NEWS revision 42542f5f
1Snapshot 2.99.916 (2014-09-08)
2==============================
3Quick update for MST in UXA - we need to hook up the RandR outputs for
4dynamicaly added connectors.
5
6
7Snapshot 2.99.915 (2014-09-08)
8==============================
9A significant change to UXA to enable MST and to keep it working on recent
10kernels needs some soaking before a major release, and the usual
11plethora of bugfixes. One other feature is the support for hardware
12rotations on very recent kernels.
13
14 * Handle rotated scanouts that are all clear correctly. The region to
15   update was not being computed correctly, overdrawing the wrong CRTC.
16   https://bugs.freedesktop.org/show_bug.cgi?id=81820
17
18 * Reset cursor images after rotation and size changes
19   https://bugs.freedesktop.org/show_bug.cgi?id=81886
20   https://bugs.freedesktop.org/show_bug.cgi?id=82273
21   https://bugs.freedesktop.org/show_bug.cgi?id=82337
22
23 * Handle stale DRI2 buffers and Client errors more gracefully
24
25 * Fallback if we fail to render a glyph onto a too-large surface, e.g.
26   direct rendering of glyphs onto extended desktops with gen2/gen3.
27   https://bugs.archlinux.org/task/40949
28
29 * Further work to enable tiled rendering onto large surfaces in
30   severely aperture and memory constrained devices (e.g. gen2/gen3)
31
32 * Honour the Primary option from xorg.conf and mark that as the RandR
33   primary.
34   https://bugs.freedesktop.org/show_bug.cgi?id=82193
35   https://bugs.freedesktop.org/show_bug.cgi?id=82205
36
37 * Fix another Client-Window DRI2 close race
38   https://bugs.freedesktop.org/show_bug.cgi?id=82979
39
40 * Fix incorrect discarding of GPU damage when copying over the TearFree
41   scanout - which caused a flicker in rapidly updated elements (like
42   simple video panes).
43   https://bugs.freedesktop.org/show_bug.cgi?id=81973
44
45
46Snapshot 2.99.914 (2014-07-23)
47==============================
48And a brown paper bag to hide the rebuilding from the tarball with
49'autoreconf -fi' error that arose from not distributing the libobj/
50directory.
51
52
53Snapshot 2.99.913 (2014-07-23)
54==============================
55This should be it... A few fixes from testing the new code, we should be
56ready for the final release. However, we do have one standout feature in
57this snapshot, we now officially recognise HD Graphics 5300/5500/5600,
58Iris Graphics 6100 and Iris Pro Graphics 6200/P6300 (formerly known as
59Broadwell).
60
61 * Check the window actually covers the CRTC before doing a single CRTC flip,
62   and then restore the right framebuffer after completing CRTC flips.
63   Otherwise we would detect an error and disable an output under TearFree
64   Regression in 2.99.912
65   https://bugs.freedesktop.org/show_bug.cgi?id=80191
66
67 * Fix framebuffer creation on kernels older than 3.11
68   Regression in 2.99.912
69
70 * Check that the damage still exists after implicit reduction
71   Regression in 2.99.912
72   https://bugs.freedesktop.org/show_bug.cgi?id=77436
73
74 * Fix direction flags for fallback composited CopyAreas which
75   caused scrolling corruption in a few configurations
76   Regression from 2.20.0
77   https://bugs.freedesktop.org/show_bug.cgi?id=79843
78
79 * Do not throw away damage if there is no redundant copy
80   https://bugs.freedesktop.org/show_bug.cgi?id=79992
81
82 * Check clipping on PolyRect before discarding the clipped damage
83   Regression from 2.99.903
84   https://bugs.freedesktop.org/show_bug.cgi?id=79992
85
86 * Fix hints for GLXPixmapa, as these are never swapped and so
87   miss invalidating the hints on SwapBuffers with the result
88   that they are often presumed blank
89   Regression in 2.99.912
90   https://bugs.freedesktop.org/show_bug.cgi?id=79999
91
92 * Fix incoherent choice of source bo when constructing 8x8 tiles,
93   incorrect pattern origin when extracting, and then fix the
94   alignment of colour patterns for BLT operations
95   https://bugs.freedesktop.org/show_bug.cgi?id=80033
96
97 * Disable blending with the render engine on snoopable buffers
98   https://bugs.freedesktop.org/show_bug.cgi?id=80253
99
100 * Restore throttling to prevent client lag under heavy GPU load
101   Regression from 2.21.10
102   https://bugs.freedesktop.org/show_bug.cgi?id=77436
103
104 * Use ClientGone for notifications on shared DRI2 windows to prevent
105   rare crashes due to use-after-free of the swap requests
106   https://bugs.freedesktop.org/show_bug.cgi?id=80157
107
108 * Ensure the mmaped CPU bo is idle before migrating damage
109   https://bugs.freedesktop.org/show_bug.cgi?id=80560
110
111 * Fix incorrect clipping by the render engine for large DRI2 windows
112
113 * Ensure that the aperture tiling fallbacks are bounded
114
115 * Validate parameter to xf86-video-intel-backlight-helper more carefully
116   (CVE-2014-4910)
117
118 * Fix slaved scanouts for reverse optimus, though rotated slaves will
119   require further patches to Xorg.
120   https://bugs.freedesktop.org/show_bug.cgi?id=81383
121
122 * Fix build without Composite extension.
123
124 * Fix build without gettline().
125
126 * UXA: Allocate and resize frontbuffer consistently to pass sanity checks
127   https://bugs.freedesktop.org/show_bug.cgi?id=80088
128
129 * UXA: Report cached backlight value when the output is off (like sna)
130   https://bugzilla.redhat.com/show_bug.cgi?id=1032978
131
132 * UXA: Mark outputs as off before the kernel does (like sna)
133   This will prevent the internal panel from starting up blank in some
134   multi-monitor configurations
135   https://bugzilla.redhat.com/show_bug.cgi?id=1103806
136
137Note that the DRI2 exchange mechanism introduced in 2.99.912 exposes bugs
138in some compositors, at least kwin and comptom, which discard DRI2 buffer
139invalidates rather than resourcing their texture. For example,
140https://bugs.kde.org/show_bug.cgi?id=336589
141
142Note that the improved triple buffering introduced in DRI2 requires a patch
143to Xorg (now released upstream) to prevent crashes with DRI_PRIME.
144https://bugs.freedesktop.org/show_bug.cgi?id=80001
145
146Note that DRI3/Present require tracking the relevant development trees for
147mesa and the xserver as they are very much still under early testing. Also
148be aware that Mesa provides no support for explicit fencing so Damage
149tracking between compositors and clients is unserialised.
150
151
152Snapshot 2.99.912 (2014-06-10)
153==============================
154A final round of features. We have everything from support for variable
155cursor sizes, support for the DRI3 and Present extensions, improved DRI2
156support, support for Xserver 1.16, userptr from kernel 3.16, and
157precursory support for DP multistream transport,
158
159 * Avoid discarding dirty pixels when promoting a migration to cover the
160   whole pixmap.
161   Regression in 2.99.911
162   https://bugs.freedesktop.org/show_bug.cgi?id=77063
163   https://bugs.freedesktop.org/show_bug.cgi?id=77178
164
165 * Avoid overextending degenerate lines (and consequentially accessing
166   pixels outside of our damaged area).
167   https://bugs.freedesktop.org/show_bug.cgi?id=77074
168
169 * Fix subpixel glyph rendering on gen2 devices (830-865 chipsets)
170   Regression in 2.99.911
171   https://bugs.freedesktop.org/show_bug.cgi?id=77201
172
173 * Share the global pixman glyph cache between ZaphodHeads
174   https://bugs.freedesktop.org/show_bug.cgi?id=54707
175
176 * Light up all connected outputs, even if their status is unknown, on
177   takeover from fbcon. This prevents loss of display after a resume
178   on recent kernels, for example.
179   https://bugs.freedesktop.org/show_bug.cgi?id=77768
180
181 * Show the video overlay (when supported by the hardware) across all
182   outputs.
183   https://bugs.freedesktop.org/show_bug.cgi?id=77802
184
185 * Do not discard damage when performing "BLT" spans inplace with the CPU.
186   Regression from 2.20.10
187
188 * Avoid discarding IO buffers too early during their preparation for a
189   new batch
190   https://bugs.freedesktop.org/show_bug.cgi?id=79238
191
192 * Fix fallback handling for displaying large scaled framebuffers (that
193   are too large to be scaled by the GPU in a single pass)
194   https://bugs.freedesktop.org/show_bug.cgi?id=79320
195
196 * Listen to external modifications of backlight value and propagate
197   the notifications to RandR clients. This should make the GUI report
198   ACPI keypresses to change the backlight correctly.
199   https://bugs.freedesktop.org/show_bug.cgi?id=79699
200
201 * UXA: fix pageflips with 3 heads.
202
203 * UXA: do not report a BadMatch error for DRI2GetMsc - as clients are
204   often unprepared and die when they get the unexpected error.
205
206
207Snapshot 2.99.911 (2014-03-19)
208==============================
209Hans de Geode has been working on making the Xserver work without
210privileges under the supervision of systemd/logind. This necessitated a
211few new features for us: server fds (where we are passed which fd to use
212to talk to our device by the Xserver who may in turn receive it from
213logind or other host) and a small backlight helper so that we can continue
214to provide a RandR backlight property when running without root privileges.
215
216 * Flush when changing blend modes on Ironlake, or else single glyphs
217   are sometimes rendered incorrectly
218   Regression from 2.20.15
219   https://bugs.freedesktop.org/show_bug.cgi?id=74882
220
221 * Fix pixmap offsets for pixman fallbacks onto Composite redirected
222   windows
223   https://bugs.freedesktop.org/show_bug.cgi?id=73811
224
225 * Fix blending onto 8-bit destinations, typically used for generating
226   masks in complex Render operations, on gen2
227   https://bugs.freedesktop.org/show_bug.cgi?id=75818
228
229 * Handle failure to create DRI bo more carefully. For example on gen3,
230   the DDX supports pixmaps that are much, much larger than OpenGL can
231   use and do not support the tiling modes that we request for OpenGL.
232   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1289049
233
234 * Fix a bookkeeping bug with proxy buffers that are marked active but
235   not actually inserted into a request (so they end up permanently
236   active and confuse everybody).
237   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1289923
238
239 * Actually turn off displays with DPMS off for UXA.
240   Regression from 2.99.903, but requires
241
242   kernel commit c9976dcf55c8aaa7037427b239f15e5acfc01a3a
243   Author: Chris Wilson <chris@chris-wilson.co.uk>
244   Date:   Sun Sep 29 19:15:07 2013 +0100
245
246       drm/i915: Only apply DPMS to the encoder if enabled
247
248   instead for correct behaviour on Haswell.
249   https://code.google.com/p/chromium/issues/detail?id=341135
250
251
252Snapshot 2.99.910 (2014-02-10)
253==============================
254Another latent bug exposed by recent changes merit another snapshot for
255final testing.
256
257  * Only discard damage when overwriting the dirty CPU bo, instead
258    of discarding damage that will be shown!
259
260  * Reset operation state when switching between glyph caches.
261    https://bugs.freedesktop.org/show_bug.cgi?id=74494
262
263  * Fully reinitialise pixmaps allocated from the freed cache. Fixes
264    a potential issue (crash or misrendering) when using some compositors.
265    https://bugs.freedesktop.org/show_bug.cgi?id=74550
266
267  * Do not expose the TexturedVideo adaptor in UXA when it is disabled
268    either due to a hung GPU or explicitly disabled by the user.
269
270  * Restore the pipe stall when changing CC state on gen6, otherwise
271    the GPU may not flush intermediate results from all EU resulting
272    in render corruption (usually the occasional black box).
273    Regression from 2.99.906
274    https://bugs.freedesktop.org/show_bug.cgi?id=7237
275
276
277Snapshot 2.99.909 (2014-02-01)
278==============================
279Pass the brown paper bag. Ridiculously stupid bug in last minute coding.
280
281  * Add Xv support using glamor acceleration in addition to adaptors
282    provided by the UXA backend.
