e7ac06caee53689b0ec50ef6a7a0ba37 25948Merge: 5793e875 a0d0fcd3 25949Author: Dave Airlie <airlied@clockmaker.usersys.redhat.com> 25950Date: Thu Aug 23 19:42:19 2007 +1000 25951 25952 Merge branch 'master' into randr-merge 25953 25954 Conflicts: 25955 25956 src/radeon.h 25957 src/radeon_bios.c 25958 src/radeon_display.c 25959 src/radeon_dri.c 25960 src/radeon_driver.c 25961 src/radeon_modes.c 25962 src/radeon_probe.h 25963 src/radeon_video.c 25964 25965commit 53a67e31904bec9a3aa1bd24de8034dcafea1d2a 25966Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25967Date: Tue Aug 21 21:20:41 2007 -0400 25968 25969 RADEON: Fix color problem on pre-R3xx chips tv-out 25970 25971commit e9719e8e02eef46717ae9b4d8c7998466dac30cb 25972Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25973Date: Tue Aug 21 21:17:20 2007 -0400 25974 25975 RADEON: more tv out fixes and clean up 25976 25977commit 36c22a49580d86a6518b67f31a78bd53d39491af 25978Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 25979Date: Tue Aug 21 20:28:39 2007 -0400 25980 25981 RADEON: fix tv-out on R3xx 25982 25983 R3xx apparently needs the tv clock forced on. 25984 25985commit 9470bd67731059f26859ed5f0bea3ade09e2c80c 25986Author: Alex Deucher <alex@botch2.com> 25987Date: Mon Aug 20 20:54:06 2007 -0400 25988 25989 RADEON: Add DefaultConnectorTable option 25990 25991 This option skips the parsing the BIOS connector table 25992 and falls back to chip specific defaults. 25993 Also remove man page section for the now gone bioshotkeys 25994 option. 25995 25996commit a0d0fcd3bab765b4db25e04884fd8a342abb9c66 25997Author: iLisa Wu <liswu@ati.com> 25998Date: Mon Aug 20 15:44:07 2007 +0200 25999 26000 Fix crash with no valid mode in xorg.conf's modelist and empty Virtual 26001 26002 If the resolution defined in xorg.conf failed to find a matching mode in the 26003 supported modelist, and no virtual desktop dimensions are defined the xorg.conf 26004 either, virtual X and Y dimension will be set to 0 which will cause Xserver 26005 crash. 26006 26007 (Novell bugzilla #296856, closed) 26008 26009commit b275febdb0918e8cebdffbb433b0eeb3ff4d3746 26010Author: Alex Deucher <alex@samba.(none)> 26011Date: Sun Aug 19 20:55:32 2007 -0400 26012 26013 RADEON: turn off TVCLK when blanking tv encoder 26014 26015commit a90d675832ddb02c81ace010ccbf02619b70edac 26016Author: Alex Deucher <alex@botch2.com> 26017Date: Thu Aug 16 21:55:14 2007 -0400 26018 26019 RADEON: fix Xv clipping and overlay sourcing 26020 26021 - Basically just copied from the intel driver. I'm planning to push 26022 this to the server soon, but add it now to get things working 26023 and to provide compat for older servers. 26024 26025 - Overlay crtc source control attribute now called XV_CRTC 26026 The old attribute XV_SWITCHCRT has been removed. If anyone cares, 26027 we can add it back as an alias to XV_CRTC 26028 XV_CRTC: -1 auto, 0 crtc0, 1 crtc1 26029 26030commit 5793e8753d11432bf95c7c6dd80c811e16aba058 26031Author: Alex Deucher <alex@botch2.com> 26032Date: Wed Aug 15 19:26:36 2007 -0400 26033 26034 RADEON: Remove RADEONRestoreMode() 26035 26036 Since we no longer use it to write modes all it did 26037 was restore some of the regs for a console restore. 26038 Just move the relevant bits into RADEONRestore() and 26039 remove it. 26040 26041commit 6f011aaabaf18d66ffc255ad76aaf938b2396302 26042Author: Lisa Wu <Lisa.Wu@amd.com> 26043Date: Wed Aug 15 19:17:51 2007 -0400 26044 26045 RADEON: fix console restore on r3xx and r4xx 26046 26047 When restoring the console we need to: 26048 1) we need to restore DAC registers after all other registers are 26049 restored and CRTCs are enabled. 26050 2) we need to enable CRTC2 registers before CRTC1 registers 26051 26052commit c8dad98abb042c6abbbee18f9ae4db72084bc513 26053Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26054Date: Wed Aug 15 00:55:24 2007 -0400 26055 26056 RADEON: make sure crtc routing is correct in r300_detect_tv() 26057 26058 Also remove some unused variable 26059 26060commit f8cd74435f0072dbf5f6e83d67d2d5e1f4e82c91 26061Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26062Date: Tue Aug 14 23:50:57 2007 -0400 26063 26064 RADEON: minor cleanup 26065 26066commit 6f398cd07ea734dd66a8eac71b629e59123d75b8 26067Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26068Date: Tue Aug 14 23:42:32 2007 -0400 26069 26070 RADEON: Implement improved tv load detection for r300 26071 26072 The previous implementation resulted in false positives 26073 on occasion. This method works much more reliably. 26074 Based on beos code by Thomas Kurschel 26075 26076commit 366a1d4c240ac93622caff97b652696db99bf2e6 26077Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26078Date: Tue Aug 14 23:25:37 2007 -0400 26079 26080 RADEON: Turn off tv encoder in disableoutputs() 26081 26082commit 5b4a04c23e3f1ec2490418b111f417c16463c709 26083Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26084Date: Tue Aug 14 23:11:52 2007 -0400 26085 26086 RADEON: Make sure RMX is always programmed for crtc1 26087 26088 If the RMX registers are left as programmed by the bios 26089 this can lead to a blank screen when crtc1 is feeding a 26090 DAC. Fix found and reported by Lisa Wu <liswu@ati.com> 26091 Fixes bug 11985 26092 26093commit db2a828b2f21b92cd654b309d137204334975b89 26094Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26095Date: Tue Aug 14 22:47:29 2007 -0400 26096 26097 RADEON: Add load detection 26098 26099 Based on the beos driver by Thomas Kurschel and the 26100 existing load detection code in this driver. 26101 26102commit 6fc3ddbbb6fbbee1f6076c776e5b46c0c772b6d4 26103Author: Alex Deucher <alex@botch2.com> 26104Date: Sun Aug 12 12:00:20 2007 -0400 26105 26106 RADEON: order the VGA and DVI ports correctly 26107 26108commit b7738d2ef82e1759adf78e5db1291f8739b4166f 26109Author: Alex Deucher <alex@botch2.com> 26110Date: Sat Aug 11 17:50:42 2007 -0400 26111 26112 RADEON: fix connector setup when there's no bios tables 26113 26114commit d0895f67e327bb268fd59fcfd8fc22678d804f57 26115Author: Luc Verhaegen <libv@skynet.be> 26116Date: Thu Aug 9 12:11:40 2007 +0200 26117 26118 AGPFastWrite risk reduction. 26119 26120 Actively warn the user in the log about the effects of AGPFastWrite and 26121 sanitise AGPFastWrite handling while we're here. 26122 26123commit c01000bc684f6a23a38a52f0808182cefaa686bb 26124Author: Alex Deucher <alex@botch2.com> 26125Date: Wed Aug 8 00:07:16 2007 -0400 26126 26127 RADEON: fix typo in comment 26128 26129commit 5c3f49e651c36f3bd14fa29359e24825d8f7f77f 26130Author: Lisa Wu <liswu@ati.com> 26131Date: Wed Aug 8 00:05:47 2007 -0400 26132 26133 RADEON: make sure RADEON_CRTC2_CRT2_ON bit state is properly accounted for 26134 26135 - when the TV DAC is used for crtc1 make sure to keep this bit set when 26136 initializing crtc2. 26137 - fixes bug 11894 26138 26139commit 08fe7ad00fba523775e95b5e6295fe23a5119d60 26140Author: Alex Deucher <alex@botch2.com> 26141Date: Tue Aug 7 23:16:05 2007 -0400 26142 26143 RADEON: fix ConnectorTable option after connector table re-work 26144 26145 We may want to add the possiblity of more connectors to this option 26146 26147commit f56b90a60393ed187c0e39e149b2a3a32331ea52 26148Author: Alex Deucher <alex@botch2.com> 26149Date: Tue Aug 7 23:08:02 2007 -0400 26150 26151 RADEON: Make sure the default TV standard is supported 26152 26153commit a88a0e77eec514325b4d07bf7be9fb04e6f8e244 26154Author: Alex Deucher <alex@botch2.com> 26155Date: Tue Aug 7 00:43:20 2007 -0400 26156 26157 RADEON: minor fix to legacy bios connector table 26158 26159commit 3752808d6c08a9727370ef8d79088e787791e131 26160Author: Alex Deucher <alex@botch2.com> 26161Date: Sun Aug 5 16:13:39 2007 -0400 26162 26163 RADEON: Always set MonType in radeon_detect() 26164 26165 - Always set the the MonType in radeon detect even if no 26166 monitor is detected so users can still force disconnected 26167 outputs on. For DVI-I users will have to set the 26168 dvi_monitor_type attribute to force digital or analog, 26169 everything else should just work. 26170 26171commit b4ec3e436afb5bc99ec755cbd96eee9a1ec492a5 26172Author: Alex Deucher <alex@botch2.com> 26173Date: Sun Aug 5 15:37:04 2007 -0400 26174 26175 RADEON: add tv out properties 26176 26177commit 1fc2a1120e7c05938e2bd72d3c7837ecff8bc9da 26178Author: Alex Deucher <alex@botch2.com> 26179Date: Sun Aug 5 14:41:57 2007 -0400 26180 26181 RADEON: switch DVI-I monitor type attribute to string 26182 26183commit 7d8eb3751d74bd8a1fb9fa2d2fcb9c4c895f6ba4 26184Author: Alex Deucher <alex@botch2.com> 26185Date: Sun Aug 5 14:31:18 2007 -0400 26186 26187 RADEON: make backlight attribute lower case 26188 26189commit e71bb88bc9052af8866fb6945dbc06dbb0e6d1c0 26190Author: Alex Deucher <alex@botch2.com> 26191Date: Sun Aug 5 14:26:15 2007 -0400 26192 26193 RADEON: switch RMX attribute to string type 26194 26195 Still not actually hooked up. 26196 26197commit 284323135ec04635dfa9cabd5790a35b953abca2 26198Author: Alex Deucher <alex@botch2.com> 26199Date: Sun Aug 5 04:35:24 2007 -0400 26200 26201 RADEON: typo 26202 26203commit d1abdad167aa24ac970c69422435df443c82ebd6 26204Author: Alex Deucher <alex@botch2.com> 26205Date: Sun Aug 5 03:45:02 2007 -0400 26206 26207 RADEON: fixes 26208 26209 - fix output ordering 26210 - set tv-out to return un-connected for now in radeon_detect() 26211 26212commit 0cca25d8d6a0cb0d29b68e6cd9c699d1390aede1 26213Author: Alex Deucher <alex@botch2.com> 26214Date: Sun Aug 5 03:19:24 2007 -0400 26215 26216 RADEON: Fix DVI-I support in ATOM bios connector table parsing 26217 26218commit ba5496ae7973786802962bf649dd91c219531749 26219Author: Alex Deucher <alex@botch2.com> 26220Date: Sun Aug 5 02:27:32 2007 -0400 26221 26222 RADEON: refactor output init to handle multiple DVI or VGA 26223 26224 - refactor output init to handle multiple DVI or VGA with the new 26225 bios table parsing 26226 26227commit 2ec22783ddf4c522df9e5fd1b2003854486d7a2b 26228Author: Alex Deucher <alex@botch2.com> 26229Date: Sun Aug 5 01:39:35 2007 -0400 26230 26231 RADEON: attempt to do the right thing for standards other than PAL or NTSC 26232 26233commit cc8e1d95f1b90a259beea4e8cc4d7e29af660919 26234Author: Alex Deucher <alex@botch2.com> 26235Date: Sun Aug 5 01:14:36 2007 -0400 26236 26237 RADEON: Major rework of BIOS table parsing 26238 26239 - greatly simplify ATOM and legacy connector table parsing 26240 - use bios tables to detect LVDS and TV outputs 26241 - add support for TV table parsing (legacy only) 26242 26243commit 1de52d91ff3a04b9b587b858e1e5be40d3a7fd0a 26244Author: Dave Airlie <airlied@linux.ie> 26245Date: Sat Aug 4 17:58:58 2007 +1000 26246 26247 update configure.ac for 6.6.193 release 26248 26249commit 371001c0433db1d17e468f3ea99ea57f922145a3 26250Author: Dave Airlie <airlied@linux.ie> 26251Date: Sat Aug 4 17:51:19 2007 +1000 26252 26253 radeon: remove unused variables 26254 26255commit 90946c53d7a4b23d03270ad6da0450759a11de3d 26256Author: Dave Airlie <airlied@linux.ie> 26257Date: Sat Aug 4 17:51:06 2007 +1000 26258 26259 ati: add -Wall for gcc 26260 26261 As per luc's "suggestion" in Novell bug. 26262 26263 This may generate warnings on 64-bit until we get rid of CARD32 26264 26265commit e30a145934df8f6a7f71290d6c75e4239f9d52f7 26266Author: Dave Airlie <airlied@linux.ie> 26267Date: Sat Aug 4 17:44:46 2007 +1000 26268 26269 radeon: cleanup some pieces of the dpms/blank register programming 26270 26271commit a156db5e8b037ed12a448f70045453baf9d0c504 26272Author: Luc Verhaegen <libv@skynet.be> 26273Date: Sat Aug 4 17:37:18 2007 +1000 26274 26275 Clean up PortInfo to CRTC mapping 26276 26277 Also sanitise blanking and DPMS functions 26278 26279 Fixes from Novell Bug 264720, and fd.o 10772 26280 26281commit b61a49f2a5401560f85e11bcdd005287433cad12 26282Author: Alex Deucher <alex@botch2.com> 26283Date: Fri Aug 3 19:27:59 2007 -0400 26284 26285 RADEON: Change indexing of TV constants table in preparation for standard re-work 26286 26287commit 288fa627274cb399059262d4f8bd844fc220a042 26288Author: Alex Deucher <alex@botch2.com> 26289Date: Thu Aug 2 02:37:16 2007 -0400 26290 26291 RADEON: avoid a divide by 0 and only save tv out regs if the chip has them 26292 26293commit b66a1bc7994b33d349c1519761e431959311c85f 26294Author: Alex Deucher <alex@botch2.com> 26295Date: Thu Aug 2 02:11:20 2007 -0400 26296 26297 RADEON: fix tv-out enable/disable 26298 26299commit d86592c8d5ce45d81d8a726c263e870e94fbcf11 26300Author: Alex Deucher <alex@botch2.com> 26301Date: Thu Aug 2 00:50:51 2007 -0400 26302 26303 RADEON: add missing break 26304 26305commit 971feb34843225030fff05b3f9d3801534fbf2d4 26306Author: Alex Deucher <alex@botch2.com> 26307Date: Thu Aug 2 00:50:04 2007 -0400 26308 26309 RADEON: move tv dac enable to enabledisplay() 26310 26311commit 98d7e00437bea78e03180eb30ff30de3455d9d1a 26312Author: Alex Deucher <alex@botch2.com> 26313Date: Thu Aug 2 00:20:50 2007 -0400 26314 26315 RADEON: limit tv modes to the only one we can program at the moment 26316 26317commit 5c549c1d42f7bbc556942af13aff2661fae856f2 26318Author: Alex Deucher <alex@botch2.com> 26319Date: Wed Aug 1 23:45:07 2007 -0400 26320 26321 RADEON: remove unused elements 26322 26323commit b03978028fd975eb6946503d3a56a49c5a67f339 26324Author: Alex Deucher <alex@botch2.com> 26325Date: Wed Aug 1 22:39:16 2007 -0400 26326 26327 RADEON: convert hard coded tv out values to calculations 26328 26329commit cf54222f1fa37366b2c2b39c82f8afc02f32e63c 26330Author: Alex Deucher <alex@botch2.com> 26331Date: Tue Jul 31 02:01:49 2007 -0400 26332 26333 RADEON: more fixes... 26334 26335commit 22d460d3ad991223aa1fbd7e5edeb45e36c65dc0 26336Author: Alex Deucher <alex@botch2.com> 26337Date: Tue Jul 31 01:28:05 2007 -0400 26338 26339 RADEON: fix name of tv output 26340 26341commit 4822a2b837334f408f962646ab5ea4f8b0335ac9 26342Author: Alex Deucher <alex@botch2.com> 26343Date: Tue Jul 31 01:18:40 2007 -0400 26344 26345 RADEON: tv-out fixes. works now. tested on rv350. 26346 26347 VT siwtch is busted, and xrandr doesn't play nice yet. 26348 uncomment code in radeon_output.c to test. 26349 26350commit fe494c9db2995bb8ce7a028ecf9626e0cb0cf506 26351Author: Alex Deucher <alex@botch2.com> 26352Date: Sun Jul 29 15:26:34 2007 -0400 26353 26354 RADEON: add info about tv out code and authorship 26355 26356commit 6b9b7a7bdc290d07de9b226691ec8025af8db896 26357Author: Alex Deucher <alex@botch2.com> 26358Date: Sun Jul 29 15:23:14 2007 -0400 26359 26360 RADEON: Initial pass at integrated tv out support 26361 26362 Based on the GATOS tv-out support by Federico Ulivi <fulivi@lycos.com> 26363 and information from ati with substantial rework by myself. 26364 26365 Code is not actually hooked up yet. 26366 26367commit 8d043db1817d94edeb72ab208dfea60026715d48 26368Author: Alex Deucher <alex@botch2.com> 26369Date: Wed Jul 25 20:37:58 2007 -0400 26370 26371 RADEON: Compute PLL VCO gain 26372 26373commit 62f06d89da3f7160d5e4df8d7ce6fe1a94e9d07c 26374Author: Alex Deucher <alex@botch2.com> 26375Date: Wed Jul 25 20:22:25 2007 -0400 26376 26377 RADEON: write out saved vclk and pixclk values 26378 26379commit 165a07cbbfcd94e3d1fac434b8fada8d29428a09 26380Author: Brice Goglin <Brice.Goglin@ens-lyon.org> 26381Date: Wed Jul 25 19:01:02 2007 +0200 26382 26383 Minor fixes in the manpages 26384 26385 - ati.man 26386 . add references to radeon and r128 manpages in SEE ALSO 26387 (Xavier Bestel in Debian bug #386001). 26388 . remove the manpage suffix from atimisc reference in the text 26389 since there is no such manpage for now. 26390 - radeon.man 26391 . typo in "specifying" (reported by A. Costa in Debian bug #432059). 26392 . replace a non-ascii character with the corresponding groff escape 26393 sequence so that it works whatever the locale (Julien Cristau). 26394 26395commit 9cc3ab8320162f371bba15dc131f23c5de2013fc 26396Author: Alex Deucher <alex@botch2.com> 26397Date: Wed Jul 25 02:24:14 2007 -0400 26398 26399 RADEON: hack around console restore hang on r4xx 26400 26401 r4xx seems to hang when unblanking the crtc(s) when 26402 restoring the console. This needs to be investigated 26403 further. 26404 26405commit 193d58b0ea9de3e326017e50b88e2f3726854506 26406Author: Alex Deucher <alex@botch2.com> 26407Date: Wed Jul 25 01:27:58 2007 -0400 26408 26409 RADEON: fix randr on r4xx 26410 26411 Seems the way we parse the ATOM connector table results 26412 in reversed connectors. 26413 26414commit 262e32bd9ea4123b116362b33b5798753b2568fe 26415Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26416Date: Thu Jul 19 23:48:57 2007 -0400 26417 26418 RADEON: fix palette when depth != 24 26419 26420commit 18b00b47a483e7854727e99126808ddf361e7a4a 26421Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26422Date: Thu Jul 19 19:01:42 2007 -0400 26423 26424 R128: don't clip modes to panel on laptops when only using the CRT port 26425 26426 fixes bug 5832 26427 26428commit 882fe7631586b0a7919f808588a2ea4fb555f7e8 26429Author: Brice Goglin <brice.goglin@ens-lyon.org> 26430Date: Thu Jul 19 18:44:27 2007 -0400 26431 26432 MACH64: avoid crash in Xv code 26433 26434 fixes bug 11054 26435 26436commit 7b3e22e96cfc98621bd20fa76317e6d8f7242165 26437Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26438Date: Thu Jul 5 20:37:10 2007 -0400 26439 26440 RADEON: set default values for crtc_offset_cntls 26441 26442commit 2a35ccfb8ffee9f0ef6f8d4f0eeb80a471543caa 26443Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26444Date: Thu Jul 5 20:04:04 2007 -0400 26445 26446 RADEON: Fix tiling on r1xx and r2xx 26447 26448 spotted by Andrew Randrianasulu <randrik@mail.ru> 26449 fixes bug 11357 26450 26451commit ac9cbaf32176cf144bc694ac879e7c9e2920f762 26452Author: Michel Dänzer <michel@tungstengraphics.com> 26453Date: Mon Jul 2 10:06:02 2007 +0200 26454 26455 radeon: Fully zero-initialize info->CRT2pScrn->monitor. 26456 26457 This prevents modes from getting rejected due to the uninitialized maxPixClock 26458 value. 26459 26460 Fixes http://bugs.freedesktop.org/show_bug.cgi?id=9816 . 26461 26462commit bdcae622100c81a4d9a53938542b64908bacd195 26463Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26464Date: Fri Jun 29 01:15:57 2007 -0400 26465 26466 RADEON: save crtc on/off state for console restore 26467 26468 - prevents a possible hang if console is only using 1 crtc 26469 26470commit 21be0d3a1f4700572c6425800596785d6850626f 26471Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26472Date: Fri Jun 29 00:38:50 2007 -0400 26473 26474 RADEON: simplify console restore 26475 26476 - still need to track crtcs for blank/unblank when restoring text console 26477 26478commit d9bf28b55d98fea2d285c9c46362aaf2175d0a46 26479Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26480Date: Thu Jun 28 23:52:28 2007 -0400 26481 26482 RADEON: factor out surface_cntl init into one function 26483 26484commit 9f193985627be8e6ea1418a424e825ddbc4957b2 26485Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26486Date: Thu Jun 28 23:43:13 2007 -0400 26487 26488 RADEON: move crtc base setups to new functions 26489 26490commit 0f361e9e80a29d287fa42436c32c657e3c102539 26491Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26492Date: Thu Jun 28 23:08:07 2007 -0400 26493 26494 RADEON: fix corruption after 3D apps run 26495 26496commit 1d4630067d293d26284d5fe230debef3913ec6cf 26497Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26498Date: Thu Jun 28 22:24:04 2007 -0400 26499 26500 RADEON: set info->CurrentLayout.displayWidth in RADEONSetPitch() 26501 26502commit 2754d1ba01fc3367019487e0c0f59d74c950aaa7 26503Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26504Date: Thu Jun 28 22:15:22 2007 -0400 26505 26506 RADEON: Switch cursor back to memcpy() 26507 26508commit 7901bcafa92dccd319ddb5de4627d806a39f15f9 26509Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26510Date: Mon Jun 25 18:39:14 2007 -0400 26511 26512 RADEON: make sure we unblank in RADEONRestore() 26513 26514 reported by Andrew Randrianasulu (bugs 11357, 11146) 26515 26516commit 6a3599d6155d073a3116c4b13bdf04b44bb9e087 26517Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26518Date: Mon Jun 25 18:20:54 2007 -0400 26519 26520 RADEON: make sure we set the stride properly 26521 26522commit 771c37fcac8104b894120cc3fc0154b31b0318d1 26523Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26524Date: Mon Jun 25 17:53:26 2007 -0400 26525 26526 RADEON: attempt to fix cursor on big endian (only tested on x86) 26527 26528commit 5b5b90c2cea7e36895354f5872acd3fc769653f9 26529Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26530Date: Mon Jun 25 16:27:54 2007 -0400 26531 26532 RADEON: fix possible segfault on mobility chips with MM tables 26533 26534 spotted by Stefan Buehler 26535 26536commit 01bd5eb713500d3e3d4351865d460c8d1e476454 26537Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26538Date: Mon Jun 25 16:16:18 2007 -0400 26539 26540 RADEON: fix possible segfault on mobility chips with MM tables 26541 26542 reported by Stefan Buehler 26543 26544commit 52cc1dc1491559a9055f3ba6dd54064bb382ad86 26545Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26546Date: Mon Jun 25 15:49:19 2007 -0400 26547 26548 RADEON: tiling and FB size fixes 26549 26550 - fix tiling on older radeons 26551 - allow the user to specify FB size using Virtual line in config 26552 Current default is 1600x1200 26553 - reset ecp_div for Xv in randr crtc mode set 26554 26555commit 3bb46c03fddd3bd79bdab887366aeeced0290a3a 26556Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26557Date: Fri Jun 22 01:12:38 2007 -0400 26558 26559 RADEON: bump crtc max to 1200 so 1600x1200 will work out of the box 26560 26561 - a better fix will be forthcoming... 26562 26563commit b9fac5695bf8f123b391e3b162c4865258a685dd 26564Author: Andrew Randrianasulu <randrik@mail.ru> 26565Date: Fri Jun 22 01:01:14 2007 -0400 26566 26567 RADEON: Fix VT switch hangs 26568 26569 - re-order DRI resume functions in EnterVT() 26570 - fixes bug 11287 26571 26572commit 92e65d5e0d6817ff4c9a08020a0a9b3a8c3c98b0 26573Author: Andrew Randrianasulu <randrik@mail.ru> 26574Date: Fri Jun 22 00:55:00 2007 -0400 26575 26576 RADEON: Fix VT switch hangs 26577 26578 - reorder RADEONDRISetVBlankInterrupt() and RADEONDRIResume() 26579 - see bug 11287 26580 26581commit ec1fd2c388ec2c4033644776ee8588405a3d25f1 26582Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26583Date: Tue Jun 19 00:39:59 2007 -0400 26584 26585 RADEON: Split FP reg restore to only touch regs relevant to the output 26586 26587 - only restore output specific regs 26588 - restore fp2_gen_cntl on R200 DAC restore as R200 DAC2 is connected via 26589 DVO 26590 26591commit 0f5a92667281ca3f80218bc67d4031125a5e4793 26592Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26593Date: Mon Jun 18 23:35:21 2007 -0400 26594 26595 RADEON: only touch LVDS on mobility chips 26596 26597commit b72ff160f908bf3aa9f64705377e92d80360a4f7 26598Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26599Date: Mon Jun 18 21:51:45 2007 -0400 26600 26601 RADEON: only touch LVDS on mobility chips 26602 26603commit 80313621ffa8c409ae63fc1b28c15fc4abdc3a7b 26604Author: Henry Zhao <henryz@localhost.localdomain> 26605Date: Fri Jun 15 17:00:05 2007 -0700 26606 26607 Update CRT2pScrn->monitor->Last to reflect the 26608 last mode of CRT2pScrn. See bug 11278. 26609 26610commit 5868af53204e8fc210735e68e891e15085285629 26611Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26612Date: Tue Jun 12 23:53:48 2007 -0400 26613 26614 RADEON: yet another LVDS... 26615 26616commit f19a6f7ee5bf4ec632e7813359f167599c08e823 26617Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26618Date: Tue Jun 12 23:44:41 2007 -0400 26619 26620 RADEON: additional LVDS off fix (missed in previous commit) 26621 26622commit 4f2a1ba9c2fc20042d67132f986a86a9783a245e 26623Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26624Date: Tue Jun 12 23:41:24 2007 -0400 26625 26626 RADEON: additional LVDS off fix (missed in previous commit) 26627 26628commit 9fc416ce08cc5d74faa7b184e80ce43ed2f4fbee 26629Author: Michel Dänzer <michel@tungstengraphics.com> 26630Date: Tue Jun 12 23:35:09 2007 -0400 26631 26632 radeon: Fix panel size detection from registers with stretched mode programmed. 26633 26634commit c4045cbc5e6975520efef422ce9df2a7d99a7ff4 26635Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26636Date: Tue Jun 12 23:32:18 2007 -0400 26637 26638 RADEON: make sure to clear RADEON_LVDS_ON when turning LVDS off 26639 26640 - fixes bug 3483 26641 26642commit 090888d3baf61a70640d28e99afdce20952cdea0 26643Author: Tormod Volden <bugzi06.fdo.tormod@xoxy.net> 26644Date: Tue Jun 12 23:30:18 2007 -0400 26645 26646 RADEON: limit PanelPwrDly to 2000 ms on ATOM bios as per legacy bios 26647 26648 - fixes bug 11238 26649 26650commit 44748a826b52924a6e050215d8d62755e0ac7fb2 26651Author: Tormod Volden <bugzi06.fdo.tormod@xoxy.net> 26652Date: Tue Jun 12 23:23:49 2007 -0400 26653 26654 RADEON: limit PanelPwrDly to 2000 ms on ATOM bios as per legacy bios 26655 26656 - fixes bug 11238 26657 26658commit c292fc64499ff4cc135c07deda99cf4169f8fef4 26659Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26660Date: Tue Jun 12 23:20:18 2007 -0400 26661 26662 RADEON: make sure RADEON_LVDS_ON is cleared when turning LVDS off 26663 26664 - fixes bug 3483 26665 26666commit c6a3286d6f6ddda89115d98d45665dadd78bf41d 26667Author: Michel Dänzer <michel@tungstengraphics.com> 26668Date: Mon Jun 11 09:39:38 2007 +0200 26669 26670 radeon: Fix panel size detection from registers with stretched mode programmed. 26671 26672commit 7a6b3c0f77c69019268f585c51c3dcbcc99014d4 26673Author: Michel Dänzer <michel@tungstengraphics.com> 26674Date: Mon Jun 11 09:39:38 2007 +0200 26675 26676 radeon: Don't loop indefinitely if no mode matches detected panel size. 26677 26678commit f31fd9ce598841c505a0b5ed32bf124f49fea332 26679Author: Michel Dänzer <michel@tungstengraphics.com> 26680Date: Mon Jun 11 09:39:38 2007 +0200 26681 26682 radeon: Fix some more ErrorFs when setting mode. 26683 26684commit 9dd9f09a9021653e5590d27bd7aa0e7af85416ad 26685Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26686Date: Thu Jun 7 15:32:31 2007 -0400 26687 26688 RADEON: fix some issues with bios table init code 26689 26690 - fix some issues with bios table init code 26691 - re-org preinit to better handle bios table init of cards 26692 26693commit 9ad311eeb0ac2b70a862dbb9de278154ba9142ec 26694Author: Tilman Sauerbeck <tilman@code-monkey.de> 26695Date: Thu Jun 7 11:03:02 2007 +0200 26696 26697 Fixed 'make dist'. 26698 26699commit 2527f2b69aa7dffa3ba4359c45955e3185bdf0e6 26700Author: Alex Deucher <alex@botch2.com> 26701Date: Wed Jun 6 00:08:45 2007 -0400 26702 26703 RADEON: implement support for posting cards based on x86 bios tables (untested) 26704 26705 This is based on the netbsd radeonfb driver by Garrett D'Amore. 26706 The code is not hooked up yet, but should allow you to post cards 26707 without needing to execute bios code. This should be useful on 26708 non-x86 platforms, for posting secndary cards, and for suspend/resume. 26709 Works on legacy bioses only (no ATOM support). 26710 26711commit e67d1420bf65055ecb6fdfe6b1b1f53aae83854a 26712Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26713Date: Tue Jun 5 22:31:50 2007 -0400 26714 26715 RADEON: fixup RADEONDisplayVideo() to better handle the crtc being used 26716 26717commit 7886405308e3288d5c86b6f2c7dbfa8ff865139c 26718Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26719Date: Mon Jun 4 21:16:56 2007 -0400 26720 26721 RADEON: fix indenting 26722 26723commit f54ad565a69deb52547fd04e123f56fc1294cd0a 26724Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26725Date: Mon Jun 4 21:13:16 2007 -0400 26726 26727 RADEON: no need to go through all the crtcs, we've got what we need 26728 26729commit 687879bff716ad01f9f158860deb8ba770faab99 26730Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26731Date: Mon Jun 4 21:08:40 2007 -0400 26732 26733 RADEON: add support for RMX on DVI, make sure RMX is only enabled on crtc1 26734 26735commit e6161e472ff266f69547704a61040228a8704b06 26736Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26737Date: Mon Jun 4 20:32:55 2007 -0400 26738 26739 RADEON: remove some cruft 26740 26741commit be0ce38232ea4f5679c5829ab663939144c6e617 26742Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26743Date: Mon Jun 4 20:22:15 2007 -0400 26744 26745 RADEON: number the outputs if there are more than one of the same 26746 26747commit 8ec617f6493dd0aea5d11f92e3d58c3feef8c8fd 26748Author: Dave Airlie <airlied@linux.ie> 26749Date: Sun Jun 3 17:32:43 2007 +1000 26750 26751 radeon: disable irqs at server start until 3D app starts 26752 26753commit 2d40fa55e8d7a1cfb204d66ca4a4d95a3b13d5b5 26754Author: Dave Airlie <airlied@linux.ie> 26755Date: Sun Jun 3 17:32:43 2007 +1000 26756 26757 radeon: disable irqs at server start until 3D app starts 26758 26759commit dcb64a4d3947e5a9fbda4b72e29a5b6102370f07 26760Author: Dave Airlie <airlied@linux.ie> 26761Date: Sun Jun 3 17:10:49 2007 +1000 26762 26763 radeon: disable vbl interrupts when no 3d is running on a new enough drm 26764 26765commit 73d8e3ec8536b4777490b7ba457566f02233811f 26766Author: Dave Airlie <airlied@linux.ie> 26767Date: Sun Jun 3 17:07:59 2007 +1000 26768 26769 randr-1.2 increase dri minor version for mesa to know we can do vbl on both 26770 26771commit 63f0d4ed0a98830ecbe18c6e4174689111a59b68 26772Author: Dave Airlie <airlied@linux.ie> 26773Date: Sun Jun 3 16:49:14 2007 +1000 26774 26775 randr-1.2: add support for vblank on both heads 26776 26777 Also disable vbl when 3D isn't running 26778 26779commit d7775c1b38b5bdc439a27ec2c3c3a03a5b24cf57 26780Author: Dave Airlie <airlied@linux.ie> 26781Date: Sun Jun 3 12:46:31 2007 +1000 26782 26783 randr-1.2: make native mode preferred 26784 26785 If we have a native mode from the BIOS make it preferred mode 26786 26787commit 5a9516fe4f17854acd4fbf3a8eadf5139081dbd4 26788Author: Paul TBBle Hampson <Paul.Hampson@pobox.com> 26789Date: Sat Jun 2 14:28:18 2007 -0400 26790 26791 RADEONProbePLLParameters sets pll->reference_div, and 26792 RADEONGetPanelInfoFromReg uses it. 26793 26794 This ensures that the former is called before the latter. 26795 (this should fix randr on ppc) 26796 26797commit 802804461ebdca9a951e7e562ec68fd08d8eae01 26798Author: Matthieu Herrb <matthieu@deville.herrb.com> 26799Date: Thu May 31 22:23:57 2007 -0600 26800 26801 Fix build whithout XF86DRI 26802 26803commit 0e0946e0aa2527794e07473199e851bbfbc47cb1 26804Author: Michel Dänzer <michel@tungstengraphics.com> 26805Date: Wed May 30 18:56:53 2007 +0200 26806 26807 radeon: Lots of warning fixes. 26808 26809 Move code where it's used, remove unused variables, etc. 26810 26811commit 31c1be420d5277dd15505bd73e6144827a0580cd 26812Author: Dave Airlie <airlied@nx6125b.(none)> 26813Date: Wed May 30 17:49:01 2007 +1000 26814 26815 remove these syncs, at least on rs480 it doesn't break 26816 26817 probably requires testing on other r300 based cards, with the syncs in 26818 we hang when moving the cursor into the second CRTC. 26819 26820commit 7fc02657c4d740941fbda5a8823cf45de3eca3f8 26821Merge: 800bf532 4c61c0ee 26822Author: Dave Airlie <airlied@nx6125b.(none)> 26823Date: Wed May 30 17:27:22 2007 +1000 26824 26825 Merge branch 'origin' into randr-1.2-test 26826 26827 Conflicts: 26828 26829 src/radeon_cursor.c 26830 src/radeon_display.c 26831 src/radeon_driver.c 26832 26833commit 800bf53279e2c2bf854682bbfd6fa16d03afed00 26834Author: Matthieu Herrb <matthieu@deville.herrb.com> 26835Date: Tue May 29 23:31:13 2007 -0600 26836 26837 Fix build without XF86DRI 26838 26839commit 4c61c0ee91a2ffeefce30972a584486f1df1d1ae 26840Author: Matthieu Herrb <matthieu@deville.herrb.com> 26841Date: Tue May 29 21:35:35 2007 -0600 26842 26843 Fix build without XF86DRI 26844 26845commit 5337e7bd0069a3f2c4ab22b21a19471427ad3d81 26846Author: Dave Airlie <airlied@nx6125b.(none)> 26847Date: Wed May 30 08:10:44 2007 +1000 26848 26849 radeon: add bios quirk for nx6125 monid 26850 26851commit bff809dc8ed07ac39e9b576a87916486a5e37156 26852Author: Dave Airlie <airlied@nx6125b.