283
284  * Fix overeager discarding of CPU damage
285    Regression in 2.99.908 :(
286    https://bugs.freedesktop.org/show_bug.cgi?id=74327
287
288
289Snapshot 2.99.908 (2014-01-31)
290==============================
291A couple of regressions dashed the hopes that .907 was to be the final
292release candidate, so time to start the cycle again after applying a few
293more bugfixes.
294
295  * Fix invalid pageflipping of GLXPixmaps by UXA.
296    https://bugs.freedesktop.org/show_bug.cgi?id=73282
297
298  * Consistently treat DPMS Suspend/Standy as Off, otherwise we may
299    incorrectly restore the backlight in UXA.
300
301  * Fix disabling the backlight in UXA when querying the value whilst off
302    https://bugs.freedesktop.org/show_bug.cgi?id=73181.
303
304  * Invalidate between every operation if rendering into the source or mask.
305    Fixes regression in 2.99.907 on Ivybridge (seen in KDE)
306    https://bugs.freedesktop.org/show_bug.cgi?id=73208
307
308  * Check for available batch buffer state before restoring state on gen4,
309    otherwise we overwrite surface state with commands and cause a GPU hang.
310    https://bugs.freedesktop.org/show_bug.cgi?id=73348
311
312  * Prevent an invalid free by TearFree
313    https://bugs.freedesktop.org/show_bug.cgi?id=73469
314
315  * Fix confusion in TearFree when it tried to pageflip to a stale bo
316    https://bugs.freedesktop.org/show_bug.cgi?id=70905
317
318  * Initialise TearFree contents with the current frontbuffer upon creation,
319    otherwise old content may be visible for significant fractions of a
320    second after resume or DPMS on
321    https://bugs.freedesktop.org/show_bug.cgi?id=73842
322
323  * Nullify the old pointer into a CPU mmapping of the bo when promoting it
324    on the GPU. Eventually that old pointer becomes invalid and will cause
325    the Xserver to crash.
326    Fixes regression in 2.99.906
327    https://bugs.freedesktop.org/show_bug.cgi?id=73351
328
329  * Restore the serialNumber on the GC correctly after falling back to
330    software. This was seen to cause corruption with Wine.
331    https://bugs.freedesktop.org/show_bug.cgi?id=73856
332
333  * Do not ignore CPU damage on a pixmap unless the Composite operation
334    will completely overwrite it. This was observed in xfce4.
335    https://bugs.freedesktop.org/show_bug.cgi?id=69528
336
337  * Skip fake outputs during initial probe as they can cause a NULL pointer
338    dereference.
339    https://bugs.freedesktop.org/show_bug.cgi?id=73981
340
341  * Fix rendering of dashed zero-width lines onto 24bit buffers
342    https://bugzilla.redhat.com/show_bug.cgi?id=1059152
343
344  * Cap the freed Pixmap cache and reuse it more often
345    Fixes a regression in 2.99.907 that caused the memory used by X to
346    grow until it was cleaned up during server regreneration
347    https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1272338
348
349
350Snapshot 2.99.907 (2013-12-30)
351==============================
352The highlight here is that things seem to be quietening down on the bug
353reporting front. All is not quiet as you will see below, and maybe it is
354just the holiday season that is approaching*, but there has been a
355remarkable falloff in bug reports. Lets hope this trend holds and we can
356make a stable release shortly!
357
358* This was intended to be sent a couple of weeks ago and the holiday season
359has been very quiet....
360
361Oh, and acceleration support for Intel's next generation of
362integrated processor graphics has landed, codenamed Broadwell.
363
364 * Fix potential X server infinite recursion (crash) from a gen2 bug fix
365   Regression in 2.99.906
366   https://bugs.freedesktop.org/show_bug.cgi?id=71605
367
368 * Workaround a missing pipeline flush within Ivybridge, that would
369   leave black rectangles randomly over the output
370   https://bugs.freedesktop.org/show_bug.cgi?id=68410
371
372 * Fix tiled fills. gen2-3 and gen4+ had two different bugs that
373   both broke rendering with small 8x8 patterns in some circumstances
374   https://bugs.freedesktop.org/show_bug.cgi?id=71260
375
376 * Fix reads from a cropped video image using a packed pixel format.
377
378 * Another clear the clear hint after DRI2 SwapBuffers, like the bug fixed
379   in 2.99.903. Failure to clear the hint would cause read backs of the
380   frontbuffer (Xvnc) to be blank.
381   https://bugs.freedesktop.org/show_bug.cgi?id=72194
382
383 * Disable VSync on Baytrail
384   https://bugs.freedesktop.org/show_bug.cgi?id=69869
385
386 * Handle partial uploads with TearFree correctly
387   https://bugs.freedesktop.org/show_bug.cgi?id=72343
388   https://bugs.freedesktop.org/show_bug.cgi?id=72430
389
390 * Avoid recusing through DRI event handlers whilst processing TearFree,
391   leading to a double free (and memory corruption)
392   https://bugs.freedesktop.org/show_bug.cgi?id=72690
393
394 * Hide the gen4 render corruption by crippling the GPU
395   https://bugs.freedesktop.org/show_bug.cgi?id=55500
396
397
398Snapshot 2.99.906 (2013-11-13)
399==============================
400Several stability fixes required after the recent tweaking of the core
401mechanics to handle the updated TearFree and attempting to make static
402analyzers happy.
403
404 * Fix damage handling when rendering to a partially damaged GPU surface.
405   Regression in 2.99.905
406   https://bugs.freedesktop.org/show_bug.cgi?id=70527
407
408 * Use asprintf() instead of sprintf()
409   Regression in 2.99.905
410   https://bugs.freedesktop.org/show_bug.cgi?id=70835
411
412 * Improve accounting for fence overallocation on older gen2/3, and
413   improve the tiling mechanism to fit into the same aperture constraints
414   https://bugs.freedesktop.org/show_bug.cgi?id=70924
415
416 * Add an extra GPU flush on Sandybridge to fix some rare font corruption
417
418 * Rasterise lines through all clip boxes
419   https://bugs.freedesktop.org/show_bug.cgi?id=70802
420
421 * Fix regression from stricter handling of failures to move a GC to the GPU
422   Regression in 2.99.905
423   https://bugs.freedesktop.org/show_bug.cgi?id=71415
424
425 * Fix various fail along the memcpy_xor paths, including inadequate error
426   handling and integer overflow
427   https://bugs.freedesktop.org/show_bug.cgi?id=70527
428
429 * Fix outside-of-target stipple uploads
430   https://bugs.launchpad.net/bugs/1247785
431
432 * Fix clip detection for long glyphs
433   Incomplete bug fix (causing a regression) in 2.99.905
434   https://bugs.freedesktop.org/show_bug.cgi?id=70527
435
436 * Fix VSync for the render engine (Xv) on Haswell
437   https://bugs.freedesktop.org/show_bug.cgi?id=70527
438
439
440Snapshot 2.99.905 (2013-10-23)
441==============================
442The highlight for this snapshot is the extension of TearFree to support
443transformed outputs, along with some polishing to eliminate its impact
444upon input and output latency. As always, thanks to everyone who have
445been testing, reporting bugs and helping to improve the stability before
446release.
447
448 * Prevent a crash when starting with a user specified mode or position
449
450 * Prevent some crashes in UXA after allocation failure
451
452 * Stop marking the user's preferred backlight value as 0 if the
453   backlight property is queried whilst the connector is disabled
454   https://bugs.freedesktop.org/show_bug.cgi?id=70406
455
456 * Pad GETCONNECTOR ioctl for compatability between 32/64-bit userspace
457   and kernel
458
459 * Handle long glyph runs correctly
460   https://bugs.freedesktop.org/show_bug.cgi?id=70541
461
462 * Fix clipping of stippled rectangles against clip regions
463   https://bugs.freedesktop.org/show_bug.cgi?id=67865
464
465 * Support TearFree rendering of rotated outputs
466   https://bugs.freedesktop.org/show_bug.cgi?id=22969
467
468
469Snapshot 2.99.904 (2013-10-09)
470==============================
471There is one more feature planned to be completed for 3.0, so time for a
472snapshot beforehand to push out the bug fixes from the last week.
473
474 * Fix video output using sprites when changing the image size
475
476 * Apply more restrictive tile constaints for 915g class devices
477   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1232546
478
479 * Ensure all overlapping rectangles are drawn for XRenderFillRectangles
480   https://bugs.freedesktop.org/show_bug.cgi?id=66313
481
482 * Fix trapezoid clipping against the left-edge
483   https://bugs.freedesktop.org/show_bug.cgi?id=69469
484
485 * Prevent discarding active upload buffers, causing glitches in chromium
486   https://bugs.freedesktop.org/show_bug.cgi?id=66990
487
488 * Prevent specifying a negative timeout to select if the BlockHandler
489   takes too long to update the display
490
491 * Promote the Ironlake pipecontrol to be a full pipeline flush to
492   prevent render cache corruption
493   https://bugs.freedesktop.org/show_bug.cgi?id=51422
494
495 * Never pass an invalid trapezoid to pixman
496   https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/1197921
497
498 * Prevent out-of-bounds access by overassigning work amongst threads
499   https://bugs.freedesktop.org/show_bug.cgi?id=70204
500
501 * Make sure the current mode is always listed amongst the output modes
502   https://bugs.freedesktop.org/show_bug.cgi?id=70132
503
504 * Build fixes for 1.14.99.2
505
506
507Snapshot 2.99.903 (2013-09-28)
508==============================
509Lots more stabilization work, not yet peaceful enough to christen 3.0. We
510have everything ranging from build fixes for systems like Suse Linux
511Enterprise Desktop that like to backport hardware enablement to ancient
512packages, to fixes for hardware enabling, and some more performance tuning.
513
514 * Fix VSync on Haswell.
515   https://bugs.freedesktop.org/show_bug.cgi?id=69119
516
517 * Disable Y-tiling on gen4 - it too frequently leads to instability.
518   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1222203
519
520 * Disable same EDID detection based on property id - the kernel likes
521   to reuse ids leading to a failure to detect the right modes on a monitor
522   change.
523
524 * Avoid issuing multiple DPMS requests to the same encoder (alised to
525   multiple connectors) to avoid upsetting Haswell and leaving the
526   screens blank.
527
528 * Honour the user preferrence for the initial mode, even if they are
529   being silly
530   https://bugzilla.novell.com/show_bug.cgi?id=841696
531
532 * Clear the clear hint when apply DRI updates - to prevent some
533   screenshots from GL windows being left blank.
534   https://bugs.freedesktop.org/show_bug.cgi?id=69730
535
536 * Prevent a NULL dereference from trying to undo an non-existent buffer
537   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1228677
538
539 * Handle out-of-memory conditions far more gracefully. If the system is
540   hard against the memory wall, then the kernel will start issuing
541   SIGBUS even for CPU mmaps. Untrapped these will cause X to die.
542   https://bugs.freedesktop.org/show_bug.cgi?id=67889
543
544
545Snapshot 2.99.902 (2013-09-07)
546==============================
547We do not condone or support Canonical in the course of action they have
548chosen, and will not carry XMir patches upstream.
549-The Management
550
551Snapshot 2.99.901 (2013-09-04)
552==============================
553What's this? A pre-release snapshot of an upcoming 3.0 major release?
554What could be the big new feature about to land? SNA by default!
555Not only that, we also have a preview of the XMir integration patches.
556
557Whilst we consider the SNA acceleration method to be ready to take over
558from UXA (the current acceleration method), it still remains possible to
559switch over to the old method with the following xorg.conf snippet:
560  Section "Device"
561    Identifier "Device0"
562    Driver "intel"
563    Option "AccelMethod" "UXA"
564  EndSection
565
566The promise of SNA is that it improves the user experience by providing a
567faster, more fluid, desktop, that is both more power efficient and less
568prone to crashes than UXA. If it still fails to meet your expections,
569please file bugs!
570
571
572Release 2.21.15 (2013-08-21)
573============================
574Some build fixes for the BSDs and alternate compiler, and conflicting
575configure options and a critcial fix for some PRIME setups.
576
577 * Don't fail to configure if DRI1 is not available and the user asks
578   for both DRI and KMS-only
579   [Regression from 2.21.14]
580
581 * Lots of miscellaneous fixes for older gcc, other compilers and BSD.
582
583 * Initial framework support for hosted X.
584
585 * Improve transition from fbcon to X with multiple outputs and extended
586   desktops.
587
588 * Ensure the framebuffer exists before checking it against the CRTC
589   constraints.
590   https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/121234
591   [Regression from 2.21.10]
592
593 * Add RGB overlay support for Ironlake and later.
594
595 * Don't release the overlay buffer too early, i.e. before the client
596   disables the Xv image.
597
598 * Tweak the ring switching logic to reduce use of semaphores on
599   Sandybridge and later - a delicate balancing act between trying to
600   use the faster engine for the task and preventing concurrent use of
601   multiple rings.
602
603
604Release 2.21.14 (2013-08-04)
605============================
606A few updates for the top-end Haswell systems (notably GT3 and GT3e
607systems), in particular, setting the appropriate thread counts for the
608larger GPU and handling a few scaling issues with the emedded DRAM. More
609important than the performance tuning though is a fix for a critical
610issue encountered by a few people on gen4/gen5 systems.
611
612 * Initial performance tuning for HSW:GT3e
613
614 * Fix a crash with Planetary Annihilation
615
616 * Disable triple buffering for compositors
617
618 * Reserve space in the exec buffer array for deferred VBOs.
619   https://bugs.freedesktop.org/show_bug.cgi?id=67504
620
621
622Release 2.21.13 (2013-07-27)
623============================
624A minor release to repair the build for non-Linux systems, and to undo a
625few more regressions.
626
627 * Avoid potential memory corruption with allocations of very small
628   depth 1 bitmaps.
629
630 * Fix source clipping whilst computing copy extents and exposures.
631   [Regression from 2.21.12]
632   https://bugs.freedesktop.org/show_bug.cgi?id=66970
633
634 * Use /proc/cpuinfo if cpuid4 is not available (old hardware or old
635   gcc), and repair the build on systems without cpuid.
636   [Regression from 2.21.12]
637
638 * Fix performance regression on Ironlake from inadvertently flushing
639   after every operation.
640   [Regression from 2.21.11]
641   https://bugs.freedesktop.org/show_bug.cgi?id=67157
642
643 * Fix conflict handling when probing initial KMS configuration for
644   ZaphodHeads.
645   [Regression from 2.21.11]
646   https://bugs.freedesktop.org/show_bug.cgi?id=67176
647
648 * Stop being overly restrictive and rejecting stale DRI2 buffers.
649   (DRI2 is inherently racy in that the client may have completed and
650   submitted rendering to buffers that are now invalid on the server.
651   This race is magnified by bugs within the Xserver where it forgets
652   to notify the DRI2 clients of certain invalidation events.)
653   https://bugs.freedesktop.org/show_bug.cgi?id=67210
654
655 * Handle a failure to use a GTT mmap for a pixmap upload and try
656   an alternative method before giving up.
657
658
659Release 2.21.12 (2013-07-14)
660============================
661In this release, we clear up the teething troubles from preserving the KMS
662configuration, notably external connections on Haswell and plugging in new
663outputs after startup were broken. Besides these regression fixes, there
664are a couple of fixes for some long standing issues, such as incorrect
665rendering on gen2, an infinite loop with very, very large pixmaps and a
666slight improvement to the tempermental gen4.
667
668 * Allow untiled scanouts again (required for large extended desktops on
669   gen2 and gen3).
670   [Regression from 2.21.11]
671
672 * Use the correct count of the number of dirty damage boxes for the quick
673   check on whether the existing damage contains the requested area.
674   The danger is that we may get a false result and skip migration and so
675   cause pixmap corruption (in the unlikely event that the application
676   frequently causes fallbacks).
677   [Regression from 2.21.11]
678   https://bugs.freedesktop.org/show_bug.cgi?id=66430
679
680 * Fix initial connection probing for multi-function encoders, such as the
681   external connections on Haswell.
682   [Regression from 2.21.11, initial connection probing]
683   https://bugs.freedesktop.org/show_bug.cgi?id=66488
684
685 * Fix gen2 rendercopy into a8 surfaces, for example, glyph uploads into
686   the glyph cache.
687
688 * Fix detection of user overrides for initial connection configuration.
689   The code used the xorg-server-1.15 values, having missed the
690   introduction of ZoomModes into that release.
691   [Regression from 2.21.11, initial connection probing]
692
693 * Always initialise the gamma ramp, even on unconnected CRTCs.
694   https://bugs.freedesktop.org/show_bug.cgi?id=66563
695   [Regression from 2.21.11, initial connection probing]
696
697 * Some more tuning of the gen4 vertex corruption workaround. The root
698   cause behind the GPU using incorrect texture coordinates is still
699   not solved, but by reducing the maximum number or rectangles in
700   flight through the GPU we reduce the likelihood of corruption.
701
702 * Fix compilation with gcc-4.5
703   [Regression from 2.21.11]
704
705 * Avoid integer overflow when performing tiled uploads and operations
706   on very large (>28k pixels wide or tall pixmaps)
707   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1200766
708
709
710Release 2.21.11 (2013-06-30)
711============================
712An eventful week. What started with a regression with some builds of
713firefox on some machines lead ultimately to the discovery of an older
714kernel bug. Aside from the work to fix the image bug and a few other
715older bugs that were reported and resolved this week, there is also a
716(hopefully) subtle change to the initial configuration of displays. In
717the absence of user overrides in xorg.conf, the DDX will try to preserve
718the same display configuration as used by the kernel, which hopefully
719will be the same configuration as setup by the BIOS. The result should
720be a boot sequence that does not resize at all (aka fastboot) - until
721the display manager takes over and loads a completely different
722configuration!
723
724 * Add reference counting of drmMaster for ZaphodHeads
725   https://bugs.freedesktop.org/show_bug.cgi?id=66041
726
727 * Add a GPU flush before changing blend modes on Ironlake
728   https://bugs.freedesktop.org/show_bug.cgi?id=51422
729
730 * Fix occasional missing images for inplace uploads
731   [regression from 2.21.10]
732   https://bugs.freedesktop.org/show_bug.cgi?id=66059
733
734 * Add missing utility files to the tarball and remove a few unused ones
735
736 * Initialise PolyPoint operand state before calling miWideDash
737   https://bugs.freedesktop.org/show_bug.cgi?id=66104
738
739 * Fix redirection handling for rendering into large surfaces
740   https://bugs.freedesktop.org/show_bug.cgi?id=66168
741   https://bugs.freedesktop.org/show_bug.cgi?id=66249
742
743 * Fix compilation of UXA with xorg-xserver < 1.10
744   [regression from 2.20.0]
745
746 * Fix consideration of gradients for deciding when to migrate render
747   operations
748   [performance regression from 2.21.10, the bug itself is older]
749   https://bugs.freedesktop.org/show_bug.cgi?id=66297
750
751Also fixed this week was:
752
753commit 22fd5ca947b58901927d100d2b1aa0f1672b3435
754Author: Chris Wilson <chris@chris-wilson.co.uk>
755Date:   Fri Jun 28 16:54:08 2013 +0100
756
757    drm/i915: Only clear write-domains after a successful wait-seqno
758
759which affects kernels 3.7 - 3.10, coming to a stable kernel near you soon.
760
761
762Release 2.21.10 (2013-06-22)
763============================
764Fixes missing support for Xv (with the textured video adaptor) on
765Haswell, and an old bug with wide monitors on various generation. Along
766with a few other regression fixes and performance tweaks for various
767corner cases.
768
769 * Do not lose track of fast pageflips across mode changes
770   [regression from 2.20.8]
771
772 * Fix listing of Visuals for Xv
773   [regression from 2.21.8]
774   https://bugs.freedesktop.org/show_bug.cgi?id=65479
775
776 * Improve coherency of concurrent CPU accesses to a pixmap
777   https://bugs.freedesktop.org/show_bug.cgi?id=61628
778
779 * Set sampler swizzling for textured video on Haswell
780   https://bugs.freedesktop.org/show_bug.cgi?id=65699
781
782 * Apply scanout stride limits
783   https://bugs.freedesktop.org/show_bug.cgi?id=65099
784
785 * Undo the self-copy for cloned pixmaps for loimpress's animations
786   [regression from 2.21.7]
787   https://bugs.freedesktop.org/show_bug.cgi?id=65665
788
789
790Release 2.21.9 (2013-06-06)
791===========================
792Consolidating the copy-on-write support, hopefully cleaning up the last of
793the regressions.
794
795 * Restore vsync on textured videos.
796   [regression from 2.21.8]
797   https://bugs.freedesktop.org/show_bug.cgi?id=65048
798
799 * Fix incorrect ordering of possible_clones with certain outputs, which
800   can lead to attempting to incorrectly clone 2 outputs and failing to
801   light them up.
802   [regression from 2.20.10]
803
804 * Fix performance regression from not promoting large fills to the GPU
805   [regression from 2.21.7]
806
807 * Undo the pixmap clone before performing a DRI2CopyRegion
808   [regression from 2.21.7]
809   https://bugs.freedesktop.org/show_bug.cgi?id=65250
810
811
812Release 2.21.8 (2013-05-27)
813===========================
814A quick release to cleanup a few regressions from the introduction of
815copy-on-write support, notably hitting wine applications and a memory
816leak for firefox.
817
818 * Only mark a PolyFillRect operation as replacing if it is unclipped
819   https://bugs.freedesktop.org/show_bug.cgi?id=64841
820
821 * Prevent potential NULL dereference of damage when checking COW support
822
823 * Fix invalidation of clone after dirtying the pixmap via the CPU
824
825 * Prevent discarding an operation before requiring it for a fallback
826
827 * Fix memory leak from replacing the clone under certain circumstances
828   https://bugs.freedesktop.org/show_bug.cgi?id=64978
829
830
831Release 2.21.7 (2013-05-21)
832===========================
833A couple of weeks turned into a month and a couple of weeks... Amidst
834the usual bug fixes, we have added the complete set of Haswell PCI IDs -
835hopefully future proofing ourselves against being surprised by new
836products. We can also now use the correct term for the top of the range
837Haswell variants, GT3.
838
839 * Fix several assertion failures hit by Jiri Slaby.
840
841 * Allow XvMC to also target overlay/sprite planes.
842
843 * Throw in a paranoid MI_FLUSH between BLT and RENDER operations on
844   Ironlake.
845   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1168066
846
847 * Prevent reuse of old framebuffers after a resize.
848   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1157678
849
850 * Fix compilation with --enable-valgrind and no --enable-debug
851
852 * Improve partial migration of render sources.
853
854 * Fix origin of trapezoids.
855   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1178020
856
857 * Introduce copy-on-write support for cloning pixmaps. The ultimate
858   goal here is to efficiently support the TearFree mode of operation,
859   but this provides immediate benefits with firefox - most importantly
860   because of the inefficient way it now implements scrolling.
861
862
863Release 2.21.6 (2013-04-06)
864===========================
865A surprising highlight of this release is a little refresh to the KMS
866support for OpenBSD. OpenBSD now has its own KMS implementation which is
867mostly compatible with the interface in Linux, with one or two tweaks
868supplied by Mark Kettenis. This release continues to cleanup behaviour
869for Haswell.
870
871 * Workaround a failure by the xserver to invalidate DRI buffers
872   following a pixmap change for XComposite redirection.
873   https://bugs.freedesktop.org/show_bug.cgi?id=62614
874
875 * Fix computation of clip extents for stippling
876   https://bugs.freedesktop.org/show_bug.cgi?id=62618
877
878 * Support KMS on OpenBSD, by Mark Kettenis
879
880 * Clean up sockets upon CloseScreen (making ourselves better behaved
881   for muxed setups).
882
883 * Fix the tests for AVX/AVX2 support in CPUID and remember to check for
884   OS support as well.
885
886 * Report a monotonic UST value for undisplayed drawables rather than 0
887   by Daniel Kurtz
888
889 * Fix video playback on gen4 through a complex clip (more gen4 GPU woes)
890   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1162046
891
892
893Release 2.21.5 (2013-03-21)
894===========================
895Haswell reintroduces a command to load the scanline window from the
896command stream and so requires its own specialised wait-for-vsync routine
897- failure to do so was then causing hangs when trying to do tearfree video
898or use a compositor.