(none)> 26853Date: Wed May 30 08:02:26 2007 +1000 26854 26855 rs480: more unknown regs 26856 26857 Hardcode the values from a working fglrx run, this works for me now 26858 26859 I've no idea what it might do for anyone else 26860 26861commit 104105fee5c3945d3f210e6a4cb73ab492c61543 26862Author: Dave Airlie <airlied@nx6125b.(none)> 26863Date: Tue May 29 19:09:33 2007 +1000 26864 26865 rs480: make second crtc work with magic number in magic register. 26866 26867 I've no idea why or what this does. 26868 26869commit 5aa603bcabbb077dec169c48438c2e2ebe1195d7 26870Author: Dave Airlie <airlied@nx6125b.(none)> 26871Date: Tue May 29 07:23:24 2007 +1000 26872 26873 rs480: only has single dac 26874 26875commit c52322354fe64725733842b3356798c50e7735d5 26876Merge: dd6a966e 975da595 26877Author: Dave Airlie <airlied@nx6125b.(none)> 26878Date: Tue May 29 07:21:48 2007 +1000 26879 26880 Merge branch 'origin' 26881 26882commit c2637a01f9aca4032262c66ade305f5fe2c54294 26883Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26884Date: Sat May 26 18:26:35 2007 -0400 26885 26886 RADEON: remove some dead code from the last commit 26887 26888commit bbb769c4107bfcae682e46d026e54cbfb67d62cd 26889Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26890Date: Sat May 26 18:25:06 2007 -0400 26891 26892 RADEON: more rotation work (still not there) 26893 26894 - once again borrowed heavily from intel 26895 26896commit a69f90be9384244744fabfa76469ede9cd26ac98 26897Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26898Date: Sat May 26 17:54:48 2007 -0400 26899 26900 RADEON: first pass at rotation (not working yet) 26901 26902 - based heavily on intel 26903 26904commit 3b619d88a65a5801c85d120f2d39704194433f10 26905Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26906Date: Sat May 26 17:06:18 2007 -0400 26907 26908 RADEON: clean up and add comment regarding clones 26909 26910commit 089ae4afd23322bf602d340c531c36a4c8b461ed 26911Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26912Date: Sat May 26 15:49:11 2007 -0400 26913 26914 RADEON: Switch SetOutputType() to use names rather than numbers 26915 26916commit c4cef0c9481257a744ac99dfc7beb988ce51a8ee 26917Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26918Date: Sat May 26 13:12:37 2007 -0400 26919 26920 RADEON: make sure we have the lock when SYNCing 26921 26922commit 899c62e987d5c7524817ba85675e39a8a28e0232 26923Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26924Date: Sat May 26 12:54:01 2007 -0400 26925 26926 RADEON: Make sure LVDS and FP2 routing info gets written 26927 26928commit 18857184ffa6847815d349c020b003f8401e36ee 26929Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26930Date: Sat May 26 12:27:00 2007 -0400 26931 26932 RADEON: fix handling of DRI lock 26933 26934commit 3a61453efb4f04492cef823b6dd1273b55c6a785 26935Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26936Date: Sat May 26 01:38:09 2007 -0400 26937 26938 RADEON: implement backlight control for LVDS 26939 26940 This code is currently disabled as I'm not sure which 26941 laptops actually use this method for backlight control. 26942 My laptop seems to use another method as adjusting the 26943 backlight level doesn't seem to touch LVDS_GEN_CNTL. 26944 Maybe just macs? 26945 26946commit ceec3f62257bafe4771e75d3c4f1d2a517d7acf8 26947Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26948Date: Fri May 25 23:29:51 2007 -0400 26949 26950 RADEON: implement auto/analog/digital output property for DVI-I (untested) 26951 26952commit e3e9c608651e7cbb9851489274815f47d2a8dbbb 26953Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26954Date: Fri May 25 19:26:24 2007 -0400 26955 26956 RADEON: more clean of last commits 26957 26958commit a7d2c9c0f6155c4e65a2e73f3832c0d8ca7af5fc 26959Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26960Date: Fri May 25 19:09:11 2007 -0400 26961 26962 RADEON: fixup last commit and add support for ext TMDS detect 26963 26964commit e50ca35ce2ea79dadb38ce14e459eed836452ff7 26965Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26966Date: Fri May 25 18:56:32 2007 -0400 26967 26968 RADEON: clean up non-DDC probe, add detect for DVI 26969 26970commit abb9b57e7b4162ab6c21fd1e809d24f13f7e1ea1 26971Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26972Date: Wed May 23 21:37:58 2007 -0400 26973 26974 RADEON: clean up some logic 26975 26976commit 679236defd28bea8874004377436347950ff7bec 26977Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26978Date: Wed May 23 21:18:40 2007 -0400 26979 26980 RADEON: add output properties (not functional yet) 26981 26982commit 975da595f032c145ad74079ff8edeaead779dc7b 26983Author: Michel Dänzer <michel@tungstengraphics.com> 26984Date: Tue May 22 10:56:47 2007 +0200 26985 26986 radeon: Provide new DRI texOffsetStart hook when available with EXA. 26987 26988commit e38cad5633c974467f6417051eda8d487ac3686e 26989Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 26990Date: Tue May 22 00:09:46 2007 -0400 26991 26992 RADEON: Only add valid connectors 26993 26994commit 8275151baac22c34149cef0b7d922771d24abc3e 26995Author: Michel Dänzer <michel@tungstengraphics.com> 26996Date: Mon May 21 10:25:48 2007 +0200 26997 26998 radeon: HW cursor cleanup. 26999 27000 Don't needlessly turn the HW cursor on/off in RADEONLoadCursor*(). 27001 27002 Besides cleaning up the code, this semms to avoid some HW cursor related 3D 27003 lockups, see https://bugs.freedesktop.org/show_bug.cgi?id=10815 . My best 27004 guess is that this is because the engine is now always idled before touching 27005 the CRTC registers. 27006 27007commit 137e3fc1899078af0f72303ab0a4e6cf35804a7b 27008Author: Michel Dänzer <michel@tungstengraphics.com> 27009Date: Mon May 21 10:25:48 2007 +0200 27010 27011 radeon: Suppress debugging output by default. 27012 27013 It can be enabled at runtime by increasing the log verbosity level. 27014 27015 Also change the prefix from (**) to (II) to make grepping the log file for 27016 defaults overridden by xorg.conf more useful again. 27017 27018 Turn some MC related debugging output into normal informational output as it's 27019 useful for recognizing corner cases that can cause stability issues. 27020 27021commit aed193a47a939451d9a6d05b02653b1d73e1d523 27022Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27023Date: Sun May 20 21:13:59 2007 -0400 27024 27025 RADEON: fix crtc1 (un)blanking code after last commit 27026 27027 - crtc1 MC control is on CRTC_GEN_CNTL 27028 - fix indenting 27029 27030commit 870c8043068a0f44b53d1148371b2cc1e3970a7b 27031Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27032Date: Sun May 20 21:03:10 2007 -0400 27033 27034 RADEON: several updates 27035 27036 - move output init from InitCrtc() functions to the output mode_set() 27037 - take the crtc off the MC when blanking 27038 - move EnableDisplay() to output commit() 27039 - clean up some dead code 27040 - don't enable dacs in initcrtc() functions, this is taken care of in EnableDisplay() 27041 27042commit 09bfc8ed000f95ede5b73f2bad69edc1a4d9bac6 27043Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27044Date: Sun May 20 18:06:22 2007 -0400 27045 27046 update to 6.6.192 for rc release 27047 27048commit 764cb73e8dec4040cdd418d249fc504399fca3ee 27049Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27050Date: Sun May 20 17:26:26 2007 -0400 27051 27052 Fix regular/"xinerama"/zaphod dualhead mode 27053 27054 - logic in RADEONUnblank() was wrong 27055 - Calling RADEONSetupConnectors() on second instance screwed up the port info 27056 - still seem to be HW cursor issues with zaphod mode 27057 27058commit 9a147fef8e0e2ede2a0008c4ecfbd9b00c8dc5f6 27059Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27060Date: Fri May 18 23:45:11 2007 -0400 27061 27062 RADEON: if connector is VGA set TMDS to none 27063 27064commit bbd6faff4c6acb48970d774375c8a61861405f96 27065Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27066Date: Fri May 18 23:30:19 2007 -0400 27067 27068 RADEON: Change default LVDS i2c line for powerpc 27069 27070commit 7940ec364f3cbf02ba64b92c74cbaad4555baf38 27071Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27072Date: Fri May 18 00:23:08 2007 -0400 27073 27074 RADEON: cleanup 27075 27076commit f71bfde7352ef858c1041037d7dc77c237e315a4 27077Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27078Date: Fri May 18 00:12:03 2007 -0400 27079 27080 RADEON: more cleanup 27081 27082commit f711e266d6927dec648f6ff26f15c6f48643f78c 27083Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27084Date: Thu May 17 23:59:07 2007 -0400 27085 27086 RADEON: turn off all outputs in screeninit(). We'll turn on the ones we want later 27087 27088commit bfc1c372d7475b7fa6bffb8a79dec1dc2f11ec59 27089Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27090Date: Thu May 17 23:45:29 2007 -0400 27091 27092 RADEON: More cleanup 27093 27094commit c60b3bc9e3e7463bdb42e2478be8cc3f22c63c68 27095Author: Paul TBBle Hampson <Paul.Hampson@pobox.com> 27096Date: Thu May 17 23:26:54 2007 -0400 27097 27098 Fix for infinite loop in RADEONGetLVDSInfo 27099 27100commit 11289f8206cab1a94bd64a3938cf9af50f19497e 27101Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27102Date: Tue May 15 00:48:31 2007 -0400 27103 27104 RADEON: remove some debugging code 27105 27106commit fba1a11e287ebf04cf311645d31299896bad7283 27107Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27108Date: Tue May 15 00:44:29 2007 -0400 27109 27110 RADEON: Move DAC regs to their own Save() function 27111 27112commit 9ed00e959fcdab2739a7e64ca1e303a6faf28f15 27113Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27114Date: Tue May 15 00:36:05 2007 -0400 27115 27116 RADEON: Lots of small fixes 27117 27118 - remove extra crtc2_base set 27119 - remove some debugging code 27120 - clean up RestoreMode() 27121 - make sure a DDC reg gets assigned for LCD_DDC 27122 - make sure we adjust the right frame in AdjustFrame() 27123 - rename RADEONMapControllers() to RADEONPrintPortMap() to 27124 reflect what it actually does now 27125 - make i2c bus name match the DDC port 27126 - remove or comment out un-needed code 27127 27128commit 5106c8fe5f74865cac8b00937739a4efde9fc254 27129Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27130Date: Sun May 13 21:48:26 2007 -0400 27131 27132 RADEON: Further cleanup 27133 27134commit 224a73e41e9be344d5644203e7ebd5a3a8272604 27135Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27136Date: Sun May 13 21:41:59 2007 -0400 27137 27138 RADEON: remove remnants of "old" multi-head support 27139 27140 The old screen based multi-head code is broken at this point 27141 and I have no intention of fixing it, so to clean things up 27142 I've removed it. 27143 27144commit 9c2f20a83a5fed14225f4c3ebcd8ca41e9d4bd48 27145Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27146Date: Sun May 13 21:13:23 2007 -0400 27147 27148 RADEON: additional cleanups 27149 27150commit e187321ab8dd58d2b2fe92c062d070ba4820a2bf 27151Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27152Date: Sun May 13 19:49:10 2007 -0400 27153 27154 RADEON: add new ConnectorTable option and re-add PanelSize option 27155 27156commit e60a7bcaf5611ad4706b1d1442ce4cae49145f42 27157Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27158Date: Sun May 13 18:14:29 2007 -0400 27159 27160 RADEON: remove mergedfb, etc. cruft from man page 27161 27162commit 51caa5dccc4a2e14d1b41bb7868ad20468dd4cfe 27163Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27164Date: Sun May 13 18:10:02 2007 -0400 27165 27166 RADEON: remove old mergedfb and dualhead options that are no longer used 27167 27168commit c5da9d4040cb08598d171d20f84d3f6c20a033e0 27169Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27170Date: Sun May 13 18:05:38 2007 -0400 27171 27172 RADEON: remove more dead code 27173 27174commit e776fa9ecc53d63f916a5447a76fb2bb3b824167 27175Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27176Date: Sun May 13 15:43:39 2007 -0400 27177 27178 RADEON: re-org randr code 27179 27180 - move crtc stuff to radeon_crtc.c 27181 - move output stuff to radeon_output.c 27182 27183commit 58ce388452b7bc790c438d75c9cf4a0f69f0d7b2 27184Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27185Date: Sun May 13 15:05:01 2007 -0400 27186 27187 RADEON: Remove dead code 27188 27189commit 66e8e6c8348d007930730e90295588efe8108844 27190Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27191Date: Sun May 13 14:25:03 2007 -0400 27192 27193 RADEON: several fixes 27194 27195 - give better DDC names 27196 - disable DAC if connector is DVI-D 27197 - fix indentation in ProbeDDCModes() 27198 27199commit 1c16c2ce9c5b02b03d23da965127d82eea4c4039 27200Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27201Date: Sun May 13 13:54:04 2007 -0400 27202 27203 RADEON: fix VT switch 27204 27205commit 7c66e903368f77ecc4d3bd1c9f08d2adbd85e83c 27206Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27207Date: Sun May 13 13:37:59 2007 -0400 27208 27209 RADEON: move crtc offset handling into init/save/restore functions 27210 27211commit aec078eb0740651fba8ec602e8239bd679efc8ad 27212Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27213Date: Sun May 13 11:57:57 2007 -0400 27214 27215 RADEON: fix up DDCConnected() 27216 27217commit 0550c37ecc434b8075fb3c367d100ff27625bb64 27218Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27219Date: Sun May 13 11:44:50 2007 -0400 27220 27221 RADEON: don't need to pass pScrn to EnableDisplay() 27222 27223commit 117220527de9fd3158f600645bcfcaf46847f45f 27224Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27225Date: Sun May 13 11:37:35 2007 -0400 27226 27227 RADEON: remove hardcoded output limit and PortInfo stuff 27228 27229commit 7e5c29961ac2a9e9dbe5d6d2d73d11cd018d62b5 27230Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27231Date: Fri May 11 18:00:40 2007 +0200 27232 27233 RADEON: Fix RMX after the last commit 27234 27235commit ab5603edd8fc3ef0560bdfb6a6d9c6af2a49d1e5 27236Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27237Date: Fri May 11 17:34:35 2007 +0200 27238 27239 RADEON: Move LVDS, TMDS, DAC properties to the output rec 27240 27241commit 94eb0681de0641e490f06486468617a727fefe86 27242Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27243Date: Fri May 11 14:39:48 2007 +0200 27244 27245 RADEON: switch output dpms to use RADEONEnableDisplay() 27246 27247commit 673ede5578d5d9caf2adf0445fe1e684b034eea5 27248Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27249Date: Fri May 11 14:28:17 2007 +0200 27250 27251 RADEON: re-arrange output mode setting 27252 27253commit 0070a7d787adaae99f7bc2659be4b0f49f439db5 27254Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27255Date: Fri May 11 13:06:15 2007 +0200 27256 27257 RADEON: more re-org 27258 27259 - move crtc mode setting around 27260 - add dri lock/unlock to crtc lock/unlock calls 27261 27262commit bba456232ac9a6218aa7fbd504d6093fa72860cf 27263Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27264Date: Thu May 10 15:33:51 2007 +0200 27265 27266 RADEON: fix VT switch 27267 27268commit 1779a12a947401e5c6bcf784b47e9b3c80d37204 27269Author: Jesse Barnes <jbarnes@jbarnes-mobile.amr.corp.intel.com> 27270Date: Thu May 10 06:22:35 2007 -0700 27271 27272 Add cscope files to .gitignore 27273 27274commit 0cb23277666db3b30438c6f88840d861e04df414 27275Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27276Date: Thu May 10 15:20:56 2007 +0200 27277 27278 RADEON: randr driver re-org checkpoint 27279 27280 - split the mode setting per-crtc 27281 - reduce start up flicker 27282 27283commit 33c370b1d8350945f80ac12097d3e91243a400f2 27284Author: Jesse Barnes <jbarnes@jbarnes-mobile.amr.corp.intel.com> 27285Date: Wed May 9 16:16:39 2007 -0700 27286 27287 RADEON: 27288 - fix an ugly modesetting bug: if we happened to set the mode on 27289 CRTC1 before CRTC2, CRTC2's RestoreMode function would clobber 27290 CRTC1's CRTC_OFFSET register since we never updated 27291 ModeReg.crtc_offset... so make AdjustFrame use ModeReg and pull 27292 the call to it up before RestoreMode, seems to work ok here. 27293 27294commit 6263248a0044777a352e4ee7380b4b8f9afd091b 27295Author: Jesse Barnes <jbarnes@jbarnes-mobile.amr.corp.intel.com> 27296Date: Wed May 9 14:52:00 2007 -0700 27297 27298 RADEON: 27299 - use fixup_mode hook to set RADEON_USE_RMX flag so panel scaling 27300 works 27301 - use valid_mode hook to prune invalid default modes from list 27302 - use adjusted_mode in crtc_mode_set (using adjusted_mode from 27303 fixup hook) 27304 27305commit d2497009e395800fbde5777465f3087a54b94418 27306Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27307Date: Wed May 9 22:23:45 2007 +0200 27308 27309 randr checkpoint 27310 27311 - server still hangs if you start with external monitor connected 27312 - RMX not working 27313 - more than 2 outputs now possible (untested) 27314 27315commit 4488f0737d5268168eab41440b7a3b5732efb15e 27316Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27317Date: Wed May 9 16:41:13 2007 +0200 27318 27319 more randr re-work 27320 27321 - remove RADEONQueryConnectedDisplays(); randr takes care of this now 27322 - print edid after randr detection 27323 27324commit 4d992386e2ab9d8c50f0484445564325dfb42930 27325Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27326Date: Wed May 9 16:12:21 2007 +0200 27327 27328 enable all outputs on each crtc on mode restore 27329 27330commit 2618cf2aa8ed76411b943eb90c95869814c2f151 27331Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27332Date: Wed May 9 15:48:40 2007 +0200 27333 27334 More re-org to allow more than 2 outputs 27335 27336 - Move radeon output crtc map into crtc_set_mode 27337 - in modeinit, set up all outputs attached to crtc 27338 27339commit 61b9e79cbeee6f735a4c82ec8a802aee85d8b890 27340Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27341Date: Wed May 9 15:11:23 2007 +0200 27342 27343 start to clean up MapControllers 27344 27345commit 6a724dd798c1a7b461672993c02be83a7bccded6 27346Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27347Date: Wed May 9 14:48:17 2007 +0200 27348 27349 minor cleanup 27350 27351commit 03860fed24b4f76cc7f1f4210ec8f8040fa04777 27352Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27353Date: Tue May 8 19:27:48 2007 +0200 27354 27355 remove some old mergedfb cruft 27356 27357commit 76670f665ebec7cdf40a04bf9379cb3ad4417507 27358Merge: 83f81ed5 a3ee4220 27359Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27360Date: Tue May 8 18:41:25 2007 +0200 27361 27362 Merge branch 'master' into randr-1.2 and fix conflicts 27363 27364commit 83f81ed5e3c33c94c80500316c37a7cbfc51f41f 27365Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27366Date: Mon May 7 22:54:16 2007 +0200 27367 27368 RADEON: Fix cursor handling for randr 27369 27370commit 5fc21cd34436919300018d8d4850fc67db284eb3 27371Author: Jesse Barnes <jbarnes@jbarnes-mobile.amr.corp.intel.com> 27372Date: Mon May 7 13:31:20 2007 -0700 27373 27374 - fix randr 1.2 on pre-RV350 chips by enabling DAC_MACRO_CNTL writes 27375 - set dac_cntl on non-primary crtcs 27376 - set XCRT_CNT_EN in CRTC_EXT_CNTL just because 27377 - fix warnings in calls to xf86PrintModeline (wants screen index not 27378 pScrnInfo) 27379 27380commit 99b3df154317f0209618e532282a3e7ad091c00f 27381Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net> 27382Date: Mon May 7 19:02:35 2007 +0200 27383 27384 RADEON - update randr cursor handling, LVDS setup fix 27385 27386 - quick fix for the cursor handling to update to the latest 27387 server bits 27388 - make sure connector type is CONNECTOR_PROPRIETARY for LVDS. 27389 27390commit a3ee42207aab77d93655a82fdcb32be38268b85f 27391Author: Dave Airlie <airlied@linux.ie> 27392Date: Wed Apr 25 09:52:22 2007 +1000 27393 27394 radeon: another attempt at fixing the mergedfb refresh rate 27395 27396 This attempts to keep it inside the 32-bit limit when multiplying things out 27397 later in the randr tree. 27398 27399 Let me know if I screwed this up.. 27400 27401commit 3828237200fc16d4d32664fb8358950c213d4897 27402Author: Dave Airlie <airlied@nx6125b.(none)> 27403Date: Sun Apr 22 11:36:00 2007 +1000 27404 27405 radeon: add support for DDC on some laptop chipsets 27406 27407 I noticed fglrx has DDC for the panel in the rs480 laptop, however radeon 27408 didn't pick it up, so I valgrinded fglrx and spotted 0x1a0/0x1a4 accesses 27409 I actually noticed this before from the BIOS but never figured it out. 27410 27411 So now I get DDC from the LCD on this laptop. 27412 27413commit dd6a966e862b774a8e8b9e1a085309219673efad 27414Author: Dave Airlie <airlied@nx6125b.(none)> 27415Date: Sun Apr 22 11:36:00 2007 +1000 27416 27417 radeon: add support for DDC on some laptop chipsets 27418 27419 I noticed fglrx has DDC for the panel in the rs480 laptop, however radeon 27420 didn't pick it up, so I valgrinded fglrx and spotted 0x1a0/0x1a4 accesses 27421 I actually noticed this before from the BIOS but never figured it out. 27422 27423 So now I get DDC from the LCD on this laptop. 27424 27425commit c81ed9bd7b37c9d02141d10f6c7bad3d0c57032f 27426Author: Dave Airlie <airlied@linux.ie> 27427Date: Sat Apr 21 18:58:40 2007 +1000 27428 27429 radeon: fix build on older server 27430 27431commit 406eec71116a58d42288a7f1c809a92d5bda7350 27432Author: Dave Airlie <airlied@linux.ie> 27433Date: Sat Apr 21 18:56:28 2007 +1000 27434 27435 radeon: fix build since patches for IBM don't actually build 27436 27437commit ad119960095b4b64f4c6793f65950c9967ce4989 27438Author: Matthias Hopf <mhopf@suse.de> 27439Date: Thu Apr 19 11:54:46 2007 +0200 27440 27441 Disable RMX for IBM Lewis server. 27442 27443 Due to the hardware layout RMX ddc_mode has to be set. 27444 If ddc_mode is set, RADEONValdiateFPModes() shouldn't be called. 27445 Bugzilla #10620 (3). 27446 27447commit 16ef77df4ebaf5ea13baa82972aaf98e71ac32ee 27448Author: Matthias Hopf <mhopf@suse.de> 27449Date: Wed Apr 18 17:36:15 2007 +0200 27450 27451 Set sync polarity restriction flags even for non-"digital separate" monitors. 27452 27453 According to Lisa Wu, this is correct regarding the VESA EEDID standard. 27454 Bugzilla #10620 (2), original patch by Lisa Wu @ATI 27455 27456commit 0abce69f0d826a7ca1a41d963cd4730b6e01c145 27457Author: Matthias Hopf <mhopf@suse.de> 27458Date: Wed Apr 18 17:32:52 2007 +0200 27459 27460 Fix inconsistent use of Mode lists. 27461 27462 Some scans used to only check every second entry, some stopped at the entry 27463 before the last entry. 27464 Bugzilla #10620 (1), original patch by Lisa Wu @ATI 27465 27466commit aea801cf9a5ce519a53d6fffd9a3a2e526ec79ea 27467Author: Matthias Hopf <mhopf@suse.de> 27468Date: Fri Apr 13 16:16:05 2007 +0200 27469 27470 Fix crash if MergedFB and secondary head not found 27471 27472 If the secondary head isn't found (Monitor unplugged etc.) but MergedFB 27473 is configured, the driver segfaults because it tries to access the mode 27474 list private structures, which are not filled in. 27475 27476commit 07ddffb32e6293c77b32c94b87ec468caef3d6f5 27477Author: Matthias Hopf <mhopf@suse.de> 27478Date: Wed Apr 11 14:36:51 2007 +0200 27479 27480 Fixed typo in mode list generation. 27481 27482commit 4effa67ea75736a31b9e78a7b35acf74b067c43e 27483Author: Dave Airlie <airlied@linux.ie> 27484Date: Mon Apr 9 22:08:31 2007 +1000 27485 27486 radeon: add support for enabling direct rendering on RS480 27487 27488 Thanks to Matthew Garrett and Ubuntu for the hw loan to get this working. 27489 27490 Still no 3D driver support but at least you should get CP acceleration for 27491 2D now. 27492 27493commit 6b25a4c48796e022a093f3072574ffe9709ecaf4 27494Author: Michel Dänzer <michel@tungstengraphics.com> 27495Date: Wed Apr 4 10:47:29 2007 +0200 27496 27497 radeon: Link nearest modes by default for clone mode. 27498 27499 This makes sure all modes of both CRTCs will be available by default with 27500 MergedFB. 27501 27502commit 3c892f163ec1fa9be6e733aab091c9b718f41efc 27503Author: Michel Dänzer <michel@tungstengraphics.com> 27504Date: Wed Apr 4 09:52:37 2007 +0200 27505 27506 radeon: Always sort modes when adding to list. 27507 27508 This makes sure mode lists will always be sorted from larger to smaller. 27509 27510commit 3a8190ccc79969925257e7b980b78d79053d208d 27511Author: Michel Dänzer <michel@tungstengraphics.com> 27512Date: Wed Apr 4 09:47:07 2007 +0200 27513 27514 radeon: Don't shrink virtual size based on secondary modes. 27515 27516commit 9b1e97284ce185d358ca756a235d2cee346fa53f 27517Author: Henry Zhao <henryz@localhost.localdomain> 27518Date: Sat Mar 31 23:01:52 2007 -0800 27519 27520 10205: Radeon driver's own mode validation code does not work properly 27521 27522commit 9c2dcd19be8fc2cc29e637d1e9748e66196e3900 27523Author: Henry Zhao <henryz@localhost.localdomain> 27524Date: Sat Mar 31 20:10:03 2007 -0800 27525 27526 9337: EDID modes do not participate in validation for CRT monitor 27527 27528commit 1acd6d6fa42acec07fb11aeb189f492ddb021cb4 27529Author: Michel Dänzer <michel@tungstengraphics.com> 27530Date: Fri Mar 30 11:06:10 2007 +0200 27531 27532 radeon: Guard some MergedFB specific code with info->MergedFB tests. 27533 27534 Fixes https://bugs.freedesktop.org/show_bug.cgi?id=10442 . 27535 27536commit 1a71106c0e4fe5f650239dc694163fdf52d33663 27537Author: Michel Dänzer <michel@tungstengraphics.com> 27538Date: Tue Mar 27 10:13:21 2007 +0200 27539 27540 radeon: Fix typo. 27541 27542commit f87e57d4d773a019d1cc8a10425c57480430f6a4 27543Author: Dave Airlie <airlied@linux.ie> 27544Date: Tue Mar 27 18:08:54 2007 +1000 27545 27546 radeon: fix up crtc debug dereference problem 27547 27548commit 66b4a571a4e7960da6807d3f30955aa08e89ccc6 27549Author: Dave Airlie <airlied@linux.ie> 27550Date: Tue Mar 27 17:00:37 2007 +1000 27551 27552 update number to 6.6.191 for rc release 27553 27554commit fca30a6b581cb6c1466ab1bc316df8fed5d82b60 27555Author: Alex Deucher <alex@botch2.com> 27556Date: Mon Mar 26 23:26:51 2007 -0400 27557 27558 fix NULL dereference 27559 27560 when IsSecondary is true, crtc1 is NULL 27561 Noticed by Sverre Froyen. 27562 27563commit 97d8d1ed10d069343f2b3172ba64ca421821a602 27564Author: George Sapountzis <gsap7@yahoo.gr> 27565Date: Sat Mar 24 20:02:12 2007 +0200 27566 27567 Move atichip.c from ati to atimisc. 27568 27569commit aa8f5b02ebc9be60df48722588261627d6a457e8 27570Author: George Sapountzis <gsap7@yahoo.gr> 27571Date: Sat Mar 24 19:53:02 2007 +0200 27572 27573 [mach64] Use Mach64Chipsets[] instead of ATIChipNames[]. 27574 27575 atimisc is PCI-only now, we can get the chip name with xf86TokenToString(). 27576 27577commit 39e896a1e688ea2d2d21f88c1c5d34c5810aac1c 27578Author: George Sapountzis <gsap7@yahoo.gr> 27579Date: Sat Mar 24 19:47:18 2007 +0200 27580 27581 [mach64] Set pATI->ChipRevision correctly, instead of overriding. 27582 27583 There is no need to override pATI->ChipRevision for GX/CX, as it is only 27584 reported with a printf. 27585 27586commit f046a910ca117279fbabc6281b2e23439ec9ea4e 27587Author: George Sapountzis <gsap7@yahoo.gr> 27588Date: Fri Mar 23 22:19:17 2007 +0200 27589 27590 Drop probing by driver name from "Device" section. 27591 27592 atimisc is PCI-only now, so we only need to scan the PCI space. 27593 27594commit 9cd175d9cd4ed710fccb303664c77519ecaf1e21 27595Author: George Sapountzis <gsap7@yahoo.gr> 27596Date: Fri Mar 23 22:12:48 2007 +0200 27597 27598 Fold FillIn() back to Probe(). 27599 27600commit d7a8cd0e476034796fc38e25a28cd28d05ea4a13 27601Author: George Sapountzis <gsap7@yahoo.gr> 27602Date: Fri Mar 23 22:10:03 2007 +0200 27603 27604 Drop now unneeded _X_EXPORT's. 27605 27606commit 58626d8b78f26f0d9c480033d3c3a12e541342b1 27607Author: George Sapountzis <gsap7@yahoo.gr> 27608Date: Fri Mar 23 21:30:19 2007 +0200 27609 27610 Move {atimach64,r128,radeon}_probe.c from ati to subdrivers. 27611 27612 Subdrivers are now loaded from the wrapper Probe function rather than at screen 27613 creation time. 27614 27615 The wrapper Identify callback only prints chip families now, chip lists are 27616 printed when a subdriver is loaded. This also avoids duplication of subdriver 27617 Identify callbacks. 27618 27619 Unknown radeons should still get a list of known radeon and then probe fails... 27620 27621 Probe for atimisc last to avoid needless loading in most cases (r128, radeon). 27622 27623commit 2e3d43af1e5077cc61dd8668551a6291368d9ed2 27624Author: George Sapountzis <gsap7@yahoo.gr> 27625Date: Mon Mar 19 19:05:04 2007 +0200 27626 27627 [mach64] Compute MMIO address once for each case (PIO, MMIO). 27628 27629 For an MMIO-only driver, the extra MMIO address probing is not needed since we 27630 already compute the MMIO address during probe and do not make any BIOS calls. 27631 27632 For a PIO driver, this extra probing could even result in a wrong MMIO address 27633 since the checking is done by reading through PIO, not MMIO. The MMIO address 27634 is later corrected but having the extra probe there is useless. 27635 27636commit e7defc0e930f1b2d83623a769b2bfdb05c66a5fe 27637Author: George Sapountzis <gsap7@yahoo.gr> 27638Date: Mon Mar 19 19:04:10 2007 +0200 27639 27640 [mach64] Move setting of PIO and MMIO addresses up, after BIOS init. 27641 27642 Partially revert "Minor refactoring of pATI->Block0Base computation." 27643 27644 This parially reverts commit 2c8b33b761e4411451aea1eea3a89d629aeea40d. 27645 27646commit 3dfdba195f1ff794ab00585ab80e7ea6ac92fcf4 27647Author: George Sapountzis <gsap7@yahoo.gr> 27648Date: Sun Mar 18 16:40:07 2007 +0200 27649 27650 [mach64] Make mapping code friendlier to pci-rework source-code compatibility. 27651 27652commit 321db6d87c9110654ab386e5ec270116eca1e04f 27653Author: George Sapountzis <gsap7@yahoo.gr> 27654Date: Sun Mar 18 16:15:20 2007 +0200 27655 27656 [mach64] Add macros for pci-rework source-code compatibilty. 27657 27658commit dfd07b6e99020d1db43d7ce0cae4423d8c6b1f05 27659Author: George Sapountzis <gsap7@yahoo.gr> 27660Date: Sun Mar 18 16:38:26 2007 +0200 27661 27662 [mach64] Drop checking that BAR base + size fits in 32 bit. 27663 27664 I can't see how these tests could fail on 32-bit. OTOH they are potentially 27665 problematic with 64-bit PCI addresses. 27666 27667commit e674338a98c50800637b8ebc01adf3aec2a3eb38 27668Author: George Sapountzis <gsap7@yahoo.gr> 27669Date: Sun Mar 18 15:50:41 2007 +0200 27670 27671 [mach64] Single mapping per BAR. 27672 27673 - We reset write-combining of the FB when MMIO registers fall in the linear 27674 aperture. 27675 27676 - We set MMIOInLinear in all three aperture mappings since the mapping code 27677 depends on it. 27678 27679 - We map the linear aperture in its entirety for MMIOInLinear also. 27680 This simplifies mapping of HW cursor and the LE aperture on BE archs, as 27681 they always fall in the linear aperture. 27682 27683 - I assumed that BAR[2] is page-aligned (we know it's 4K-aligned for sure). 27684 27685commit e7969de8ec572d8801806f9525a9a081a58a03f1 27686Author: George Sapountzis <gsap7@yahoo.gr> 27687Date: Sun Mar 18 15:35:25 2007 +0200 27688 27689 [mach64] Drop poking the VGA MMIO aperture. 27690 27691 We will not be able to handle this when we do mappings for PCI BARs only and 27692 it should be handled with a CPIO driver anyway. 27693 27694commit 166c760a86165330175023e07c4b2bd6891633c5 27695Author: Michel Dänzer <michel@tungstengraphics.com> 27696Date: Tue Mar 20 09:16:02 2007 +0100 27697 27698 Fix advertised minimum minor version of the DRI module. 27699 27700 We don't automatically require bumped minor versions. 27701 27702commit 1bdd376dbd57de8925244f0808f974d6d8cff39d 27703Author: Michel Dänzer <michel@tungstengraphics.com> 27704Date: Tue Mar 20 09:14:39 2007 +0100 27705 27706 radeon: Only sync to hardware when really necessary with EXA. 27707 27708 In particular, don't sync again after accelerated DownloadFromScreen, which 27709 syncs implicitly. This avoids calling into the kernel when it's not necessary, 27710 which can be relevant in some situations. 27711 27712commit 3cfa3a5c8daf03aaad6fc30d275709f6eb717d29 27713Author: Michel Dänzer <michel@tungstengraphics.com> 27714Date: Tue Mar 20 09:13:24 2007 +0100 27715 27716 radeon: Unify code to release the CP. 27717 27718commit e174d8df8c801fad95e5f79cff69187c200bee6e 27719Author: Michel Dänzer <michel@tungstengraphics.com> 27720Date: Tue Mar 20 09:12:51 2007 +0100 27721 27722 radeon: Minor BlockHandler cleanups. 