899
900 * Prevent buffer leak if a non-fullscreen Window is closed with multiple
901   pending swap events.
902
903 * Fix offset transformation for fallback gradient paths.
904   https://bugs.freedesktop.org/show_bug.cgi?id=62198
905
906 * Prevent Glamor from crashing if misconfigured.
907   Thanks to Michel Dänzer.
908
909 * Prevent UXA from crashing if torn down during PreInit.
910   Thanks to Aaron Plattner.
911
912 * Prevent miscompilation with different functional units having different
913   compiler flags. Some functions were expected to be inlined and so
914   recompiled with the current target. However, some compilers were
915   choosing to emit subroutine calls instead without noticing that the
916   ABI was different between the caller and callee - causing corruption.
917   https://bugs.freedesktop.org/show_bug.cgi?id=62198
918
919 * Fix rendering of CompositeTriFan with recent Xorg.
920
921 * Apply the video src-offset fix highlighted in the last release!
922   A typo prevented the fix from working for gen4+.
923   https://bugs.freedesktop.org/show_bug.cgi?id=62343
924
925 * Fix rendering of multiple glyphs to very large destination surfaces
926   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1156387
927
928 * Fix scanline waits for Haswell
929   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1156679
930
931
932Release 2.21.4 (2013-03-11)
933===========================
934More bugs, more fixes, more releases. A minor new feature being introduced
935is the runtime detection of CPU instructions sets along with specialised
936paths to take advantage of the available CPU.
937
938 * Honour LinearFramebuffer for clumsy PowerXpress integration
939
940 * Disable read-read optimisations of mappings whilst the root
941   cause of a corruption issue remains elusive.
942   https://bugs.freedesktop.org/show_bug.cgi?id=61628
943
944 * Disable 8-bpp framebuffers in UXA as the regressions therein
945   remain unfixed.
946
947 * Restart vertex checks after lock contention
948   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1124576
949
950 * Handle Screen Pixmap recreation whilst exported via PRIME.
951
952 * Correct application of scale factors to video source offsets.
953   https://bugs.freedesktop.org/show_bug.cgi?id=61610
954
955 * Chain up CloseScreen so that resources are actually freed across regen
956   https://bugs.freedesktop.org/show_bug.cgi?id=56608
957
958 * Fix Haswell CRW PCI-IDs
959
960 * Handle batch submission failure during DRI copies
961   https://bugs.freedesktop.org/show_bug.cgi?id=61708
962
963 * Probe for kernel support of requested Screen depth
964   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1135403
965
966 * Correct GPU limits for early gen2 and gen3 architectures and prevent
967   an infinite recursion for particular image sizes.
968
969
970Release 2.21.3 (2013-02-20)
971===========================
972A few minor bugfixes, another point release.
973
974 * Fix tracking of DRI pixmaps and their backing bo across reparenting. If
975   we tried to execute a SwapBuffers after a Window was reparented, but
976   before the DRI client has updated its references, then we would end up
977   manipulating an exported pixmap without a flush flag set. In the worst
978   case, this would culminate in a segfault in the driver.
979   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1127497
980
981 * Restore the gen4 workarounds for flickering rendering - a few cases still
982   remain, as the root cause persists.
983   https://bugs.freedesktop.org/show_bug.cgi?id=60402
984
985 * Double check that the device has KMS enabled before claiming. This allows
986   X to gracefully fallback to VESA/fbdev rather than bailing out.
987   https://bugs.freedesktop.org/show_bug.cgi?id=60987
988
989 * Fix the UXA render programs for projective transforms on Ivybridge.
990
991
992Release 2.21.2 (2013-02-10)
993===========================
994Pass the brown paper bags, I need half a dozen or so. That seemingly
995innocuous build fix with xorg-1.13 happened to have the little side-effect
996of breaking glyph rendering with xorg-1.12 and older on 64-bit machines.
997
998
999Release 2.21.1 (2013-02-10)
1000===========================
1001A fix for a potential GPU hang on 945gm (GMA950) and earlier chipsets,
1002along with backporting SNA to the packages found in stable distributions
1003like Debian 6.0 (Squeeze).
1004
1005 * Cleanup compilation warnings from deblint, thanks to Paul Menzel
1006
1007 * Minor build improvements by Damien Lespiau.
1008
1009 * Disable generating span geometry for non-rectilinear spans on gen4
1010   in order to work around and prevent one class of render corruption.
1011
1012 * Prevent cache thrashing and severe performance degradation on LLC
1013   machines for streaming texture updates. However, note the effect was
1014   only observed on just one particular laptop.
1015
1016 * Fix alignment of subsurface proxies for old chipsets.
1017   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1120108
1018
1019 * Repair build against Xserver-1.6 and contemporary packages.
1020
1021
1022Release 2.21.0 (2013-02-01)
1023===========================
1024A few new features:
1025
1026 * Enable render acceleration for Haswell GT1/GT2.
1027
1028 * Enable multi-threaded rasterisation of trapezoids and fallback composition
1029
1030 * Utilise a new kernel interface (v3.9) for processing relocations
1031
1032along with a few older features from the 2.20.x series:
1033
1034 * PRIME support for hotplug GPUs and hybrid systems
1035
1036 * Support for IvyBridge GT1 machines, aka HD2500 graphics.
1037
1038 * Stable 830gm/845g support, at last!
1039
1040As usual we have a large number of bug fixes since the last release:
1041
1042 * Prevent a stray relocation being left after a buffer is removed from
1043   a batch, leading to GPU hangs.
1044
1045 * Make the driver more robust against its own failures to submit batches
1046   by falling back to software rendering.
1047
1048 * Fix emission of scanline waits for secondary pipes on gen6/7. Otherwise
1049   you may encounter GPU hangs in MI_WAIT_FOR_EVENT.
1050
1051 * Fix a missing corner pixel when drawing rectangles with PolyLines
1052   https://bugs.freedesktop.org/show_bug.cgi?id=55484
1053
1054 * Don't try to use Y-tiling colour buffers with mesa/i915c as mesa
1055   doesn't support them and wil fallback to software rendering
1056
1057 * Ensure that any cached mmaps are invalidated for a SwapBuffers
1058   https://bugs.freedesktop.org/show_bug.cgi?id=60042
1059
1060 * Correctly handle the composition of rotated displays too large for the
1061   3D pipeline
1062   https://bugs.freedesktop.org/show_bug.cgi?id=60124
1063
1064 * Fix the computation of the planar video frame size
1065   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1104180
1066
1067
1068Release 2.20.19 (2013-01-20)
1069============================
1070A quick release as the last broke USB DisplayLink slave outputs badly. The
1071performance of those displays was unusable due to an inadvertent change that
1072caused us to flush the entire scanout over the USB for every drawing
1073operation.
1074
1075 * Implement the GNOME Build API. A couple of minor changes to make
1076   integrators and distributors lives a little easier, or at least more
1077   consistent.
1078
1079 * Correctly offset inplace trapezoids for subwindows, such as the GTK+
1080   close button after it has a background image uploaded.
1081
1082 * Explicitly prevent ring-switching for synchronized rendering to
1083   scanouts (for vsync).
1084
1085 * Clip dirty region to slave pixmaps (otherwise UDL is nigh unusuable)
1086   https://bugs.freedesktop.org/show_bug.cgi?id=59539
1087
1088
1089Release 2.20.18 (2013-01-16)
1090============================
1091A bunch of miscellaneous fixes for assertion failures and various
1092performance regressions when mixing new methods for offloads, along with
1093a couple of improvements for rendering with gen4.
1094
1095 * Remove use of packed unnormalized texture coordinates on gen4/5 as
1096   these GPUs do not support unnormalized coordinates in the sampler.
1097
1098 * Remove dependency upon x86 asm for cross-building to unsupported
1099   architectures.
1100   https://bugs.gentoo.org/show_bug.cgi?id=448570
1101
1102 * Apply damage around PRIME updates in the correct order.
1103
1104 * Correctly read the initial backlight level for when the user
1105   overrides UXA's choice of backlight controller.
1106
1107 * Throttle UXA and prevent it queuing work much faster than the GPU can
1108   complete it. This realised itself in impossible performance figures and
1109   the entire display freezing for several seconds whlist the GPU caught
1110   up. One side effect is that it also caused the DDX to consume more
1111   memory than was required as it could not recycle buffers quick
1112   enough, and in some cases this produces a marked improvement in
1113   performance. Also note on gen2/3 this requires a new libdrm [2.4.41]
1114   in order to prevent a bug causing the DDX to fallback to swrast.
1115
1116Release 2.20.17 (2012-12-26)
1117============================
1118A minor update to prepare for co-operating with the kernel over managing
1119stability on 830gm/845g. On this pair of chipsets, the kernel will perform
1120an extra copy of the batchbuffer into reserved memory, which prevents them
1121from randomly dying. However, that extra copy does have a noticeable
1122impact upon throughput, so we also have a mechanism for userspace to
1123opt-out of the kernel workaround and take responsibility for ensuring its
1124batches are coherent.
1125
1126 * Build fixes against xorg-1.14
1127   https://bugs.freedesktop.org/show_bug.cgi?id=58552
1128   https://bugs.freedesktop.org/show_bug.cgi?id=58406
1129
1130 * Fixed the origin of cropped (textured) video windows (Xv and XvMC)
1131   https://bugs.freedesktop.org/show_bug.cgi?id=23033
1132
1133 * Fix potential corruption when using images larger than ~1GiB
1134
1135
1136Release 2.20.16 (2012-12-15)
1137============================
1138Rejoice! We have found a trick to make 830gm/845g stable at long last.
1139Ever since the switch to GEM and dynamic video memory, those early
1140second generation chipsets have been plagued by instability. The lack of
1141flushing cachelines from the CPU to GMCH was eventually solved by using
1142an undocmented bit, but 830/845 were still hanging under memory pressure.
1143These deaths were all due to garbage finding its way into the command
1144streamer, and they go away if we take a leaf out of the original driver
1145and never reuse those pages for anything else. So for the first time
1146ever, I have been able to complete running the test suite on an 845g,
1147even whilst thrashing the page and buffer caches!
1148
1149 * Run the SF stage as single-threaded on gen4 to workaround a few issues
1150   https://bugs.freedesktop.org/show_bug.cgi?id=57410
1151
1152 * Keep the scanout SURFACE_STATE separate to avoid overriding its
1153   memory access control on gen6/7 (i.e. writes to the scanout need to
1154   be kept out of the render cache)
1155
1156 * Tune batch flushing after an operation to an exported surface under a
1157   compositor.
1158
1159 * Make sure the source is on the CPU for inplace composition of trapezoids
1160   using the CPU
1161   https://bugs.freedesktop.org/show_bug.cgi?id=56825
1162
1163 * Immediately flush in the block hander after a split batch to reduce
1164   latency between the two halves of an operation.
1165   https://bugs.freedesktop.org/show_bug.cgi?id=51718
1166
1167 * Install a fallback config if we fail to install the desired config
1168   at VT switch (i.e. booting, after resume with 3 incompatible pipes on
1169   Ivybridge)
1170
1171 * Pin batches to avoid CS incoherence on 830/845
1172   https://bugs.freedesktop.org/show_bug.cgi?id=26345
1173
1174
1175Release 2.20.15 (2012-12-03)
1176============================
1177And lo, enabling more of the common acceleration paths for gen4 revealed
1178another lurking bug - something is wrong with how we prepare Y-tiling
1179surfaces for rendering. For the time being, we can surreptiously disable
1180them for gen4 and avoid hitting GPU hangs.
1181
1182 * Avoid clobbering the render state after failing to convert the
1183   operation to use the blitter.
1184   https://bugs.freedesktop.org/show_bug.cgi?id=57601
1185
1186 * Disable shadow tracking upon server regeneration, and so fix a crash
1187   if you restart the server whilst a RandR transform (e.g. rotation) is
1188   in effect.
1189   https://bugs.freedesktop.org/show_bug.cgi?id=52255
1190   https://bugs.freedesktop.org/show_bug.cgi?id=56608
1191
1192
1193Release 2.20.14 (2012-11-26)
1194============================
1195The highlight of this release is gen4, from 965g to gm45. Quite an old
1196bug surfaced in the shader assembly, sparking a chance to review a few
1197design choices within that backend and experiment on fresh ways to
1198workaround the remaining issues.