27723 27724 Don't flush indirect buffer in BlockHandler; it's done in LeaveServer. 27725 27726 Also set the EXA engine mode to unknown only at the end. 27727 27728commit 113fb4b61e709a9b54fc2ef73efce06011e771c1 27729Author: Michel Dänzer <michel@tungstengraphics.com> 27730Date: Tue Mar 20 09:12:09 2007 +0100 27731 27732 radeon: Disable CP line acceleration on RV280s. 27733 27734 There have been several reports of stability issues with things like the 27735 xscreensaver hack hypercube. 27736 27737commit 4651d00b183cb498879d605c4b93cd3a0c63cb33 27738Author: Michel Dänzer <michel@tungstengraphics.com> 27739Date: Tue Mar 20 09:10:19 2007 +0100 27740 27741 radeon: Make sure 3D clients will re-upload textures to video RAM after LeaveVT. 27742 27743 Walk the SAREA texList and bump the age of every active object, so their owners 27744 will consider them kicked out when they grab the HW lock next time. 27745 27746commit 2d2fb54ba370c1df9ef5102e83c17a7ff5c55403 27747Author: Michel Dänzer <michel@tungstengraphics.com> 27748Date: Thu Mar 15 17:26:56 2007 +0100 27749 27750 atimisc: Fix strict aliasing violation flagged by -Wall. 27751 27752commit cfb82631fb683a0dc9bbc8f2d214646a8f1bca0b 27753Author: Michel Dänzer <michel@tungstengraphics.com> 27754Date: Thu Mar 15 17:12:43 2007 +0100 27755 27756 radeon: Fix build warnings. 27757 27758commit 50a524445ed1ea3c4fab56de772921ba86a21d35 27759Author: George Sapountzis <gsap7@yahoo.gr> 27760Date: Thu Mar 8 17:43:48 2007 +0200 27761 27762 Move atioption.c from ati to atimisc. 27763 27764commit 52684241ba1ffb393cb655ecc09cbd0cef5b1717 27765Author: George Sapountzis <gsap7@yahoo.gr> 27766Date: Thu Mar 8 07:07:02 2007 +0200 27767 27768 Move atiprobe.c from ati to atimisc. 27769 27770commit 2a1cd107a593630001799d6cd9e72c64222553b2 27771Author: Jesse Barnes <jbarnes@jbarnes-mobile.amr.corp.intel.com> 27772Date: Tue Mar 6 14:35:50 2007 -0800 27773 27774 Add prepare/commit hooks to output and crtc func table 27775 27776commit 703c6fc0142ffc600285c13fe6dafecf988c0a1d 27777Author: Jesse Barnes <jbarnes@jbarnes-mobile.amr.corp.intel.com> 27778Date: Tue Mar 6 10:12:42 2007 -0800 27779 27780 Update for new CRTC resize hooks. 27781 27782commit 1e270b411cf691c69194932b864ee2c9db292263 27783Author: George Sapountzis <gsap7@yahoo.gr> 27784Date: Sun Mar 4 18:21:49 2007 +0200 27785 27786 [mach64] Bug 5586: overlay scaler limited to 720 pixels on ATI Rage Pro. 27787 27788commit 54f9b8c1d01ecb97e2d4d247918f0efc86ed6037 27789Author: George Sapountzis <gsap7@yahoo.gr> 27790Date: Sun Mar 4 18:16:40 2007 +0200 27791 27792 [mach64] Xv: use single surface and encoding structs. 27793 27794commit e1e55b533d3ae528c8da37dcb77ed906d05697fb 27795Author: Dave Airlie <airlied@linux.ie> 27796Date: Sun Mar 4 19:11:34 2007 +1100 27797 27798 radeon: add option to set pci aperture size from config file 27799 27800 This requires a drm > 1.26 to work 27801 27802commit 58ee31d015cf8bec0edca62a46faec0b3505be8c 27803Author: Dave Airlie <airlied@linux.ie> 27804Date: Sun Feb 25 23:29:09 2007 +1100 27805 27806 fix typo 27807 27808commit 3cfe94d5438961b869766dacbcd13fde8b770ca2 27809Merge: 31c018ca 3b43788c 27810Author: Dave Airlie <airlied@linux.ie> 27811Date: Sun Feb 25 23:27:19 2007 +1100 27812 27813 Merge branch 'radeon-randrv12-v4' 27814 27815 Conflicts: 27816 27817 src/radeon_display.c 27818 src/radeon_driver.c 27819 27820commit 3b43788c45f51ad2d3e8e64383c412f4ddd97207 27821Author: Dave Airlie <airlied@linux.ie> 27822Date: Sun Feb 25 23:17:31 2007 +1100 27823 27824 cleanup radeon code against master server 27825 27826commit 24c6fa7cfac5602ba9e6e2f331bcac52fab258e5 27827Author: Dave Airlie <airlied@linux.ie> 27828Date: Sun Feb 25 21:43:54 2007 +1100 27829 27830 make radeon randr build against master server 27831 27832commit 4374895d393bf170134ccffc188521fe515d5a77 27833Author: George Sapountzis <gsap7@yahoo.gr> 27834Date: Sat Feb 24 21:25:13 2007 +0200 27835 27836 [mach64] Consolidate adjustments of mode timings, part 3. 27837 27838 Cosmetic. 27839 27840commit 8b9a4ec8af90b2f9f0d0a3a7f6627061df7e01b3 27841Author: George Sapountzis <gsap7@yahoo.gr> 27842Date: Sat Feb 24 21:22:49 2007 +0200 27843 27844 [mach64] Consolidate adjustments of mode timings, part 2. 27845 27846 Factor out to seperate function. 27847 27848commit bb226c9e7218d2a65c056fe74cb7eece0550670f 27849Author: George Sapountzis <gsap7@yahoo.gr> 27850Date: Sat Feb 24 21:20:53 2007 +0200 27851 27852 [mach64] Consolidate adjustments of mode timings, part 1. 27853 27854 Move LCD block from atimode.c to atimach64.c . 27855 27856commit 31c018ca4a18ce426b29006f103f56eee7f985fa 27857Author: Roland Scheidegger <sroland@tungstengraphics.com> 27858Date: Wed Feb 21 13:47:37 2007 +0100 27859 27860 add missing part of last commit... 27861 27862commit df07fa14da73e92d1a6ee0173468ab5c075d1cbb 27863Author: Roland Scheidegger <sroland@tungstengraphics.com> 27864Date: Wed Feb 21 13:05:42 2007 +0100 27865 27866 fix alignment issues with planar yuv and a bug with packed uyvy 27867 27868 respect that all source planar yuv planes are already dword aligned. 27869 Some attempts to fix up odd widths and odd heights (which are a bit strange 27870 for 4:2:0 formats). They still don't quite work 100% correctly (at the borders) 27871 but neither do they with packed yuv formats. 27872 While here, fix totally broken packed UYVY format by inserting missing break... 27873 Both bugs reported by Felipe Contreras. 27874 27875commit dfcb431adfbbaaee0d262d32735585555a0cbde4 27876Author: Dave Airlie <airlied@linux.ie> 27877Date: Tue Feb 20 21:20:28 2007 +1100 27878 27879 radeon: move some i2c registers out into common register file 27880 27881 These regs are more useful than just multimedia 27882 27883commit 7a25512415bb40b772491bb4a773dfe02ac8b71a 27884Author: Alex Deucher <alex@botch2.com> 27885Date: Fri Feb 16 00:44:11 2007 -0500 27886 27887 Add "DRI" option 27888 27889 This option allows you to disable the DRI per card. It also 27890 removes the "RN50Force3D" option as it is now covered by this 27891 option. RN50 users should set this to TRUE if they want to force 27892 the DRI on. 27893 27894commit fa30ec6d5cd9bf4eb1a960592ca7311175219e4b 27895Author: George Sapountzis <gsap7@yahoo.gr> 27896Date: Tue Feb 13 15:35:32 2007 +0200 27897 27898 [mach64] PreInit: maxPitch, minor cosmetic. 27899 27900commit 1777dcc1956d910073e13322767d9a3de41a949b 27901Author: George Sapountzis <gsap7@yahoo.gr> 27902Date: Tue Feb 13 15:03:40 2007 +0200 27903 27904 [mach64] PreInit: pitchInc is local. 27905 27906commit 9cdf4fcc3a93a4e20244286392dc31a0b8e6a10a 27907Author: George Sapountzis <gsap7@yahoo.gr> 27908Date: Tue Feb 13 11:53:42 2007 +0200 27909 27910 [mach64] PreInit: Use goto bail. 27911 27912 Match ATILock - ATIUnlock. 27913 27914commit 482b85144fee42b6543dcc97a75899e363e05e9e 27915Author: George Sapountzis <gsap7@yahoo.gr> 27916Date: Mon Feb 12 14:27:37 2007 +0200 27917 27918 [mach64] Enable RENDER acceleration on the Pro variants. 27919 27920commit 63248f0b4308a4487cda3aa22daa36e3e0d38d14 27921Author: Dave Airlie <airlied@linux.ie> 27922Date: Mon Feb 12 19:37:36 2007 +1100 27923 27924 fix LVDS by moving bios reading around 27925 27926commit 6a25f620d40cbb063de94aa6b5267296cd5a0670 27927Author: David Airlie <airlied@asimov.stargames.com.au> 27928Date: Mon Feb 12 15:59:43 2007 +1100 27929 27930 remove assignments to pOutput before it is created 27931 27932commit 9234d8045c5fefbd1a781209409e55a13e3e5370 27933Author: David Airlie <airlied@asimov.stargames.com.au> 27934Date: Mon Feb 5 11:53:18 2007 +1100 27935 27936 remove all the locking glxgears runs fine over screen changes without it 27937 27938commit eda9fb15e8bc042661a91a3d8c921006dfb3ddd9 27939Author: George Sapountzis <gsap7@yahoo.gr> 27940Date: Tue Nov 21 04:02:41 2006 +0200 27941 27942 Consolidate atixv.c w/ atimach64xv.c, part 2. 27943 27944 Move {Initialize,Close}XVideo from atixv.c to atimach64xv.c . 27945 27946commit 67b240ddd4a6549a23885afc723b6efa7c52db4c 27947Author: George Sapountzis <gsap7@yahoo.gr> 27948Date: Tue Oct 31 16:44:10 2006 +0200 27949 27950 Consolidate atixv.c w/ atimach64xv.c, part 1. 27951 27952 Merge ATIXVInitializeAdaptor() in ATIMach64XVInitialiseAdaptor(). 27953 27954commit 4dac7896549392f7378cad9620618cec28f41a16 27955Author: George Sapountzis <gsap7@yahoo.gr> 27956Date: Sun Oct 29 03:55:39 2006 +0200 27957 27958 Consolidate aticursor.c w/ atimach64cursor.c 27959 27960commit 23b12b9ff55224e8c3c45eb58ea5fbf2f747d82e 27961Author: George Sapountzis <gsap7@yahoo.gr> 27962Date: Tue Oct 31 16:10:07 2006 +0200 27963 27964 Consolidate atiaccel.c w/ atimach64accel.c, part 2. 27965 27966 Merge ATIInitializeAcceleration() in ATIMach64AccelInit(). 27967 27968commit bc3b222b57d7edf7bdce00a53e8abd9047c314a4 27969Author: George Sapountzis <gsap7@yahoo.gr> 27970Date: Sat Oct 28 19:29:12 2006 +0300 27971 27972 Consolidate atiaccel.c w/ atimach64accel.c, part 1. 27973 27974 Move ATIResizeOffscreenLinear() from atiaccel.c to atimach64xv.c and make it 27975 static. It is only used by XV memory management functions. 27976 27977commit d876889add41c8052d4f3feaa21ddf384f5a7054 27978Author: George Sapountzis <gsap7@yahoo.gr> 27979Date: Tue Nov 21 04:47:06 2006 +0200 27980 27981 Drop non-PCI support from atividmem.c 27982 27983 - we require a PCI device (pVideo != NULL) 27984 - map the VGA framebuffer with xf86MapDomainMemory() 27985 27986commit 2c8b33b761e4411451aea1eea3a89d629aeea40d 27987Author: George Sapountzis <gsap7@yahoo.gr> 27988Date: Tue Nov 21 04:51:02 2006 +0200 27989 27990 Minor refactoring of pATI->Block0Base computation. 27991 27992commit a659b7235332e19659c56cc9d7af362bd8885a29 27993Author: George Sapountzis <gsap7@yahoo.gr> 27994Date: Sun Feb 4 11:28:54 2007 +0200 27995 27996 Cleanup checks for pATI->PCIInfo. 27997 27998 Fixes a bug where shadowfb would not be enabled by default in the non-accel 27999 case because pATI->PCIInfo is now set after ATIProcessOptions() is called. 28000 28001commit 9d77aabdff919360f0c9333105436c31f1f5749a 28002Merge: ff8ea19f 57822be7 28003Author: George Sapountzis <gsap7@yahoo.gr> 28004Date: Mon Feb 5 19:16:51 2007 +0200 28005 28006 Merge branch 'mach64-pci-1' 28007 28008commit d7317604c843e21c1df048a9253ed55a1957cccd 28009Author: David Airlie <airlied@linux.ie> 28010Date: Fri Feb 2 14:54:00 2007 +1100 28011 28012 get rid of mergetype and displaytype from radeon.h 28013 28014commit 15c68ac19ff9971ae02cf52036ba36ddea1a6759 28015Author: David Airlie <airlied@linux.ie> 28016Date: Fri Feb 2 14:30:16 2007 +1100 28017 28018 rename some variables to output from connector 28019 28020commit 72ef23ef7dd159d827e8e122482a58928d532268 28021Author: David Airlie <airlied@linux.ie> 28022Date: Fri Feb 2 14:26:29 2007 +1100 28023 28024 fixup mode probing a bit neater 28025 28026commit 96acf6b2b242454345cc4b9cfc7ca07e0b597b43 28027Author: David Airlie <airlied@linux.ie> 28028Date: Fri Feb 2 14:26:19 2007 +1100 28029 28030 bring over setproperty from intel code 28031 28032commit 76bc53f9b153880730ab61dcd2b6e4e7717e4058 28033Author: David Airlie <airlied@linux.ie> 28034Date: Fri Feb 2 14:15:03 2007 +1100 28035 28036 move i2c buses into outputs 28037 28038commit 47fb9ce657f018177a35b449a4d716dc03be9327 28039Author: David Airlie <airlied@linux.ie> 28040Date: Fri Feb 2 11:45:40 2007 +1100 28041 28042 fixup callers to crtcsetmode fixes my cursors 28043 28044commit 720730b1b0de632488d3b9818210ec5e9c7f07ae 28045Author: David Airlie <airlied@linux.ie> 28046Date: Thu Feb 1 16:43:38 2007 +1100 28047 28048 add locks for crtcs and some missing lines from intel update 28049 28050commit 6748732658850ea506f623a3622aa7135513ffd0 28051Author: David Airlie <airlied@linux.ie> 28052Date: Thu Feb 1 16:26:31 2007 +1100 28053 28054 update to latest intel codebase modulo using their mode set 28055 28056 and it breaks my cursor 28057 28058commit a77f08298dc7e097025e3f7f92e3665c0ef30095 28059Author: David Airlie <airlied@linux.ie> 28060Date: Mon Jan 29 16:12:24 2007 +1100 28061 28062 radeon: always reset cursor regs 28063 28064commit ff8ea19fcdce099732f9359e53cd62b9a04bfa6d 28065Author: Alex Deucher <alex@botch2.hsd1.va.comcast.net> 28066Date: Tue Jan 30 21:11:09 2007 -0500 28067 28068 add option to force the DRI on for RN50 chips 28069 28070 Allow user to force the DRI on for RN50 chips. 28071 3D is not guaranteed to work on these chips, 28072 however in some cases it does. 28073 fixes bug 9802. 28074 28075commit 57822be75740f339445f2375d44632560f4bbe57 28076Author: George Sapountzis <gsap7@yahoo.gr> 28077Date: Tue Nov 21 03:08:26 2006 +0200 28078 28079 Drop pATI->Chipset, no longer used. 28080 28081commit 8b7c88e9ed549777413136c0f2ccc61747b010d8 28082Author: George Sapountzis <gsap7@yahoo.gr> 28083Date: Thu Dec 28 14:46:10 2006 +0200 28084 28085 Add option "probe_sparse" to force probing sparse I/O bases. 28086 28087commit 3654b786f1ad1f4e6bfb2fff5ff057254958f7c0 28088Author: George Sapountzis <gsap7@yahoo.gr> 28089Date: Wed Dec 27 17:59:14 2006 +0200 28090 28091 Sane probe, based on work by Luc Verhaegen <libv@skynet.be>. 28092 28093commit 0117c9bde0d0bbb487bc259cf55e8066ebd63f6a 28094Author: George Sapountzis <gsap7@yahoo.gr> 28095Date: Wed Dec 27 17:49:45 2006 +0200 28096 28097 Drop checking of sparse I/O bases conflicts 28098 28099 This will be handled with an option somehow. 28100 28101commit 87cf2f1874ca8a7cc492d2e2b6918ba69a8aa713 28102Author: George Sapountzis <gsap7@yahoo.gr> 28103Date: Mon Aug 14 17:55:14 2006 +0300 28104 28105 Minor refactoring of module (un)loading for TV_OUT. 28106 28107commit 0b31333c9147af21c50493872779da12a3668e5d 28108Author: George Sapountzis <gsap7@yahoo.gr> 28109Date: Mon Nov 20 22:08:49 2006 +0200 28110 28111 Drop support for non-programmable clocks. 28112 28113 From README.ati: 28114 28115 Clocks for supported programmable clock generators: 28116 28117 The driver currently supports all programmable clock generators known 28118 to exist on Mach64 adapters. 28119 28120 Clocks for unsupported programmable clock generators: 28121 28122 This case is unlikely to occur, but is documented for the sake of 28123 completeness. 28124 28125 Thus: 28126 28127 - check for (pATI->ProgrammableClock > ATI_CLOCK_FIXED) && 28128 (pATI->ProgrammableClock < ATI_CLOCK_MAX) 28129 28130 - drop "probe_clocks" option 28131 28132 - pATIHW->ClockUnmap is no longer used 28133 28134 - pATIHW->ClockMap is only used with NewHW.crtc which is always ATI_CRTC_MACH64 28135 and has the identity map, so drop it 28136 28137 - (pATI->ProgrammableClock != ATI_CLOCK_INTERNAL) => (pATI->depth <= 8) 28138 28139commit c9a9ea5ebd1c27881c1fac19842cfa3af55c126e 28140Author: George Sapountzis <gsap7@yahoo.gr> 28141Date: Mon Nov 20 15:09:41 2006 +0200 28142 28143 Banked memory is no longer needed. 28144 28145 - (pATI->BankInfo.BankSize = 0) in all cases, cull pATI->BankInfo 28146 28147 - only keep the minimal pATIHW.SetBank interface for save/restore 28148 28149 - clean ATISwap() a little, (NewHW.crtc != ATI_CRTC_VGA) 28150 28151 - (UseSmallApertures == TRUE) <=> pATI->VGAAdapter 28152 28153commit 383e963e275d351ea3631c352f5795340162d69f 28154Author: George Sapountzis <gsap7@yahoo.gr> 28155Date: Sun Jan 28 00:44:31 2007 +0200 28156 28157 Always use the linear aperture. 28158 28159 - drop (pATI->OptionLinear == FALSE) 28160 28161 - AcceleratorVideoRAM is always set, i.e. VGAVideoRAM is not used 28162 28163 - pATI->LinearBase is always set 28164 28165 - xf86LinearVidMem() is now checked in atipreinit() for both CPIO and MMIO 28166 28167commit 9b126f45db27496c1ad16db65a61fe641a018983 28168Author: George Sapountzis <gsap7@yahoo.gr> 28169Date: Mon Nov 20 20:38:59 2006 +0200 28170 28171 We always use the accelerator CRTC within the xserver. 28172 28173 - cull (pATI->NewHW.crtc != ATI_CRTC_MACH64). 28174 28175commit 0abcb2b2c84a50e64bb6aa9e94760659ecf33add 28176Author: George Sapountzis <gsap7@yahoo.gr> 28177Date: Sun Jan 28 00:19:15 2007 +0200 28178 28179 Clean pATI->Adapter and pATI->VGAAdapter. 28180 28181 - cull (pATI->Adapter != ATI_ADAPTER_MACH64) 28182 28183 - treat pATI->VGAAdapter as Bool 28184 28185commit 817b4b5338dba4d8eb862f1f79ef0edb75aad7f4 28186Author: George Sapountzis <gsap7@yahoo.gr> 28187Date: Wed Dec 27 21:16:18 2006 +0200 28188 28189 Keep PCI mach64 only, drop: 28190 28191 - Chip < ATI_CHIP_88800GXC 28192 - Chipset != ATI_CHIPSET_ATI 28193 - Adapter != ATI_ADAPTER_MACH64 28194 28195 - depth < 8 28196 28197commit d350860e29f043e98bfb1da74b26280f1755ab6f 28198Author: George Sapountzis <gsap7@yahoo.gr> 28199Date: Wed Dec 27 20:56:45 2006 +0200 28200 28201 Keep PCI mach64 only (atipreinit.c, atimode.c), drop: 28202 28203 - Chip < ATI_CHIP_88800GXC 28204 - Chipset != ATI_CHIPSET_ATI 28205 - Adapter != ATI_ADAPTER_MACH64 28206 28207 - depth < 8 28208 28209 atimode.c only: 28210 28211 - NewHW.crtc != ATI_CRTC_MACH64 28212 28213 This allows to drop VGACalculate(), VGAWonderCalculate() cruft early. 28214 28215commit 5fa77f2d122e7267911a15235338d8d3f9eece2e 28216Author: George Sapountzis <gsap7@yahoo.gr> 28217Date: Wed Dec 27 14:50:48 2006 +0200 28218 28219 Drop probing for VGA, VGAWonder, 8514, Mach32. 28220 28221 - ChipHasSUBSYS_CNTL 28222 - Coprocessor 28223 - SharedAccelerator 28224 - SharedVGA <=> (VGAAdapter != ATI_ADAPTER_NONE) 28225 28226commit 1b2a9e1735da63bbb8cb41939668f789cee22579 28227Author: George Sapountzis <gsap7@yahoo.gr> 28228Date: Mon Nov 20 02:35:56 2006 +0200 28229 28230 Drop AVOID_NON_PCI. 28231 28232 This was not set anyway, because configure.ac would compute ATIMISC_NON_PCI and 28233 then test ATI_AVOID_NON_PCI to set AVOID_NON_PCI... 28234 28235commit 2fb84dd2fc1a663c7b9d77e706189439ffe08381 28236Author: Alex Deucher <alex@botch2.hsd1.va.comcast.net> 28237Date: Thu Jan 25 02:01:42 2007 -0500 28238 28239 I think we can say r300 support is not experimental 28240 28241commit e155659f4b006edff31f1fda88b32bf608dbeece 28242Author: Alex Deucher <alex@botch2.hsd1.va.comcast.net> 28243Date: Thu Jan 25 01:59:27 2007 -0500 28244 28245 update man page. fixes bug 6457 28246 28247 Most cards support 2D and 3D now 28248 28249commit fa12fe1cc90dd745f3eea35a07d4f3efd652e7b4 28250Author: Dave Airlie <airlied@linux.ie> 28251Date: Tue Jan 23 19:37:51 2007 +1100 28252 28253 remove unused restore palette function 28254 28255commit 4d3649b84a3325043c38cece4499de0095cebd71 28256Author: Dave Airlie <airlied@linux.ie> 28257Date: Tue Jan 23 19:37:21 2007 +1100 28258 28259 remove isused and isactive 28260 28261commit 9149e763865598c307cbefc753ff6ebdeeaf32ae 28262Author: David Airlie <airlied@asimov.stargames.com.au> 28263Date: Tue Jan 23 15:20:11 2007 +1100 28264 28265 radeon: fixup problem with cursor not going off 28266 28267 Also fixup secondary dac detect to return unknown for now 28268 28269commit 2a13a3b641d9acf4f50472e1c8ba07633c3b78d6 28270Author: David Airlie <airlied@asimov.stargames.com.au> 28271Date: Tue Jan 23 10:36:06 2007 +1100 28272 28273 fixup init disp bandwidth 28274 28275 This probably break old-style dualhead badly 28276 28277commit cbd84bed13582e82f8b2e84aa152602474c09cd4 28278Author: Dave Airlie <airlied@linux.ie> 28279Date: Mon Jan 22 22:19:46 2007 +1100 28280 28281 removed even more mergedfb remains 28282 28283commit 6296882135b183425a219efb75374d3b172de370 28284Author: Dave Airlie <airlied@linux.ie> 28285Date: Mon Jan 22 22:14:48 2007 +1100 28286 28287 remove more mergedfb relics 28288 28289commit 26e1bab2c7a4ad52710746265b3da495b54dab36 28290Author: Dave Airlie <airlied@linux.ie> 28291Date: Mon Jan 22 22:11:55 2007 +1100 28292 28293 radeon: fixup some of the video code after randr 28294 28295 This may not be complete, I need to test on dual-head system later 28296 28297commit 9008e1caa45e4a18e6f3289c7b17036730fe578a 28298Author: Dave Airlie <airlied@linux.ie> 28299Date: Mon Jan 22 21:16:47 2007 +1100 28300 28301 remove unused merged code 28302 28303commit e29a32d711553fcb4ca9928122ac285fe0b0c1a7 28304Author: Dave Airlie <airlied@linux.ie> 28305Date: Mon Jan 22 21:09:53 2007 +1100 28306 28307 add back overlay CRTC chooser this may not work yet 28308 28309commit aa9d04ba94a3cd60b248231da517e2817591fc69 28310Author: Dave Airlie <airlied@linux.ie> 28311Date: Sun Jan 21 20:29:56 2007 +1100 28312 28313 remove alot more mergedfb support 28314 28315commit e8be0056e8ff666f63a294691661f5dab289203e 28316Author: Dave Airlie <airlied@linux.ie> 28317Date: Thu Jan 18 16:41:25 2007 +1100 28318 28319 remove mergedfb 28320 28321commit a648050a3cc60f92b1ca0b3d707aadf93d076d91 28322Author: Dave Airlie <airlied@linux.ie> 28323Date: Wed Jan 17 17:28:07 2007 +1100 28324 28325 move some cursor code around for show/hide 28326 28327commit d2ecfb507282726122bb8b0d17fd3637d0ae7d46 28328Author: Dave Airlie <airlied@linux.ie> 28329Date: Wed Jan 17 17:16:27 2007 +1100 28330 28331 make dri work again 28332 28333commit 369f7c85ceff983defb7657b80ec9cd3e5440b07 28334Author: Dave Airlie <airlied@linux.ie> 28335Date: Wed Jan 17 11:14:09 2007 +1100 28336 28337 comment out unblanking on VT switch 28338 28339commit bdb66a2042f02c4b57bd3c0181a00b39fcbdb232 28340Author: Dave Airlie <airlied@linux.ie> 28341Date: Sun Jan 14 21:07:05 2007 +1100 28342 28343 radeon: get randr-1.2 mode switching mostly working on my laptop 28344 28345 The main change is to fix the dac macro + cntl register writes for rv350 and 28346 above, this still has a problem with resetting the same mode after connect 28347 disconnect cycle, need to talk to keithp 28348 28349commit aa0e7337815d4daca4df0671768621330b759011 28350Author: Dave Airlie <airlied@linux.ie> 28351Date: Sun Jan 14 19:17:32 2007 +1100 28352 28353 use more randr-ish names for variables 28354 28355commit 9d37f23aea43b74a7ec640e6b03617c8392e2572 28356Author: Dave Airlie <airlied@linux.ie> 28357Date: Sun Jan 14 19:10:11 2007 +1100 28358 28359 radeon: move blank/unblank to use randr 28360 28361commit 6bd4fe42789c38e7e804826715214ce6badcca6a 28362Author: Dave Airlie <airlied@linux.ie> 28363Date: Sun Jan 14 18:38:59 2007 +1100 28364 28365 disable dpms on/off as it was writing regs before we wanted to 28366 28367commit 3949288ed26a91e180b178f4796f2f7e1bdc8ed6 28368Author: Dave Airlie <airlied@linux.ie> 28369Date: Sun Jan 14 16:26:39 2007 +1100 28370 28371 radeon: destroy and fixup LVDS crtc 28372 28373commit f5ffb9fb3bdd7cec020240c5831ae92ee744972c 28374Author: Timo Jyrinki <timo.jyrinki@hut.fi> 28375Date: Fri Jan 12 23:08:50 2007 +1100 28376 28377 ati: fix detection of some newer radeons using ati wrapper 28378 28379 This is mainly due to the cards having a different resource 1. 28380 Fixes 6796 28381 28382commit 1e690c57ad1dfaeb760ef6e799faa6398e9a0551 28383Author: Dave Airlie <airlied@linux.ie> 28384Date: Fri Jan 12 23:03:10 2007 +1100 28385 28386 always create a second CRTC I think this should solve any problems 28387 28388commit 25b36a4c56422ad8e25a4a1c55055c2f062213ca 28389Author: Dave Airlie <airlied@linux.ie> 28390Date: Fri Jan 5 16:13:16 2007 +1100 28391 28392 minor randr1.2 fixups 28393 28394commit abe8791e4f9fa3e88273897b351387cd33822734 28395Author: Dave Airlie <airlied@linux.ie> 28396Date: Fri Jan 5 15:09:43 2007 +1100 28397 28398 hook up DPMS through xf86 28399 28400commit 8a9c68c2234b91ed38555f892afdad30b9e4b455 28401Author: Dave Airlie <airlied@linux.ie> 28402Date: Fri Jan 5 15:09:15 2007 +1100 28403 28404 set binding for new heads 28405 28406commit d6a2b8aeb5a97ee907fd45b574bc6e4ab4b3aede 28407Author: Dave Airlie <airlied@linux.ie> 28408Date: Fri Jan 5 09:33:54 2007 +1100 28409 28410 fix LUT 28411 28412commit ef1d36e56dec1fec37cee2dfd9cb5bf8ce2c485c 28413Author: Roland Scheidegger <sroland@tungstengraphics.com> 28414Date: Wed Jan 3 15:56:23 2007 +0100 28415 28416 fix uninitialized mode regs (bug 9495) 28417 28418 Since the reorganization of the mode setting code, the mode registers relying 28419 on state already set (by bios) were not read, thus clearing out all bits the 28420 driver does not touch. At the very least, this could lead to completely 28421 nonfunctional to misbehaving dvi output (see bug 9495). Fix this by using the 28422 SavedReg values, which also makes it more obvious that those are bits which 28423 were not set by the driver previously, but come from register readback. 28424 (cherry picked from 9506f7015a1e442f4ca0bd3bfae555ec7e8a5f37 commit) 28425 28426commit 55aa832157bdebcba2d58896777942d108c352b0 28427Author: Dave Airlie <airlied@linux.ie> 28428Date: Thu Jan 4 17:57:31 2007 +1100 28429 28430 fix cursor handling 28431 28432commit 4f8a7cafdc77e98dc44f9eced876560b1ee01117 28433Author: Dave Airlie <airlied@linux.ie> 28434Date: Thu Jan 4 16:23:39 2007 +1100 28435 28436 hook up crtc modesetting 28437 28438commit 52f749c8a613ee316044abe82156ee270412ced8 28439Author: Dave Airlie <airlied@linux.ie> 28440Date: Thu Jan 4 16:23:26 2007 +1100 28441 28442 fix typo 28443 28444commit d7ff61c6822cbede7f5b59b411048d33dbae9ee4 28445Author: Dave Airlie <airlied@linux.ie> 28446Date: Thu Jan 4 13:38:26 2007 +1100 28447 28448 hook up detection of output 28449 28450commit 50439d7e9f4f6f9933bacd59f8bb3e655a35dbc2 28451Author: Dave Airlie <airlied@linux.ie> 28452Date: Thu Jan 4 13:21:57 2007 +1100 28453 28454 fixup some function calls 28455 28456commit e067bfee92f2b0877108355619b2fb9188a9d15a 28457Author: Dave Airlie <airlied@linux.ie> 28458Date: Thu Jan 4 13:21:45 2007 +1100 28459 28460 add more files for back compat 28461 28462commit f1f34627ffbe2136ac3e023c01c0430412919ded 28463Author: Dave Airlie <airlied@linux.ie> 28464Date: Thu Jan 4 12:50:03 2007 +1100 28465 28466 make restore mode non-static 28467 28468commit 249ec67c296b34d0c7c1d78602628d2f7ce6a045 28469Author: Dave Airlie <airlied@linux.ie> 28470Date: Thu Jan 4 12:43:03 2007 +1100 28471 28472 more minor fixes to get radeon up to speed 28473 28474commit a43c1d55f5f855d9e6ae939dd4eec1c607b6d514 28475Author: Dave Airlie <airlied@linux.ie> 28476Date: Thu Jan 4 12:29:51 2007 +1100 28477 28478 bring radeon randr code inline with intel randr code 28479 28480commit c54a31565e3ea50f1af7b81b1bf0c0e437099f47 28481Author: Dave Airlie <airlied@linux.ie> 28482Date: Thu Jan 4 10:39:01 2007 +1100 28483 28484 radeon: fixup fastwrites to be always off 28485 28486 My 8500 in i845 doesn't work with fastwrites even setup by the firmware. 28487 28488commit dda2bbfdc2e0da6e8b9cb8d1bf78b91f1bb8131f 28489Author: Dave Airlie <airlied@linux.ie> 28490Date: Thu Jan 4 10:29:54 2007 +1100 28491 28492 radeon: don't touch Controller[1] on single CRTC cards. 28493 28494 This might fix bug 9371 28495 28496commit 9506f7015a1e442f4ca0bd3bfae555ec7e8a5f37 28497Author: Roland Scheidegger <sroland@tungstengraphics.com> 28498Date: Wed Jan 3 15:56:23 2007 +0100 28499 28500 fix uninitialized mode regs (bug 9495) 28501 28502 Since the reorganization of the mode setting code, the mode registers relying 28503 on state already set (by bios) were not read, thus clearing out all bits the 28504 driver does not touch. At the very least, this could lead to completely 28505 nonfunctional to misbehaving dvi output (see bug 9495). Fix this by using the 28506 SavedReg values, which also makes it more obvious that those are bits which 28507 were not set by the driver previously, but come from register readback. 28508 28509commit d5c6306385cbc7b13a132ae72ed04e04e24d9886 28510Author: George Sapountzis <gsap7@yahoo.gr> 28511Date: Tue Dec 19 09:45:24 2006 +0200 28512 28513 Use portable XORG_RELEASE_VERSION macro. 28514 28515commit 97df0a32b14e34f22dce9cd3b375c1ac5fb1c5eb 28516Author: Michel Dänzer <michel@tungstengraphics.com> 28517Date: Tue Dec 19 11:28:21 2006 +0100 28518 28519 radeon: Advertise 2D engine limits to EXA when we're not using the 3D engine. 28520 28521 Also round up to the maximum width and height, as that's what EXA compares. 28522 28523commit 295823d0879a5b574bb79843a6acd43adb9259e5 28524Author: Michel Dänzer <michel@tungstengraphics.com> 28525Date: Tue Dec 19 11:19:20 2006 +0100 28526 28527 radeon: Default to leaving AGP transfer mode and fast writes unchanged. 28528 28529 Based on the assumption that firmware should have set up the card and host 28530 bridge appropriately for these settings, this may actually be safer, at least 28531 for the transfer rate; leaving fast writes enabled is hopefully safe as well, 28532 it certainly is on my sytem. 28533 28534 See https://bugs.freedesktop.org/show_bug.cgi?id=9284 . 28535 28536commit 81f3b4070b70483d6da4f7587e52a64ac69c8ca0 28537Author: Eamon Walsh <ewalsh@tycho.nsa.gov> 28538Date: Mon Dec 18 14:06:15 2006 -0500 28539 28540 Revert "Convert callers of LookupWindow() to dixLookupWindow()." 28541 28542 This reverts commit 48ff33a1770f3684cd50184db8f1944a456d9974. 28543 28544commit 87592ffb717da1f0a1767a38918d16d60953599c 28545Author: Dave Airlie <airlied@linux.ie> 28546Date: Sun Dec 17 20:10:58 2006 +1100 28547 28548 add setdpi call 28549 28550commit 41cd50487bc85708a3f791dfa70bf85223d91a65 28551Author: Dave Airlie <airlied@linux.ie> 28552Date: Sun Dec 17 20:08:07 2006 +1100 28553 28554 add displayWidth set 28555 28556commit 555e0f988c571aa47bc62ec6d9612a71f1a1f59b 28557Author: Dave Airlie <airlied@linux.ie> 28558Date: Sun Dec 17 19:59:19 2006 +1100 28559 28560 add randr pre init 28561 28562commit 3e2c72d7ac80ed874bcf2887a5253dac1c9bb02c 28563Author: Dave Airlie <airlied@linux.ie> 28564Date: Sun Dec 17 19:49:08 2006 +1100 28565 28566 fixup even more remnants of old code 28567 28568commit 4356b031886e00b3ed5ac1dbceeadebd7fc29fc2 28569Author: Dave Airlie <airlied@linux.ie> 28570Date: Sun Dec 17 19:47:45 2006 +1100 28571 28572 fix remnant of old code 28573 28574commit 17d39502a8dbea395051007aebffdf23644e9ee1 28575Author: Dave Airlie <airlied@linux.ie> 28576Date: Sun Dec 17 19:44:39 2006 +1100 28577 28578 use getddcmodes properly 28579 28580commit bf0b364cbd8682f297e2d110ece1f72e7151340d 28581Author: Dave Airlie <airlied@linux.ie> 28582Date: Sun Dec 17 17:31:08 2006 +1100 28583 28584 fix up output namings 28585 28586commit 4962a5430844114de864418d803f3182b90792ff 28587Author: Dave Airlie <airlied@linux.ie> 28588Date: Sun Dec 17 16:58:07 2006 +1100 28589 28590 remove some code that moved into server 28591 28592commit b47013f41249516ae82e22bd75caf573da2cc13c 28593Author: Dave Airlie <airlied@linux.ie> 28594Date: Sat Dec 16 12:02:37 2006 +1100 28595 28596 hook up mode detect to new randr code 28597 28598commit 48ff33a1770f3684cd50184db8f1944a456d9974 28599Author: Eamon Walsh <ewalsh@tycho.nsa.gov> 28600Date: Fri Dec 15 17:43:39 2006 -0500 28601 28602 Convert callers of LookupWindow() to dixLookupWindow(). 28603 28604commit 62a5e7565b66d7b7d65609c034f34b55bd266617 28605Author: Dave Airlie <airlied@linux.ie> 28606Date: Fri Dec 15 19:48:42 2006 +1100 28607 28608 bring over latest code from Intel git repo. 28609 28610 this code doesn't work yet just realigns with intel work 28611 28612commit 29124400c7f193317d41d8cfd748371a239cfea1 28613Author: Dave Airlie <airlied@linux.ie> 28614Date: Sun Dec 10 18:00:17 2006 +1100 28615 28616 hook up randr crtc setting 28617 28618commit 120c854f185c1e7711cf0dee19303fdb8894d49d 28619Author: Roland Scheidegger <sroland@tungstengraphics.com> 28620Date: Fri Dec 8 02:51:52 2006 +0100 28621 28622 radeon: avoid unnecessary OUTPLL/INPLL calls when displaying video 28623 28624 It is not necessary to always emit a OUTPLL/INPLL pair when we display 28625 a video frame. On some chips there are erratas for which the workarounds 28626 cause a 10ms delay by those calls. This is related to #5876 though those 28627 affected may suffer from other slowness issues too. 28628 28629commit ccd37b3e88cdcfe03b7c707b6082339eb8f11e6b 28630Author: Michel Dänzer <michel@tungstengraphics.com> 28631Date: Thu Dec 7 18:41:48 2006 +0100 28632 28633 radeon: Move Xv option handling code into RADEONPreInitXv(). 28634 28635 This unclutters RADEONPreInit() somewhat, but more importantly moves comparison 28636 against info->ChipFamily after that's initialized. 28637 28638commit 02be8c04a268eaa71bdfe891331758715fef4e71 28639Author: Michel Dänzer <michel@tungstengraphics.com> 28640Date: Thu Dec 7 18:31:58 2006 +0100 28641 28642 radeon: Make cache flushing for R300 series less invasive. 28643 28644 Instead of calling the DRM CP idle ioctl, just emit the cache flush commands 28645 into the CP stream. 28646 28647commit be46f7b6f7a729695a0affe2328077666283af12 28648Author: Michel Dänzer <michel@tungstengraphics.com> 28649Date: Thu Dec 7 16:27:10 2006 +0100 28650 28651 Fix radeon compile warnings. 28652 28653commit 59c5b4f69c97fb233144677ab18095f88aed006b 28654Author: Daniel Stone <daniel@fooishbar.org> 28655Date: Wed Dec 6 18:48:55 2006 +0200 28656 28657 Makefile.am: re-add ChangeLog to CLEANFILES 28658 28659commit 7611f14dc697b75cd7bb6493bb65b3a7ead379b8 28660Author: Daniel Stone <daniel@fooishbar.org> 28661Date: Wed Dec 6 18:47:33 2006 +0200 28662 28663 Makefile.am: fix changelog hook 28664 Make changelog hook safe for all situations. 28665 28666commit f2a9d2d61addc4e3d5e128d8078a04c9a136be6a 28667Author: Roland Scheidegger <sroland@tungstengraphics.com> 28668Date: Wed Dec 6 00:16:41 2006 +0100 28669 28670 radeon: fix up packed yuv broken by planar yuv patch 28671 28672 fix the forgotten leftuv value for packed yuv which is 28673 needed to get correct uv starting pixel (fixes broken clipping / 28674 non-null src start pixel as tvtime uses) 28675 28676commit d2aa57a4bdd4a0deb2c659bb07f3f57d1116f0d1 28677Author: Roland Scheidegger <sroland@tungstengraphics.com> 28678Date: Tue Dec 5 01:30:13 2006 +0100 28679 28680 radeon: use overlay scaler native planar yuv capability. 28681 28682 Radeons can do planar yuv just fine, there is no need to convert all data 28683 to packed yuv manually. This saves some cpu cycles as well as some 28684 (graphic card) ram. 28685 28686commit 5d5fa1b86e5179b061f0db47fe0227d1b84c37f8 28687Author: Dave Airlie <airlied@linux.ie> 28688Date: Mon Dec 4 18:53:33 2006 +1100 28689 28690 update radeon driver to Intel driver xf86Crtc interfaces 28691 28692commit 51d1cf19e71dd5de47f2c6467f4a1685eefd9e1e 28693Author: Dave Airlie <airlied@linux.ie> 28694Date: Sun Dec 3 16:30:01 2006 +1100 28695 28696 Add radeon randr 1.2 initial attempt at support 28697 28698 This doesn't do a huge amount yet 28699 28700commit eb17c9aed2144701ad7bd1042b2905446e4d708a 28701Author: airlied <airlied@optimus.localdomain> 28702Date: Sat Nov 18 15:48:02 2006 +1100 28703 28704 move radeon over to not using X's mode management like the intel driver. 28705 28706 this builds and at least seems to pick a mode.. might be near to hooking 28707 up randr1.2 soon 28708 28709commit 941b5120916ebff69ec0ee4ca7a46105d5306e5d 28710Author: airlied <airlied@optimus.localdomain> 28711Date: Sat Nov 18 10:41:12 2006 +1100 28712 28713 make radeon modes use generic code 28714 28715commit 608427283ea59cdfe3d85e1cf32719bc3809dc16 28716Author: airlied <airlied@optimus.localdomain> 28717Date: Sat Nov 18 10:35:31 2006 +1100 28718 28719 add more functions from i830 driver 28720 28721 This adds more generic functions from the i830 driver to the radeon 28722 28723commit d5b8cafc185b7d9fb909cb18a08615f81d89eaf8 28724Author: airlied <airlied@optimus.localdomain> 28725Date: Sat Nov 18 10:22:38 2006 +1100 28726 28727 add generic mode handling code from Intel driver 28728 28729commit 9cfa82e1670ad85746926995972a535ddf03ee07 28730Author: Dave Airlie <airlied@linux.ie> 28731Date: Sun Dec 3 15:56:38 2006 +1100 28732 28733 move monitor detection into single function 28734 28735commit 032a00d155ecdcdcab3ca0d047f552fe6b338a97 28736Author: airlied <airlied@optimus.localdomain> 28737Date: Sat Nov 18 10:23:12 2006 +1100 28738 28739 make DDC function more generic 28740 28741commit 2632ae24b3b50c69567e92611ce7c2cd2ee98baa 28742Author: Dave Airlie <airlied@linux.ie> 28743Date: Sun Dec 3 15:22:38 2006 +1100 28744 28745 re-program bios 4 scratch so we know if the LCD panel is there post startup 28746 28747commit ddec94e8f2f5746bad0e998d15303e771803d22b 28748Author: Dave Airlie <airlied@linux.ie> 28749Date: Sun Dec 3 11:07:23 2006 +1100 28750 28751 move crtc to connector mapping to a connector to crtc mapping. 28752 28753 Still not allowing any clone modes, but heading in the correct direction 28754 I hope... there is a chance this will regress something from superpatch.. 28755 28756commit 218cb7f8db35d0ce8f9fc4ece72bd5ec3f52106f 28757Author: Dave Airlie <airlied@linux.ie> 28758Date: Sun Dec 3 09:53:35 2006 +1100 28759 28760 radeon: dynamically allocate connectors 28761 28762commit f22fd02bc7f7ef67e33c9bd93eae2d7488669b5e 28763Author: Dave Airlie <airlied@linux.ie> 28764Date: Sat Dec 2 19:48:00 2006 +1100 28765 28766 make CRTCs malloc'ed rather than part of Ent 28767 28768 This is leading towards randr-1.2 believe me :-) 28769 28770commit 13efd1f75fbd29a4d1f69d7d5a5bb2b8d90ad820 28771Author: Dave Airlie <airlied@linux.ie> 28772Date: Sat Dec 2 19:16:41 2006 +1100 28773 28774 move HasCRTC2 to pRADEONEnt from info 28775 28776commit 3933b86390eb2f4198d490acd4914fb470c40f47 28777Author: Michel Dänzer <michel@tungstengraphics.com> 28778Date: Fri Dec 1 18:00:17 2006 +0100 28779 28780 Make log file output clear about origin of used overlay scaler buffer width. 28781 28782commit 563dfb52f5b24d2f765fae0394e599c11dcf0b5b 28783Author: Michel Dänzer <michel@tungstengraphics.com> 28784Date: Fri Dec 1 17:56:19 2006 +0100 28785 28786 Increase default overlay scaler buffer width for RV350. 28787 28788 Works with 1920x1080 video on my M10. 28789 28790commit 984ea3d9b7c27d8183ca94ad9d7905f03613fe10 28791Author: Michel Dänzer <michel@tungstengraphics.com> 28792Date: Fri Dec 1 17:42:23 2006 +0100 28793 28794 Fix build warning. 28795 28796commit 9f5ea3981449f29ff204eb154166e8fc813205fa 28797Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 28798Date: Wed Nov 29 16:46:57 2006 +0100 28799 28800 radeon: bug #1462, predownscale to make HD video work. 28801 28802 Use the overlay scaler's predownscale capability to make videos with large 28803 horizontal resolution work if it exceeds the scaler buffer width. Make the 28804 scaler buffer width user-configurable since we don't know it for all chips, 28805 and using predownscaling may otherwise reduce quality even if it wouldn't 28806 be needed. This should fix bug #1462. 28807 28808commit 35a5eaf5adf103ef57f8d4590dae25b0c50bd780 28809Author: airlied <airlied@optimus.localdomain> 28810Date: Tue Nov 28 17:31:05 2006 +1100 28811 28812 fixup EXA after previous commit 28813 28814commit ba047ae863aac869595a31ea94f102084636fd29 28815Author: airlied <airlied@optimus.localdomain> 28816Date: Tue Nov 28 17:26:10 2006 +1100 28817 28818 radeon: initial move framebuffer away from zero prep. 28819 28820 In theory the driver should be able to handle the front buffer not at VRAM 0 28821 In practice it didn't.. this is cleanup for at least XAA parts of the driver 28822 to allow for the frontbuffer to move. It has to re-organise a large part of ScreenInit so things happen in the correct order otherwise some things get setup in-correctly. (not sure EXA with fb not at 0 works yet...) 28823 28824commit 0d42cab2d889070ab6de05990e2b19724595f7af 28825Author: Alex Deucher <alex@t41p.(none)> 28826Date: Wed Nov 22 13:44:49 2006 -0500 28827 28828 clean up previous VIP commit. 28829 28830commit 8e0f57571228729cc40ed59266484ecc43819f44 28831Author: Alex Deucher <alex@t41p.(none)> 28832Date: Wed Nov 22 13:38:01 2006 -0500 28833 28834 Don't mess with the VIP bus on mobility chips. Seems to cause fan 28835 problems on asus laptops (see bug 7463). We don't support video-in 28836 on any laptops at the moment anyway. 28837 28838commit e4850610596a08082a528c38186268b72012e78c 28839Author: Alain Péteut <alain.peteut@gmx.ch> 28840Date: Wed Nov 22 12:54:55 2006 -0500 28841 28842 enable i2c for PCI_CHIP_RV250_If. allows AIW fuctionality to work. 28843 patch and testing from Alain Péteut 28844 28845commit ff38c2c9cf0c6acbdaf53b681ec721487cf411f8 28846Author: Michel Dänzer <michel@tungstengraphics.com> 28847Date: Sun Nov 19 17:11:46 2006 +0100 28848 28849 radeon: Bug #8989: Fix page flipping artifacts with XAA. 28850 28851 We still have to force-sync the pages on enabling page flipping with XAA as the 28852 second page may have been clobbered by the offscreen pixmap cache since they 28853 were last synced. 28854 28855commit 2dd0892817090709b2fb7ffba920595ada00fcad 28856Author: Michel Dänzer <michel@tungstengraphics.com> 28857Date: Sun Nov 19 17:09:11 2006 +0100 28858 28859 Fix compile warnings. 28860 28861commit 8c1b9269aaac746989bdb9be02915722bcf6fec0 28862Author: airlied <airlied@optimus.localdomain> 28863Date: Fri Nov 17 18:47:50 2006 +1100 28864 28865 forgot SetPitch function is called 28866 28867commit 3a0338bb487739a1740ddfb4d36d53cfa7f798c6 28868Author: airlied <airlied@optimus.localdomain> 28869Date: Fri Nov 17 18:37:00 2006 +1100 28870 28871 split mode handling code into separate modes source file 28872 28873commit 13e0a95564220d42ad8060b43b0aaaba3002157a 28874Author: David Airlie <airlied@asimov.stargames.com.au> 28875Date: Mon Nov 13 09:35:18 2006 +1100 28876 28877 fix up issue with bindings and mergedfb 28878 28879commit 63444770ff30afeb134befed5bb3a5c28b7bc4e9 28880Author: George Sapountzis <gsap7@yahoo.gr> 28881Date: Thu Nov 9 22:01:28 2006 +0200 28882 28883 [mach64] Prefer double-buffered visuals for GLX. 28884 28885 So that 3D apps apropriately fail in DRI... 28886 28887commit 0411f68d71c75cda38ad46ad5c0f5cc3cda64464 28888Author: George Sapountzis <gsap7@yahoo.gr> 28889Date: Thu Nov 9 22:00:56 2006 +0200 28890 28891 [mach64] Use asynchronous DMA mode by default. 28892 28893commit 8dd668db05e8f1282fa0e29a91e38c7cbdc97aa1 28894Author: Dave Airlie <airlied@linux.ie> 28895Date: Sun Nov 12 16:33:21 2006 +1100 28896 28897 radeon: abstract init display bandwidth from mergedfb 28898 28899commit 45a9f434b77ad892efcdba267a935135b972dd42 28900Author: Dave Airlie <airlied@linux.ie> 28901Date: Sun Nov 12 16:12:56 2006 +1100 28902 28903 add radeon Init2 function, to abstract some mergedfb mode setting 28904 28905commit d0828804aec0852afaca44ee41c4792edc40dbb4 28906Author: Dave Airlie <airlied@linux.ie> 28907Date: Sun Nov 12 15:59:02 2006 +1100 28908 28909 add initial controller binding, remove uses of mergedfb test 28910 28911commit 78061ca460c9d6fd338ccf273bbfa152f471622e 28912Author: Dave Airlie <airlied@linux.ie> 28913Date: Mon Oct 23 16:49:59 2006 +1000 28914 28915 radeon: separate static controller setup from non-static 28916 28917commit f917f65e8d2b13d8b9bfc4320f4a935f5a5c4bff 28918Author: George Sapountzis <gsap7@yahoo.gr> 28919Date: Thu Nov 2 21:16:53 2006 +0200 28920 28921 Officially require libdrm 2.2.0 28922 28923 Note that this is only for the headers at compile-time. The driver still 28924 uses libdrm 1.2.0 from Xserver at run-time. 28925 28926commit ca8c148942371c9445f32985625ce38eabc016e1 28927Author: George Sapountzis <gsap7@yahoo.gr> 28928Date: Thu Oct 26 15:40:46 2006 +0300 28929 28930 Handle .99 in version patch level. 28931 28932commit 23cec754158c94b43225b2070fc0a0b003e7bd25 28933Author: George Sapountzis <gsap7@yahoo.gr> 28934Date: Thu Oct 26 15:39:49 2006 +0300 28935 28936 [mach64] EXA: reserve DRI buffers only when DRI is active. 28937 28938commit 9cd78af1d06f753e7bd7e1acaff25398bde45f6e 28939Author: Dave Airlie <airlied@linux.ie> 28940Date: Mon Oct 23 16:52:07 2006 +1000 28941 28942 radeon: set preferred modes 28943 28944commit 914a5f0117ccb76292bdd81ea4a2a79d33319550 28945Author: Michel Dänzer <michel@tungstengraphics.com> 28946Date: Mon Oct 23 10:13:24 2006 +0200 28947 28948 radeon: Pitch doesn't matter when height == 1. 28949 28950commit 2bcb51d66edaa944379cf8c8ca1ba91fffdc20a8 28951Author: Michel Daenzer <michel@tungstengraphics.com> 28952Date: Sun Oct 22 17:26:28 2006 +0200 28953 28954 Bug #6756: Attempt to fix repeat picture acceleration. 28955 28956 Always use normalized texture coordinates on R200 and fall back if a POT 28957 texture pitch doesn't match the HW's implicit pitch. 28958 28959commit dabffb8335027b60ca1fc554423e196dfb9acd6d 28960Author: Henry Zhao <henry.zhao@sun.com> 28961Date: Tue Oct 17 21:17:47 2006 -0400 28962 28963 improved results for mergedfb xrandr refresh rates - bug 6966 28964 28965commit 831fc4a009841f68fe161b6e0db5b212d7d8079b 28966Author: Henry Zhao <henry.zhao@sun.com> 28967Date: Sun Oct 15 17:12:06 2006 -0400 28968 28969 fix bug 6966 (Henry Zhao, Shem Multinymous) 28970 28971commit 5425be996692425e57884e5a4c159fc79510e89f 28972Author: Alex Deucher <alex@t41p.(none)> 28973Date: Sun Oct 15 16:49:24 2006 -0400 28974 28975 clean up magic number 28976 28977commit 457178f418983041ea59df703a17513a47d17044 28978Author: Alex Deucher <alex@t41p.(none)> 28979Date: Sun Oct 15 16:40:08 2006 -0400 28980 28981 Disable all ouputs on initial mode set. The ones we want will get 28982 enabled by RADEONEnableDisplay(). 28983 28984commit 3fb3f410931a13542fdd4c51c4c9ecada7b37d76 28985Author: Alex Deucher <alex@t41p.(none)> 28986Date: Sun Oct 15 13:29:48 2006 -0400 28987 28988 Rework the output init code. Divide into separate functions 28989 for each output. 28990 28991commit 45e13f3c8786e75276cce85de4af653713d9b4d8 28992Author: Alex Deucher <alex@t41p.(none)> 28993Date: Sun Oct 15 13:18:01 2006 -0400 28994 28995 Remove paneloff option. This is already covered by monitorlayout and 28996 shouldn't be needed with xrandr++. 28997 Also remove commented out Bioshotkey code. 28998 28999commit 4abb5b7afea8d37e95461335e52b16168e555dab 29000Author: Alex Deucher <alex@t41p.(none)> 29001Date: Sun Oct 15 13:10:22 2006 -0400 29002 29003 Remove Bioshotkeys option. Leave the bios as configured at boot. 29004 Most distros enable this option by defalt anyway, and with xrandr++ 29005 it won't be an issue anyway. 29006 29007commit 2933140b5f1f8a6e270460d8918a14943bd0f174 29008Author: Michel Dänzer <michel@tungstengraphics.com> 29009Date: Sun Oct 15 17:05:02 2006 +0200 29010 29011 radeon: Use xf86XVFillKeyHelperDrawable() when available. 29012 29013commit 4edbfb9eb952f9b3f5f597749192c17baf15cdce 29014Author: Michel Dänzer <michel@tungstengraphics.com> 29015Date: Sun Oct 15 17:03:13 2006 +0200 29016 29017 radeon: Track screen damage all the time for page flipping. 29018 29019 This way we don't ever have to forcedly synchronize the pages so we hopefully 29020 never clobber one page's contents when we shouldn't. 29021 29022commit c9c2ac2b42398f5bfad909f3289e2545bb7b3888 29023Author: Michel Dänzer <michel@tungstengraphics.com> 29024Date: Sat Oct 14 15:49:22 2006 +0200 29025 29026 radeon: Don't try to synchronize pages when the CP isn't running. 29027 29028 Also move the DamageEmpty() call into RADEONDRIRefreshArea(). 29029 29030commit c19915a401ccf69046bed03c8d5067f2e72a094e 29031Author: Dave Airlie <airlied@linux.ie> 29032Date: Fri Oct 6 09:55:40 2006 +1000 29033 29034 always turn on DAC for scanning the CRT connect. 29035 29036 Ubuntu 63994 - myself and BenH discussed this as a good idea on irc 29037 29038commit b8547aa89f93a7fefbefb1c0d2743999930ac785 29039Author: Dave Airlie <airlied@linux.ie> 29040Date: Wed Oct 4 01:44:24 2006 +1000 29041 29042 update version number to 6.6.99 in preparation for a 6.7 branch at some stage 29043 29044commit 23a6f97e08fd49e1cae03cd97cae67a5f06b7634 29045Author: Alex Deucher <alex@samba.(none)> 29046Date: Tue Oct 3 13:48:09 2006 -0400 29047 29048 FP timing regs required for both internal and external TMDS 29049 29050commit 829be884702043fe4c9e349488d2585bf0561ae9 29051Author: Alex Deucher <alex@samba.(none)> 29052Date: Tue Oct 3 13:41:21 2006 -0400 29053 29054 remove bogus registers 29055 29056commit cd900580dbc8928ebc28514162b4c1dbfe9fd5ae 29057Author: George Sapountzis <gsap7@yahoo.gr> 29058Date: Tue Oct 3 00:32:22 2006 +0300 29059 29060 [mach64] Fix build for !defined(USE_EXA). 29061 29062commit 22860a3c0685a25f982983654303ad02aedc02ec 29063Author: George Sapountzis <gsap7@yahoo.gr> 29064Date: Sun Jul 16 19:23:56 2006 +0300 29065 29066 Bug 6242: [mach64] Use private DMA buffers. 29067 29068 Map the DMA buffers read-only. This eliminates a security problem where a 29069 client can alter the contents of the DMA buffer after submission to the DRM. 29070 29071commit cb53fe7c1e45937746e43437ae6adb0355306ae9 29072Author: Alex Deucher <alex@t41p.(none)> 29073Date: Sun Oct 1 18:46:05 2006 -0400 29074 29075 fix LVDS plus internal TMDS 29076 29077commit 145c716de6d42778e447581578f3e4db13c9a653 29078Author: David Woodhouse <dwmw2@infradead.org> 29079Date: Sun Oct 1 18:22:14 2006 +0200 29080 29081 Fix radeon vs fbdev pitch calculation. 29082 29083 Fetch the correct displayWidth from fbdevhw after setting the mode. 29084 29085commit d43ad88fa3913437f6987ab5ab46a38f0cb555a8 29086Author: Michel Dänzer <michel@tungstengraphics.com> 29087Date: Sat Sep 30 19:25:20 2006 +0200 29088 29089 radeon: Add support for page flipping with EXA. 29090 29091 Also use the damage layer directly instead of via shadowfb and blit dirty 29092 rectangles to the second page in LeaveServer in order to try and improve the 29093 tradeoff between performance and correctness. 29094 29095commit 146ea328b7962f56c6ab47fbe2adc7561b36189c 29096Author: Michel Dänzer <michel@tungstengraphics.com> 29097Date: Sat Sep 30 19:22:26 2006 +0200 29098 29099 radeon: Fix function name typo which could probably lead to an unresolved symbol. 29100 29101 Also remove declaration of an unused variable. 29102 29103commit 6671c1b01bf29d8f1cacf9306ef658b967d8a3cf 29104Author: Michel Dänzer <michel@tungstengraphics.com> 29105Date: Sat Sep 30 18:29:45 2006 +0200 29106 29107 radeon: Always call RADEONGetAccessibleVRAM(). 29108 29109 This fixes issues with Option "UseFBDev" and the new memmap code. 29110 29111 OTOH, there is potential for this to break IGPs, but that's what Bugzilla and 29112 git-bisect are for... 29113 29114commit c979c7f87fc449cc9631820101edff82c3013e66 29115Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 29116Date: Sat Sep 30 15:14:42 2006 +0200 29117 29118 radeon: Check ROM more thoroughly before treating it as an x86 BIOS. 29119 29120 This prevents the ROMs on some Mac cards getting treated as x86 BIOSen, 29121 resulting in garbage values. 29122 29123commit 37c114acbabe35881c39642e293519e2635907aa 29124Author: Michel Dänzer <michel@tungstengraphics.com> 29125Date: Sat Sep 30 15:11:42 2006 +0200 29126 29127 Remove some dead code. 29128 29129 Fix two compile warnings. 29130 29131commit 6d7b0763173f56050411d7bdf9a583b9bfc050c4 29132Author: Adam Jackson <ajax@benzedrine.nwnk.net> 29133Date: Fri Sep 29 18:50:44 2006 -0400 29134 29135 Undo unintended configure.ac change. 29136 29137 Dammit, git, grow a sensible UI already. 29138 29139commit f96ddd32d3a9e64dc0adec5c50386d6db5b182ee 29140Author: Ronald Wahl <rwahl@gmx.de> 29141Date: Fri Sep 29 17:51:42 2006 -0400 29142 29143 Hook up DFP power to DPMS events. 29144 29145commit f6b42149c5a3fee663575fe6bbe78b62695bdbc8 29146Author: Alex Deucher <alex@samba.(none)> 29147Date: Fri Sep 29 00:44:36 2006 -0400 29148 29149 tvdac handling was reversed between r200 and other chips 29150 29151commit c0660495080719c052d6393ede707755929102cd 29152Author: Dave Airlie <airlied@linux.ie> 29153Date: Wed Sep 27 18:12:52 2006 +1000 29154 29155 radeon: check DDC info for both connectors not just the pScrn one. 29156 29157 Otherwise we disable DDC before we've checked the CRT if LVDS is in use 29158 29159commit 5b98545618141e86164f3ce66469f9e16b1cf6d4 29160Author: Dave Airlie <airlied@linux.ie> 29161Date: Wed Sep 27 18:10:30 2006 +1000 29162 29163 radeon: don't move connectors around if LVDS connector is connector 1 29164 29165 On my laptop this allow me to plug or not plug a monitor without affecting 29166 my lvds 29167 29168commit 74c725bd86348b5bf9bba0c50e8a6a7885d7ff7b 29169Author: Dave Airlie <airlied@linux.ie> 29170Date: Mon Sep 25 11:35:32 2006 +1000 29171 29172 radeon: cleanup some bits from superpatch 29173 29174commit 9827afaa40720a58e2fc029c0bf92ad2fe223d11 29175Merge: 10b4b46c ad8259c8 29176Author: Dave Airlie <airlied@linux.ie> 29177Date: Mon Sep 25 11:25:17 2006 +1000 29178 29179 Merge branch 'radeon-sp' 29180 29181commit ad8259c814629e741ed5567923f40879cc0c7051 29182Author: Dave Airlie <airlied@linux.ie> 29183Date: Mon Sep 25 11:25:03 2006 +1000 29184 29185 radeon : add back tv-dac register write 29186 29187commit 10b4b46c16ff3748856b732f9a39de40ba197112 29188Author: Dave Airlie <airlied@linux.ie> 29189Date: Mon Sep 25 09:08:47 2006 +1000 29190 29191 remove all CVS annotations from the ATI driver files 29192 29193commit 976dec04786e9b050945c290840b4cdc978c0b23 29194Author: Dave Airlie <airlied@linux.ie> 29195Date: Mon Sep 25 09:00:21 2006 +1000 29196 29197 radeon: put back in reverseddc code 29198 29199 BenH claims it is needed, agd5f can fix it up later if it isn't. 29200 29201commit 980fb2f6bd7641c8c57769b0c67e3561903e89a9 29202Author: Dave Airlie <airlied@linux.ie> 29203Date: Sat Sep 23 08:31:56 2006 +1000 29204 29205 radeon: add some missing bits of superpatch for crtc registers 29206 29207commit 1ba4f36821e60289cad937abbb0edb273c88436c 29208Author: Dave Airlie <airlied@linux.ie> 29209Date: Sat Sep 23 08:25:42 2006 +1000 29210 29211 radeon: finish radeon_driver.c import for register changes 29212 29213commit 547543bbefe605a453bfa5ae6d063ae02c5f040e 29214Author: Alex Deucher <alexdeucher@gmail.com> 29215Date: Sat Sep 23 08:21:59 2006 +1000 29216 29217 radeon: re-organise FP and CRTC register setting routines 29218 29219commit fd978140bcb7670f28c684c06c2b6c611c26bef4 29220Author: Dave Airlie <airlied@linux.ie> 29221Date: Sat Sep 23 08:09:57 2006 +1000 29222 29223 radeon: initial CRTC register setting alignment 29224 29225commit 406f4911e7ce821002c111d6bcdec35f5b56c943 29226Author: Dave Airlie <airlied@linux.ie> 29227Date: Sat Sep 23 07:59:24 2006 +1000 29228 29229 radeon: setup the radeon init function to use secondary 29230 29231commit 452f0f8079d65679905ed5178a256534ef0db0e4 29232Author: Dave Airlie <airlied@linux.ie> 29233Date: Sat Sep 23 07:41:57 2006 +1000 29234 29235 radeon: bring over final set of changes for radeon_display.c from susperpatch 29236 29237 This brings over Alex's final set of changes... 29238 29239 i'm unsure about the disappearance of reverseddc - alex to review 29240 29241commit a8b9aa69259a0706d0688d2656ece5a292799242 29242Author: Dave Airlie <airlied@linux.ie> 29243Date: Sat Sep 23 07:38:25 2006 +1000 29244 29245 radeon: align tabs and make display detection like superpatch 29246 29247commit 8d4ae98c528a2309dc3c545893dbba2e7f76916c 29248Author: Dave Airlie <airlied@linux.ie> 29249Date: Sat Sep 23 07:26:44 2006 +1000 29250 29251 radeon: align some output with superpatch 29252 29253commit 65b54847c5e79b2ec3cb1a08b265229443c9deb9 29254Author: Dave Airlie <airlied@linux.ie> 29255Date: Sat Sep 23 07:22:31 2006 +1000 29256 29257 radeon: change function name to Displays rather than Monitors 29258 29259commit 7e16d9fff96dcdff8c25dbbe2a51be0023b7985a 29260Author: Dave Airlie <airlied@linux.ie> 29261Date: Sat Sep 23 07:21:15 2006 +1000 29262 29263 radeon: make GetConnectedMonitors a static void 29264 29265commit 6f63b3dc8b3e2299f5f02abc1c565176d2d46cee 29266Author: Dave Airlie <airlied@linux.ie> 29267Date: Sat Sep 23 07:19:43 2006 +1000 29268 29269 radeon: some minor alignments and cleanups for info-> 29270 29271commit c9f3ced3b1c194cacf1ab9436ee6c2d4efd3a8e9 29272Author: Alex Deucher <alexdeucher@gmail.com> 29273Date: Sat Sep 23 07:16:10 2006 +1000 29274 29275 radeon: add reverse display option to switch displays on non-mobility cards 29276 29277commit e752443c540d9a94522050c61ef7c33d912fece5 29278Author: Dave Airlie <airlied@linux.ie> 29279Date: Fri Sep 22 07:10:10 2006 +1000 29280 29281 radeon: flesh out mapcontrollers 29282 29283commit 7a0e4eb47655c4701fc3a4b33b6c8e9f378ccc9e 29284Author: Dave Airlie <airlied@linux.ie> 29285Date: Fri Sep 22 07:05:56 2006 +1000 29286 29287 radeon: add some info to the map controllers 29288 29289commit 75bf69c868be6164d0041a32f52244f1ef83e050 29290Author: Dave Airlie <airlied@linux.ie> 29291Date: Fri Sep 22 07:03:52 2006 +1000 29292 29293 radeon: add simple map controllers call 29294 29295commit 8ba5abd1329347c7fbf4b4d09c95da0d23b4884a 29296Author: Dave Airlie <airlied@linux.ie> 29297Date: Fri Sep 22 06:58:41 2006 +1000 29298 29299 radeon: move comment 29300 29301commit e8d77b7c36fa791a2350a6f4b15728b15d8f3de0 29302Author: Dave Airlie <airlied@linux.ie> 29303Date: Fri Sep 22 06:55:51 2006 +1000 29304 29305 radeon: refactor blank/unblank 29306 29307commit 8534723f5710e2916b4a22efe22a6b7365cc8db6 29308Author: Alex Deucher <alexdeucher@gmail.com> 29309Date: Fri Sep 22 06:36:03 2006 +1000 29310 29311 radeon: make DVI_I default connector 29312 29313commit d7fc79fd9f094ac462de0883af7e6030cb6195a6 29314Author: Dave Airlie <airlied@linux.ie> 29315Date: Fri Sep 22 06:35:34 2006 +1000 29316 29317 radeon: add enable display function 29318 29319commit 4a54886d510f26b29d27e5c9a73647554291b1a6 29320Merge: 70264d64 9fa176c7 29321Author: Dave Airlie <airlied@linux.ie> 29322Date: Fri Sep 22 06:00:04 2006 +1000 29323 29324 Merge branch 'master' into radeon-spm1 29325 29326 Conflicts: 29327 29328 src/radeon_driver.c 29329 29330commit 9fa176c7989030e7340cc9b85d0b6a7b34303091 29331Author: Alex Deucher <alex@samba.(none)> 29332Date: Tue Sep 19 13:58:17 2006 -0400 29333 29334 Provide all resolutions (instead only the largest one) in MergedFB clone 29335 mode without config file (Henry Zhao) 29336 29337commit e742aeb28c7d9d6e75932c408bcc7c44af52e303 29338Author: Alex Deucher <alex@samba.(none)> 29339Date: Tue Sep 19 13:45:33 2006 -0400 29340 29341 fix large negative refresh rates with mergedfb (Henry Zhao) 29342 fixes bug 6966 29343 29344commit 9c0c805704f1df1b66adc2a6b9313597ae5b86cc 29345Author: Dave Airlie <airlied@linux.ie> 29346Date: Mon Sep 18 19:41:21 2006 +1000 29347 29348 radeon: fix rn50 memmap 29349 29350 This should fixup the rn50/m6/m7 memory map problem that has been reported 29351 since BenH's memory mapping changes 29352 29353commit 70264d645724718a48ca229fedc04e33681b7d31 29354Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29355Date: Mon Sep 18 09:56:25 2006 +1000 29356 29357 radeon; put back old TV DAC code for now 29358 29359commit 53eb843ad8ec5a292dff0844330ecfa9b7b2a0cb 29360Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29361Date: Mon Sep 18 07:55:21 2006 +1000 29362 29363 radeon: move code around in InitCrtcRegisters 29364 29365commit bcae4352079109a9dd16117bf9ec19dd215180bf 29366Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29367Date: Mon Sep 18 07:33:48 2006 +1000 29368 29369 radeon: cleanup DPMS set on/off code to use new controller structure 29370 29371commit 87622f56b3dd7e2d2f00ad71e5b569eac31fb2b1 29372Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29373Date: Mon Sep 18 07:15:54 2006 +1000 29374 29375 radeon: cleanup DacPowerSet 29376 29377commit a417fa081926b2f84e2cbdc355cc7462d1ebdfad 29378Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29379Date: Mon Sep 18 07:15:41 2006 +1000 29380 29381 radeon: fixup InitFPRegisters function call to not pass info 29382 29383commit 8bff57a938e92dfc8729e40f7aa9ad9a96262131 29384Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29385Date: Mon Sep 18 07:02:35 2006 +1000 29386 29387 radeon: copy of overupdated bios hotkeys code from Alex 29388 29389commit 76352c684d5e840c9f7ac0781e3ecb32951b32db 29390Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29391Date: Mon Sep 18 06:54:28 2006 +1000 29392 29393 radeon: minor cleanup of masking for crtc registers 29394 29395commit a7b97697ed13ac25844890439ffba31524e03a16 29396Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29397Date: Mon Sep 18 06:53:23 2006 +1000 29398 29399 radeon: initialise controller structures 29400 29401commit 362ee3778764633cb4c9388aa329c5b715885038 29402Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29403Date: Mon Sep 18 06:41:14 2006 +1000 29404 29405 radeon: move InitFPRegisters up to align with superpatch 29406 29407commit 11fb9a14741746c752b0001c1b1d39a54e69657a 29408Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29409Date: Mon Sep 18 06:35:50 2006 +1000 29410 29411 radeon: add tv dac adj lookup 29412 29413commit a15ac59c4d1a15380910be2a7438111b09161c81 29414Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29415Date: Mon Sep 18 06:35:28 2006 +1000 29416 29417 radeon: re-add ATI patch for rn50 modes 29418 29419commit d80ea2d67b63ea2f2ce0277952af186b69b19e7d 29420Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29421Date: Mon Sep 18 06:30:30 2006 +1000 29422 29423 radeon: re-add DAC power on/off code around detect 29424 29425commit 0a93936f9f048b8837f1c765fea7ccb61f072697 29426Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29427Date: Mon Sep 18 06:25:15 2006 +1000 29428 29429 radeon: restore secondary head behaviour 29430 29431commit 6da5a0bfdedb2062ec835f45514c22917abdcf4c 29432Author: Alex Deucher <alexdeucher@gmail.com> 29433Date: Mon Sep 18 06:16:58 2006 +1000 29434 29435 radeon: move to using Alex's controller structure 29436 29437commit cd1bc1d1febf0a79d835c7dc8b353123278dec51 29438Author: Alex Deucher <alexdeucher@gmail.com> 29439Date: Mon Sep 18 02:44:09 2006 +1000 29440 29441 radeon: add save/restore of some new registers 29442 29443 move some registers about 29444 29445commit 9b5c4cf8a98b0c20bb821f5003495481fbd0c10c 29446Author: Alex Deucher <alexdeucher@gmail.com> 29447Date: Mon Sep 18 02:28:32 2006 +1000 29448 29449 radeon: add pre init controller function 29450 29451commit 311e776c4429485d29aaf11d350cab8e6fe7564b 29452Author: Alex Deucher <alexdeucher@gmail.com> 29453Date: Mon Sep 18 02:21:49 2006 +1000 29454 29455 radeon: add missing tv-out registers to register list 29456 29457commit 65cb77c79c2c5cb536494200d4fba6815d7e771c 29458Author: Alex Deucher <alexdeucher@gmail.com> 29459Date: Mon Sep 18 02:21:29 2006 +1000 29460 29461 radeon: restructure files along lines of Alex's superpatch 29462 29463commit 25fa71ca0a24190cc8be2e80769443e0e09303b3 29464Author: Henry Zhao <henryz@ati.com> 29465Date: Mon Sep 18 03:15:54 2006 +1000 29466 29467 Power up DAC around check for CRT connectedness 29468 29469commit 39fcac431eba60adf2d4e357cfc811dbae36a4e1 29470Author: Henry Zhao <henryz@ati.com> 29471Date: Mon Sep 18 03:09:43 2006 +1000 29472 29473 Bump up clock rate for ES1000 to allow higher resolutions 29474 29475commit 7d9e2b450f0fa42553156f57572a6dab70a0ea13 29476Author: Dave Airlie <airlied@starflyer-ubuntu.(none)> 29477Date: Tue Sep 12 07:04:23 2006 +1000 29478 29479 radeon: this stops mergedfb working on my ATI M7 development board 29480 29481 I've no idea what this is for, it came from an ATI codedrop and it may 29482 fix some laptop or production system, but it looks to me more like a dirty 29483 hack than a true fix, as I have an M7 which I need to specify CRT for the 29484 second head and it breaks mergedfb for me. 29485 29486commit ad1778eb1eb59de14ec309b99131931cbccf64e9 29487Author: Michel Dänzer <michel@tungstengraphics.com> 29488Date: Sun Sep 10 23:32:44 2006 +0200 29489 29490 Revert "* [ Bug 6966 ] large negative refresh rates reported" 29491 29492 This reverts caaed927a07ffbac68b08246185ef93c1e7bb98c commit. 29493 29494 See https://bugs.freedesktop.org/show_bug.cgi?id=8137 . 29495 29496commit 6e92981b024b59c2237f949321f6b2591dada519 29497Author: Tilman Sauerbeck <tilman@code-monkey.de> 29498Date: Sun Sep 3 21:11:10 2006 +0200 29499 29500 Removed the warning re. the r300 DRI driver status. 29501 29502commit caaed927a07ffbac68b08246185ef93c1e7bb98c 29503Author: Henry Zhao <henryz@localhost.localdomain> 29504Date: Thu Aug 31 13:27:45 2006 -0700 29505 29506 * [ Bug 6966 ] large negative refresh rates reported 29507 * [ Bug 6309 ] radeon/r128 fails to read hsync/vsync rates when rang 29508 * descriptor is missing 29509 * [ Bug 5473 ] Blank screen with Radeon Mobility X700 (Acer Ferrari 4005) 29510 * Reboot should not be needed for external CRT to function when connected 29511 later on Ferrari 4000 29512 * Modes need to be sorted, and sorted correctly 29513 * Provide all resolutions (instead only the largest one) in MergedFB clone 29514 mode without config file 29515 * Bump up clock rate for ES1000 to allow higher resolution 29516 29517commit 6d2028b84ff2b82c849b74776258f8e52c952e36 29518Author: Dave Airlie <airlied@linux.ie> 29519Date: Fri Aug 25 10:05:44 2006 +1000 29520 29521 and of course it takes two commits to do a simple commit 29522 29523commit e90c6feddbcd719fda3bbc3d48fd55045e9d5adc 29524Author: Dave Airlie <airlied@linux.ie> 29525Date: Fri Aug 25 09:59:58 2006 +1000 29526 29527 add changelog hook from nv driver 29528 29529commit 1e06212f792d1ed1f7545e16824b86fbd34fa284 29530Author: Adam Jackson <ajax@benzedrine.nwnk.net> 29531Date: Thu Aug 24 19:52:30 2006 -0400 29532 29533 Enable CPIO in the atimisc driver on ia64 platforms. 