1199
1200 * Avoid using inplace XOR'ed uploads for very large buffers
1201   https://bugs.freedesktop.org/show_bug.cgi?id=57031
1202
1203 * Fix the gen4/5 opacity shader
1204   https://bugs.freedesktop.org/show_bug.cgi?id=57054
1205
1206 * Queue a pending vblank request after flip completion
1207   https://bugs.freedesktop.org/show_bug.cgi?id=56423
1208
1209 * Avoid migrating an uninitialised pixmap for use as a render source
1210   https://bugs.freedesktop.org/show_bug.cgi?id=47597
1211
1212 * Improve handing of texture fallbacks for 830/845.
1213   https://bugs.freedesktop.org/show_bug.cgi?id=57392
1214
1215
1216Release 2.20.13 (2012-11-11)
1217============================
1218Nothing but bug fixes. Many thanks to everyone who took the time to
1219report their issues, and for their help in improving the driver.
1220
1221 * Sanity check the platform probe points to our expected i915 device
1222   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1069031
1223
1224 * Prevent 16-bit overflow for computing the sample area to upload of
1225   sources for render operations
1226   https://bugs.freedesktop.org/show_bug.cgi?id=56324
1227
1228 * Clamp the drawable box for migration to prevent 16-bit overflow
1229   https://bugs.freedesktop.org/show_bug.cgi?id=56591
1230
1231 * Disable RandR hotplug events if Xinerama is enabled and thereby prevent
1232   a crash upon hotplug
1233   https://bugs.freedesktop.org/show_bug.cgi?id=55260
1234
1235 * Call ValidatePicture before attempting to flatten the alphamaps
1236   https://bugs.freedesktop.org/show_bug.cgi?id=56367
1237
1238 * Clip the trapezoid correctly if it ends on the boundary pixel
1239   https://bugs.freedesktop.org/show_bug.cgi?id=56395
1240
1241 * Make sure the pipeline choice is propagated to the scanline wait
1242   across a batch flush
1243   https://bugs.freedesktop.org/show_bug.cgi?id=47597
1244
1245 * Set the valid drawable box when choosing placement of BLT composite ops
1246   https://bugs.freedesktop.org/show_bug.cgi?id=47597
1247
1248 * Prevent use-after-free when promoting a partial-GPU bo to a full-GPU bo
1249   https://bugs.freedesktop.org/show_bug.cgi?id=56591
1250
1251 * gen4 opacity spans require the per-rectangle workaround
1252   https://bugs.freedesktop.org/show_bug.cgi?id=55500
1253
1254 * Prevent use of invalid damage pointers when redirecting rendering
1255   https://bugs.freedesktop.org/show_bug.cgi?id=56785
1256
1257
1258Release 2.20.12 (2012-10-20)
1259============================
1260More bug reports, more bug fixes! Perhaps the headline feature is
1261that with a secure batches, coming to a 3.8 kernel near you, we may
1262finally have the ability to perform updates to the scanout synchronized
1263to the refresh rate on later SandyBridge and IvyBridge chipsets. It comes
1264at quite a power cost as we need to keep the GPU out of its power saving
1265modes, but it should allow legacy vsync to function at last. But this
1266should allow us to address a longstanding issue with tearing on
1267SandyBridge+.
1268
1269 * Fix component-alpha rendering on IvyBridge, for example subpixel
1270   antialiased glyphs.
1271   https://bugs.freedesktop.org/show_bug.cgi?id=56037
1272
1273 * Flush before some "pipelined" state changes on gen4. The evidence is
1274   that the same flushes as required on gen5+ are also required for gen4.
1275   https://bugs.freedesktop.org/show_bug.cgi?id=55627
1276
1277 * Prevent a potential crash when forcing a stall on a busy CPU bo
1278   https://bugs.freedesktop.org/show_bug.cgi?id=56180
1279
1280[Release 2.20.11 contained a typo causing UXA to fail immediately.]
1281
1282Release 2.20.10 (2012-10-14)
1283============================
1284The last couple of weeks have been fairly retrospective, a dive into
1285prehistory tidying up the earlier generations which lay languishing as
1286the core progressed and lead to a number of annoying core bugs being
1287fixed.
1288
1289 * Release DRM master earlier during shutdown so switching between
1290   multiple X servers works automatically.
1291   https://bugs.freedesktop.org/show_bug.cgi?id=55446
1292
1293 * Suppress error propagation from DRI2GetMSC and behave as if the pipe
1294   was simply off to avoid unexpected errors in the clients
1295   https://bugs.freedesktop.org/show_bug.cgi?id=55395
1296
1297 * A few fixes to i8xx batch emission, ensuring that the GPU is always
1298   in a valid state.
1299   https://bugs.freedesktop.org/show_bug.cgi?id=55455
1300
1301 * Prevent a use-after-free during UXA shutdown due to inspecting
1302   the glamor flags after the glamor interface had been freed.
1303
1304 * Prevent a crash combining TearFree and rotations.
1305   https://bugs.freedesktop.org/show_bug.cgi?id=55527
1306
1307 * Correct a missing damage upload along PutImage after using the CPU bo
1308   as a source for the GPU.
1309   https://bugs.freedesktop.org/show_bug.cgi?id=55508
1310
1311 * Fix compilation for older glibc without O_CLOEXEC
1312   https://bugs.freedesktop.org/show_bug.cgi?id=55577
1313
1314 * Fix out-of-tree builds failing to recompile the gen4 assemblies
1315   https://bugs.freedesktop.org/show_bug.cgi?id=55645
1316
1317 * Fix non-standard build host configuration handling for intel-gen4asm
1318   https://bugs.freedesktop.org/show_bug.cgi?id=55646
1319
1320 * Fix a potential batch buffer overflow when replacing the last BLT fill
1321   operation with a copy
1322   https://bugs.freedesktop.org/show_bug.cgi?id=55700
1323
1324 * Flush the render pipeline more frequently on Ironlake as not all
1325   pipelined state changes are.
1326   https://bugs.freedesktop.org/show_bug.cgi?id=51422
1327
1328 * Detect when we need to read the destination for the background raster
1329   op during fallbacks.
1330   https://bugs.freedesktop.org/show_bug.cgi?id=55810
1331
1332 * Avoid a potential deference of an invalid CPU mmap after doing an
1333   inplace tiled upload.
1334   https://bugs.freedesktop.org/show_bug.cgi?id=55812
1335
1336 * Prevent sign extension when packing the upload data for CopyPlane
1337   https://bugs.freedesktop.org/show_bug.cgi?id=55823
1338
1339 * Fix some render corruption with a UDL slave output and pageflipping
1340
1341
1342Release 2.20.9 (2012-09-29)
1343===========================
1344And so it came to pass that a critical bug was uncovered in UXA. The
1345kernel does not like to pageflip when the pipe is off, yet due to the
1346delayed nature of a pageflip and the relaxed checking performed by UXA,
1347we could request a pageflip after turning off the display (DPMS). The
1348kernel rejected that pageflip and the error handling path failed to
1349restore sanity, and when the screen came back it was stuck on the image
1350seen before it went to sleep. (Note that there are also some related
1351kernel bugs, but this update should prevent the most conspicious of the
1352freezes.) Many thanks to Timo Aaltonen for his efforts in tracking down
1353the issue.
1354
1355In other news:
1356
1357  * Prepare for xorg-1.14, the api is being tweaked again.
1358
1359  * Handle early FreeScreen in UXA.
1360    https://bugs.freedesktop.org/show_bug.cgi?id=55346
1361
1362  * Reenable XvMC support
1363
1364  * Do not replace the GPU bo when uploading into the shadow/CPU copy
1365    https://bugs.freedesktop.org/show_bug.cgi?id=54978
1366
1367  * Fix use of an uninitialised GC when drawing glyphs to a depth=1 pixmap
1368
1369
1370Release 2.20.8 (2012-09-16)
1371===========================
1372Another new small feature, another new release. And a few more bugs
1373fixed as well! But what is this new feature, do I hear you ask? Why, it
1374is nothing less than enabling the ValleyView SDV! The lucky person to
1375have their hands on one will now be able to enjoy X in full TechniColor.
1376
1377For the rest of us, a few more bugs were fixed with interesting
1378combinations of software and rendering patterns.
1379
1380 * Add an extra layer of defence against trying to use a non-GEM device
1381   with UXA. This should already be taken care of with the new probe,
1382   but the extra sanity check already existed in the code but was doing
1383   nothing.
1384   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/962892
1385
1386 * Fix computation of valid CRTCs bitmask for ZaphodHead
1387   Some systems can only handle certain outputs on certain pipes
1388   (Screens), and this information was not being propagated through to X
1389   and randr correctly, breaking valid configurations and not detecting
1390   invalid configurations correctly.
1391
1392 * Disable global glyph caching with ZaphodHeads.
1393   The glyph privates need to be Screen private, but at the moment are
1394   global leading to conflicts and invalid rendering with multiple heads.
1395   https://bugs.freedesktop.org/show_bug.cgi?id=54707
1396
1397 * Prevent direct read back of unmappable buffers
1398   https://bugs.freedesktop.org/show_bug.cgi?id=54808
1399
1400 * Tile large uploads whilst replacing the alpha channel
1401   https://bugs.freedesktop.org/show_bug.cgi?id=54808
1402
1403 * Correct the source offset when converting a RENDER composite
1404   operation into a BLT composite operation.
1405   https://bugs.freedesktop.org/show_bug.cgi?id=54868
1406
1407 * Correct a minor typo in flattening alphamaps which caused the
1408   replacement pixmap to often end up with a height of zero.
1409   https://bugs.gentoo.org/show_bug.cgi?id=434860
1410
1411 * Don't discard the CPU damage if only part of is being replaced.
1412   This is most evident when combining software renders like Opera with
1413   accelerated rendering.
1414   https://bugs.freedesktop.org/show_bug.cgi?id=54937
1415
1416Release 2.20.7 (2012-09-08)
1417===========================
1418Continuing the flurry of releases, this week saw the release of
1419xorg-1.13 bringing with it the first steps towards hotpluggable gpu
1420support and the ability to offload DRI rendering onto other GPUs. In
1421light of that advance, this release includes the support code by Dave
1422Airlie to integrate PRIME into xf86-video-intel and make those new
1423features available.
1424
1425Aside from landing PRIME, a few other bugs were fixed since 2.20.6:
1426
1427 * Fix framebuffer leak on server regeneration
1428
1429 * Fix texture cache flushing on IvyBridge and Kwin with plastique
1430   https://bugs.freedesktop.org/show_bug.cgi?id=54488
1431
1432 * Redirect large solid fills on SandyBridge+. By large I mean greater
1433   than 8192 or 16384 pixels on SandyBridge and IvyBridge respectively.
1434   https://bugs.freedesktop.org/show_bug.cgi?id=54134
1435
1436 * Fix up backlight option handling in the manpage and uxa.
1437   https://bugs.freedesktop.org/show_bug.cgi?id=54397
1438
1439 * Unbreak ZaphodHeads.
1440   https://bugs.freedesktop.org/show_bug.cgi?id=52438
1441
1442Release 2.20.6 (2012-09-02)
1443===========================
1444A serious bug that caused a crash on SandyBridge and IvyBridge when
1445mixing CPU and GPU operations on the same buffer, and an annoyance from
1446bad scheduling of windowed swapbuffer updates causing low framerates and
1447jitter. Plus the usual smattering of assertion fixes and a long standing
1448issue with incoherent page access to a streaming buffer.
1449
1450 * Low frame rates in Blobby Valley when "fullscreen"
1451   https://bugs.freedesktop.org/show_bug.cgi?id=54274
1452
1453 * Incoherent concurrent access with the CPU and GPU
1454   https://bugs.freedesktop.org/show_bug.cgi?id=51422
1455   https://bugs.freedesktop.org/show_bug.cgi?id=52299
1456
1457 * Add Option "Backlight" to override automatic selection of the
1458   backlight interface.
1459
1460 * Avoid overwriting the composite operation info when testing if we
1461   can transfer the operation to the BLT.
1462
1463Release 2.20.5 (2012-08-26)
1464===========================
1465Another silly bug found, another small bugfix release. The goal was for
1466the driver to bind to all Intel devices supported by the kernel.