29534 29535 This fixes many mach64s, but potentially breaks mach32. Mach64 is widespread 29536 on ia64 machines but PCI mach32's are quite rare on any platform. 29537 29538commit dfa47b09b5b8d7eeaa3f8ae629ebb3c7efb51511 29539Author: Adam Jackson <ajax@benzedrine.nwnk.net> 29540Date: Thu Aug 24 19:50:46 2006 -0400 29541 29542 Disable DRI on xpress 200{,m} since we know it's broken. 29543 29544commit 1dcd9b1f75ba2307beb42556aa7f2f945171d5fb 29545Author: Adam Jackson <ajax@benzedrine.nwnk.net> 29546Date: Thu Aug 24 19:42:26 2006 -0400 29547 29548 Disable DRI on RN50 in a more pleasant way. 29549 29550 Rather than publish a non-existant driver name for RN50, just don't bother 29551 initializing it in the first place. 29552 29553commit 64b18e246586b1043a12f381a9d4c834464ed9b5 29554Author: Dave Airlie <airlied@linux.ie> 29555Date: Fri Aug 25 09:40:04 2006 +1000 29556 29557 update version numbers for release 29558 29559commit be229d97e1f0282f2099c7ace7cb257a2738a58f 29560Author: Dave Airlie <airlied@linux.ie> 29561Date: Fri Aug 25 09:39:46 2006 +1000 29562 29563 make colortiling default to on all cards 29564 29565commit 750ed51bc3d5119e2d2a91a05ae5a971dcd3b88a 29566Author: George Sapountzis <gsap7@yahoo.gr> 29567Date: Mon Aug 7 17:09:02 2006 +0300 29568 29569 [mach64] RENDER support: drop caching DP_WRITE_MASK. 29570 29571 Caching DP_WRITE_MASK causes artifacts, e.g. with evas_xrender_x11_test. 29572 29573commit b2beea2fa8949874d3d57fb9b43fe85cc08a8bff 29574Author: George Sapountzis <gsap7@yahoo.gr> 29575Date: Sat Aug 5 18:36:24 2006 +0300 29576 29577 [mach64] RENDER support: save/restore, cache texture registers. 29578 29579commit 43aaed99950640c3695b3c2b91faabf00c6338a7 29580Author: George Sapountzis <gsap7@yahoo.gr> 29581Date: Sat Aug 12 22:12:07 2006 +0300 29582 29583 [mach64] RENDER support (bug #6877). 29584 29585 RENDER accleration is disabled by default. It is strongly recommended that the 29586 patch from bug #6772 and the 'exa-damagetrack branch' are merged in the xserver 29587 before enabling RENDER acceleration. To enable RENDER acceleration, add the 29588 following in xorg.conf: 29589 29590 Section "Device" 29591 [...] 29592 29593 Option "AccelMethod" "exa" 29594 Option "RenderAccel" "true" 29595 EndSection 29596 29597commit d113f40fd97408661193895ff3aa82262f100143 29598Author: Dave Airlie <airlied@linux.ie> 29599Date: Sat Aug 12 17:16:33 2006 +1000 29600 29601 radeon: make some local data static 29602 29603 This is inspired by a patch in XFree86 from David Dawes, it isn't the 29604 same as we have different gamma curves. 29605 29606commit a1da2b5e0a38feae01107676a76bdc5c185ad34b 29607Author: George Sapountzis <gsap7@yahoo.gr> 29608Date: Sat Aug 5 03:18:34 2006 +0300 29609 29610 [mach64] Halve max w/h for hardware-accelerated operations. 29611 29612 EXA hits these limits for some operations on offscreen pixmaps and GTPRO seems 29613 to support the reduced limits. 29614 29615commit a66c8d2f9855caa10142b39f155dd8232b1f2389 29616Author: George Sapountzis <gsap7@yahoo.gr> 29617Date: Sat Aug 5 03:16:56 2006 +0300 29618 29619 [mach64] EXA: add memcpy-based UTS/DFS. 29620 29621 EXA hits more optimized paths when it does not have to fallback because of 29622 missing UTS/DFS. 29623 29624commit 0fb8503d60cf646c91a806f9ed322f9ceff18baa 29625Author: George Sapountzis <gsap7@yahoo.gr> 29626Date: Sat Aug 5 03:16:15 2006 +0300 29627 29628 [mach64] XAA: minor refactoring of memory manager setup. 29629 29630 - DRI: factor out to separate function and add comments on layout 29631 - no DRI: factor out to separate function, maxScanlines is always equal to 29632 ATIMach64MaxY 29633 29634commit e203d86643d5d70bf18248712d05b72b79aee705 29635Author: George Sapountzis <gsap7@yahoo.gr> 29636Date: Sat Aug 5 03:15:42 2006 +0300 29637 29638 [mach64] EXA support. 29639 29640commit dc1e289a611a17090e6dc7ae8a8d3f26d20df4eb 29641Author: Samuel Thibault <samuel.thibault@ens-lyon.org> 29642Date: Sat Jul 29 01:05:35 2006 +0300 29643 29644 Bug #6623: Fix I2C bus Mach64 initialisation failure. 29645 29646commit b377f403e0cd5253402ef3945e5944d1a2763a94 29647Author: Marc Aurele La France <tsi@xfree86.org> 29648Date: Sat Jul 29 01:03:34 2006 +0300 29649 29650 Fix warnings for Option:1 bit fields. 29651 29652commit 91b8b0ca41ad0b9659f0982a05148cab8558e9d5 29653Author: Marc Aurele La France <tsi@xfree86.org> 29654Date: Sat Jul 29 01:02:48 2006 +0300 29655 29656 Bug #807: copy throttling. 29657 29658 Fix corruption associated with the engine randomly not waiting for a copy 29659 operation to commit its results. 29660 29661commit 7800407a2db4b7307192bf3b43fffb9c4748b9c0 29662Author: Marc Aurele La France <tsi@xfree86.org> 29663Date: Sat Jul 29 01:00:56 2006 +0300 29664 29665 Bug #807: invalidate read-back cache. 29666 29667 Fix corruption when scaling 1-pixel wide pixmaps, in which case the first CPU 29668 read of the framebuffer will return stale data. 29669 29670 This bug always manifests itself in mach64 EXA (to be committed soon), with 29671 this patch mach64 EXA passes rendercheck. 29672 29673commit 49b11c540d37152eee4907ab30353ede01e4fb56 29674Author: George Sapountzis <gsap7@yahoo.gr> 29675Date: Sat Jul 29 00:52:28 2006 +0300 29676 29677 Bug #807: split out TestRegisterCaching block for DP regs. 29678 29679commit 290c5aaa580428bc4748a47e4b9bcc22b90fc8d4 29680Author: George Sapountzis <gsap7@yahoo.gr> 29681Date: Sat Jul 29 00:51:53 2006 +0300 29682 29683 Bug #807: split out TestRegisterCaching block for DP regs. 29684 29685commit fba8c839b47e2c8d6a6d65950c3431ff5b870aa0 29686Author: Michel Dänzer <michel@tungstengraphics.com> 29687Date: Sun Jul 23 18:51:47 2006 +0200 29688 29689 Bug #6819: Fix broken rendering with *8b8g8r8 source formats. 29690 29691 Use the corresponding texture formats with R200 family. 29692 29693 Unfortunately, it looks like the R100 family can't accelerate these source 29694 formats, so we have to fall back for them. 29695 29696commit 60ebe97e2d5d27c2bf4213e201cef74503809cb6 29697Author: Tilman Sauerbeck <tilman@code-monkey.de> 29698Date: Mon Jul 3 19:37:28 2006 +0200 29699 29700 Move variable declarations before statements to be C89 compatible 29701 29702commit 6900ff3c5e574881b1470e2363b200e99ab8bf3e 29703Author: Alex Williamson <alex.williamson@hp.com> 29704Date: Mon Jul 3 13:09:47 2006 +0200 29705 29706 Bug #6872: Disable digital output while reading ROM with Radeon 7000. 29707 29708 This is a workaround for ATI Radeon 7000 erratum F1. 29709 29710commit 920ac1f56fad6f48a30e65836ce054815f270738 29711Author: Michel Dänzer <michel@tungstengraphics.com> 29712Date: Mon Jul 3 01:21:18 2006 +0200 29713 29714 Do even less cache flushing when the previous engine mode is known. 29715 29716 OTOH, flush everything when the previous engine mode is unknown, and mark the 29717 engine mode as unknown in a couple more cases. 29718 29719commit dd0dd4bab76907b4b1b6d814f24e73e57d71f587 29720Author: Eric Anholt <anholt@freebsd.org> 29721Date: Mon Jul 3 01:10:10 2006 +0200 29722 29723 Bug #6762: Remove gratuitous cache flushing. 29724 29725commit 22f165f290f57e638d198f24a37f62fc16b72bae 29726Author: Michel Dänzer <michel@tungstengraphics.com> 29727Date: Mon Jul 3 00:39:45 2006 +0200 29728 29729 Build and warning fixes for !defined(USE_XAA). 29730 29731commit ea5105230c7c3519730b8d3c250e9ba9a4e4374a 29732Author: Michel Dänzer <michel@tungstengraphics.com> 29733Date: Mon Jul 3 00:03:22 2006 +0200 29734 29735 Reserve HW cursor memory statically instead of from EXA offscreen. 29736 29737 As a result, quite a bit of code can be removed that attempted to deal with the 29738 EXA offscreen memory disappearing, with varying success. Because the HW cursor 29739 memory is now always immediately after the front buffer and before the back 29740 buffer, this also fixes bug #6808. 29741 29742 As the HW cursor memory is now reserved statically with EXA and XAA, change 29743 RADEONUseHWCursor() to check info->cursor instead of info->cursor_offset, for 29744 which 0 might become valid with HW cursor enabled in the future. 29745 29746 Change info->cursor_offset from unsigned long to CARD32 as the corresponding 29747 register is 32 bits wide, and to avoid issues with printf. 29748 29749commit 66586f905857ea5bf37cb5d34949cd450a9ed067 29750Author: Michel Dänzer <michel@tungstengraphics.com> 29751Date: Sun Jul 2 23:38:20 2006 +0200 29752 29753 Clean up some redundant code. 29754 29755commit 1159c85861dc507ebf323aca82e32fcc67726420 29756Author: Michel Dänzer <michel@tungstengraphics.com> 29757Date: Sun Jul 2 18:24:28 2006 +0200 29758 29759 Bug #6548: Fix playback of interlaced video with more than 1023 lines. 29760 29761 This will still break with 2048 lines... 29762 29763commit 11bedb7c845fc6074383890d5525315aebe05f2c 29764Author: Michel Dänzer <michel@tungstengraphics.com> 29765Date: Sun Jul 2 17:03:23 2006 +0200 29766 29767 Bug #6939: Disable RENDER acceleration by default on RN50. 29768 29769 Based on patch from Weixing Zhang, but still allows enabling it with Option 29770 "RenderAccel". 29771 29772commit 4a0c22e32901a510c7aad070be25c31519c284e3 29773Author: Weixing Zhang <anzhang@ati.com> 29774Date: Sun Jul 2 16:55:25 2006 +0200 29775 29776 Bug #6939: Disable 3D acceleration on RN50 by setting invalid ClientDriverName. 29777 29778commit 9c4a1e3300d30b23806988aa9328fc6ef92e2663 29779Author: Frank Gevaerts <frank@gevaerts.be> 29780Date: Thu Jun 29 18:21:22 2006 +0200 29781 29782 Fix build for !defined(USE_EXA). 29783 29784commit ee0544b670567491b7770f469a64a6cd85519ba1 29785Author: Michel Dänzer <michel@tungstengraphics.com> 29786Date: Thu Jun 29 18:19:58 2006 +0200 29787 29788 Remove unused variables. 29789 29790commit 15c556f5a8c9069d3dd00ce139d46c39b7298951 29791Author: Michel Dänzer <michel@tungstengraphics.com> 29792Date: Thu Jun 29 18:18:21 2006 +0200 29793 29794 Don't declare unused variables on __powerpc__ and __alpha__. 29795 29796commit 055ac2836ec0afd51cf554d043a59068c36a07d3 29797Author: Michel Dänzer <michel@tungstengraphics.com> 29798Date: Thu Jun 29 18:14:45 2006 +0200 29799 29800 Fix strict aliasing violation flagged by gcc -Wall. 29801 29802commit 4b1904017caa976c138594a86e75feaf470e72b5 29803Author: Dave Airlie <airlied@linux.ie> 29804Date: Mon Jun 26 20:34:34 2006 +1000 29805 29806 radeon: force CP and VIP clocks on some r300 and rv100 chips. 29807 29808 fglrx appears to do this on r300 and Jerome Glisse has spent most of his 29809 life tracking down what caused the 9800 to lockup. This is my attempt to 29810 fix this. Please report if it works for you. 29811 29812 Signed-off-by: Dave Airlie <airlied@linux.ie> 29813 29814commit c5d8c8db3e75976687bfb364f0a7b235c7ae8793 29815Author: Michel Dänzer <michel@tungstengraphics.com> 29816Date: Sat Jun 24 16:01:19 2006 +0200 29817 29818 Bug #7274: Fix corruption with 'small' accelerated DownloadFromScreen transfers. 29819 29820commit f25ba3dc0e7850f0b1b78f54c0e7211bf2e81b3a 29821Author: Tilman Sauerbeck <tilman@code-monkey.de> 29822Date: Fri Jun 16 11:38:18 2006 +0200 29823 29824 Remove obsolete FIXME. 29825 29826commit 764013ec61b9043f1487513658e374569f7154a9 29827Author: Tilman Sauerbeck <tilman@code-monkey.de> 29828Date: Fri Jun 16 10:56:02 2006 +0200 29829 29830 Use AC_CHECK_HEADER() to check for exa.h 29831 29832commit 3c701d49b3eb7adb82267cec53163e475913aca3 29833Author: Tilman Sauerbeck <tilman@code-monkey.de> 29834Date: Fri Jun 16 09:59:29 2006 +0200 29835 29836 Use AC_MSG_RESULT() instead of echo for check results. 29837 29838commit dbef62615f38aebf4cdc579e3b26a7e7c719abc4 29839Author: Tilman Sauerbeck <tilman@code-monkey.de> 29840Date: Fri Jun 16 08:50:49 2006 +0200 29841 29842 conditionally declare pPCI and PciReg, too 29843 29844commit a01c01787cfbe3706ae7747cde1657abd79ecdaa 29845Author: Michel Dänzer <michel@tungstengraphics.com> 29846Date: Wed Jun 21 09:38:47 2006 +0200 29847 29848 Bug #7283: Build fixes for !defined(XF86DRI). 29849 29850 Thanks to Matthieu Herrb for pointing out some of these. 29851 29852commit 79d1c573133e64ddf4446b6bb0eaab4a5a6c73ab 29853Author: Michel Dänzer <michel@tungstengraphics.com> 29854Date: Sun Jun 18 17:39:09 2006 +0200 29855 29856 Implement accelerated EXA DownloadFromScreen hook. 29857 29858 x11perf -getimage numbers only increase by about 20-30% on my PowerBook with an 29859 M9, but by about 100 times(!) with a PCIe X550. I suspect the former could 29860 perform better with PCI as opposed to AGP transfers, which would also remove 29861 the need to disable this by default with AGP. 29862 29863commit 2bcdc7cfd013bd457d35f927390e5cc3544f30f1 29864Author: Michel Dänzer <michel@tungstengraphics.com> 29865Date: Sun Jun 18 16:01:00 2006 +0200 29866 29867 Remove #include <fbdevhw.h>, not used here. 29868 29869commit 25a5ccb63a9e3074b4c8eaf48844877d84ff76bb 29870Author: Michel Dänzer <michel@tungstengraphics.com> 29871Date: Sun Jun 18 15:26:55 2006 +0200 29872 29873 Prefer visuals with stencil when it's hardware acclerated. 29874 29875 Also make the formatting of the if statements consistent. 29876 29877commit 8fcc436f19f9c1371295af2d0c45608fb43c584b 29878Author: Michel Dänzer <michel@tungstengraphics.com> 29879Date: Sun Jun 18 15:16:08 2006 +0200 29880 29881 Take part of the blame for the radeon driver. 29882 29883commit c8366d7b457565a06acdf6b40c5448c7f13f1641 29884Author: Michel Dänzer <michel@tungstengraphics.com> 29885Date: Sun Jun 18 14:24:14 2006 +0200 29886 29887 Add Option "DepthBits". 29888 29889 This allows getting higher depth buffer precision and stencil hardware 29890 acceleration in depth 16 or saving some video RAM at the cost of these features 29891 in depth 24. 29892 29893commit 23a21d00db6f3b88b54c44648f968018a03677d3 29894Author: Tilman Sauerbeck <tilman@code-monkey.de> 29895Date: Fri Jun 16 16:37:07 2006 +1000 29896 29897 [PATCH] move the declaration of pMach64 into the correct ifndef statement 29898 Signed-off-by: Tilman Sauerbeck <tilman@code-monkey.de> 29899 Signed-off-by: Dave Airlie <airlied@linux.ie> 29900 29901commit c6b33db9d2c751e34c2152b0bb7e0e84a812589f 29902Author: Dave Airlie <airlied@linux.ie> 29903Date: Fri Jun 16 16:22:03 2006 +1000 29904 29905 ati: prepare for release 29906 29907 Update some versions for release 29908 29909commit 9c534b5330534fc6cf392936a9b200a5128c787c 29910Author: Michel Dänzer <michel@tungstengraphics.com> 29911Date: Thu Jun 15 16:09:55 2006 +0200 29912 29913 Some more ignores. 29914 29915commit 3d68698647fb17a6f5ae9a225f552cb0130fa946 29916Author: Michel Dänzer <michel@tungstengraphics.com> 29917Date: Thu Jun 15 16:07:15 2006 +0200 29918 29919 Fix semantic glitches pointed out by gcc -Wall. 29920 29921 The radeon driver builds warning-free with gcc -Wall here now. 29922 29923commit ab2b48e0df328fd08a2412911126b3184a601aea 29924Author: Michel Dänzer <michel@tungstengraphics.com> 29925Date: Thu Jun 15 16:06:13 2006 +0200 29926 29927 Remove another unused variable. 29928 29929commit 56bd5a370472663af5607cd327f757367b238db0 29930Author: Michel Dänzer <michel@tungstengraphics.com> 29931Date: Thu Jun 15 15:35:05 2006 +0200 29932 29933 Fix some more cosmetic warnings. 29934 29935commit ec93e08b24f7c0e583970d08bfe18ad142e6b047 29936Merge: 051c49d0 6ca98624 29937Author: Michel Dänzer <michel@tungstengraphics.com> 29938Date: Thu Jun 15 15:03:27 2006 +0200 29939 29940 Merge branch 'master' of ssh+git://git.freedesktop.org/git/xorg/driver/xf86-video-ati 29941 29942commit 6ca986242c80bcfe7ca2d435a772bbdf034892c4 29943Author: Dave Airlie <airlied@linux.ie> 29944Date: Thu Jun 15 21:24:46 2006 +1000 29945 29946 remove unused variable 29947 29948commit 9a5f5bc0d3f5e62b79355993ca6246382394daee 29949Author: Dave Airlie <airlied@linux.ie> 29950Date: Thu Jun 15 21:24:16 2006 +1000 29951 29952 remove all printf specifier warnings on Linux 29953 29954commit 051c49d0f002c72d1d8f15f6df264d8b093d4e95 29955Merge: a840afc7 dfac8191 29956Author: Michel Dänzer <michel@tungstengraphics.com> 29957Date: Thu Jun 15 13:10:44 2006 +0200 29958 29959 Merge branch 'master' of ssh+git://git.freedesktop.org/git/xorg/driver/xf86-video-ati 29960 29961commit dfac8191dbda29ef7f6ce33d7356fee89c17d59b 29962Author: Dave Airlie <airlied@linux.ie> 29963Date: Thu Jun 15 21:05:16 2006 +1000 29964 29965 fix last ati*.[ch] format specifiers 29966 29967commit 33525540f3a9d6760c6edcfe765d2a004a6b894f 29968Author: Dave Airlie <airlied@linux.ie> 29969Date: Thu Jun 15 21:03:01 2006 +1000 29970 29971 fix more printf formats 29972 29973commit 6bcf6f29af24b6725e097dd72af8e8b9f6d0676c 29974Author: Dave Airlie <airlied@linux.ie> 29975Date: Thu Jun 15 20:59:33 2006 +1000 29976 29977 some more unused variable cleanups 29978 29979commit a840afc75a0453b4694e3f987ebdbddc6844a169 29980Merge: c90c49da d511ef48 29981Author: Michel Dänzer <michel@tungstengraphics.com> 29982Date: Thu Jun 15 12:46:58 2006 +0200 29983 29984 Merge branch 'master' of ssh+git://git.freedesktop.org/git/xorg/driver/xf86-video-ati 29985 29986commit d511ef486a15b582e5d1708a17a00701f5cf1ccd 29987Author: Dave Airlie <airlied@linux.ie> 29988Date: Thu Jun 15 20:39:22 2006 +1000 29989 29990 ati/r128: add includes to get C function decls 29991 29992 Also fix a void pointer cast 29993 29994commit f46c3bebe821ee0031770d98d04bd406d1abb9ed 29995Author: Dave Airlie <airlied@linux.ie> 29996Date: Thu Jun 15 20:34:56 2006 +1000 29997 29998 fix some more unused variables and bad parameter passing 29999 30000commit 6c3f81683abab8ff6b8f2fcaf01a2a1a63af5ef6 30001Author: Dave Airlie <airlied@linux.ie> 30002Date: Thu Jun 15 20:25:00 2006 +1000 30003 30004 ati: fix -Wall in atidri.c 30005 30006 atidri.c has some parameter passing bugs, fixed those and cleaned other 30007 warnings. 30008 30009commit 58c6aac0669101481ef6b8217e5cede4573ece42 30010Author: Dave Airlie <airlied@linux.ie> 30011Date: Thu Jun 15 19:58:11 2006 +1000 30012 30013 theatre: fix all -Wall warnings in theatre* 30014 30015 This fixes all the warnings, it doesn't clean up the horrible mess that is 30016 theatre*.[ch]. 30017 30018commit c90c49da3844f93d2d29070b15bb7def72b98b0f 30019Author: Michel Dänzer <michel@tungstengraphics.com> 30020Date: Thu Jun 15 10:29:46 2006 +0200 30021 30022 Collapse .gitignore files into single toplevel file, sorted with LC_ALL=C. 30023 30024 Suggested by Keith Packard. 30025 30026commit b7a3e4c9b33f72ae447f99c0c765f24fce988142 30027Author: Michel Dänzer <michel@tungstengraphics.com> 30028Date: Thu Jun 15 09:45:12 2006 +0200 30029 30030 Fix some implicit function declarations. 30031 30032commit 556ddc9c41020271e92389a3900d097ec443c625 30033Author: Michel Dänzer <michel@tungstengraphics.com> 30034Date: Thu Jun 15 09:30:43 2006 +0200 30035 30036 .cvsignore -> .gitignore 30037 30038commit 4708327969a21fc171dcd80a1b506028cc8e4f90 30039Author: Dave Airlie <airlied@linux.ie> 30040Date: Thu Jun 15 07:37:03 2006 +1000 30041 30042 ati: remove Changelog from ATI driver. 30043 30044commit ecbec8379c24bd6c16ced674d5e6affcd3985e46 30045Author: Michel Daenzer <michel@daenzer.net> 30046Date: Thu Jun 8 17:29:12 2006 +0000 30047 30048 Bug #5678: Fix setting of transfer mode for AGPv3 devices (Tilman 30049 Sauerbeck). 30050 30051commit 874ff105bc63c4fc242fbabaec04fde2cb81fdd4 30052Author: Daniel Stone <daniel@fooishbar.org> 30053Date: Sat Jun 3 10:50:33 2006 +0000 30054 30055 Bug #1889: Stop the CCE before an engine reset. (Conn) 30056 30057commit f9db5595fe4d62c54b3a501b2af64b4e2b7035cd 30058Author: Michel Daenzer <michel@daenzer.net> 30059Date: Sun May 21 14:33:34 2006 +0000 30060 30061 Derive ATI_VERSION_{MAJOR,MINOR,PATCH} from package version in order to 30062 hopefully confuse bug triagers a little less. 30063 30064commit 25ba85cab8b3d831da646fef19f01769410fc958 30065Author: Alex Deucher <agd5f@yahoo.com> 30066Date: Wed May 10 02:46:37 2006 +0000 30067 30068 - remove obsolete option "OverlayOnCRT2" (bug 6488) 30069 30070commit ede75fe2b1850a4d9f03c724b7c1d1ed65e60f2f 30071Author: Michel Daenzer <michel@daenzer.net> 30072Date: Tue May 9 14:40:23 2006 +0000 30073 30074 Drop support for radeon DRM <= 1.7. The 2.4 kernel has version 1.7, and 30075 various problems up to system hangs on X server startup have been 30076 reported with that and DRI enabled. 30077 30078commit bc93a46c4fe499f23b9c0ff52e2e36b4c587f7d6 30079Author: Michel Daenzer <michel@daenzer.net> 30080Date: Mon May 8 09:29:34 2006 +0000 30081 30082 Bugzilla #4640 <https://bugs.freedesktop.org/show_bug.cgi?id=4640> Patch 30083 #5273 <https://bugs.freedesktop.org/attachment.cgi?id=5273>: 30084 Add cases for FOURCC_RGBA32 and FOURCC_RGB24. This prevents clients from 30085 receiving incorrect data sizes with these formats. (Jan Schmidt) 30086 30087commit 2b0cdd9448a24ea067b0d78f319b99c1041df2e0 30088Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30089Date: Wed May 3 04:25:27 2006 +0000 30090 30091 Enable decoding of version 4 TMDS PLL config tables, they actually work, 30092 the problem in the comment only concerned pre-production cards that Hui 30093 tested apparently. Also set back the default value to what it was 30094 initially, fixes some more cards that don't have a proper table (Mac 30095 Mini for example) 30096 30097commit 36c04b88416905b9ae305583e8e34fefdbfafc50 30098Author: Michel Daenzer <michel@daenzer.net> 30099Date: Mon May 1 17:49:41 2006 +0000 30100 30101 Bugzilla #6755 <https://bugs.freedesktop.org/show_bug.cgi?id=6755> Patch 30102 #5536 <https://bugs.freedesktop.org/attachment.cgi?id=5536>: 30103 Change HostDataBlit interface to take dst_offset_pitch and coordinates 30104 instead of just a destination pointer, as the latter is not sufficient 30105 with tiling. Also, use HW clipping to avoid overwriting destination 30106 data outside of the specified width. 30107 Adapt to new HostDataBlit interface. This fixes corruption with 30108 UploadToScreen to the front buffer (from exaPutImage). 30109 30110commit b90d3710e67caad0686efab491c2b76ae877920c 30111Author: Michel Daenzer <michel@daenzer.net> 30112Date: Sat Apr 29 21:30:23 2006 +0000 30113 30114 Bugzilla #1992 <https://bugs.freedesktop.org/show_bug.cgi?id=1992> Patch 30115 #4988 <https://bugs.freedesktop.org/attachment.cgi?id=4988>: Fix Option 30116 "DisplayPriority" "HIGH" for R300 family. 30117 30118commit 30f3ce5c5ba205a23748def11de019ee727897f3 30119Author: Michel Daenzer <michel@daenzer.net> 30120Date: Sat Apr 29 20:34:57 2006 +0000 30121 30122 Properly de-initialize EXA driver data, so EXA can be initialized again in 30123 the next generation. 30124 30125commit d705429c230c9cc1645ae540a51dba62cae9f23e 30126Author: Michel Daenzer <michel@daenzer.net> 30127Date: Sat Apr 29 20:18:48 2006 +0000 30128 30129 Fix tests for (un)forcing HW cursor. The broken tests caused problems when 30130 switching to console with 3D clients running. 30131 30132commit 6142cb079d4a5844e4863c800baf4ae19c636af5 30133Author: Michel Daenzer <michel@daenzer.net> 30134Date: Sat Apr 29 20:12:38 2006 +0000 30135 30136 Prefer double-buffered visuals. As a consequence, the root window's visual 30137 is double-buffered by default, which is desirable in a number of 30138 situations, in particular when running a GLX compositing manager with 30139 AIGLX. 30140 30141commit eb841974d041b19461b7fb726687df8938b2e39b 30142Author: Eric Anholt <anholt@freebsd.org> 30143Date: Fri Apr 28 02:55:08 2006 +0000 30144 30145 Bug #6761: Fix font rendering in non-DRI mode by replacing RB2D_DSTCACHE_* 30146 with RB3D_DSTCACHE_*. The RB2D versions are read-only mirrors of the 30147 RB3D. This is masked when DRI is enabled because the DRM uses the right 30148 registers in its flushing. 30149 30150commit 85e2ee6de6fd347aad66785ec7119f92013cd994 30151Author: Dave Airlie <airlied@linux.ie> 30152Date: Wed Apr 26 08:38:47 2006 +0000 30153 30154 Make the ATI driver use the PCI BAR size to determine the MMIO register 30155 aperture size. Otherwise it tries to overmap things if they are up near 30156 the end of memory. 30157 30158commit 13b19edf5accb7427436d3f0b1c9a9efb4fe98fa 30159Author: Eric Anholt <anholt@freebsd.org> 30160Date: Wed Apr 26 07:26:15 2006 +0000 30161 30162 Add support for some component-alpha operations (InReverse, OutReverse) 30163 necessary for testing upcoming general component alpha acceleration 30164 assistance in EXA. Tested with rendercheck on an r100. 30165 30166commit 44a45bcb82d49349f4fdfae3e4b8f24f98d99b46 30167Author: Eric Anholt <anholt@freebsd.org> 30168Date: Tue Apr 25 23:31:56 2006 +0000 30169 30170 Use a CPPFLAGS with XORG_CFLAGS so that the exa test can actually find its 30171 headers and not fail. Also correct some nearby spelling. 30172 30173commit cc1825b503ce7e785cbb4611fd177477a1081b6e 30174Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30175Date: Mon Apr 24 07:44:52 2006 +0000 30176 30177 Fix enabling of external VGA output on some cards with dual entity setup 30178 instead of MergedFB. This was broken due to bogus manipulations of a 30179 register between the 2 entities. This fix is a big "violent" in that it 30180 always enables the analog output regardless of what is plugged. It 30181 should work though (at the expense of some additional power 30182 consumption). A proper fix should be easy to do once Alex gets his 30183 output mapping rework in. 30184 30185commit a66145bc9d487f7e19237cf292f12750cb959f61 30186Author: Dave Airlie <airlied@linux.ie> 30187Date: Fri Apr 21 00:38:44 2006 +0000 30188 30189 Bugzilla 6672: fix interrupts on radeon since memmap changes 30190 30191commit f1f057df9aaba42986d9e20afe6f609a90318333 30192Author: Dave Airlie <airlied@linux.ie> 30193Date: Fri Apr 21 00:34:43 2006 +0000 30194 30195 Add support for backing up the PCIE GART table on VT switch. This makes 30196 suspend/resume work a lot better on PCIE cards. 30197 30198commit 31396df3b0101571bb85a4a66b67f5a8b0c992a9 30199Author: Adam Jackson <ajax@nwnk.net> 30200Date: Sat Apr 8 01:19:17 2006 +0000 30201 30202 Bump to 6.6.0 and release. This module's versioning is completely screwed, 30203 so until we split out mach64/r128/radeon please just increment the 30204 package number whenever you make significant changes. 30205 30206commit d706f494ed22afd369991e8c75a2d4f5aed15305 30207Author: Aaron Plattner <aplattner@nvidia.com> 30208Date: Fri Apr 7 23:06:02 2006 +0000 30209 30210 Add a DrawablePtr argument to the XV functions to pave the way for 30211 redirected video. 30212 30213commit f94685aec26b65f60c00a4b1bcd8f26bd80eb6e4 30214Author: Eric Anholt <anholt@freebsd.org> 30215Date: Tue Apr 4 07:42:36 2006 +0000 30216 30217 Bug #5632: Fix up ati.4 to explain that it is a wrapper driver for the 30218 other three. While here, touch up radeon.4, and add an authors list 30219 gleaned from checking copyright headers. 30220 30221commit 58c9b1d84f514755945e7a0d25fd5530af34ce95 30222Author: Eric Anholt <anholt@freebsd.org> 30223Date: Sun Apr 2 21:45:12 2006 +0000 30224 30225 Add missing .cvsignore lines. 30226 30227commit f1ce6170ef21e1f9769337c3138cd8d65c33e40a 30228Author: Alex Deucher <agd5f@yahoo.com> 30229Date: Sat Apr 1 23:02:40 2006 +0000 30230 30231 - Fix dpi when switching from clone to dualhead with MergedFB. 30232 - Add ConstantDPI option to force a particlar dpi across mode changes Both 30233 based on Thomas Winischhofer's sis code. 30234 30235commit 607f18cfbfc2a7b5509f8b567bf510fce31b361e 30236Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30237Date: Mon Mar 27 06:12:57 2006 +0000 30238 30239 Fix a regression of my latest fix that broke color tiling on some setups. 30240 RADEONRestoreMemMapRegisters() must not clear CRTC_OFFSET_CNTL and 30241 friends unconditionally but only when there is a map change or it will 30242 clear them after the DRI updates the AGP location, thus turning tiling 30243 back off on the CRTC. Also remove a duplicate call to 30244 RADEONInitCommonRegisters(). 30245 30246commit bf1c4471137f10b7de9d79519ca2240058997e18 30247Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 30248Date: Thu Mar 23 18:04:17 2006 +0000 30249 30250 commit the actual changes instead of just the Changelog... 30251 30252commit aa5fff9bae4818bbf99e8d9caa8f2ce9905294ca 30253Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 30254Date: Thu Mar 23 17:59:25 2006 +0000 30255 30256 Use programmable tap coefficients for (hopefully) better video image 30257 quality when downscaling. Remove unnecessary code programming the tap 30258 coefficients twice. Disable some code which is supposed to improve 30259 filtering quality but causes very visible image quality problems at 30260 least with rv250 when upscaling. Fix potentially broken FOURCC_RGB16T 30261 format (untested if it was broken, untested if it works now...). Change 30262 some magic numbers to (new) macro defines. 30263 30264commit 17c52ff15ebe2cf220d5988ee13b599c41799886 30265Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30266Date: Thu Mar 23 01:37:15 2006 +0000 30267 30268 Call RADEONRestoreCommonRegisters() before RADEONRestoreCrtc2Registers() 30269 instead of after. I can't find any good reason why we did it backward 30270 until now and it definitely causes problems as the workaround for VT 30271 switch in RADEONRestoreCommonRegisters() will break output to the 30272 secondary head in various situations. Also does an unrelated minor 30273 update to a comment. 30274 30275commit 673e4d7501861066fdd0874919a4524b820b944b 30276Author: Kristian Høgsberg <krh@redhat.com> 30277Date: Thu Mar 23 00:06:56 2006 +0000 30278 30279 Add atipciids.h. 30280 30281commit 5dcb4152dca6ac1fe9342f44eb48c0198b15772e 30282Author: Kristian Høgsberg <krh@redhat.com> 30283Date: Wed Mar 22 22:30:14 2006 +0000 30284 30285 Convert use of xf86fopen() and other xf86 wrapped libc symbols to use libc 30286 symbols directly. The xf86* versions aren't supposed to be used 30287 directly. 30288 Drop libc wrapper; don't include xf86_ansic.h and add includes now missing. 30289 30290commit 21acdf371e3cf913c5affbd3e86641cfb115311c 30291Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30292Date: Sat Mar 18 00:08:24 2006 +0000 30293 30294 Clear the offsets when updating the memory map instead of when enabling the 30295 CRTCs and try to make that code more reliable. Doesn't disable the 30296 CRTCs beforehand. 30297 30298commit 323ecb92e40d71c5ef994b41b6d8dedba6dd6203 30299Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 30300Date: Fri Mar 17 03:00:53 2006 +0000 30301 30302 Fix various small cosmetic issues. Change a driver message, get the order 30303 right for requesting drm versions, replace the use of some numbers with 30304 the respective macro defines in radeon_video.c, and add some more macro 30305 defines. None of that really matters. 30306 30307commit dc552edf191e14249421f8b27146874f16b80c1a 30308Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30309Date: Thu Mar 16 21:53:58 2006 +0000 30310 30311 Add missing return statement (the driver would not start randomly), add 30312 RV410 to the list of new generation cards for the memory map setup and 30313 fix a bogus message. 30314 30315commit 901409f47cc990721c1648b1c08e23450e6a4452 30316Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30317Date: Thu Mar 16 04:32:22 2006 +0000 30318 30319 Fix incorrect CRTC2_OFFSET values when using old-style dual head (not 30320 MergedFB). Fixes Xinerama for me. 30321 30322commit 4d6656062129da0489eb4bc898871379ba891d8f 30323Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30324Date: Thu Mar 16 03:09:34 2006 +0000 30325 30326 Add various workarounds that seem to fix some remaning lockup scenarios I'm 30327 experiencing with the driver when setting the memory map. Some of the 30328 magic delays are a bit dodgy but they seem to work, I suppose I can't 30329 do better now without help from ATI. Also removed some really too noisy 30330 debug messages. 