1467Unfortunately we were too successful and started claiming Pouslbo,
1468Medfield and Cedarview devices which are still encumbered by propietary
1469IP and not supported by this driver.
1470
1471Bugs fixed since 2.20.4:
1472
1473 * Only bind to Intel devices using the i915 kernel module
1474
1475 * Regression in the bitmap-to-region code, e.g. icewm window buttons
1476   https://bugs.freedesktop.org/show_bug.cgi?id=53699
1477
1478Release 2.20.4 (2012-08-18)
1479===========================
1480Continuing the small bugfix releases, the only real feature is initial
1481enabling for Haswell for the purpose of rendering verification and
1482validation - by no means is it complete!
1483
1484Bugs fixed since 2.20.3:
1485
1486  * Some potential errors along failure paths found by a static analyser
1487  with the help of Zdenek Kablac.
1488
1489  * Eliminate zero-sized rectangles from PolyFillRectangles as the code
1490  assumes that they did not exist and so caused corruption.
1491
1492  * Remove the UXA warning for failing to tile the front buffer if it is
1493  disallowed by hardware, and so expected.
1494
1495  * Fix the validation of the XV pipe parameter.
1496
1497  * Fix 8x8 tiled pattern fills
1498  https://bugs.freedesktop.org/show_bug.cgi?id=53353
1499
1500  * Fix compile failure when using --with-builderstring
1501
1502  * Restore w/a flush for gen4 fill/copy/video, fortunately rare
1503  operations as at least for fill/copy we prefer to use the BLT.
1504  https://bugs.freedesktop.org/show_bug.cgi?id=53119
1505
1506  * Restore preferred use of the RENDER ring for SNB+ DRI copies.
1507
1508Release 2.20.3 (2012-08-04)
1509===========================
1510Just a minor bugfix for gen4 chipsets (965gm, gm45 and friends) that
1511crept into 2.20.2. As an added bonus, the pessimistic workaround for a
1512GPU hang on gen4 has been relaxed and the shaders have been overhauled
1513which should pave the way to eliminating the last of the uncommon CPU
1514operations, along with immediately realising a small perforamnce
1515improvement.
1516
1517Bugs fixed since 2.20.2:
1518
1519 * Update DPMS bookkeeping after modeset
1520   https://bugs.freedesktop.org/show_bug.cgi?id=52142
1521
1522 * Avoid overlapping gpu/cpu damage after ignoring cpu damage in the
1523   consideration of placement for the operation.
1524
1525 * Enable acceleration by default on 830gm/845g. The GMCH on this pair
1526   of chipsets is notoriously incoherent, so the GPU is almost certainly
1527   going to hang at some point, though unlikely to hang the system and
1528   should automatically disable acceleration (and thence behave
1529   identically as if the acceleration was disabled from the start).
1530   Option "NoAccel" can be used to disable all 2D acceleration and
1531   Option "DRI" can be used to disable all 3D acceleration.
1532   https://bugs.freedesktop.org/show_bug.cgi?id=52624
1533
1534 * Fix vertex bookkeeping for gen4 that was causing corruption in the
1535   command stream.
1536
1537Release 2.20.2 (2012-07-27)
1538===========================
1539For the last 9 months, since 2.16.901, we have been shipping a driver that
1540does not work on IvyBridge GT1 systems (HD2500 graphics); we were telling
1541the GPU to use an invalid number of threads for the pixel shader and this
1542in turned caused the GPU to hang.
1543
1544Also fixed since the last release just a few days ago:
1545
1546* Support for the gmux backlight controller on Apple laptops
1547  https://bugs.freedesktop.org/show_bug.cgi?id=52423
1548
1549* Fix X -configure not to list this driver as matching any Intel device,
1550  just the VGA class devices will do!
1551
1552* A crash in SNA when repeatedly switching xrandr rotations
1553
1554* Corruption in SNA observed in kwin on IvyBridge
1555  https://bugs.freedesktop.org/show_bug.cgi?id=52473
1556
1557Release 2.20.1 (2012-07-22)
1558===========================
1559A week in, grab the brown paper bags, for it is time to reveal a couple
1560of critical bugs that spoilt the 2.20.0 release.
1561
1562Firstly we have the restoration of DRI for i810. I am sure that the
1563solitary user will be overjoyed in a couple of years when a new xserver
1564is forced upon him. That enjoyment will be short-lived when as no actual
1565acceleration remains, not even shadow, for the chipset.
1566
1567Perhaps a little more wildly felt, I hope!, will be that the SNA
1568fallbacks were broken on 64-bit machines if they required clipping. One
1569little misplaced cast of a pointer, and the screen is filled with
1570corruption.
1571
1572Among the other tweaks this week:
1573
1574* A bug affecting gen4 handling of trapezoids was fixed, and CPU
1575  overhead reduced.
1576  https://bugs.freedesktop.org/show_bug.cgi?id=52158
1577
1578* A fix for a bug causing corruption of a DRI2 unredirected client
1579  window that was resized whilst under a compositor.
1580
1581* Support for snoopable buffers on non-LLC architectures, coming to
1582  a future kernel. The aim to accelerate transfers between the CPU
1583  and the GPU, in particular to dramatically improve readback
1584  performance, and to further minimise clflushes.
1585
1586* Improvement to the composite performance on GT2 SandyBridge and
1587  IvyBridge devices, in particular the render copy is significantly
1588  improved.
1589
1590* Improved handling for when acceleration is disabled, including
1591  permitting DRI2 to remain supported even if the X server believes
1592  the GPU wedged.
1593
1594* Shadow support was dropped from UXA as it was neither complete nor
1595  correct, use SNA instead.
1596
1597Release 2.20.0 (2012-07-15)
1598===========================
1599First the big news, a new acceleration method that aims to be faster and
1600consume far less CPU than UXA is now available for selection at runtime.
1601This snazzy new architecture can be selected through use of
1602  Option "AccelMethod" "sna"
1603in your xorg.conf. Whilst it has been under development for some time, it
1604has not yet had the same degree of widespread testing of UXA, so tread
1605lightly. Try it and if you spot anything that can be improved, please do
1606report a bug.
1607
1608Otherwise we have the usual smattering of bug fixes for UXA:
1609
1610 * Use a white source whilst adding glyphs to the glyph mask
1611   (This fixes blank glyphs if using a font that mixes ARGB and A glyphs.)
1612
1613 * Avoid fallbacks for glyph-to-dst in e.g. gnome-terminal
1614   https://bugs.freedesktop.org/show_bug.cgi?id=50508
1615
1616 * Force unused outputs off when VT switching
1617   https://bugs.freedesktop.org/show_bug.cgi?id=50772
1618
1619 * Copy the fbcon across to the Screen pixmap at startup.
1620   (This patch has been kicking around in the distributions for years.)
1621
1622 * Many missed malloc failures checks and forgotten frees found by a static
1623   analyzer. Thanks Zdenek Kabelac!
1624
1625 * Leak of the back buffer when terminating an application after pageflipping
1626   https://bugs.freedesktop.org/show_bug.cgi?id=50670
1627
1628 * Double check that the pipe is on before emitting a WAIT_ON_EVENT.
1629   In conjunction with an uptodate kernel, this should eliminate any
1630   hangs when changing resolutions or adding/removing displays.
1631   https://bugs.freedesktop.org/show_bug.cgi?id=50668
1632
1633 * Update to new Xorg APIs. Future proofing for the next generation of
1634   hotplug Xorg display servers.
1635
1636Many thanks to everyone who has reported a bug and otherwise helped to
1637improve the driver.
1638
1639Release 2.19.0 (2012-04-29)
1640===========================
1641More stability fixes for UXA and support for another variant of IvyBridge.
1642Given the severity of the stability fixes, I strongly recommend everybody
1643to upgrade to 2.19.0.
1644
1645* Prevent waiting on scanlines whilst not in control of the VT and therefore
1646  whilst referencing foreign CRTC configurations.
1647
1648* Pixmap (and bo leak) during fallback glyph composition
1649
1650* Remove broken acceleration for rendering glyphs directly upon the
1651  destination pixmap, exposed by cairo-1.12.0 (and coincidentally fix
1652  another Pixmap leak upon fallback handling).
1653
1654* Add support for Ivy Bridge GT2 Server chipset [PCI id 0x016a]
1655
1656* Remove broken damage flushing with CompositeRectangles
1657  https://bugs.freedesktop.org/show_bug.cgi?id=32547
1658
1659* Fix crash upon server start with multiple monitors
1660  https://bugs.freedesktop.org/show_bug.cgi?id=47395
1661
1662* Fix composition issues resulting from overly aggressive Pixmap reuse
1663  https://bugs.freedesktop.org/show_bug.cgi?id=47345
1664
1665Release 2.18.0 (2012-02-24)
1666===========================
1667Time passes, a few more bugs have crept out of the woodwork that are a
1668compelling reason to update.
1669
1670Bugs fixed in this release (compared to 2.17.0)
1671-----------------------------------------------
1672
1673* Limit maximum object size so that all of the source, mask and
1674  destination can be mapped into the aperture simultaneously by basing the
1675  limit on the mappable aperture size rather than the size of the total
1676  GATT.
1677
1678* Incorrect clipping of polygons
1679  https://bugs.freedesktop.org/show_bug.cgi?id=43649
1680  Regression from 2.15.901
1681
1682* Limit number of VMA cached to avoid hitting the per-process VMA limit
1683  There still is a residual bug in that we seem to have so many objects
1684  floating around in the first place and that still leads to exhaustion
1685  of system limits.
1686  https://bugs.freedesktop.org/show_bug.cgi?id=43075
1687  https://bugs.freedesktop.org/show_bug.cgi?id=40066
1688
1689* Latency in processing user-input during continuous rendering
1690
1691Release 2.17.0 (2011-11-16)
1692==============================
1693A few months have passed, and we have accumulated a surprising number of
1694bug fixes. Oops! We would strongly encourage everyone to upgrade.
1695
1696Bugs fixed in this snapshot (compared to 2.16.902)
1697--------------------------------------------------
1698
1699* Video clobbering composite batch state
1700  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=635953
1701
1702Snapshot 2.16.902 (2011-11-13)
1703==============================
1704This is the second release candidate in preparation for the upcoming
17052.17.0 release. We will appreciate any feedback we can get from
1706testing of this snapshot, and thanks to everyone who tested and reported
1707upon issues in the last snapshot.
1708
1709A very old bug demonstrating that nobody has ever used alpha maps, and a
1710couple of more recent mistakes were found.
1711
1712Bugs fixed in this snapshot (compared to 2.16.901)
1713--------------------------------------------------
1714
1715 * Incorrect reuse of surface bindings within a batch for multiple formats
1716   https://bugs.freedesktop.org/show_bug.cgi?id=40926
1717
1718 * Nothing was rendered for text with procedural sources
1719   https://bugs.freedesktop.org/show_bug.cgi?id=31819
1720
1721 * Handle fallbacks involving alpha maps
1722
1723Snapshot 2.16.901 (2011-10-30)
1724==============================
1725This is the first release candidate in preparation for the upcoming
17262.17.0 release. We will appreciate any feedback we can get from
1727testing of this snapshot.
1728
1729The basis of this release are to push out a couple of critical stability
1730fixes for SandyBridge and IvyBridge. There is very little else to see
1731here. Those 200+ other commits do not exist...
1732
1733Bugs fixed in this snapshot (compared to 2.16.0)
1734--------------------------------------------------
1735
1736 * Workaround blitter hang on SandyBridge and IvyBridge
1737   https://bugzilla.kernel.org/show_bug.cgi?id=27892
1738   https://bugs.freedesktop.org/show_bug.cgi
1739
1740 * Workaround pipe control issues on SandyBridge
1741
1742 * Use correct maximum PS thread count on IvyBridge
1743
1744 * Protect against failed pixmap allocation for XV
1745   https://bugs.freedesktop.org/show_bug.cgi?id=40439
1746
1747Release 2.16.0 (2011-08-09)
1748==============================
1749A new quarter, a new release! The key feature of this release, looking past
1750the bug fixes, is the enabling of IvyBridge acceleration. We have also fixed
1751many bugs and graphical glitches and would encourage everyone to upgrade.