30331 30332commit 36799e3ea7ddde5fce528e8bc17ea942933a9111 30333Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30334Date: Wed Mar 15 04:03:37 2006 +0000 30335 30336 Fix various issues with DRI & server recycle by putting DRICloseSreen in 30337 the normal CloseScreen callback chain at a spot that matches the call 30338 to DRIFinishScreenInit in order to unwrap at the right time. 30339 30340commit 50923b6a9325592edc8b02b17646f96b9b8c4a01 30341Author: Luc Verhaegen <libv@skynet.be> 30342Date: Wed Mar 15 00:46:33 2006 +0000 30343 30344 Add some autoconf to keep this driver from building EXA when a slightly 30345 older server with older exa is installed. 30346 The way in which this problem was treated by the actually related 30347 developers explains a lot about this driver. 30348 30349commit 74f4ed4275866475b3c182fae962df2fc0946b85 30350Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30351Date: Mon Mar 13 02:57:17 2006 +0000 30352 30353 Fix typo in the function setting the color space transformation factors for 30354 old radeon 30355 30356commit 65625c6c2053792cd14b54c904bf46f360ef1565 30357Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30358Date: Mon Mar 13 00:09:44 2006 +0000 30359 30360 Earliest fix introduced a bug due to another chicken & egg problem in card 30361 detection. Now split PreInitConfig into PreInitChipType and PreInitVRAM 30362 so that PreInitDRI can be done just in between. Restores proper AGPx4 30363 functionality (when enabled) 30364 30365commit 3a290115af41b0ecb94eb201d7e7abb42ebbfec6 30366Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30367Date: Sun Mar 12 06:53:27 2006 +0000 30368 30369 Stop using xf86PciInfo.h, instead use a local copy of the PCI IDs we need 30370 in atipciids.h so we can update the ATI driver independently of the 30371 server when new chips are added 30372 30373commit d6b8f0a5eeb332563ec3ec7082f03be1797d26d8 30374Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 30375Date: Sat Mar 11 04:19:47 2006 +0000 30376 30377 Add #ifdef XF86DRI guards around references to info->allowPageFlip and 30378 info->depthMoves, since they are only defined if building with DRI. 30379 30380commit 69fe90bc501df9bd9e74b894bfca79a4025725bd 30381Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 30382Date: Fri Mar 10 13:00:49 2006 +0000 30383 30384 Use "FBTexPercent" option for XAA too (only for increasing memory reserved 30385 for textures). 30386 30387commit 8033d7423267fc1b3725df6a077dd910b586daaf 30388Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30389Date: Thu Mar 9 23:26:27 2006 +0000 30390 30391 More memory map fixes !!! This time, we add a way to get the DRM kernel 30392 module version early, in order to properly limit the framebuffer size 30393 when a module that doesn't support the new map is loaded. That will 30394 improve backward compatibility. I also made the minimum required kernel 30395 module version 1.3 which hopefully shouldn't be a problem for anybody, 30396 and cleaned up the ordering in which we test for DRM version to enable 30397 features 30398 30399commit c637939d3c4ce5cdddbc5200d5ffbca4c3927050 30400Author: Michel Daenzer <michel@daenzer.net> 30401Date: Thu Mar 9 15:41:16 2006 +0000 30402 30403 RN50: Skip modes that exceed memory bandwidth. 30404 Bugzilla #5766 <https://bugs.freedesktop.org/show_bug.cgi?id=5766> Patch 30405 #4636 <https://bugs.freedesktop.org/attachment.cgi?id=4636> 30406 - Acknowledge that RN50 only has one CRTC, and use this to distinguish it 30407 from RV100. 30408 - Fix detection of RN50 memory type and bus width. 30409 - Model RN50 memory bandwidth limits by capping the pixel clock range based 30410 on memory clock, bpp and memory bus width. (ATI Technologies Inc.) 30411 30412commit 890b95a3ddad7634f3aea635cb236b221e902acc 30413Author: Eric Anholt <anholt@freebsd.org> 30414Date: Thu Mar 9 06:06:25 2006 +0000 30415 30416 Update drivers that support EXA for the new EXA ABI. This consists of 30417 moving all the accel and card members into the driver ptr, filling in 30418 the exa_major/ exa_minor fields, and always using LoadSubModule so we 30419 can check the module version up front. Only tested on ATI. 30420 30421commit e19ccfa18fee046c54d7a9401e4e0098b2300598 30422Author: Michel Daenzer <michel@daenzer.net> 30423Date: Fri Mar 3 16:41:41 2006 +0000 30424 30425 Add Option "FBTexPercent" to override the amount of video RAM reserved for 30426 OpenGL textures with EXA. 30427 30428commit ca00035eccb6e3495b27f8e64f2d016d8d2a9ce2 30429Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30430Date: Thu Mar 2 02:05:17 2006 +0000 30431 30432 Fix the fix ... I didn't interpret PciInfo->size properly and forgot that 30433 the function works in Kb not bytes... Ooops. 30434 30435commit b2145aea36bb035bff048366c607b967d70fff49 30436Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30437Date: Wed Mar 1 21:35:14 2006 +0000 30438 30439 Extend the alignement workaround to post-rv280 chips as well (thanks Hui) 30440 and also limit the CPU accessible memory to the size of the PCI BAR 30441 size (yeah, it was passed to the driver, I just got blind for a while 30442 it seems) 30443 30444commit 163581a82fe9395672c31b41cd67aa44158f5201 30445Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30446Date: Tue Feb 28 23:34:03 2006 +0000 30447 30448 Workaround for rv280 bug which needs the framebuffer memory mapping to be 30449 aligned to its own size. Also add back limit of 128Mb of mapped fb 30450 space for now until I fully sort out what's up with dual function 30451 cards. I would really need to access the PCI BAR sizes here but that 30452 information seem to be missing from the PCI info passed to the driver. 30453 30454commit 5ff8fb2bea1fe7d14db53c19ff3a3a56668ec417 30455Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 30456Date: Tue Feb 28 00:52:53 2006 +0000 30457 30458 Add the newly added (and tons of previously forgotten ones) pci ids to the 30459 ati wrapper too (probably not required for it to work as everything 30460 should default to radeon anyway there). 30461 30462commit 1574b41a0f894f9b6e30d3ee94e596695a4c64f0 30463Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30464Date: Sun Feb 26 10:01:32 2006 +0000 30465 30466 Fix page flipping with XAA. The new code ended up calling ShadowFBInit() 30467 before XAAInit(), which breaks ShadowFB. 30468 30469commit 1f43a584a684af8078631425886cce236500e0e6 30470Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 30471Date: Sat Feb 25 01:57:05 2006 +0000 30472 30473 Add pci ids known to exist (see #4284 for instance). There are still 30474 entries which probably don't really exist (cancelled cards and such), 30475 leave them as-is. Fix the name of some entries, mostly based on the 30476 .inf file of the newest catalyst driver. Use own family id for rv410 30477 and rs400, though there is no different code (yet?). 30478 30479commit 5c141bb15d1163e04c012a0cdf0699d534f0be37 30480Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30481Date: Thu Feb 16 23:27:44 2006 +0000 30482 30483 Memory map and misc fixes. This completely reworks the way the driver 30484 detects available & accessible video memory and initializes the card 30485 internal memory map. Along the way, I re-ordered some of the stuffs in 30486 pre-init in order to remove some dodgy codepath, fixed various init, 30487 exit and vt switch issues, and overall made the driver more robust. The 30488 full memory map fix requires an updated DRM. The X driver no longer 30489 tries to position the AGP aperture, this becomes entirely DRM 30490 responsibility. 30491 30492commit 8c6e5a6eb50d371fd4ec700c26aa628cb5f45731 30493Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30494Date: Wed Feb 15 23:36:55 2006 +0000 30495 30496 The engine setup would trigger a bogus line write before the proper 30497 addresses are setup, thus causing the card to try to bus master over 30498 the system bus to some stale location (usually 0 but could depend on 30499 whatever was used before X). Remove that useless init bit, and always 30500 setup some good enough engine source & destination pointers in case we 30501 have some other spurrious engine activity going on. 30502 30503commit d4ec0223029d8fe03a836a5a6e9a7cc8897b4c90 30504Author: Benjamin Herrenschmidt <benh@kernel.crashing.org> 30505Date: Mon Feb 13 05:30:13 2006 +0000 30506 30507 Fix EXA acceleration of textures with byteswap or tiling From John Clemens 30508 <john@deater.net> 30509 30510commit 79c60ebbc34aaf0ac0b0723898d5b63015547178 30511Author: Eric Anholt <anholt@freebsd.org> 30512Date: Tue Feb 7 04:39:59 2006 +0000 30513 30514 Fix a copy'n'paste-o in the EXA render implementation that resulted in 30515 [ax]8b8g8r8 pict formats not being supported on r200, but no other 30516 issues. 30517 30518commit 8d74f3d49e73626f5d658ba6586a7c1487dae141 30519Author: Daniel Stone <daniel@fooishbar.org> 30520Date: Thu Jan 19 14:20:37 2006 +0000 30521 30522 Bug #5656: Set all Radeon XPRESS 200 R[CS]4xx-based chips to have IsIGP = 30523 1, fixing modesetting issues. 30524 30525commit e490f4b86e6bba70f7c884fca3cf41d159323689 30526Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 30527Date: Fri Jan 13 01:37:43 2006 +0000 30528 30529 XF86_VERSION_CURRENT -> XORG_VERSION_CURRENT 30530 30531commit ac296d9137b02753bc79a9930d168855adea9c95 30532Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 30533Date: Fri Jan 13 01:02:01 2006 +0000 30534 30535 When failing to parse MetaModes, reset CRT2pScrn to NULL after free'ing it 30536 so we don't crash when trying to use it later (such as in 30537 RADEONFreeRec). 30538 30539commit dd98ce78407c9eddd7adaad1d73d6910e6949b87 30540Author: Adam Jackson <ajax@nwnk.net> 30541Date: Sun Jan 8 19:02:16 2006 +0000 30542 30543 Bug #5523: Make radeon compile without USE_XAA. (Tilman Sauerbeck) 30544 30545commit a99429b2bc6407ff2a094e6ee30ce2032f734966 30546Author: Kevin E Martin <kem@kem.org> 30547Date: Wed Dec 21 02:29:59 2005 +0000 30548 30549 Update package version for X11R7 release. 30550 30551commit 96e225b950c74e64f3f788be8a8489c5075ea55a 30552Author: Adam Jackson <ajax@nwnk.net> 30553Date: Mon Dec 19 16:25:52 2005 +0000 30554 30555 Stub COPYING files 30556 30557commit 6bea785ea0bdf01673a6ddc45765ae7372fdd371 30558Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 30559Date: Mon Dec 19 09:07:50 2005 +0000 30560 30561 Typo fixes, mailing list & url updates, and other changes to prepare for 30562 X11R6.9 & 7.0 releases. 30563 30564commit 07d5cf8fc37eb72dbb344081ff07d59dbf568e9c 30565Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 30566Date: Thu Dec 15 23:15:37 2005 +0000 30567 30568 Bugzilla #5238 <https://bugs.freedesktop.org/show_bug.cgi?id=5238> Patch 30569 #3978 <https://bugs.freedesktop.org/attachment.cgi?id=3978> r128 man 30570 page dualhead note update (Alex Deucher) 30571 30572commit 6db5cbfbf6d405d9871892d1255dd7442e4539e9 30573Author: Kevin E Martin <kem@kem.org> 30574Date: Thu Dec 15 00:21:32 2005 +0000 30575 30576 Update package version number for final X11R7 release candidate. Bump 30577 driver version number. 30578 30579commit 726fd3a81dfbd1b0c8e4ede104a48523daecab92 30580Author: Adam Jackson <ajax@nwnk.net> 30581Date: Wed Dec 14 19:41:28 2005 +0000 30582 30583 Bug #1760: Fix ati wrapper multihead to be dlloader friendly. 30584 30585commit 328a2d74dabe32ce35e879799412666feedb0eef 30586Author: Adam Jackson <ajax@nwnk.net> 30587Date: Fri Dec 9 18:25:02 2005 +0000 30588 30589 Bug #3869: Don\'t walk off the end of an array in connector detection. 30590 (Eric Anholt) 30591 30592commit 03961a05438e916b6725f4e547f23a84ead13c5e 30593Author: Adam Jackson <ajax@nwnk.net> 30594Date: Fri Dec 9 18:15:31 2005 +0000 30595 30596 Bug #1106: Fix memory allocation for Render. (Michel Daenzer) 30597 30598commit 091e5c584124f9b849c0f46802b2f0d8b0cc2388 30599Author: Kevin E Martin <kem@kem.org> 30600Date: Thu Dec 8 17:55:18 2005 +0000 30601 30602 Add configure options to allow hard-coded paths to be changed. 30603 30604commit e38e2df15c9a4b08dee6386a225e98827bb35aad 30605Author: Kevin E Martin <kem@kem.org> 30606Date: Thu Dec 8 17:54:40 2005 +0000 30607 30608 Allow hard-coded paths to be configurable. 30609 30610commit b8203604bf9bd3813668313a1c8c17040b9d93ce 30611Author: Kevin E Martin <kem@kem.org> 30612Date: Tue Dec 6 22:48:33 2005 +0000 30613 30614 Change *man_SOURCES ==> *man_PRE to fix autotools warnings. 30615 30616commit a86920e812bfd40ec9facfc234220e83066dd263 30617Author: Kevin E Martin <kem@kem.org> 30618Date: Sat Dec 3 05:49:32 2005 +0000 30619 30620 Update package version number for X11R7 RC3 release. 30621 30622commit 010adc8988846e6f76f111a9fd32eca021d12fbc 30623Author: Kevin E Martin <kem@kem.org> 30624Date: Sat Dec 3 05:45:17 2005 +0000 30625 30626 Bump version numbers for X11R6.9/X11R7 RC3. 30627 30628commit a5efa6724e7fd8318e262060b084630632dfcb44 30629Author: Kevin E Martin <kem@kem.org> 30630Date: Fri Dec 2 02:16:07 2005 +0000 30631 30632 Remove extraneous AC_MSG_RESULT. 30633 30634commit feeef3cba1f4736a96daf3002ec902ced2874198 30635Author: Adam Jackson <ajax@nwnk.net> 30636Date: Wed Nov 30 18:51:22 2005 +0000 30637 30638 Bump libdrm dep to 2.0. 30639 30640commit 7ca2b3fa30fedd07951f7c6e73ed4d2a6f6606ea 30641Author: Adam Jackson <ajax@nwnk.net> 30642Date: Tue Nov 29 23:29:58 2005 +0000 30643 30644 Only build dlloader modules by default. 30645 30646commit e167e5151e32d54919e96cbfdcf932e16b1cab2a 30647Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 30648Date: Mon Nov 28 22:04:08 2005 +0000 30649 30650 Change *mandir targets to use new *_MAN_DIR variables set by xorg-macros.m4 30651 update to fix bug #5167 (Linux prefers *.1x man pages in man1 subdir) 30652 30653commit 0015b1a007bed61aaf389f8f927e4ae0568d7da7 30654Author: Eric Anholt <anholt@freebsd.org> 30655Date: Mon Nov 21 10:49:09 2005 +0000 30656 30657 Add .cvsignores for drivers. 30658 30659commit 00b8f6ccfbb5163c47e9def0e4ecc6ac7be7e8af 30660Author: Eric Anholt <anholt@freebsd.org> 30661Date: Fri Nov 11 11:05:42 2005 +0000 30662 30663 Bug #1028: Don't (incorrectly) mark the 24/8 z/stencil case as slow, which 30664 resulted in whining at runtime about the visual not being supported. 30665 30666commit 33326847ccc9cf4cd6a23a257e5392e348e6dd22 30667Author: Kevin E Martin <kem@kem.org> 30668Date: Wed Nov 9 21:31:20 2005 +0000 30669 30670 Bump version numbers for X11R6.9/X11R7 RC2. 30671 30672commit 68e11246bd6f6a641807230834dce4e7fd6911af 30673Author: Kevin E Martin <kem@kem.org> 30674Date: Wed Nov 9 21:15:14 2005 +0000 30675 30676 Update package version number for X11R7 RC2 release. 30677 30678commit 1582d3b34643babcbb794d2df8d7c9a755d569c9 30679Author: Matthias Hopf <mhopf@suse.de> 30680Date: Wed Nov 9 17:42:57 2005 +0000 30681 30682 Bug #4998: Add ';' and ',' as valid separators for MetaModes as well. 30683 30684commit 2d22d78d157e07e6ec6a3e2d03fc17f102f6e00e 30685Author: Kevin E Martin <kem@kem.org> 30686Date: Wed Nov 9 07:30:03 2005 +0000 30687 30688 Fix DoLoadableServer NO build. 30689 30690commit 999f67763a5ea9fef7b7b576262f0da079efcfee 30691Author: Michel Daenzer <michel@daenzer.net> 30692Date: Tue Nov 8 16:30:48 2005 +0000 30693 30694 bugzilla #988 (https://bugs.freedesktop.org/show_bug.cgi?id=988) Fix typo 30695 which may or may not have had a negative impact on stability with R300 30696 class cards. 30697 30698commit 5aa401efa0b8d81da418eafeb857142a33f85da9 30699Author: Michel Daenzer <michel@daenzer.net> 30700Date: Tue Nov 8 14:05:00 2005 +0000 30701 30702 bugzilla #4494 (https://bugs.freedesktop.org/show_bug.cgi?id=4494) Call 30703 RADEONDGAInit() even if EXA is enabled; the problematic DGA 30704 acceleration hooks are already disabled in that case. 30705 This fixes DGA issues such as DGA mouse (typically used by games in 30706 fullscreen mode) not working. 30707 Fix suggested by Eric Anholt. 30708 30709commit 345180131a7b2752773dc2a3ee1221ad1e05c815 30710Author: Michel Daenzer <michel@daenzer.net> 30711Date: Tue Nov 8 13:35:23 2005 +0000 30712 30713 bugzilla #4916 (https://bugs.freedesktop.org/show_bug.cgi?id=4916) Fix 30714 segfault with MergedFB and EXA enabled and HW cursor temporarily 30715 disabled. 30716 30717commit 23911421ffa7bdd2e191a72c41bf1f26c92c1a67 30718Author: Michel Daenzer <michel@daenzer.net> 30719Date: Tue Nov 8 09:44:41 2005 +0000 30720 30721 bugzilla #4951 (https://bugs.freedesktop.org/show_bug.cgi?id=4951) 30722 attachment #3689 (http://bugs.freedesktop.org/attachment.cgi?id=3689) 30723 Fix HW cursor getting silently disabled with EXA. 30724 30725commit e1259c0117a190cf4c6e7be07d5aa83e03afb3a0 30726Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 30727Date: Mon Nov 7 19:28:40 2005 +0000 30728 30729 Replaced the variadic macros(gcc) by macros according to C99 standard 30730 30731commit 67bdaae04ab37f38a6d0c8dc4b6328c3ba3b3393 30732Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 30733Date: Sun Nov 6 07:41:29 2005 +0000 30734 30735 Sun bug #6325220: Xorg dumps core in dual monitor configuration with radeon 30736 driver if CRT2Position not specified in xorg.conf 30737 <http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6325220> 30738 (Stuart Krietman - Sun Microsystems) 30739 30740commit 733ebf6e4a7c7b2f1a861077d43da2a853e4c0d7 30741Author: Kevin E Martin <kem@kem.org> 30742Date: Tue Nov 1 15:08:52 2005 +0000 30743 30744 Update pkgcheck depedencies to work with separate build roots. 30745 30746commit dd0eb76337bfdaec8f456380a6ced9e0487e190f 30747Author: Adam Jackson <ajax@nwnk.net> 30748Date: Fri Oct 21 18:40:18 2005 +0000 30749 30750 Bug #4829: Remove some useless DRI symbol checks. 30751 30752commit e6c1859e3a7c84b882e0c30d0783255ed1258d9e 30753Author: Donnie Berkholz <spyderous@gentoo.org> 30754Date: Thu Oct 20 23:01:58 2005 +0000 30755 30756 Bug #4831 <https://bugs.freedesktop.org/show_bug.cgi?id=4831>. Check for 30757 x86_64 in addition to amd64 in host_cpu. 30758 30759commit 33460fb393c9591babcc43b30a73eef3971765c2 30760Author: Kevin E Martin <kem@kem.org> 30761Date: Wed Oct 19 02:48:02 2005 +0000 30762 30763 Update package version number for RC1 release. 30764 30765commit 77d5c3549db4f6a7ff9f969f885d7a4e58ab2b16 30766Author: Kevin E Martin <kem@kem.org> 30767Date: Tue Oct 18 15:46:37 2005 +0000 30768 30769 Do the following to make the drivers pass distcheck: 30770 - Only define EXTRA_DIST when there are actually man pages to be processed 30771 - Add missing man pages to keyboard and ati drivers EXTRA_DIST 30772 30773commit 35e717078c840414b0d6a0e37b3ffed0a7b102aa 30774Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 30775Date: Tue Oct 18 00:01:52 2005 +0000 30776 30777 Use @DRIVER_MAN_SUFFIX@ instead of $(DRIVER_MAN_SUFFIX) in macro 30778 substitutions to work better with BSD make 30779 30780commit ff6e5b0412fe347575992c83bb7cb47b6f96f390 30781Author: Adam Jackson <ajax@nwnk.net> 30782Date: Mon Oct 17 22:57:32 2005 +0000 30783 30784 More 1.7 braindamage: define EXTRA_DIST in terms of @DRIVER_NAME@ instead 30785 of indirectly 30786 30787commit a2e7c55850d383e3af948d445b1645d8a47811cf 30788Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 30789Date: Mon Oct 17 00:09:01 2005 +0000 30790 30791 Use sed & cpp to substitute variables in driver man pages 30792 30793commit 524fb1440760a14c53fb3f238aad6c3cdb7e6a08 30794Author: Dave Airlie <airlied@linux.ie> 30795Date: Sun Oct 9 10:36:50 2005 +0000 30796 30797 Secure the Radeon PCIE GART tables in framebuffer memory by not allowing 30798 the clients to get access to a secure area of framebuffer. 30799 30800commit e526211b86ec5343002978debfcf78f74a10c69e 30801Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk> 30802Date: Tue Oct 4 20:01:03 2005 +0000 30803 30804 Check in generated README files 30805 30806commit 431d0e144ae4c294e1f6110fc3db497ce9bd9cfd 30807Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk> 30808Date: Mon Oct 3 21:22:07 2005 +0000 30809 30810 - For all drivers that have a <driver>.sgml file, add code in their build 30811 system to build the README file at make dist time 30812 - in util/macros/xorg-macros.m4, add a new XORG_CHECK_LINUXDOC macro that 30813 will check if the required tools and files exist, and if so set a 30814 conditional. 30815 - util/modular/symlink.sh 30816 - Link all the <driver>.sgml to xf86-video-<driver>/README.sgml 30817 - Add all the README.<driver> to the list of excluded files 30818 - xc/programs/Xserver/hw/xfree86/doc/sgml/SiS.sgml: Various changes to make 30819 it spew less warnings when the text file is built. 30820 30821commit 9147c20e5ff662ac511a19e7623b406dffee2242 30822Author: Adam Jackson <ajax@nwnk.net> 30823Date: Mon Oct 3 06:14:58 2005 +0000 30824 30825 Bug #4001: Don't crash when PanelSize is unset. (Alan Coopersmith) 30826 30827commit 7e2090e3ebf0bd699d66d336353c097d658a86ae 30828Author: Matthieu Herrb <matthieu.herrb@laas.fr> 30829Date: Sat Oct 1 20:42:06 2005 +0000 30830 30831 missed one missing return value. 30832 30833commit a843e4281134de1545d538cb8970179b1d6977f6 30834Author: Matthieu Herrb <matthieu.herrb@laas.fr> 30835Date: Sat Oct 1 17:09:48 2005 +0000 30836 30837 add missing return value on success in RADEONSetupMemXAA(). 30838 30839commit 98f7d043f894c73fecaf14440c1a24b6c2fc4b06 30840Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk> 30841Date: Wed Sep 28 15:10:38 2005 +0000 30842 30843 Make ati driver distcheck 30844 radeon_commonfuncs.c \ 30845 radeon_exa_render.c \ 30846 radeon_exa_funcs.c \ 30847 30848commit e7b1af5d1677e53820dd1c0926cc8b5a4284901e 30849Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 30850Date: Tue Sep 27 23:55:15 2005 +0000 30851 30852 Move the RADEONEnablePageFlip/RADEONChangeSurfaces calls back to the end of 30853 the function (the code above will cause pixmap cache to be freed and 30854 copied around, and these functions must be called after that otherwise 30855 pixmap damage will occur). May be related to #4440, #4456 (unclear if 30856 these setups involve 3d apps) 30857 30858commit bc3c26fe365347c427e04385c1df95a5eec14f28 30859Author: Eric Anholt <anholt@freebsd.org> 30860Date: Tue Sep 27 23:34:11 2005 +0000 30861 30862 Fix Radeon MMIO Render acceleration with EXA by writing out floating- point 30863 coordinates rather than integers. 30864 30865commit 58dfb95c446b73fbc34784f0894cf89152dab043 30866Author: Adam Jackson <ajax@nwnk.net> 30867Date: Sat Sep 24 18:45:45 2005 +0000 30868 30869 Disable the {Open,Close}FullScreen DRI protocol. Remove empty FullScreen 30870 stubs from drivers, comment the non-empty ones. 30871 30872commit 65b5f0017e29903565aff0b7527d955ed016c9a6 30873Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 30874Date: Tue Sep 20 16:53:34 2005 +0000 30875 30876 Only include radeon_dri.h & radeon_sarea.h if XF86DRI is defined. 30877 Include "compiler.h" to define __inline__ on non-gcc compilers 30878 30879commit 0b7441b19a1a5df91aca6b1a02944afb636bc523 30880Author: Eric Anholt <anholt@freebsd.org> 30881Date: Sun Sep 18 02:32:23 2005 +0000 30882 30883 Break EXA ABI while we still can. Add coordinates to the UploadToScreen 30884 hook so we can upload a subset of a pixmap, and convert the current 30885 drivers to respect that. Use this support to directly UploadToScreen in 30886 exaGlyphs, providing a 47.4% +/-2.4% decrease in wall time for ls -lR 30887 programs/Xserver in an antialiased gnome-terminal on an M6 (n=3, caches 30888 hot). I would have bumped major version, only I can't tell what the 30889 EXA_VERSION_* is supposed to be doing as opposed to the module version. 30890 30891commit 64e7db2698ee48af92594c16d72e33c16372c9d7 30892Author: Eric Anholt <anholt@freebsd.org> 30893Date: Sat Sep 17 20:02:01 2005 +0000 30894 30895 - Don't try to upload 0 byte-per-pixel (PICT_a1) data using 30896 RADEONHostDataBlit. 30897 - Disable the shortcut for switching from 3d to 3d in radeon_exa.c. It 30898 appears that we do need the cache flush here, thought it's not clear 30899 why. Disable the 2d to 2d shortcut while here, since I'm unsure of what 30900 we're doing. Exposed by the following bit: 30901 - Bug #4485: Add a new routine, exaGlyphs, to handle font drawing. Glyphs 30902 were being accumulated in from non-migratable scratch pixmaps, causing 30903 the destination pixmap to move towards screen but the migration 30904 necessary for source never to happen, leading to abysmal performance. 30905 Instead, copy the scratch glyph data into a real pixmap first, then 30906 composite from that into the destination, allowing for migration. time 30907 ls -lR from programs/Xserver showed 26.9% (+/- 6.3%) decrease in wall 30908 time (n=3). 30909 - Create exaDrawableUse* wrapping exaPixmapUse*, but which are aware of 30910 windows needing backing store. Makes migration code prettier, and 30911 ensures that composited windows will be migrated as normal when we turn 30912 off cw for EXA. (issue brought up by keithp) 30913 30914commit 3c71198a088e71a105e274f176c8ceb503615d46 30915Author: Dave Airlie <airlied@linux.ie> 30916Date: Sat Sep 17 07:47:51 2005 +0000 30917 30918 Switch radeon driver to using a card type rather than having a IsPCI flag, 30919 allow for PCI/AGP/PCIE cards. Set PCI gart allocation to only happen on 30920 PCIE cards. Add Radeon PCIE card detection using capability bits. 30921 30922commit 10cf3523f236a017a1ca7e5ee57e02176b924059 30923Author: Eric Anholt <anholt@freebsd.org> 30924Date: Wed Sep 14 22:06:41 2005 +0000 30925 30926 Bug #2573: Remove the disabling of automatic cache flushing on non-r300 30927 cards. It has been implicated in hanging at least r300s and some r100s, 30928 and seemed to be the cause of some bad rendering for me. 30929 30930commit d68894a463a3bb08c97e91ff6fc4b65b29183ff6 30931Author: Eric Anholt <anholt@freebsd.org> 30932Date: Tue Sep 13 22:04:12 2005 +0000 30933 30934 Remove the hardcoded USE_EXA/USE_XAA defines now that they've been 30935 hardcoded in configure.ac (config.h). 30936 30937commit baa9b5b6e19cf9794bf810e3c582bf1d7f808663 30938Author: Dave Airlie <airlied@linux.ie> 30939Date: Mon Sep 12 09:27:07 2005 +0000 30940 30941 Fix Radeon EXA for when RENDER isn't defined 30942 30943commit 8c89610389197dc94d3bceaf44d9d2ea91848f1b 30944Author: Dave Airlie <airlied@linux.ie> 30945Date: Mon Sep 12 09:17:42 2005 +0000 30946 30947 Move GART offset/size outside XAA 30948 30949commit eba44d9747adde8385f2a242dee5a8e73ce5a205 30950Author: Daniel Stone <daniel@fooishbar.org> 30951Date: Mon Sep 12 09:04:36 2005 +0000 30952 30953 Define USE_XAA and USE_EXA for the time being. 30954 30955commit a255e7b0984890daad98b74ace0c572965fe61f5 30956Author: Eric Anholt <anholt@freebsd.org> 30957Date: Sun Sep 11 21:00:44 2005 +0000 30958 30959 Attempt to fix up the modular pieces for EXA support. (untested) 30960 30961commit 3e0d9c945a6a71cc476d27341ef18618529a91a8 30962Author: Eric Anholt <anholt@freebsd.org> 30963Date: Sun Sep 11 20:58:53 2005 +0000 30964 30965 Add support for EXA to the radeon driver. Building EXA and XAA support is 30966 controlled at compile time, plus the runtime option of Option 30967 "AccelMethod" "EXA" or "XAA". The XAA support appears to remain as 30968 before, while the EXA pieces need just a little more polishing. Notable 30969 features: 30970 - Render acceleration working on Radeon 100 and 200-series with DRI on. 30971 - DRI works with EXA Notable issues: 30972 - DGA disabled in the EXA case. 30973 - Backbuffer moves disabled in the EXA case. 30974 - No textured XVideo. 30975 - MMIO render acceleration is close but still has some issues. 30976 - Memory pressure while using Composite is really troublesome with DRI on. 30977 This patch is based on an initial patch by Zack Rusin, with significant 30978 work by Benjamin Herrenschmidt and myself. 30979 30980commit fd62082b68ac3aadd8ffc441352d75d88334904e 30981Author: Dave Airlie <airlied@linux.ie> 30982Date: Sun Sep 11 08:51:38 2005 +0000 30983 30984 Add support for allocating PCI GART table in framebuffer memory. This is 30985 needed to support PCIE Radeons using a new DRM. 30986 30987commit cb63f8d9c0563fb0eff28e2be6d4adf5666540d2 30988Author: Dave Airlie <airlied@linux.ie> 30989Date: Tue Sep 6 10:03:19 2005 +0000 30990 30991 Handle VIP timeouts more gracefully -- impose a hard limit of 10ms on 30992 waiting for VIP_BUSY so we don't lock up hard, spinning. 30993 30994commit dbd83cf18e1bf5e7e31e1ac150227bd27e8b1739 30995Author: Adam Jackson <ajax@nwnk.net> 30996Date: Sun Aug 28 20:51:14 2005 +0000 30997 30998 Bug #4148: Bump required DRM version for r300 to 1.17. 30999 31000commit ce0589ff3b9fcf8c6bab64da4dba40ba540a6a03 31001Author: Adam Jackson <ajax@nwnk.net> 31002Date: Sun Aug 28 18:10:34 2005 +0000 31003 31004 Bug #1109: Fix VGA init on Rage Mobility 7500 (Marc Le France) 31005 31006commit 250c8732a54924eeb459880f2905f8e59d038e60 31007Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 31008Date: Sun Aug 28 18:00:23 2005 +0000 31009 31010 Modified the licens type from GPL to a X/MIT one 31011 31012commit 94c0196652c3cfa0558f61b55f5df9bf50f9ac49 31013Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 31014Date: Fri Aug 26 20:04:01 2005 +0000 31015 31016 #include "extensions/dpms.h" -> <X11/extensions/dpms.h> 31017 31018commit 107e42cabe1cc41fe86b0218c3229b5f34b2df84 31019Author: Daniel Stone <daniel@fooishbar.org> 31020Date: Fri Aug 26 07:00:24 2005 +0000 31021 31022 Fix 'ugly' warning in r128. (Branden Robinson, Debian #069) 31023 31024commit d3f5911c85ccfaaf50c05ce80d56dac13fd15701 31025Author: Daniel Stone <daniel@fooishbar.org> 31026Date: Fri Aug 26 05:30:52 2005 +0000 31027 31028 Only use vgahw symbols if we're building vgahw for this architecture. TODO: 31029 Fix this for the modular tree (maybe an unconditional yes?). Define 31030 that we're using a DPMS server and include DPMS headers. (Debian #024b) 31031 31032commit 879213455740d2d653d38c712275f07610e4dc9c 31033Author: Alex Deucher <agd5f@yahoo.com> 31034Date: Fri Aug 26 02:11:24 2005 +0000 31035 31036 - fix pitch increment (Aapo Tahkola, Roland Scheidegger) 31037 31038commit 6519cc71e4387ef979de90be517145611511ff1a 31039Author: Daniel Stone <daniel@fooishbar.org> 31040Date: Thu Aug 25 03:34:58 2005 +0000 31041 31042 Fix bug in Radeon connector table parsing -- parse all entries, not just 31043 the first. (Ben Herrenschmidt) 31044 31045commit 8734a5d0bafb60f5e38b2cee10587f01fc016379 31046Author: Daniel Stone <daniel@fooishbar.org> 31047Date: Thu Aug 25 03:07:50 2005 +0000 31048 31049 Blacklist Radeon M300 from multimedia I2C code. Handle I2C timeouts more 31050 gracefully -- impose a hard limit of 10ms on waiting for I2C_GO so we 31051 don't lock up hard, spinning. 31052 31053commit 578564b4f7fce1f02d056c2b12ffb7803d268829 31054Author: Adam Jackson <ajax@nwnk.net> 31055Date: Mon Aug 22 03:05:52 2005 +0000 31056 31057 Explicit license statement. 31058 31059commit c9005c91bf6aaf4728a6606284f5d24e0b4a8282 31060Author: Daniel Stone <daniel@fooishbar.org> 31061Date: Thu Aug 18 09:03:46 2005 +0000 31062 31063 Update autogen.sh to one that does objdir != srcdir. 31064 31065commit 508a7421a1386b5b783a867d860b858a55f216d3 31066Author: Alan Hourihane <alanh@fairlite.demon.co.uk> 31067Date: Mon Aug 15 07:30:05 2005 +0000 31068 31069 Egbert's 64bit fixes for mixed 32/64bit clients 31070 31071commit 204290412fbe8b4ac1a0f4f97bc7aec2847a1df7 31072Author: Daniel Stone <daniel@fooishbar.org> 31073Date: Sun Aug 14 16:14:51 2005 +0000 31074 31075 Make DGA, CPIO, non-PCI, and TV-Out support detected per-platform, using 31076 the same semantics as in drivers/ati/Imakefile. 31077 31078commit adb093bdf3cea8108a40bcd72f09924b416de0c1 31079Author: Alex Deucher <agd5f@yahoo.com> 31080Date: Mon Aug 8 23:42:36 2005 +0000 31081 31082 - radeon XPRESS chips are actually RV3xx based 31083 31084commit 9c11018beb186b4b303524685fb62629b9ca863e 31085Author: Vladimir Dergachev <volodya@mindspring.com> 31086Date: Sun Aug 7 14:46:55 2005 +0000 31087 31088 Modified: 31089 programs/Xserver/hw/xfree86/drivers/ati/radeon_driver.c 31090 Update warning about DRI support on R300 and later cards. 31091 31092commit a28b2c76c7df2fe65a93eed34bc85c2cf5afb314 31093Author: Vladimir Dergachev <volodya@mindspring.com> 31094Date: Sun Aug 7 14:40:58 2005 +0000 31095 31096 Modified: 31097 programs/Xserver/hw/xfree86/drivers/ati/radeon_driver.c 31098 Properly configure fbLocation and related registers for R300 cards. (Dave 31099 Airlie) 31100 31101commit 692fb0f8b51abeccaf70b51fc1defc62132b15e9 31102Author: Alex Deucher <agd5f@yahoo.com> 31103Date: Sat Aug 6 17:23:16 2005 +0000 31104 31105 - missing line in mergedfb option parsing. Spotted by Christoph Bauer 31106 31107commit 14e77a39320a7a074b5c2eeeea9750b469abc978 31108Author: Ian Romanick <idr@us.ibm.com> 31109Date: Fri Aug 5 23:04:32 2005 +0000 31110 31111 I have resolved this by making all the "Direct rendering disabled" messages 31112 X_WARNING and all the "Direct rendering enabled" messages X_INFO. 31113 Ideally, we need to factor some of this stuff out to common code. There 31114 is a *LOT* of duplication among the *_dri.c files under 31115 programs/Xserver/hw/xfree86/drivers. 31116 I'm trying to clean up as much low hanging fruit as I can and got tired of 31117 waiting for ajax. :P 31118 31119commit 8e88d36f78c13025847310b411402f1d46a2e0cd 31120Author: Adam Jackson <ajax@nwnk.net> 31121Date: Sun Jul 31 17:19:27 2005 +0000 31122 31123 Post-merge fixups: 31124 - Disable building the gamma driver 31125 - Fix numerous Imakefiles to get header paths and subdirs right 31126 - Symlink client-side GLX lib source from Mesa 31127 - Add new shader subdirs to the Mesa core build 31128 - Tweak the DDX drivers to not ask for "xf86dri.h"; they don't need it, 31129 since they don't generate protocol, and it was breaking the build. 31130 31131commit fb9f712fd1044feb8ce6a99aabaeb5966123e268 31132Author: Kevin E Martin <kem@kem.org> 31133Date: Fri Jul 29 21:22:42 2005 +0000 31134 31135 Various changes preparing packages for RC0: 31136 - Verify and update package version numbers as needed 31137 - Implement versioning scheme 31138 - Change bug address to point to bugzilla bug entry form 31139 - Disable loadable i18n in libX11 by default (use --enable-loadable-i18n to 31140 reenable it) 31141 - Fix makedepend to use pkgconfig and pass distcheck 31142 - Update build script to build macros first 31143 - Update modular Xorg version 31144 31145commit ee45e1fb5da793fa14770b28fd662e69d52b2c6a 31146Author: Michel Daenzer <michel@daenzer.net> 31147Date: Fri Jul 29 19:45:14 2005 +0000 31148 31149 bugzilla #3911 (https://bugs.freedesktop.org/show_bug.cgi?id=3911) 31150 attachment #3191 (http://bugs.freedesktop.org/attachment.cgi?id=3191) 31151 Disable bus mastering while updating MC_FB_LOCATION and friends to 31152 prevent the X server from hanging on startup every now and then under 31153 some circumstances. (ATI Technologies Inc.) 31154 31155commit 44a66ec5a01c3b4dfa046a66b80a3b7e9dfff537 31156Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk> 31157Date: Fri Jul 15 15:02:47 2005 +0000 31158 31159 Install theathre drivers in @moduledir@/multimedia 31160 31161commit dbcf104689ca368523715eb5945bcd800985799b 31162Author: Kevin E Martin <kem@kem.org> 31163Date: Fri Jul 15 03:39:58 2005 +0000 31164 31165 Fix filename typo 31166 31167commit 2256fabf1cd69c950639f236f61e59e68cc8e6fa 31168Author: Kevin E Martin <kem@kem.org> 31169Date: Thu Jul 14 22:30:47 2005 +0000 31170 31171 Fix check for headers. Use {XORG,DRI}_CFLAGS subst instead of replacing 31172 CFLAGS. 31173 31174commit 520317863ffb365eca17c96b4d545adb554bbbd6 31175Author: Kevin E Martin <kem@kem.org> 31176Date: Thu Jul 14 19:35:48 2005 +0000 31177 31178 Make compiling with DRI support configurable/autodetecable. Autodetect DRI 31179 sdk headers. Expand help string for xorg-module-dir option. 31180 31181commit ab5264838ae34bb29791bc08d5a4f853a995faf1 31182Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk> 31183Date: Thu Jul 14 17:10:01 2005 +0000 31184 31185 fuck you cvs 31186 31187commit a608c4422665ce1e933483b7c2a25ab2daae5930 31188Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk> 31189Date: Thu Jul 14 16:03:36 2005 +0000 31190 31191 Try adding build system for ATI driver again. Who knows, maybe CVS works 31192 today. 31193 31194commit d715561d28d82591adf5c7504068c1818eb2245c 31195Author: Adam Jackson <ajax@nwnk.net> 31196Date: Mon Jul 11 02:29:45 2005 +0000 31197 31198 Prep for modular builds by adding guarded #include "config.h" everywhere. 31199 31200commit 5b37b695019b7a4522a96a63f1f6b035160e6a25 31201Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 31202Date: Mon Jul 4 23:44:59 2005 +0000 31203 31204 Bug #2604: Disable workaround for horrible flickering when scrolling 31205 vertically in a virtual screen when color tiling and page flip are both 31206 enabled, as it potentially does more harm than good. The flickering 31207 still needs fixing probably though. 31208 31209commit 08b886850f115f404187acd619aa2c2ca18ae3c7 31210Author: Daniel Stone <daniel@fooishbar.org> 31211Date: Fri Jul 1 22:43:11 2005 +0000 31212 31213 Change all misc.h and os.h references to <X11/foo.h>. 31214 31215commit 1c7c3372776d6d116facabbf055f27af3e96f59f 31216Author: Alex Deucher <agd5f@yahoo.com> 31217Date: Fri Jul 1 04:16:42 2005 +0000 31218 31219 - fix depth tiling for r3/4xx chips (Aapo Tahkola) 31220 - adjust limits for tiled surfaces on r3/4xx 31221 31222commit e8c2574557950bfa150806432d566cd38c47158f 31223Author: Alex Deucher <agd5f@yahoo.com> 31224Date: Fri Jul 1 02:56:04 2005 +0000 31225 31226 - Fix Support for Philips FM1236/F tuner on ATI AIW 9600 XT (Jeff Smith) 31227 Bug 3401 31228 31229commit d258e1c1566d005693260230712c9d78f1447dba 31230Author: Alex Deucher <agd5f@yahoo.com> 31231Date: Thu Jun 30 04:46:11 2005 +0000 31232 31233 - change back part of a comment I seem to have accidently changed. 31234 31235commit 6d9c02d58eae1966e8fa8d7d305849c6e57434ad 31236Author: Alex Deucher <agd5f@yahoo.com> 31237Date: Thu Jun 30 02:02:10 2005 +0000 31238 31239 - small mergedfb fix from Thomas Winischhofer 31240 31241commit 85365b791542cab221c37c9b3b4593ea0c6c9e2d 31242Author: Alex Deucher <agd5f@yahoo.com> 31243Date: Thu Jun 30 01:40:17 2005 +0000 31244 31245 - add support for colortiling on r3/4xx hardware (disabled by default). In 31246 order for pageflipping to work on r3/4xx we will need to update the the 31247 sarea and drm so that the crtc1/2 xytiling regs get updated. Initially 31248 worked out by Aapo Tahkola. 31249 31250commit 3ec99c9cd821aeacdd57a9bbc20c4e4ec4711790 31251Author: Alex Deucher <agd5f@yahoo.com> 31252Date: Thu Jun 30 00:56:01 2005 +0000 31253 31254 - Increase the priority of display 0/1 in the memory controller for r3/4xx 31255 hardware when displaypriority is set to HIGH. Fixes display problems in 31256 high res modes. Originally reported by Aapo Tahkola. 31257 31258commit a7523e6c0be7a600963e9de10cf0060d16117bf6 31259Author: Alex Deucher <agd5f@yahoo.com> 31260Date: Thu Jun 30 00:33:36 2005 +0000 31261 31262 - fix some OUTREGs that should be OUTPLLs in dynamicclocks code 31263 31264commit c7e1d31c6aec3a99061c77c4d6f57e2e874e37f1 31265Author: Adam Jackson <ajax@nwnk.net> 31266Date: Sun Jun 26 04:38:12 2005 +0000 31267 31268 Bug #3628: Fix video driver submodule visibility. 31269 31270commit 1beca962ecee5b5c51eb2fb8451fcc2a2ed4b97c 31271Author: Adam Jackson <ajax@nwnk.net> 31272Date: Sat Jun 25 21:16:53 2005 +0000 31273 31274 Bug #3626: _X_EXPORT tags for video and input drivers. 31275 31276commit 9a4895c4c2fd9614d83adfe3967f4f46dc4dfff6 31277Author: Eric Anholt <anholt@freebsd.org> 31278Date: Wed Jun 22 22:50:49 2005 +0000 31279 31280 Add missing detection of the PCI ID for 0x5b62, Radeon X600 (Vladimir 31281 Kushnir). 31282 31283commit 43e3cc7c21160f4536ed860f97d02545c267dfbc 31284Author: Alex Deucher <agd5f@yahoo.com> 31285Date: Wed Jun 15 23:40:50 2005 +0000 31286 31287 Sync up radeon mergedfb with Thomas' latest sis mergedfb changes: 31288 - non-rectangular mergedfb desktops 31289 - screen offsets 31290 31291commit b67f9f5c875efaa41bded7ed2f38b2a5fbd17597 31292Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 31293Date: Sat Jun 4 20:26:28 2005 +0000 31294 31295 Sync with 6.8.2 branch: Bug #591 31296 (https://bugs.freedesktop.org/show_bug.cgi?id=591) attachment #308 31297 (https://bugs.freedesktop.org/attachment.cgi?id=308): Fix video timing 31298 problems with Sony PCG-C1VN Vaio Picturebook notebook && co. Patch by 31299 Marc Aurele La France 31300 31301commit daae70d9083e9c15358f160f0df3176e344f1cb9 31302Author: Alan Coopersmith <Alan.Coopersmith@sun.com> 31303Date: Sat Jun 4 20:04:36 2005 +0000 31304 31305 Sync with 6.8.2 branch: Bugzilla #1306 31306 (https://bugs.freedesktop.org/show_bug.cgi?id=1306) attachment #1291 31307 (https://bugs.freedesktop.org/attachment.cgi?id=1291): Fix SEGV in 31308 "radeon" driver. Patch by Kevin E. Martin <kem@freedesktop.org> 31309 31310commit c27c193f192b117232dbe39575d3d5ab7c6a1b1e 31311Author: Felix Kuehling <fxkuehl@gmx.de> 31312Date: Fri May 27 22:33:23 2005 +0000 31313 31314 Fix for bugzilla https://bugs.freedesktop.org/show_bug.cgi?id=3414 (patch 31315 https://bugs.freedesktop.org/attachment.cgi?id=2782): Don't use RMX if 31316 we are Dell Server. 31317 31318commit 3b674815e0a6d19b0719e71261c0909b92173aed 31319Author: Adam Jackson <ajax@nwnk.net> 31320Date: Tue May 24 23:26:50 2005 +0000 31321 31322 Typo fix from bug #3163 commit 31323 31324commit 0845d1db801ae74688464bd4c41a7e40e4b50186 31325Author: Egbert Eich <eich@suse.de> 31326Date: Mon May 23 09:50:20 2005 +0000 31327 31328 Explicitely disable DPMS power states during close screen. Fix macro that 31329 takes a value and a mask of bits to modify to be behave evquvalently if 31330 the bits that should remain unchanged are set or unset in the value 31331 (Bugzilla #3369). 31332 31333commit 74b86b036def72e5180a46871b8452ba9eea9bb1 31334Author: Adam Jackson <ajax@nwnk.net> 31335Date: Wed May 18 17:47:35 2005 +0000 31336 31337 Bug #3163: Create new DRIINFO_*_VERSION macros to indicate the version 31338 number of the DRIInfoRec, disambiguating it from the XF86DRI protocol 31339 version number. Modify DRIQueryVersion to return the libdri version 31340 number, which all DDXes were requesting implicitly. Fix the DDXes to 31341 check for the DRIINFO version they were compiled against. 31342 31343commit 0f4a8fc6aa5157a8b48beb49f30511665cfa4b63 31344Author: Matthieu Herrb <matthieu.herrb@laas.fr> 31345Date: Wed Apr 20 21:07:44 2005 +0000 31346 31347 programs/Xserver/hw/xfree86/drivers/ati/theatre.h Build fix for 31348 DoLoadableServer = NO systems. 31349 31350commit 3355d12f832909b05e8b022fad784c9f7af7341f 31351Author: Daniel Stone <daniel@fooishbar.org> 31352Date: Wed Apr 20 12:25:22 2005 +0000 31353 31354 Fix includes right throughout the Xserver tree: 31355 change "foo.h" to <X11/foo.h> for core headers, e.g. X.h, Xpoll.h; 31356 change "foo.h", "extensions/foo.h" and "X11/foo.h" to 31357 <X11/extensions/foo.h> for extension headers, e.g. Xv.h; 31358 change "foo.[ch]" to <X11/Xtrans/foo.[ch]> for Xtrans files. 31359 31360commit 31cb4758eae4703d4de31a0b291847cefa855fc3 31361Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 31362Date: Sun Apr 17 23:28:09 2005 +0000 31363 31364 Added fifo_read and fifo_write in tge generic bus definition 31365 31366commit 2b363e02c7c122f07e6fc406fb8ee52d97b55cad 31367Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 31368Date: Sun Apr 17 23:26:36 2005 +0000 31369 31370 Added new options for the ATI driver: RageTheatreMicrocPath and 31371 RageTheatreMicrocType 31372 31373commit 6286efeba564c63ec198c091e4bba22027586901 31374Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 31375Date: Sun Apr 17 23:23:45 2005 +0000 31376 31377 Addeded support for FI1216ME (PAL) version of the tuner, detection of 31378 UDA1380 sound processor 31379 31380commit 9a63c216e90622cd96d4555155df06dd66c3235b 31381Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 31382Date: Sun Apr 17 23:18:26 2005 +0000 31383 31384 Use the theatre_detect to determine the chip ID and then load the 31385 coresponding theatre module 31386 31387commit 247057762813f12f171b1fa05f31c5ace4c245a7 31388Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 31389Date: Sun Apr 17 23:15:31 2005 +0000 31390 31391 Added functions that access the FIFO on RT200 31392 31393commit bcccf9259784f21d84d52c526aa68df0787118b4 31394Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 31395Date: Sun Apr 17 23:13:45 2005 +0000 31396 31397 Added RT200 registers 31398 31399commit da9d27f8c8cccf2f40a244b25ffe0d3162c41a63 31400Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 31401Date: Sun Apr 17 23:09:28 2005 +0000 31402 31403 This is the theatre chip detection module 31404 31405commit e4623bbce161c41ccac65210fd2539211e8af715 31406Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 31407Date: Sun Apr 17 23:07:54 2005 +0000 31408 31409 Now the detection of the theatre chip is in other module called 31410 theatre_detect 31411 31412commit 405b4dcb067346fce7fd9f6d8a6507a58c740ac7 31413Author: Bogdan Diaconescu <b_diaconescu@yahoo.com> 31414Date: Sun Apr 17 23:06:17 2005 +0000 31415 31416 Added the RageTheatre200 video demodulator support 31417 31418commit 310eb24a45b145efbd47196a1c3ec750cf004154 31419Author: Daniel Stone <daniel@fooishbar.org> 31420Date: Mon Apr 4 23:07:08 2005 +0000 31421 31422 Release SDA/SDL I2C lines from an asserted-low state after a DDC probe, 31423 which caused Apple Studio Display monitors to shut off after a second 31424 or so (Ben Herrenschmidt). 31425 31426commit a416667f980a8679352be897af2fb3215da2d000 31427Author: Alex Deucher <agd5f@yahoo.com> 31428Date: Thu Mar 31 23:18:10 2005 +0000 31429 31430 - Add new Radeon pci ids (ATI devrel), bug 2827 31431 31432commit 2e7365ffb767099faf3f191dc65341dbb3ef6247 31433Author: Vladimir Dergachev <volodya@mindspring.com> 31434Date: Thu Mar 31 18:26:05 2005 +0000 31435 31436 Modified: 31437 programs/Xserver/hw/xfree86/drivers/ati/radeon_video.c 31438 Patch from Antti Ajanki - provide VBI capture support for PAL and SECAM 31439 (You need GATOS km module to actually get at the data, the change 31440 merely configures the hardware to handle it correctly). 31441 31442commit 9a3c1c8e4cba2885b707db0f855e11dc6e640b49 31443Author: Michel Daenzer <michel@daenzer.net> 31444Date: Tue Mar 29 06:14:34 2005 +0000 31445 31446 bugzilla #2844 (https://bugs.freedesktop.org/show_bug.cgi?id=2844) 31447 attachment #2230 (http://bugs.freedesktop.org/attachment.cgi?id=2230) 31448 Wait for idle before uploading HW cursor data to the framebuffer to 31449 avoid lockups when moving the mouse on some machines. 31450 31451commit 97863f112850be5206ea555297354c1c104fcdbe 31452Author: Kevin E Martin <kem@kem.org> 31453Date: Tue Mar 29 03:49:04 2005 +0000 31454 31455 bugzilla #2840 (https://bugs.freedesktop.org/show_bug.cgi?id=2840) 31456 attachment #2227 (https://bugs.freedesktop.org/attachment.cgi?id=2227) 31457 Fix build problems introduced by previous Radeon PLL errata 31458 workarounds. 31459 31460commit 04cbb8d174cc5e3c7ecdd4e171170213ecb3c7ed 31461Author: Dave Airlie <airlied@linux.ie> 31462Date: Sat Mar 26 00:53:01 2005 +0000 31463 31464 bugzilla #2057 (https://bugs.freedesktop.org/show_bug.cgi?id=2057) 31465 attachment #1516 (https://bugs.freedesktop.org/attachment.cgi?id=1516) 31466 Add TVOUT Support for Mach64 (Leif Delgass, fixed up for Xorg by me). 31467 31468commit 5be4bf9000bdf58584a10a6b8e285d0f173304fa 31469Author: Alex Deucher <agd5f@yahoo.com> 31470Date: Thu Mar 24 06:45:52 2005 +0000 31471 31472 - radeon PLL access errata workarounds (Benjamin Herrenschmidt, acked by 31473 Hui Yu) 31474 31475commit 34af3da80304a13b6407dcc0f4db54d6c8bb7a41 31476Author: Michel Daenzer <michel@daenzer.net> 31477Date: Fri Mar 18 22:36:47 2005 +0000 31478 31479 Add support for production version of ATI RN50/ES1000. (ATI Technologies 31480 Inc.) 31481 31482commit b56a5f96048b920a7751318dde211e881f53b2bb 31483Author: Vladimir Dergachev <volodya@mindspring.com> 31484Date: Thu Mar 17 03:36:25 2005 +0000 31485 31486 Modified: 31487 programs/Xserver/hw/xfree86/drivers/ati/radeon_mergedfb.c Back out 31488 RadeonWaitForIdleMMIO that I put in earlier. 31489 From discussion on dri-devel we do not need to be as careful with INREG() 31490 access on Radeons as we are on mach64 chips. 31491 The fix is backed out due to: 31492 it breaks DRI on SMP systems (due to absense of the lock in mouse handling 31493 functions) 31494 I am no longer able to reproduce the lockup I saw before 31495 31496commit 01a9c1157cdfbc5f74b035383e060b8419979563 31497Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 31498Date: Tue Mar 8 23:41:37 2005 +0000 31499 31500 Fix accidentally removed depth pitch adjustment if displayWidth is not a 31501 multiple of 32 pixels (bug #788) 31502 31503commit 9b994b40710103b263364153429ebdb9ea3f58ed 31504Author: Vladimir Dergachev <volodya@mindspring.com> 31505Date: Fri Mar 4 22:21:40 2005 +0000 31506 31507 programs/Xserver/hw/xfree86/drivers/ati/radeon_dri.c Let through all chips 31508 R300 and above - presumably the filtering should take place at the DRM 31509 driver level. Correct cut'n'paste position misprint. Big thanks for 31510 Michel Danzer for pointing out both improvements. 31511 31512commit 7d37a9f1f9fffe98067c3b5bc4749cc799fb0c77 31513Author: Vladimir Dergachev <volodya@mindspring.com> 31514Date: Fri Mar 4 05:48:31 2005 +0000 31515 31516 programs/Xserver/hw/xfree86/drivers/ati/radeon_dri.c Turns out Radeon R420 31517 chips work fine with R300 microcode. Thanks go to Rune Petersen for 31518 careful investigation. 31519 31520commit f1edbb7852d85f660188498f578a0a697c2094fe 31521Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 31522Date: Tue Feb 22 00:39:15 2005 +0000 31523 31524 Fix building radeon driver with BuildXF86DRI set to NO, the breakage was 31525 caused by the tiling support (bug #2481) 31526 31527commit b853dbb50fb7c46fb5a70f349a3f264f0bb94e8c 31528Author: Vladimir Dergachev <volodya@mindspring.com> 31529Date: Sat Feb 19 16:00:12 2005 +0000 31530 31531 Modified: 31532 programs/Xserver/hw/xfree86/drivers/ati/radeon_video.c 31533 Fix an INREG without WaitForIdleMMIO. This one should be unlikely to 31534 trigger, but we should still do this in the name of correctness. 31535 31536commit 84290849b4aee336e9db0779d93fc2d7ed8881b1 31537Author: Vladimir Dergachev <volodya@mindspring.com> 31538Date: Sat Feb 19 00:57:23 2005 +0000 31539 31540 Modified: 31541 programs/Xserver/hw/xfree86/drivers/ati/radeon_mergedfb.c 31542 Add a call to RADEONWaitForIdleMMIO() within RADEONChooseCursorCRTC() as 31543 this function uses OUTREGP() which, internally, calls INREG(). 31544 Thus (due to SilkenMouse) INREG() could be called while CP engine was busy. 31545 I have observed this lockup on RV350 + R300 3d driver and have no idea 31546 why no issues were reported for R200 cards. Perhaps, few of them have 31547 enough memory to run merged fb and 3d simultaneously ? 31548 This fix is the right thing to do, however, we might also need to put fifo 31549 management calls into other cursor handling functions (in particular 31550 those that write cursor image directly to video memory). 31551 31552commit 60070063e2130af0b9d5d34eed7adbae0c0fa67b 31553Author: Michel Daenzer <michel@daenzer.net> 31554Date: Fri Feb 18 19:55:35 2005 +0000 31555 31556 Bug #2576: Add support for ATI RN50/ES1000. (ATI Technologies Inc.) 31557 31558commit 32032cdc78097840b7e2685fdec11bc0816aa0ac 31559Author: Adam Jackson <ajax@nwnk.net> 31560Date: Fri Feb 18 16:42:11 2005 +0000 31561 31562 Bug #2567: Fix reversed memset arguments. (David Krause) 31563 31564commit d5fadd6d86ae2a2f8c1bb55e36ab3d6d49649b46 31565Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 31566Date: Wed Feb 9 02:22:39 2005 +0000 31567 31568 do not enable color tiling on r300 chips and newer, since it doesn't work 31569 (yet) (Bugzilla #2504) 31570 31571commit 5ff35300e66a81ecc17c4e9396eb6d158f5c8e1b 31572Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 31573Date: Wed Feb 2 21:35:30 2005 +0000 31574 31575 fix error with color tiling and hw scrolling, it looks like rv100 doesn't 31576 need the special treatment, after all (reported by Stephane Marchesin) 31577 31578commit 7a9031683914bb1d17d223634e5dbf2e18955651 31579Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 31580Date: Mon Jan 31 20:10:28 2005 +0000 31581 31582 change option EnableColorTiling to ColorTiling 31583 31584commit a2b123f7672ea678290798786ecb408a86d8c683 31585Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 31586Date: Wed Jan 26 18:23:41 2005 +0000 31587 31588 (Stephane Marchesin, me) Add support for color (framebuffer) tiling. 31589 Convert all XAA accel code to use explicit dst and src offset pitch 31590 control to make it work. Use the new drm surface ioctls to set up the 31591 tiled surfaces (if dri is enabled). 31592 31593commit e0df1e41724cd009fe7c833d2f9555f5d87e3202 31594Author: Roland Scheidegger <rscheidegger_lists@hispeed.ch> 31595Date: Wed Jan 26 17:39:56 2005 +0000 31596 31597 fix issues with mergedfb + pageflip (fix up sarea frame values) 31598 31599commit aeb01807263c47f49abaf00031d8505e8ebda82c 31600Author: Egbert Eich <eich@suse.de> 31601Date: Mon Jan 24 15:40:21 2005 +0000 31602 31603 Don't exclude 'RADEONInfoPtr info = ...' declaration even when not building 31604 with DRI support (Bugzilla #2370) 31605 31606commit c6c30cfa354eb456638f10771a7cb02271696af2 31607Author: Vladimir Dergachev <volodya@mindspring.com> 31608Date: Thu Jan 13 17:35:49 2005 +0000 31609 31610 Modified: 31611 programs/Xserver/hw/xfree86/drivers/ati/radeon_driver.c 31612 Some of the newer Radeons (R420 for example) have MM_TABLE's with size 0x33 31613 (instead of 0x0c that we know). Furthermore, they lockup during access 31614 of I2C bus ! 31615 Add a check to validate MM_TABLE. Note: the check is not more stringent as 31616 other card manufacturers (like Asus or Tekram) might make cards with 31617 slightly different MM_TABLE. 31618 31619commit 414c035a965b9db15656058c0e9cfeb3acf0158f 31620Author: Alex Deucher <agd5f@yahoo.com> 31621Date: Fri Dec 24 18:26:19 2004 +0000 31622 31623 - Fix MCLK / SCLK calculation (Benjamin Herrenschmidt) bug 2144 31624 - formatting cleanup 31625 31626commit 7c7b278883676412f5709cf812587b1408628133 31627Author: Alex Deucher <agd5f@yahoo.com> 31628Date: Sun Dec 19 19:49:33 2004 +0000 31629 31630 - Add support for Dualhead on M3/M4 (bug 1760) 31631 31632commit 08a45118b53129f3b9c30004404b906c7d7690be 31633Author: Alex Deucher <agd5f@yahoo.com> 31634Date: Sun Dec 19 19:29:10 2004 +0000 31635 31636 - Make dualhead ram allocation cleaner and related messages clearer 31637 31638commit 97cfd7b1489a9510cd233d6dc7154c210f32b9f5 31639Author: Alex Deucher <agd5f@yahoo.com> 31640Date: Sun Dec 19 06:56:37 2004 +0000 31641 31642 - Add DRI resume support (Micha Feigin) bug 2063 31643 31644commit 5b8577ae3200d967b4316156f16028cbcfd21983 31645Author: Alex Deucher <agd5f@yahoo.com> 31646Date: Sun Dec 19 06:11:06 2004 +0000 31647 31648 - fix warnings (Ronny V. Vindenes) 31649 31650commit 447124a663a24a7fffb0cd241d18dd896a43098b 31651Author: Alex Deucher <agd5f@yahoo.com> 31652Date: Sat Dec 18 21:37:48 2004 +0000 31653 31654 - fix warning and potential error in radeon_driver.c 31655 31656commit ac6166158a58b377e7383950c169956f78b62093 31657Author: Alex Deucher <agd5f@yahoo.com> 31658Date: Fri Dec 17 19:19:32 2004 +0000 31659 31660 - Add VGAAccess option to r128 driver. Based on Benh's radeon patch. r128 31661 should now work on ppcs without usefbdev. To use this on ppc make sure 31662 you set "usefbdev" to false. (bug 2089) 31663 - Apply Benjamin Herrenschmidt's radeon patches (bug 2064) 31664 31665commit ac860ed6bbed84f6fc39a9acf53ee87646cfc8dc 31666Author: Vladimir Dergachev <volodya@mindspring.com> 31667Date: Fri Dec 17 16:50:36 2004 +0000 31668 31669 Modified: 31670 programs/Xserver/hw/xfree86/drivers/ati/radeon_accel.c 31671 programs/Xserver/hw/xfree86/drivers/ati/radeon_dri.c 31672 Move DMA robustness fix into radeon_dri.c::RADEONEnterServer() as per 31673 suggestion by Michel D�nzer. 31674 I could not trigger a lockup, even with r300_demo (possibly it has code 31675 that flushes cache inside ?), so this must be good enough.. 31676 31677commit 3ba5b525c4d6390184fdf79b28c954d5caadc6a4 31678Author: Vladimir Dergachev <volodya@mindspring.com> 31679Date: Thu Dec 16 16:26:56 2004 +0000 31680 31681 Modified: 31682 programs/Xserver/hw/xfree86/drivers/ati/radeon_accel.c 31683 Add DMA robustness fix from Nicolai Haehnle. 31684 TODO: move this fix into DRM driver or other parts of Xserver when R300 31685 support is more stable. 31686 This fix affects only R300 and later variants and results in small slowdown 31687 when using CP acceleration in comparison to the previous version of the 31688 driver. 31689 We are still a lot faster in CP mode than using MMIO register access. 31690 31691commit c7b5f4a9eb49c0edd690c972be1015c98408ffaf 31692Author: Adam Jackson <ajax@nwnk.net> 31693Date: Wed Dec 15 19:35:33 2004 +0000 31694 31695 Bug #2008: Fix gcc4 compile problem. 31696 31697commit 8c1969f22fba2e1c9ed34d5ca77bee3a9f577640 31698Author: Vladimir Dergachev <volodya@mindspring.com> 31699Date: Sun Dec 12 20:42:29 2004 +0000 31700 31701 Modified: 31702 programs/Xserver/hw/xfree86/drivers/ati/radeon_render.c 31703 programs/Xserver/hw/xfree86/drivers/ati/radeon_accel.c 31704 Make DMA blitting code more robust. 31705 31706commit 8a9077bc70fd800983f138a224b95301992a2296 31707Author: Vladimir Dergachev <volodya@mindspring.com> 31708Date: Sun Dec 12 19:28:37 2004 +0000 31709 31710 Modified: 31711 programs/Xserver/hw/xfree86/drivers/ati/radeon_driver.c 31712 programs/Xserver/hw/xfree86/drivers/ati/radeon_dri.c 31713 Enable DRM support permanently, allowing its use for any CHIP_FAMILY > 31714 R300. Fix bug in naming Mesa3d driver (it was defaulting to r200) 31715 31716commit 845df6dd96785b59c44c670009c4fcdb20e147ab 31717Author: Alex Deucher <agd5f@yahoo.com> 31718Date: Sun Dec 12 17:29:45 2004 +0000 31719 31720 - missing diff from agp 8x support 31721 31722commit 36b92808e4ff1311f91543388c18cbe8e9eb4aab 31723Author: Vladimir Dergachev <volodya@mindspring.com> 31724Date: Sun Dec 12 16:05:35 2004 +0000 31725 31726 Modified: 31727 programs/Xserver/hw/xfree86/drivers/ati/radeon_accel.c 31728 programs/Xserver/hw/xfree86/drivers/ati/radeon_driver.c 31729 programs/Xserver/hw/xfree86/drivers/ati/radeon_dri.c 31730 programs/Xserver/hw/xfree86/drivers/ati/radeon_common.h 31731 programs/Xserver/hw/xfree86/drivers/ati/radeon_version.h Add R300 (and 31732 later) DRM support, protected by X_R300_DRM option. 31733 Note: at the moment this is for 2d support only, if enabled running 31734 glxgears locks up my machine, but DMAForXv=yes works fine. 31735 31736commit 68ea67ce98b8f3ad4c17f987e0e2b92f79a96fbf 31737Author: Alex Deucher <agd5f@yahoo.com> 31738Date: Sun Dec 12 02:00:48 2004 +0000 31739 31740 - add AGP 8x support (Hui Yu, Michel Daenzer) bug 1819 31741 31742commit 93675d288d5ebb35d4724ebbfc455a703e25a311 31743Author: Alex Deucher <agd5f@yahoo.com> 31744Date: Sun Dec 12 00:14:32 2004 +0000 31745 31746 - only mess with Dynamicclocks on mobility chips (Benh, me) bug 1912 31747 - fix display detection (Hui Yu) bug 1559 31748 - Add new option "BIOSHotkeys" to allow the bios to control output 31749 switching (off by default) bug 2000 31750 - update the savage man page to document the "CrtOnly" option 31751 31752commit ed217003127028483fd2956851d438977169d154 31753Author: Daniel Stone <daniel@fooishbar.org> 31754Date: Mon Dec 6 13:27:56 2004 +0000 31755 31756 Apply Debian patch to re-POST via VBE if driver thinks it has no video 31757 memory (a symptom of incompletely coming up from D3 power state), which 31758 fixes some Dell laptops (closes: #1890). 31759 31760commit 2b8ab42b05b32710fa4f71a25e2f84192e637fbd 31761Author: Michel Daenzer <michel@daenzer.net> 31762Date: Mon Dec 6 06:27:53 2004 +0000 31763 31764 When direct rendering is enabled, use hostdata blits to transfer data from 31765 system memory to video RAM, which should reduce CPU usage especially 31766 with larger videos. Can be disabled via Option "DMAForXv" if there 31767 should be any stability issues, but it's been stable for me during a 31768 week of testing. Based on a patch by Nikolaus Meine 31769 <meine@tnt.uni-hannover.de>. Probably fix endianness issues in some 31770 newer XVideo code, untested. 31771 Also use hostdata blits to transfer RENDER image data to video RAM to avoid 31772 idling the accelerator engine. Increases RENDER performance 31773 significantly for me. 31774 These changes were only tested on an M9 in a Titanium PowerBook but should 31775 work with all Radeons where direct rendering is supported. 31776 31777commit 960d15ae3a44efe7c02a3d6d0acbabe63e07be74 31778Author: Daniel Stone <daniel@fooishbar.org> 31779Date: Sat Dec 4 22:34:54 2004 +0000 31780 31781 Disable interrupt handling after the DRM handler has been removed, 31782 preventing lockups (closes: #1886). 31783 Allow pitches with a granularity of 32, rather than 64, because some 31784 clamshell iBooks have 800-byte pitches. Change Xv allocation 31785 granularity to 8 bytes, rather than 16 bytes, also (closes: #1888). 31786 31787commit 4706dfb7804a46e3cc5aafc3644f5a1134ab3d19 31788Author: Daniel Stone <daniel@fooishbar.org> 31789Date: Sat Dec 4 22:24:06 2004 +0000 31790 31791 Retry CCE idle commands until the (higher) timeout is reached, and properly 31792 stop the engine if we totally fail -- gets rid of serious performance 31793 issues and lockups with r128 DRI. 31794 31795commit bce508a288d6b95708dbbfc6b235e2d95b6b96ad 31796Author: Markus Kuhn <Markus.Kuhn@cl.cam.ac.uk> 31797Date: Sat Dec 4 00:42:59 2004 +0000 31798 31799 Encoding of numerous files changed to UTF-8 31800 31801commit e9c9fd279ed43cc0d779266efcea69a67541c794 31802Author: Egbert Eich <eich@suse.de> 31803Date: Tue Nov 30 08:38:44 2004 +0000 31804 31805 Make Xorg the default server to install. 31806 Avoid PIO access on IA64. Some IA64 machine check if legacy ports outside 31807 the VGA range are accessed. The ATi driver however does this to probe 31808 for ISA Mach8/32/64. Since no IA64 has ISA slots this restriction 31809 should not be relevant to the user. 31810 Avoid recursive calls of xf86scanpci(). This function normally detects that 31811 it has been called before by checking if the PCI structure is filled 31812 out. So far if this was not the case (because PCI probing has failed 31813 for some reason) the function is traversed again. With the chipset 31814 specific PCI bus probing this can lead to an endless recursive loop as 31815 the post-probing code calls xf86scanpci() from within this function. 31816 The OS specific PCI code for Linux worked only if bus 0 was populated as it 31817 checked for the presence of /proc/bus/pci/00. Fixed to check for 31818 /proc/bus/pci/<bus_to_look_for> instead. 31819 31820commit fffbcb761c178e025e3d55d428a803b1284481e0 31821Author: Matthieu Herrb <matthieu.herrb@laas.fr> 31822Date: Sat Nov 27 15:14:15 2004 +0000 31823 31824 xc/programs/Xserver/hw/xfree86/driver/ati/radeon.man 31825 xc/programs/Xserver/hw/xfree86/driver/voodoo/voodo.man 31826 xc/programs/Xserver/hw/xfree86/driver/wsfb/wsfb.man XF86Config -> 31827 __xconfigfile__ 31828 31829commit ad6347d187f9afa053c755aa3a2da9d821916cb2 31830Author: Daniel Stone <daniel@fooishbar.org> 31831Date: Tue Nov 23 21:27:43 2004 +0000 31832 31833 Use fp_horz_stretch instead of fp_vert_stretch for horizontal stretch 31834 factor calculation in panel size derivation in the Radeon driver 31835 (closes: #1881). 31836 31837commit ac054e427c6ce8cd5b89459dd3f16ca20e6b2d6a 31838Author: Vladimir Dergachev <volodya@mindspring.com> 31839Date: Sun Oct 24 18:17:36 2004 +0000 31840 31841 Add registers to facilitate work with stereo 3d. 31842 31843commit 445ed4b9b61347c866ee67dc0cb110904d667b84 31844Author: Alex Deucher <agd5f@yahoo.com> 31845Date: Fri Oct 22 03:51:58 2004 +0000 31846 31847 - make the overlay work on crtc2 again 31848 - make PutVideo() MergedFB aware 31849 31850commit 54cffc3fe273cf3c9ed2a8aef833d4bd337a8a09 31851Author: Egbert Eich <eich@suse.de> 31852Date: Mon Oct 18 15:19:44 2004 +0000 31853 31854 Fixing one of the infamous 'control reaches end of non void function 31855 problems'. 31856 31857commit 7a97d363327e0b2c8646d961e00d05f99766b54f 31858Author: Egbert Eich <eich@suse.de> 31859Date: Mon Oct 18 14:21:45 2004 +0000 31860 31861 Made handling of DevelDrivers for x86-64 more conformant to other 31862 platforms. 