1752
1753Bugs fixed in this snapshot (compared to 2.15.901)
1754--------------------------------------------------
1755
1756* Build fix for xserver-1.7.7
1757
1758Snapshot 2.15.901 (2011-07-30)
1759==============================
1760This is the first release candidate in preparation for the upcoming
17612.16.0 release. We will appreciate any feedback we can get from
1762testing of this snapshot.
1763
1764The highlight of this snapshot is the full enabling of IvyBridge with
1765acceleration for Render and Xv, along with handling of the shared render
1766buffer allocations required for Mesa. Also of note is that deep-color
1767support is enabled (for all chipsets), which allows you to drive your
176830-bit monitor at its native colour depth.
1769
1770Bugs fixed in this snapshot (compared to 2.15.0)
1771--------------------------------------------------
1772
1773* Misuse of the Resource database causing crashes after DRI clients close
1774  https://bugs.freedesktop.org/show_bug.cgi?id=37700
1775
1776* Crash on large strings
1777  https://bugs.freedesktop.org/show_bug.cgi?id=36860
1778
1779* Incorrect rendering for some core drawing operations
1780  http://bugs.freedesktop.org/show_bug.cgi?id=28768
1781  http://bugs.freedesktop.org/show_bug.cgi?id=28798
1782  http://bugs.freedesktop.org/show_bug.cgi?id=28908
1783  http://bugs.freedesktop.org/show_bug.cgi?id=29401
1784
1785* Crash in Xv due to insufficient checks on batch space
1786  https://bugs.freedesktop.org/show_bug.cgi?id=36319
1787
1788Release 2.15.0 (2011-04-14)
1789==============================
1790We are pleased to announce this major release of the xf86-video-intel
1791driver, roughly on schedule at 3 months since 2.14.0. With the many bug
1792fixes in this release, we encourage everyone to upgrade to 2.15.
1793
1794The priority for this quarter has been simply to be unexciting and stabilise
1795the driver further, seeking to capitalise upon the improvements elsewhere
1796in the stack.
1797
1798Bugs fixed in this snapshot (compared to 2.14.903)
1799--------------------------------------------------
1800
1801* Turn off relaxed fencing by default for older chipsets
1802  This was continuing to destabilize those system, so for the release
1803  we disabled the feature. If you wish to help us debug this, you can
1804  re-enable the optimisation with Option "RelaxedFencing" "True".
1805  Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=36147
1806
1807* Build fix for xserver-1.7.7
1808
1809* KDE glitches on SNB
1810  [Technically fixed in the previous snapshot, but I'm really pleased
1811   that this got fixed in time for the release!]
1812  https://bugs.freedesktop.org/show_bug.cgi?id=35808
1813
1814
1815Snapshot 2.14.903 (2011-04-11)
1816==============================
1817This is the third release candidate in preparation for the upcoming
18182.15.0 release. We will appreciate any feedback we can get from
1819testing of this snapshot.
1820
1821There was a bit of churn since 2.14.902 as a potential fix for a
1822performance regression was tried but had to reverted when it was found to
1823cause glitches running Compiz on SandyBridge. Otherwise, there were just a
1824couple of fixes for building against old xservers and running on an
1825obscure chipset.
1826
1827Bugs fixed in this snapshot (compared to 2.14.902)
1828--------------------------------------------------
1829
1830* Prevent issuing an invalid scanline wait command
1831  https://bugs.freedesktop.org/show_bug.cgi?id=35576
1832
1833* The 946GZ in not a 945, but a 965.
1834  https://bugs.freedesktop.org/show_bug.cgi?id=35854
1835
1836* Fix tile sizes for gen2 (finally).
1837
1838* Allow building of recent dri2 changes against old xservers.
1839
1840
1841Snapshot 2.14.902 (2011-03-29)
1842==============================
1843This is the second release candidate in preparation for the upcoming
18442.15.0 release. We will appreciate any feedback we can get from
1845testing of this snapshot.
1846
1847As befits testing of release candidates, no major regression was found and
1848a couple more bugs have been fixed.
1849
1850Bugs fixed in this snapshot (compared to 2.14.901)
1851--------------------------------------------------
1852
1853* Clients disappearing with pending swaps
1854
1855* Incorrect clipping of Xv output on i915 across extended desktops
1856  https://bugs.freedesktop.org/show_bug.cgi?id=35346
1857
1858* Introduction of a LinearFramebuffer option. (Defaults to tiled for
1859  performance and power saving.)
1860
1861
1862Snapshot 2.14.901 (2011-03-02)
1863==============================
1864This is the first release candidate in preparation for the upcoming
18652.15.0 release. We will appreciate any feedback we can get from
1866testing of this snapshot.
1867
1868Still no further along my grandiose plans to improve Render performance,
1869aside from the performance tuning lower in the stack, instead we have had
1870a steady stream of bug fixes.
1871
1872Bugs fixed in this snapshot (compared to 2.14.0)
1873------------------------------------------------
1874
1875* Green pixels within partially off-screen video playback
1876  https://bugs.freedesktop.org/show_bug.cgi?id=24767
1877
1878* Defer creation of the glyph cache to generation startup
1879  https://bugs.freedesktop.org/show_bug.cgi?id=33412
1880
1881* Incorrect maximum addresses for video decoder state
1882  https://bugs.freedesktop.org/show_bug.cgi?id=34017
1883
1884* Failure to handle oversized temporary surfaces
1885  https://bugs.freedesktop.org/show_bug.cgi?id=34399
1886
1887* Relaxed tiling corruption on gen2
1888
1889* Crash when destroying a foreign DRI drawable
1890  https://bugs.freedesktop.org/show_bug.cgi?id=34787
1891
1892
1893Release 2.14.0 (2011-01-07)
1894===========================
1895We are pleased to announce this major release of the xf86-video-intel
1896driver, on schedule at 3 months since 2.13.0. With the many bug fixes
1897in this release, we encourage everyone to upgrade to 2.14.
1898
1899The most notable feature of this release is the inclusion of acceleration
1900support, both RENDER and XVIDEO, for Sandy Bridge; note that this requires
1901a 2.6.37 kernel. In improving the acceleration code paths to support Sandy
1902Bridge, we were also able to improve the performance of the general i965+
1903RENDER acceleration. For example on i3, glyph (aa10 and rgb10) performance
1904is about 3x faster, as is performing small copies between windows, with the
1905improvement becoming negligible once you hit around 100 pixels square.
1906
1907
1908Snapshot 2.13.903 (2010-01-04)
1909==============================
1910This is the second release candidate in preparation for the upcoming
19112.14.0 release, and if no showstoppers turn up in wider testing will
1912become the next release. Besides the usual bug fixes, we've included
1913patches by Mario Kleiner to improve vblank handing which both improve the
1914timing accuracy and remove a few potential races -- provided you also
1915have the corresponding fixes in the kernel.
1916
1917Bug fixes
1918---------
1919
1920 * Crash on hot-unpluging with an active fullscreen GL application, e.g. a
1921   compositing WM or video player
1922   https://bugs.freedesktop.org/show_bug.cgi?id=32770
1923
1924 * Freeze upon maximising client-side software renderers, e.g. flash.
1925   https://bugs.freedesktop.org/show_bug.cgi?id=31367
1926
1927 * Fix confusion over G35 generation
1928   https://bugs.freedesktop.org/show_bug.cgi?id=32478
1929
1930
1931Snapshot 2.13.902 (2010-12-10)
1932==============================
1933This is the first release candidate in preparation for the upcoming
19342.14.0 release. Sandybridge support should be now be ready for use,
1935and we will appreciate any feedback we can get from this snapshot.
1936Improving the code base for Sandybridge should also have had knock
1937on effects for earlier chips as well, in particular Ironlake and to
1938a lesser extent i965.
1939
1940New requirements compared to 2.13
1941---------------------------------
1942 * Librdrm >= 2.4.23
1943
1944Bug fixes
1945---------
1946
1947 * Crash after failing to allocate memory
1948   https://bugs.freedesktop.org/show_bug.cgi?id=31487
1949
1950 * EFAULT during pwrite
1951   https://bugs.freedesktop.org/show_bug.cgi?id=29752
1952
1953 * Mark outputs as on and restore backlight after mode-set
1954
1955 * A hang with i965+ with newer kernels.
1956
1957
1958Snapshot 2.13.901 (2010-11-05)
1959==============================
1960This is an intermediate snapshot of ongoing driver development. The
1961primary purpose of this snapshot is to capture some recent
1962improvements, (particularly in Sandybridge support), for further
1963testing.
1964
1965Release 2.13.0 (2010-09-30)
1966===========================
1967We are pleased to announce this major release of the xf86-video-intel
1968driver, on schedule at 3 months since 2.12.0. With the many bug fixes
1969in this release, we encourage everyone using 2.12 to upgrade to 2.13.
1970
1971[This release is functionally identical to the earlier 2.12.902
1972release candidate.]
1973
1974New requirements compared to 2.12
1975---------------------------------
1976 * Librdrm >= 2.4.22
1977
1978Bug fixes
1979---------
1980 * Attempt to fix infinite MI_WAIT_FOR_EVENT while watching video
1981   https://bugs.freedesktop.org/show_bug.cgi?id=28964
1982
1983   [Various new checks have been added to the video code here, but the
1984   bug fix hasn't yet been verified by the original reporter.]
1985
1986 * Fix buffer-object leak
1987   https://bugs.freedesktop.org/show_bug.cgi?id=26946
1988
1989 * Fix memory leak on server reset
1990
1991 * Fix crash due to unchecked pixmap allocation
1992   https://bugs.freedesktop.org/show_bug.cgi?id=29187
1993
1994 * Fix for video artifacts when using dualscreen
1995   https://bugs.freedesktop.org/show_bug.cgi?id=29213
1996
1997 * Fix for incorrect characters in gnome-terminal when using compiz
1998   https://bugs.freedesktop.org/show_bug.cgi?id=28438
1999
2000 * Fix for hanging, full-screen applications, (flash, compiz, etc.)
2001   https://bugs.freedesktop.org/show_bug.cgi?id=29584
2002
2003 * Fix selection of backlight device on multi-GPU systems
2004   https://bugs.freedesktop.org/show_bug.cgi?id=29273
2005
2006 * Fix to avoid crash with extremely large glyphs
2007   https://bugs.freedesktop.org/show_bug.cgi?id=29430
2008
2009 * Fix for eDP panels incorrectly being given only a single, valid mode
2010   https://bugs.freedesktop.org/show_bug.cgi?id=30069
2011
2012 * Fix GPU hang involving clipped SRC copies
2013   https://bugs.freedesktop.org/show_bug.cgi?id=30120
2014
2015 * Fix to compile for 1.6 series X server.
2016
2017 * Fix to retry framebuffer allocation after an initial failure.
2018
2019 * Fix to disable dri2 after fallbacks are forced on.
2020
2021Snapshot 2.12.902 (2010-09-28)
2022==============================
2023This is the second release candidate in preparation for the upcoming
20242.13.0 release. The comes exactly one week after 2.12.902 and includes
2025only a handful of changes.
2026
2027One significant change is that the drive now requires libdrm 2.4.22
2028(or newer), as several people reported compilation failures of
2029xf86-video-intel 2.12.901 with libdrm 2.4.21.
2030
2031Other changes include a fix to compile for the 1.6 series X server, a
2032fix to retry framebuffer allocation after an initial failure, and a
2033fix to disable dri2 after fallbacks are forced on.
2034
2035Snapshot 2.12.901 (2010-09-21)
2036==============================
2037This is the first release candidate in preparation for the upcoming
20382.13.0 release. We will appreciate any feedback we can get from
2039testing of this snapshot.
2040
2041Bugs fixed in this snapshot (compared to 2.12.0)
2042------------------------------------------------
2043 * Attempt to fix infinite MI_WAIT_FOR_EVENT while watching video
2044   https://bugs.freedesktop.org/show_bug.cgi?id=28964
2045
2046   [Various new checks have been added to the video code here, but the
2047   bug fix hasn't yet been verified by the original reporter.]