31863 Compress all font encodings (Stefan Dirsch). 31864 Fixed warnings. 31865 Turn on forwarding XNSpotLocation event to XIM server in OffTheSpot and 31866 Root mode (bugzilla #1580, James Su). 31867 Added another compose key combination for the Euro symbol (Stefan Dirsch). 31868 Added support for UTF-8 in ja_JP, ko_KR and zh_TW locales (Mike Fabian). 31869 Changed default encoding for ru from KOI8-R to ISO8859-5 (Mike Fabian). 31870 This is the encoding that is also used by glibc. We may break other 31871 libcs - lets see who complains. 31872 Added explanation for DESTDIR to install to a different directory than /. 31873 Added some early bailouts to atiprobe if PCI structure pointer is NULL to 31874 prevent sig11. 31875 XV support doesn't depend on 2D acceleration any more. This patch removes 31876 this limitation from the NSC driver. This is a patch that I have 31877 committed to XFree86 a while ago but never ported over to X.Org. 31878 Matthieu Herrb contributed some build fixes. 31879 Fixing SetDPMSTimers() so that DPMS*Time == 0 disables the appropriate 31880 timer. This takes advantage of the fact that TimerSet() with a timeout 31881 argument 0 behaves like TimerCanel(). 31882 Use /dev/xconsole (named pipe) or devpts for system logger (Werner Fink). 31883 Create missing links for backward compatibility to XFree86 (Stefan Dirsch). 31884 Changed comment to mention xorg. 31885 Changed cursor for the 'kill' action from XC_draped_box to XC_Pirate. If 31886 you don't like it we can change it back (original author unkown). 31887 Added 'pic' to the man page rendering command pipeline (Werner Fink). 31888 Added missing return value (Stefan Dirsch, Roland Mainz) 31889 31890commit c7ab340fecb94a89c25d6bc7f686b7c97cd3067f 31891Author: Alex Deucher <agd5f@yahoo.com> 31892Date: Sun Oct 17 22:45:57 2004 +0000 31893 31894 - Fix DFP detection in r128 driver (bug 1386, Klaus Kusche) 31895 - Fix EmulateWheelInertia defaults (bug 1115, Andrew Pimlott) 31896 31897commit ace6e1b7933cb415b8efc41dffa335938f9c89b5 31898Author: Vladimir Dergachev <volodya@mindspring.com> 31899Date: Sun Oct 17 17:57:33 2004 +0000 31900 31901 Modified: 31902 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_driver.c On my Inspiron 31903 8600 both internal and external ports are marked DAC_PRIMARY in BIOS. 31904 So be extra careful - only swap when the first port is not DAC_PRIMARY 31905 31906commit 22405e81b812c239b5d8653b402b5f846e6e10b1 31907Author: Alex Deucher <agd5f@yahoo.com> 31908Date: Sun Oct 17 16:52:02 2004 +0000 31909 31910 - Add Thomas' latest sis MergedFB enhancements to the radeon driver 31911 31912commit 186efa31b2cb7285717678bb1c4df4cd9e182395 31913Author: Vladimir Dergachev <volodya@mindspring.com> 31914Date: Wed Oct 13 23:02:42 2004 +0000 31915 31916 Modified: 31917 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_video.c 31918 xc/programs/Xserver/hw/xfree86/drivers/i2c/fi1236.h Squash annoying warning 31919 about fi1236_dump_status 31920 31921commit 43b24861e36023613a99627c57ac44721ed1a548 31922Author: Egbert Eich <eich@suse.de> 31923Date: Tue Oct 12 19:13:43 2004 +0000 31924 31925 Set fbdev mode as the default mode on PPC (Olaf Hering). 31926 Added support for IBM space saver keyboard (Stefan Dirsch). Added support 31927 for Cherry CyMotion Master XPress (Marcus Schaefer). 31928 Change order of SetDisplayDevice(), HWRestore(), UnbindGART() and 31929 RestoreBIOSMemSize() to be exactly opposite to the Save procedure in 31930 EnterVT() (Matthias Hopf, Alan Hourihane). 31931 Fix text mode restauration by removing the assumption that the register 31932 which determines which head is programmed is set. to the active head by 31933 the BIOS (Mark Vojkovich). 31934 When I wrote the resource code 5 years ago I made some assumptions which 31935 turned out to be false: I've assumed that the bus number of the PCI 31936 hostbridge would be the PCI bus the bridge links to. This is not 31937 correct. Fixing this assumption is not easy. However I hope that the 31938 attached patch will make the system work regardless as it 'ignores' 31939 host bridges from which the target bus is not known. This should not 31940 matter at all as we really don't care about host bridges (unless we 31941 have bridge specific code which retrieves information about the 31942 bridge). 31943 Fixed server crash on reset when a structure allocated in PreInit() was 31944 freed on CloseScreen(). 31945 Fixed ring buffer lock ups that happened because the structure that 31946 contained ringbuffer data was not zeroed after allocation. 31947 Fixed numerous warnings due to signed unsigned comparisons. 31948 programs/Xserver/hw/xfree86/drivers/nv/nv_driver.c: 31949 (NVBacklightEnable): Changed the order in which the sequencer registers 31950 and the backlight control registers are written. The sequencer control 31951 register need to be written first otherwise DPMS screen blanking 31952 produces vertical bars on a mobile device. lib/Xau/Imakefile: Build 31953 libXau static library PIC so it can be linked into toolkits that would 31954 like to wrap its functionality. 31955 31956commit 515b7c4b748f70c17aee149a33ee1b41632daa1a 31957Author: Alex Deucher <agd5f@yahoo.com> 31958Date: Mon Oct 11 23:20:26 2004 +0000 31959 31960 - fix resume on radeons (bug 1220) 31961 - clean up some old cruft in the savage i2c code 31962 31963commit 07e1aaee862c4835601c704e14fcb12e28f4e78e 31964Author: Egbert Eich <eich@suse.de> 31965Date: Mon Oct 11 09:58:04 2004 +0000 31966 31967 Improving DPMS handling on VT swich and server termination/abort: previous 31968 version called the driver directly and too late. 31969 Unblank secondary screen explicitely. Don't rely on the value read during 31970 register save as the BIOS have blanked the secondary head. 31971 Checking if server isn't switched away before calling sync. Sanity check 31972 for possible bugs in aother areas of the code. 31973 Fixing default amount of of allocated video memory from AGP for i810: Use 31974 16MB if less than 192MB are installed else use 24MB (Matthias Hopf). 31975 31976commit a8c77480e7f4ddf8c45bea3ecec7b6564c0d6452 31977Author: Vladimir Dergachev <volodya@mindspring.com> 31978Date: Fri Oct 8 00:56:05 2004 +0000 31979 31980 Modified: 31981 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_video.c Remove redundant 31982 and incorrect line. 31983 31984commit ec831d4fba23dc5f4d0a1450aa63efe5ff5937a2 31985Author: Vladimir Dergachev <volodya@mindspring.com> 31986Date: Tue Oct 5 01:16:01 2004 +0000 31987 31988 Modified: 31989 xc/programs/Xserver/hw/xfree86/drivers/ati/theatre.h Patch from Matthieu 31990 Herrb - fix compilation with DoLoadableServer=NO 31991 31992commit f1dadde8f642b332a2dcd2e2a5779f25fec7a3de 31993Author: Adam Jackson <ajax@nwnk.net> 31994Date: Mon Oct 4 21:52:47 2004 +0000 31995 31996 Correct radeon man page, IGP is supported for 3D but R400 is not. 31997 31998commit e3dba9841521466629828501c4fdccbba3d6af89 31999Author: Vladimir Dergachev <volodya@mindspring.com> 32000Date: Mon Oct 4 15:24:34 2004 +0000 32001 32002 Modified: 32003 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_video.c Do not probe for 32004 Rage Theatre if no multimedia table is present and no parameters have 32005 been specified. 32006 32007commit 4aadb784928ebbd60dce172c9a9a80e8d84e943f 32008Author: Vladimir Dergachev <volodya@mindspring.com> 32009Date: Mon Oct 4 14:53:11 2004 +0000 32010 32011 Modified: 32012 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_video.c This is really 32013 embarassing - I found a part of code that had <cr> at the end of the 32014 lines. I am certain I never saw it before - talk about code rot ! 32015 32016commit db5a27081d9d3ab4cc1481a7d110d11103b3de19 32017Author: Vladimir Dergachev <volodya@mindspring.com> 32018Date: Sun Oct 3 23:54:40 2004 +0000 32019 32020 Modified: 32021 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_mm_i2c.c Remove debugging 32022 statement that cause loading of DLL modules to fail. 32023 32024commit 37a8d8a53f12f829783011c1e077fefb5729e618 32025Author: Vladimir Dergachev <volodya@mindspring.com> 32026Date: Sun Oct 3 14:41:45 2004 +0000 32027 32028 Modified: 32029 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_video.c Fix compilation 32030 with gcc 3.4.x (patch by Ronny V. Vindenes) 32031 32032commit 6dbcd503fe21c6dec27c2b97adc0f39ba0a7ab2a 32033Author: Vladimir Dergachev <volodya@mindspring.com> 32034Date: Sun Oct 3 03:07:17 2004 +0000 32035 32036 Modified: 32037 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_video.[c,h] 32038 xc/programs/Xserver/hw/xfree86/drivers/ati/theatre.c 32039 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_reg.h Port the rest of 32040 GATOS Radeon-specific code. Remove "checkpoint" noise from logs. Test 32041 with AIW Radeon 7500 32042 32043commit 7cb911291d459f08f54444d09081d3378fe73b8b 32044Author: Vladimir Dergachev <volodya@mindspring.com> 32045Date: Sun Oct 3 00:01:14 2004 +0000 32046 32047 Modified: 32048 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon.h 32049 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_driver.c 32050 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_video.[c,h] 32051 xc/programs/Xserver/hw/xfree86/drivers/ati/theatre.c Detect and initialize 32052 Rage Theatre (RT100) 32053 32054commit 6c1996174b9f7afc8f4e33c35e4864b31696d4d6 32055Author: Vladimir Dergachev <volodya@mindspring.com> 32056Date: Sat Oct 2 20:55:42 2004 +0000 32057 32058 Modified: 32059 xc/programs/Xserver/hw/xfree86/drivers/ati/Imakefile 32060 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_reg.h 32061 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_video.[c,h] 32062 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_vip.c Add VIP bus access 32063 code, hook it into initialization system. 32064 32065commit 136649a1c2065624a9cf17386054c6595aa6d85e 32066Author: Vladimir Dergachev <volodya@mindspring.com> 32067Date: Sat Oct 2 20:35:04 2004 +0000 32068 32069 Modified: 32070 xc/programs/Xserver/hw/xfree86/drivers/ati/theatre* 32071 xc/programs/Xserver/hw/xfree86/drivers/ati/generic_bus.h Add Rage Theatre 32072 (RT100) module code and hook it into the build system 32073 32074commit 7a17b1c294fa8c7af0b1cdea0618cb6060479356 32075Author: Vladimir Dergachev <volodya@mindspring.com> 32076Date: Sat Oct 2 01:59:13 2004 +0000 32077 32078 Modified: 32079 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon.h 32080 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_video.c 32081 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_video.h 32082 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_mm_i2c.c Add code to 32083 access and initialize multimedia i2c bus. Hook it up. 32084 32085commit e0123e24b6895dde7b385528da24c4e6e9d5a4cb 32086Author: Vladimir Dergachev <volodya@mindspring.com> 32087Date: Sat Oct 2 01:36:43 2004 +0000 32088 32089 Add more multimedia registers. 32090 32091commit 576f53be4f0fa54bde7167872b7508b98e5c9454 32092Author: Vladimir Dergachev <volodya@mindspring.com> 32093Date: Sat Oct 2 00:43:08 2004 +0000 32094 32095 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon_driver.c 32096 xc/programs/Xserver/hw/xfree86/drivers/ati/radeon.h Add TV-in specific 32097 options. Add TV-in specific fields to radeon per-screen structure, add 32098 code to initialize them. Preserve MM_TABLE (instead of whole VBIOS as 32099 was done in GATOS drivers) 32100 32101commit 9a9ba82172428b09312fdd6e11da04b21c8529fc 32102Author: Adam Jackson <ajax@nwnk.net> 32103Date: Wed Sep 22 04:38:03 2004 +0000 32104 32105 Bug #1252, #1253, #1255, #1256: Various typo fixes from Dave Jones. 32106 32107commit 700b0dc5148d0170c9223765d208ccef447242df 32108Author: Adam Jackson <ajax@nwnk.net> 32109Date: Tue Sep 21 18:10:34 2004 +0000 32110 32111 Convert the ATI drivers to ANSI function prototypes. 32112 32113commit 17e5f3c1169066be1004306d12b5398cf01cbd35 32114Author: Eric Anholt <anholt@freebsd.org> 32115Date: Fri Sep 17 05:16:45 2004 +0000 32116 32117 Bugzilla #1170: Fix an overcounting of DMA in BEGIN_RING that wasn't 32118 harmful. Bugzilla #1169: Add counting for 32119 BEGIN_RING/OUT_RING/ADVANCE_RING to avoid these sorts of problems in 32120 the future. 32121 32122commit 29bbee6b8a8dd1b883e07547691ff8e1289459ce 32123Author: Adam Jackson <ajax@nwnk.net> 32124Date: Fri Sep 17 03:04:50 2004 +0000 32125 32126 Bug #1192: Remove cfb support from drivers where its use is an option. 32127 Delete xf24_32bpp, as s3virge was the last user. Fix up some comments 32128 to refer to fb rather than cfb. 32129 32130commit cec06d4126f5f31d14ace73798c71ec75e544c45 32131Author: Alex Deucher <agd5f@yahoo.com> 32132Date: Thu Sep 16 22:15:05 2004 +0000 32133 32134 Update comment 32135 32136commit dbeb8d28ac3469f4a7d7d8dae979ae876d4cdcba 32137Author: Kevin E Martin <kem@kem.org> 32138Date: Fri Sep 3 23:26:19 2004 +0000 32139 32140 Updated release notes for bug #999. 32141 Update Radeon man page to note that RenderAccel is now enabled by default 32142 on certain chips. 32143 32144commit c829f667c8a79bbdd40e1f924d57638e7bd425dc 32145Author: Michel Daenzer <michel@daenzer.net> 32146Date: Mon Aug 30 04:31:50 2004 +0000 32147 32148 Hotfix for Eric Anholt's fix for bug #1156 : Clear the 32149 RADEON_NONSURF_AP0_SWP_16BPP bit for 4 bpp data, otherwise it breaks at 32150 depth 16. 32151 32152commit 677246740b5b538e73a7fbde661ec6a97319430b 32153Author: Adam Jackson <ajax@nwnk.net> 32154Date: Fri Aug 27 20:31:26 2004 +0000 32155 32156 Back out my last change, dupe. 32157 32158commit a8eaa15264ff0db1596d653e91260c9a39a8b5ee 32159Author: Adam Jackson <ajax@nwnk.net> 32160Date: Fri Aug 27 18:32:40 2004 +0000 32161 32162 Bug #1061 (documentation-only update): Document the RenderAccel option in 32163 the radeon manpage. 32164 32165commit 360dab062f895c19e1784146e6684f75a20c5b24 32166Author: Eric Anholt <anholt@freebsd.org> 32167Date: Fri Aug 27 03:36:28 2004 +0000 32168 32169 Bug #1156: Fix the build on PPC by splitting the byte-order-setting code 32170 into a separate routine. While here, fix the byte swapping in the 32bpp 32171 case. (Michel Daenzer) 32172 32173commit 9b19bbb4d37b79c97365aa492cd93931d492b3a8 32174Author: Eric Anholt <anholt@freebsd.org> 32175Date: Fri Aug 27 03:23:47 2004 +0000 32176 32177 Bug #1157: Fix Radeon line acceleration in the DMA case (Michel Daenzer). 32178 32179commit 838488b6b46359334e5e781589dbee3832909628 32180Author: Kevin E Martin <kem@kem.org> 32181Date: Fri Aug 27 02:38:56 2004 +0000 32182 32183 Fix Radeon compilation without Xinerama (Bug #1179, Simon Schubert). 32184 32185commit 5f54d4f45e2adc6b15e98152db88743c243aa128 32186Author: Kevin E Martin <kem@kem.org> 32187Date: Wed Aug 25 00:30:41 2004 +0000 32188 32189 Fix hangs during startup with Radeon driver (Bug #1007, Sjoerd Simons). 32190 Fix i810 driver startup on Sony VAIO laptop (Bug #1084, Dave Airlie). 32191 Fix several library build problem (Bug #1147). 32192 Fix build problems (Bugs #1148 and #1149). 32193 32194commit dde880e876f455f1fd66efbcc055e177886397c9 32195Author: Eric Anholt <anholt@freebsd.org> 32196Date: Tue Aug 17 20:22:55 2004 +0000 32197 32198 Reenable Radeon Render acceleration now that interactions with 3D clients 32199 should be resolved. 32200 32201commit 279efef503fdbd442a9bd00f0073d9f6006d9e61 32202Author: Eric Anholt <anholt@freebsd.org> 32203Date: Sun Aug 15 00:44:57 2004 +0000 32204 32205 Fix unresolved symbol warnings in r128 driver (Adam Jackson) 32206 32207commit 955f2b89f94aaa20b5eafd045b01e3eeef3d29ad 32208Author: Kevin E Martin <kem@kem.org> 32209Date: Thu Aug 12 05:00:22 2004 +0000 32210 32211 Fix context switching on Radeons and build problem when XF86DRI is not 32212 defined (Bug #922, Eric Anholt). 32213 32214commit 8f3ea53f26493a5f02410297a587d1d95c067094 32215Author: Adam Jackson <ajax@nwnk.net> 32216Date: Thu Aug 12 02:23:46 2004 +0000 32217 32218 oops, these didn't added somehow. 32219 32220commit 6ec74414bd33e9d8dfd5119058d15f3a6206e9d9 32221Author: Adam Jackson <ajax@nwnk.net> 32222Date: Thu Aug 12 01:03:06 2004 +0000 32223 32224 Factor out R{128,ADEON}Chipsets into their own headers so they can be 32225 sanely included in both the ati core and the r128/radeon submodules; 32226 r128 and radeon can now be directly loaded under dlloader. Teach imstt 32227 about the dlloader- frendly way to talk to fbdevhw. (Partial for bugs 32228 #393 and #400.) 32229 32230commit a5b1e4dd7adacd17b63582e477df3a585efb249a 32231Author: Adam Jackson <ajax@nwnk.net> 32232Date: Wed Aug 11 01:20:28 2004 +0000 32233 32234 Bug #978: Silence more unresolved symbol warnings from the ati driver. 32235 Patch from Stefan Dirsch. 32236 32237commit 1356884132491010a2e41eb2b5be4a227b9b8138 32238Author: Kevin E Martin <kem@kem.org> 32239Date: Wed Aug 11 01:14:14 2004 +0000 32240 32241 Fix Radeon build problem with #define DoLoadableServer NO 32242 32243commit f5c93cb015e334975c2797a23b68798372a284c9 32244Author: Adam Jackson <ajax@nwnk.net> 32245Date: Tue Aug 10 04:01:42 2004 +0000 32246 32247 Bug #978: Silence unresolved symbol warnings from the ati driver. Reported 32248 by Michel Daenzer. 32249 32250commit a2da1caf5848a78b10afd99811fd1a5677af6786 32251Author: Kevin E Martin <kem@kem.org> 32252Date: Mon Aug 9 22:37:22 2004 +0000 32253 32254 Fix install problem on platforms not using xorg.cf/xfree86.cf (Bug #339, 32255 Harold L. Hunt II, Alexander Gottwald). 32256 Fix crash when using X core font in zh_CN.UTF-8 locale (Bug #368, Yu Shao, 32257 David Dawes). 32258 Fix glXMakeCurrent(Dpy, None, NULL) crash (Bug #719, Adam Jackson). 32259 HP-PA build fix (Bug #828, Guy Martin, Paul Anderson). 32260 Fix SDK build for GATOS and Wacom driver (Bug #829, Bryan Stine). 32261 Fix attempt to read video ROM before enabling it (Bug #843, Ivan 32262 Kokshaysky, Mike A. Harris). 32263 Fix detection of primary adapter (Bug #843, Ivan Kokshaysky, Mike A. 32264 Harris). 32265 Clarify xset man page description of how to use the keyboard repeat rate 32266 settings (Bug #846, Mike A. Harris). 32267 Fix problem where print-screen key would get remapped to sys-req in certain 32268 keymaps, which broke GNOME printscreen functionality (Bug #847, Owen 32269 Taylor). 32270 Fix several render problems: 32271 - MMIO mode support 32272 - Hang on IGP chips 32273 - VT switching hang 32274 - 3D render corruption (Bug #922, Hui Yu). 32275 32276commit 3171664c6a02424736864204b731249097fa7aae 32277Author: Kevin E Martin <kem@kem.org> 32278Date: Fri Aug 6 20:07:35 2004 +0000 32279 32280 Disable Render acceleration in Radeon driver temporarily until it no longer 32281 conflicts with the 3D code. 32282 32283commit 27ef7adf3c1f6c3ca79b42b468d08b8542348cdd 32284Author: Michel Daenzer <michel@daenzer.net> 32285Date: Wed Aug 4 13:17:31 2004 +0000 32286 32287 Use info->ModeReg.surface_cntl to restore byte swapping for the framebuffer 32288 aperture on big endian machines, remove superfluous local variables and 32289 register reads. 32290 Adapt framebuffer aperture byte swapping to texture format before copying 32291 data to offscreen area and restore it afterwards on big endian 32292 machines, fixes Render acceleration there. 32293 reviewed by: Hui Yu <hyu@ati.com>, Kevin E. Martin <kem@freedesktop.org> 32294 32295commit 2efdc9f847fa9ff3096667b16df84817fc65605d 32296Author: Eric Anholt <anholt@freebsd.org> 32297Date: Wed Aug 4 10:05:37 2004 +0000 32298 32299 - Add two new XAA hooks, SetupForCPUToScreenTexture2 and 32300 SetupForCPUToScreenAlphaTexture2. These add a dstFormat argument after 32301 the previous format argument, which the driver needs to use to properly 32302 set up the destination format. Two new arrays are added for the list of 32303 destination formats supported that correspond to the previous format 32304 arrays for sources. 32305 - Make Render acceleration only occur when the new hook for that 32306 acceleration type is supplied and the dst format list is set, along 32307 with the src format list being set. Without knowing the destination 32308 format, the Render acceleration couldn't properly support all the 32309 destinations it might encounter. 32310 - Bump XAA module minor version. 32311 - Update the Radeon Render acceleration to use the new hooks when the XAA 32312 module is sufficiently new. Fix a bug in the src/dst alpha booleans for 32313 ops, and use them to set blend_cntl to support destinations without 32314 alpha. Add missing PICT_a1r5g5b5 texture format, and add list 32315 terminator. (!) 32316 32317commit 1e16c66551eed750330bb74a92208c45a7112cd0 32318Author: Matthieu Herrb <matthieu.herrb@laas.fr> 32319Date: Tue Aug 3 09:42:20 2004 +0000 32320 32321 programs/Xserver/hw/xfree86/drivers/ati/radeon_merged_fb.c Add missing 32322 argument to xf86DrvMsg(). 32323 32324commit fd440a1e2110e05d116207800c123c281fbf5e19 32325Author: Matthieu Herrb <matthieu.herrb@laas.fr> 32326Date: Tue Aug 3 09:38:03 2004 +0000 32327 32328 programs/Xserver/hw/xfree86/drivers/ati/r128_driver.c 32329 programs/Xserver/hw/xfree86/drivers/ati/r128_probe.h 32330 programs/Xserver/hw/xfree86/drivers/ati/radeon_driver.c 32331 programs/Xserver/hw/xfree86/drivers/ati/radeon_probe.h Fix Constification 32332 problems. 32333 32334commit dd0f86c4cb046c79a65144f767e97fd99e7f61c9 32335Author: Kevin E Martin <kem@kem.org> 32336Date: Sun Aug 1 23:22:01 2004 +0000 32337 32338 Fix Radeon crash on server exit or generation 32339 32340commit c64f90b673f5c487dfe059f1636b04a1c2ab7dc2 32341Author: Adam Jackson <ajax@nwnk.net> 32342Date: Sat Jul 31 01:21:18 2004 +0000 32343 32344 Change several LoaderSymbol calls introduced by the bug #400 patch to 32345 *Weak() resolver functions. 32346 32347commit b091b4b074cd61c9067ba95ef016f7a29a38c38d 32348Author: Hui YU <hyu@ati.com> 32349Date: Fri Jul 30 22:20:21 2004 +0000 32350 32351 Support for New radeon chips: R420/M18, R423, RV370/M22, RV380/M24, RS300. 32352 Add special handlings for DELL triple-head server (RV100). Misc. bug 32353 fixes for flat panel, host aperture, etc (Bug #946) 32354 32355commit d04f61bfaf6952e832d52e1cd6014e86435aebd9 32356Author: Adam Jackson <ajax@nwnk.net> 32357Date: Fri Jul 30 20:30:51 2004 +0000 32358 32359 Bug #400 (partial): Driver fixes for the dlloader. When using dlloader, all 32360 framebuffer formats except cfb and the overlay modes should work, and 32361 r128 and radeon need to be loaded from the ati driver (both issues to 32362 be fixed soon). Tested on i740, s3virge, mach64, tdfx, vesa, and vga 32363 drivers. elfloader users shouldn't be affected. 32364 32365commit 6dcb78bb818c3c1c7fe406bee5379f9bef023f4b 32366Author: Alex Deucher <agd5f@yahoo.com> 32367Date: Wed Jul 28 02:35:30 2004 +0000 32368 32369 - add gamma correction support to the radeon video overlay 32370 32371commit 26029dc9378bbd95835d397402bfc9e6b4f1a478 32372Author: Alex Deucher <agd5f@yahoo.com> 32373Date: Mon Jul 26 23:14:37 2004 +0000 32374 32375 - Add Radeon DynamicClocks option 32376 - Add small fixes and clean ups from ati's last code drop (typo_fixes, 32377 remove_fudge, laptop, xvfix) 32378 - fix possible segfault in mga_dri.c (Ryan Underwood) 32379 - Add Xv support to pre-nm2160 neomagic chipsets 32380 32381commit 9f251b16e702da486f824307e72a6f80c6987b5d 32382Author: Adam Jackson <ajax@nwnk.net> 32383Date: Mon Jul 26 19:14:58 2004 +0000 32384 32385 Bug #848: Clean up gatos notice messages. 32386 32387commit 6f1e01ecdb11913fd768b9918f516b8379b7858f 32388Author: Eric Anholt <anholt@freebsd.org> 32389Date: Fri Jul 23 05:36:14 2004 +0000 32390 32391 - Add a new option, BuildDevelDRIDrivers, which controls whether 32392 DevelDRIDrivers is non-empty. DevelDRIDrivers contains those drivers 32393 aren't built by default either due to being unusable on that 32394 architecture (but buildable for testing) or due to security concerns, 32395 as in the mach64 and savage cases. 32396 - Merge the Mach64 DDX's DRI support, under the XF86DRI_DEVEL define, which 32397 is set only when BuildDevelDRIDrivers is set. 32398 - Fix up of the sets of DRI drivers built for various architectures. 32399 - Fix build of unichrome DRI driver, which is now enabled on x86. 32400 32401commit b68322f467b7e524092af15e33ab872abb99b6a2 32402Author: Eric Anholt <anholt@freebsd.org> 32403Date: Thu Jul 22 07:03:04 2004 +0000 32404 32405 Merge and resolve conflicts from DRI/DRM/Mesa 20040721 imports. 32406 32407commit 57de5e285d6271fc4b2bc8a951a98be6308d649d 32408Author: Eric Anholt <anholt@freebsd.org> 32409Date: Fri Jul 16 04:38:54 2004 +0000 32410 32411 Bugzilla #803,804: Fix build in the !XF86DRI case after DRI and Radeon 32412 Render acceleration merges. 32413 32414commit 9b0ec803bd588d9af778bfd307ee6d958e4a1e73 32415Author: Adam Jackson <ajax@nwnk.net> 32416Date: Sat Jul 10 14:22:35 2004 +0000 32417 32418 Bugzilla #849: Fix DRI pagesize assumptions in radeon and r128 drivers. 32419 from mharris@, reviewed and applied to DRI by airlied@. 32420 32421commit ccd95bfb291a0c70272fab6822ef379b83dc298d 32422Author: Egbert Eich <eich@suse.de> 32423Date: Tue Jul 6 14:37:47 2004 +0000 32424 32425 Separated Intel drivers from default DriDrivers to avoid building them on 32426 IA64 (Egbert Eich). 32427 Fixed wrong function prototype (Egbert Eich). 32428 Don't test for generic VGA on IA64 (Egbert Eich). 32429 Fixed a segfault when accessing a structure before verifying the pointer 32430 exists (Egbert Eich). 32431 Added a showcache option for debugging (Egbert Eich). 32432 Increase default video RAM size to 16MB when DRI is enabled and more than 32433 128MB are available (Egbert Eich). Fixed lockups during mode switch. 32434 Problem was introduced when attempting to copy the behavior during 32435 LeaveVT()/EnterVT() but but forgetting to call I810DRILeave() before 32436 I810DRIEnter(). The entire DRILeave()/Enter() scenario has been 32437 commented out as it didn't seem to be necessary (Egbert Eich). 32438 Fix TweakMemorySize() (tested with i855/i865) (Egbert Eich). 32439 increased MAX_DEVICES to 128 (Egbert Eich). 32440 Use OS provided PCI config space access as default method (Egbert Eich). 32441 Added support for Linux 2.6 proc file format. 32442 Fixed unaligned accesses to pieces of the VBE info block. VESA did not 32443 align elements to size (Egbert Eich). 32444 32445commit 6bdf3d36ea3086432af5157540fde4e0aa0c44dc 32446Author: Eric Anholt <anholt@freebsd.org> 32447Date: Sun Jun 20 07:34:55 2004 +0000 32448 32449 Add file missed in last commit: Bug #748: 32450 - Add Render acceleration for Radeon 100 and 200-series cards, enabled by 32451 default. Notable performance gains include 27fps in cairogears instead 32452 of 6fps (compared to 50 with glitz), and my disks are now the 32453 bottleneck in an ls -lR in gnome-terminal. Only supported in DRI mode 32454 because the MMIO submission hangs the card so far, but the code is left 32455 in because it may be supportable soon. 32456 - Add xorg.conf options to disable Render acceleration and to force the 32457 subpixel order in the server (Hui Yu). 32458 Many thanks to Hui Yu at ATI for the code this was based off of, Michel 32459 Daenzer for bugfixes and suggestions, and LinuxFund for sponsoring the 32460 work. 32461 32462commit 8c1b37420ea890f76edb97aed61ee427a871cd3f 32463Author: Eric Anholt <anholt@freebsd.org> 32464Date: Sun Jun 20 07:26:08 2004 +0000 32465 32466 Bug #748: 32467 - Add Render acceleration for Radeon 100 and 200-series cards, enabled by 32468 default. Notable performance gains include 27fps in cairogears instead 32469 of 6fps (compared to 50 with glitz), and my disks are now the 32470 bottleneck in an ls -lR in gnome-terminal. Only supported in DRI mode 32471 because the MMIO submission hangs the card so far, but the code is left 32472 in because it may be supportable soon. 32473 - Add xorg.conf options to disable Render acceleration and to force the 32474 subpixel order in the server (Hui Yu). 32475 Many thanks to Hui Yu at ATI for the code this was based off of, Michel 32476 Daenzer for bugfixes and suggestions, and LinuxFund for sponsoring the 32477 work. 32478 32479commit 6ecf374d500afe6da494dfdd6566396ec65b6d6a 32480Author: Eric Anholt <anholt@freebsd.org> 32481Date: Wed Jun 16 09:43:59 2004 +0000 32482 32483 Merge DRI-trunk-20040613 changes in programs/Xserver/hw/xfree86/drivers, 32484 with the following notes: 32485 - Savage and Mach64 (= ati/ati*.[ch] changes) DRI not merged due to 32486 insecurity. 32487 - VIA driver converted to new drmContext and drmHandle names. 32488 - Radeon driver merge conflicted in many places, and MergedFB at least 32489 could probably use some checking at this point. 32490 32491commit bea8085e04136b0ef513c17bb65c54069ec531e1 32492Author: Eric Anholt <anholt@freebsd.org> 32493Date: Wed Jun 16 09:25:58 2004 +0000 32494 32495 DRI trunk-20040613 import 32496 32497commit 7b588cda922992a8ee2d04853ba1533cf43592a3 32498Author: Eric Anholt <anholt@freebsd.org> 32499Date: Wed Jun 16 09:25:58 2004 +0000 32500 32501 Initial revision 32502 32503commit 6cbb5bbff43d65a762d432659793333868837f0b 32504Author: Eric Anholt <anholt@freebsd.org> 32505Date: Wed Jun 16 09:22:56 2004 +0000 32506 32507 DRI XFree86-4_3_99_12-merge import 32508 32509commit 3586c3c7955365a2af0b8a286ea0d0488ddfd137 32510Author: Egbert Eich <eich@suse.de> 32511Date: Fri Apr 23 19:26:46 2004 +0000 32512 32513 Merging XORG-CURRENT into trunk 32514 32515commit 538768e0f472026325be9dd37dc4bfe7ea7d3b75 32516Author: Egbert Eich <eich@suse.de> 32517Date: Sun Mar 14 08:33:20 2004 +0000 32518 32519 Importing vendor version xf86-4_4_99_1 on Sun Mar 14 00:26:39 PST 2004 32520 32521commit 5f179e796d1cd310790d06eebb474ad67114ec84 32522Author: Egbert Eich <eich@suse.de> 32523Date: Wed Mar 3 12:12:18 2004 +0000 32524 32525 Importing vendor version xf86-4_4_0 on Wed Mar 3 04:09:24 PST 2004 32526 32527commit 493c44580191beff63575c4032f772fbbf1b06eb 32528Author: Egbert Eich <eich@suse.de> 32529Date: Thu Feb 26 13:35:52 2004 +0000 32530 32531 readding XFree86's cvs IDs 32532 32533commit 942a47fd3c5e8df85642672c01072bf97285bc30 32534Author: Egbert Eich <eich@suse.de> 32535Date: Thu Feb 26 09:23:18 2004 +0000 32536 32537 Importing vendor version xf86-4_3_99_903 on Wed Feb 26 01:21:00 PST 2004 32538 32539commit dda79b5ee3eb49f924a3213592d93ba981c99394 32540Author: Kaleb Keithley <kaleb@freedesktop.org> 32541Date: Mon Feb 23 20:35:02 2004 +0000 32542 32543 Import most of XFree86 4.4RC3. This import excludes files which have the 32544 new license. If we want to, later we can import 4.4RC3 again and pick 32545 up the files that have the new license, but for now the vendor branch 32546 is "pure." 32547 32548commit 08104f18ac497561c969694ff709da6c361ccea7 32549Author: Egbert Eich <eich@suse.de> 32550Date: Thu Jan 29 08:08:34 2004 +0000 32551 32552 Importing vendor version xf86-012804-2330 on Thu Jan 29 00:06:33 PST 2004 32553 32554commit 162d76b9c48ed17d6e645057db58c9f13436f25f 32555Author: Kaleb Keithley <kaleb@freedesktop.org> 32556Date: Fri Dec 19 20:55:06 2003 +0000 32557 32558 XFree86 4.3.99.902 (RC 2) 32559 32560commit 195772ec74f7a9b7d5bf294b3bc2c3b960e08a31 32561Author: Kaleb Keithley <kaleb@freedesktop.org> 32562Date: Thu Dec 4 22:03:16 2003 +0000 32563 32564 XFree86 4.3.99.901 (RC 1) 32565 32566commit 73105e438e90f950249cb265ba99001d4db3dc60 32567Author: Kaleb Keithley <kaleb@freedesktop.org> 32568Date: Tue Nov 25 19:28:36 2003 +0000 32569 32570 XFree86 4.3.99.16 Bring the tree up to date for the Cygwin folks 32571 32572commit 0a53864f9581196604d04fd28b4e94a9b0b73d18 32573Author: Kaleb Keithley <kaleb@freedesktop.org> 32574Date: Tue Nov 25 19:28:36 2003 +0000 32575 32576 Initial revision 32577 32578commit c56aaef79028261e0a3be0c86ffd3f16b6bdd9ba 32579Author: Kaleb Keithley <kaleb@freedesktop.org> 32580Date: Fri Nov 14 16:48:55 2003 +0000 32581 32582 XFree86 4.3.0.1 32583 32584commit d09f463a5d1ce73e0b65d5276fbcca393fa2da46 32585Author: Kaleb Keithley <kaleb@freedesktop.org> 32586Date: Fri Nov 14 16:48:55 2003 +0000 32587 32588 Initial revision 32589 32590commit d9af6dc32652502d84ea8da5d57a5ab45429c4ad 32591Author: Kaleb Keithley <kaleb@freedesktop.org> 32592Date: Fri Nov 14 15:54:49 2003 +0000 32593 32594 R6.6 is the Xorg base-line 32595