2048
2049 * Fix buffer-object leak
2050   https://bugs.freedesktop.org/show_bug.cgi?id=26946
2051
2052 * Fix memory leak on server reset
2053
2054 * Fix crash due to unchecked pixmap allocation
2055   https://bugs.freedesktop.org/show_bug.cgi?id=29187
2056
2057 * Fix for video artifacts when using dualscreen
2058   https://bugs.freedesktop.org/show_bug.cgi?id=29213
2059
2060 * Fix for incorrect characters in gnome-terminal when using compiz
2061   https://bugs.freedesktop.org/show_bug.cgi?id=28438
2062
2063 * Fix for hanging, full-screen applications, (flash, compiz, etc.)
2064   https://bugs.freedesktop.org/show_bug.cgi?id=29584
2065
2066 * Fix selection of backlight device on multi-GPU systems
2067   https://bugs.freedesktop.org/show_bug.cgi?id=29273
2068
2069 * Fix to avoid crash with extremely large glyphs
2070   https://bugs.freedesktop.org/show_bug.cgi?id=29430
2071
2072 * Fix for eDP panels incorrectly being given only a single, valid mode
2073   https://bugs.freedesktop.org/show_bug.cgi?id=30069
2074
2075 * Fix GPU hang involving clipped SRC copies
2076   https://bugs.freedesktop.org/show_bug.cgi?id=30120
2077
2078Release 2.12.0 (2010-06-24)
2079===========================
2080We are pleased to announce this major release of the xf86-video-intel
20812D driver. It contains many correctness and performance improvements
2082compared to the 2.11 releases. We encourage everyone using 2.11 to
2083upgrade to 2.12.
2084
2085[This release contains several bug fixes since the earlier 2.11.901
2086release candidate---see below for details.]
2087
2088New requirements compared to 2.11
2089---------------------------------
2090 * Libdrm >= 2.4.21
2091
2092Major performance improvements
2093------------------------------
2094There have been many performance improvements targeting the
2095915G/945G/Pineview graphics chipsets (thank to Chris Wilson!). Areas
2096improved include:
2097
2098 * Major improvements to glyph rendering
2099
2100 * Avoid software fallback when using extended desktops, (for example,
2101   video playback)
2102
2103 * Better performance when using large windows, (for example with
2104   firefox scrolling)
2105
2106 * Avoiding migration ping-pong with gigantic drawing
2107
2108The following speedups (measured with cairo-perf-trace/cairo-traces)
2109capture some of the improvements. These were measured on a Pineview
2110system with X server 1.8.1. This compares how this 2.12 release
2111improves performance compared to 2.11.0:
2112
2113	Trace			Speedup
2114	-----			-------
2115	ocitysmap		5.41x speedup
2116	firefox-talos-gfx       2.41x speedup
2117	gnome-terminal-vim      2.15x speedup
2118	poppler                 1.86x speedup
2119	firefox-planet-gnome	1.43x speedup
2120	midori-zoomed		1.39x speedup
2121	swfdec-giant-steps      1.36x speedup
2122	xfce4-terminal-a1	1.17x speedup
2123	evolution		1.09x speedup
2124	swfdec-youtube		1.06x speedup
2125
2126Other improvements
2127------------------
2128Daniel Vetter improved Xvmc and overlay, adding GEM support.
2129
2130Bug fixes
2131---------
2132There have been many conformance improvements targeting the
2133915G/945G/Pineview graphics chipsets. The rendercheck test suite
2134passes once again and there has been a significant reduction in the
2135number of errors detected by the cairo test suite (notably center
2136sampling and EXTEND_NONE behaviors are fixed).
2137
2138Other notable bug fixes include:
2139
2140 * Fix for bug 28446 Garbled fonts with Mathematica
2141   https://bugs.freedesktop.org/show_bug.cgi?id=28446
2142
2143 * Fixes for page-flipping and other DRI2-handling bugs, (Thanks Jesse
2144   Barnes, and Kristian Høgsberg)
2145
2146 * Avoid corruption when using extended desktops, (for example, video
2147   playback)
2148
2149(The bug fixes below are new since the 2.11.901 release candidate)
2150
2151 * Fix corrupted output when screen rotated
2152   https://bugs.freedesktop.org/show_bug.cgi?id=28461
2153
2154 * Fix corrupted rendering in KDE due to missing flush.
2155
2156 * Fixes i830_uxa_put_image for the following issues:
2157
2158      Bug 28569 - [i965] IGN's flash-based video player crashes X
2159      https://bugs.freedesktop.org/show_bug.cgi?id=28569
2160    
2161      Bug 28573 - [i965] Fullscreen flash and windowed SDL games fail to
2162                  update the screen
2163      https://bugs.freedesktop.org/show_bug.cgi?id=28573
2164
2165 * Fix visual corruption of scrollbar in Chromium
2166
2167 * Fix tiling limits to resolve this issue:
2168
2169      Bug 28497 - Graphics corruption after opening a specific website
2170      https://bugs.freedesktop.org/show_bug.cgi?id=28497
2171
2172Snapshot 2.11.901 (2010-06-14)
2173==============================
2174This is the first release candidate in preparation for the upcoming
21752.12.0 release. We will appreciate any feedback we can get from
2176testing of this snapshot to improve the 2.12.0 release.
2177
2178New requirements compared to 2.10
2179---------------------------------
2180 * Libdrm >= 2.4.19
2181
2182Major performance improvements
2183------------------------------
2184There have been many performance improvements targeting the
2185915G/945G/Pineview graphics chipsets (thank to Chris Wilson!). Areas
2186improved include:
2187
2188 * Major improvements to glyph rendering
2189
2190 * Avoid software fallback when using extended desktops, (for example,
2191   video playback)
2192
2193 * Better performance when using large windows, (for example with
2194   firefox scrolling)
2195
2196 * Avoiding migration ping-pong with gigantic drawing
2197
2198The following speedups (measured with cairo-perf-trace/cairo-traces)
2199capture some of the improvements. These were measured on a Pineview
2200system with a very recent X server, (from git just before 1.8.1). This
2201compares how this 2.12 release candidate improves performance compared
2202to 2.11.0:
2203
2204	Trace			Speedup
2205	-----			-------
2206	ocitysmap		5.41x speedup
2207	firefox-talos-gfx       2.41x speedup
2208	gnome-terminal-vim      2.15x speedup
2209	poppler                 1.86x speedup
2210	firefox-planet-gnome	1.43x speedup
2211	midori-zoomed		1.39x speedup
2212	swfdec-giant-steps      1.36x speedup
2213	xfce4-terminal-a1	1.17x speedup
2214	evolution		1.09x speedup
2215	swfdec-youtube		1.06x speedup
2216
2217Other improvements
2218------------------
2219Daniel Vetter improved Xvmc and overlay, adding GEM support.
2220
2221Bug fixes
2222---------
2223There have been many conformance improvements targeting the
2224915G/945G/Pineview graphics chipsets. The rendercheck test suite
2225passes once again and there has been a significant reduction in the
2226number of errors detected by the cairo test suite (notably center
2227sampling and EXTEND_NONE behaviors are fixed).
2228
2229Other notable bug fixes include:
2230
2231 * Fix for bug 28446 Garbled fonts with Mathematica
2232   https://bugs.freedesktop.org/show_bug.cgi?id=28446
2233
2234 * Fixes for page-flipping and other DRI2-handling bugs, (Thanks Jesse
2235   Barnes, and Kristian Høgsberg)
2236
2237 * Avoid corruption when using extended desktops, (for example, video
2238   playback)
2239
2240Release 2.11.0 (2010-03-29)
2241===========================
2242New in 2.11: DRI2 and page flipping
2243-----------------------------------
2244The most significant new feature of this release is support for new
2245DRI2 APIs, allowing page flipping to occur for swaps that are
2246full-screen and not rotated.
2247
2248Performance improvements in 2.11 compared to 2.10
2249-------------------------------------------------
2250Dramatically improved performance of large pixmaps on
2251memory-constrained hardware, (such as 945), by using tiling.
2252Note that this fix is new since the most recent release
2253candidate for this release (2.10.903).
2254
2255    https://bugs.freedesktop.org/show_bug.cgi?id=25375
2256
2257Elimination of software fallback with alpha-only pixmap (a fix which
2258eliminated a full second from the Moblin boot time):
2259
2260    https://bugs.freedesktop.org/show_bug.cgi?id=26189
2261
2262Notable bug fixes in 2.11 compared to 2.10
2263------------------------------------------
2264Fix for undesired black borders on some images, (caued several
2265problems with firefox):
2266
2267    http://bugs.freedesktop.org/show_bug.cgi?id=17933
2268
2269And other fixes to provide more conformance with existing
2270software-rendering, (as measures by the cairo test suite), such as the
2271sampling location for nearest-neighbor sampling.
2272
2273Other changes
2274-------------
2275Eric Anholt and Daniel Vetter both removed large piles of old and
2276useless code now that the driver requires kernel modesetting
2277(KMS). This continues to reduce the mainteance burden of the driver,
2278making it easier to isolate and fix bugs.
2279
2280Snapshot 2.10.903 (2010-03-22)
2281==============================
2282Some notable bug fixes
2283----------------------
2284* Fix unpredictable results in page-flipping code due to access of an
2285  uninitialized variable.
2286
2287* Two fixes that address a number of long-outstanding failures in the
2288  cairo test suite:
2289
2290    Remove coord-adjust for nearest centre-sampling.
2291
2292    Fill alpha on xrgb images. Closes:
2293
2294	Bug 17933 - x8r8g8b8 doesn't sample alpha=0 outside surface bounds
2295	http://bugs.freedesktop.org/show_bug.cgi?id=17933
2296
2297* Fix a regression introduced in 2.10.902:
2298
2299    Fatal server error:
2300    i915_emit_composite_setup: ADVANCE_BATCH:
2301    under-used allocation 100/104
2302
2303	Bug 21723 [i915 bisected] rendering makes X crash
2304	http://bugs.freedesktop.org/show_bug.cgi?id=27123
2305
2306Other minor fixes (avoid trying to print an error message with a freed
2307string, set a variable to NULL after freeing), and documentation
2308updates are also included.
2309
2310Snapshot 2.10.902 (2010-03-15)
2311==============================
2312Some significant bug fixes
2313--------------------------
2314Fix video color problems due to YUV plane ordering issue. This fixed a
2315regression that showed up recently (only in these release candidates).
2316
2317Fix recent regression leading to corruption under compiz:
2318
2319  https://bugs.freedesktop.org/show_bug.cgi?id=26814
2320
2321Further DRI2 improvements/fixes from Mario Kleiner and Jesse Barnes,
2322(including support for SwapBuffers request with divisor != 0).
2323
2324Fix build against X server 1.6 branch.
2325
2326Avoid flooding logs after an error occurs.
2327
2328And dramatic code cleanups
2329--------------------------
2330Eric Anholt removed several cases of old, useless code, (conditions
2331that could not occur now that we require KMS, etc.).
2332
2333Daniel Vetter eliminated piles of code by modernizing the XvMC
2334memory-management and completely removing the old i830_memory
2335allocator (i830_allocate_memory, i830_free_memory), etc.
2336
2337Snapshot 2.10.901 (2010-02-26)
2338==============================
2339The most significant new feature of this release is support for new
2340DRI2 APIs, allowing page flipping to occur for swaps that are
2341full-screen and not rotated.
2342
2343Some significant bug fixes
2344--------------------------
2345* Avoid fallback due to failing to extract pixel value from an
2346  alpha-only solid, (this fix eliminated 1 second from Mobilin boot
2347  time).
2348
2349  https://bugs.freedesktop.org/show_bug.cgi?id=26189
2350
2351* Handle full-height blits without tearing in CopyRegion
2352
2353  https://bugs.freedesktop.org/show_bug.cgi?id=22475
2354
2355Several other fixes and cleanups are included as well, (including the
2356removal of stale utility code that is now maintained in the separate
2357intel-gpu-tools module).
2358
2359Release 2.10.0 (2010-01-04)
2360===========================
2361New requirements of 2.10 compared to 2.9
2362----------------------------------------
2363 * Linux kernel with kernel-modesetting (KMS)
2364
2365   All user-modesetting code has now been removed from the driver.
2366   We highly recommend kernel version 2.6.32 or later.
2367
2368 * Libdrm >= 2.4.16
2369
2370New features in 2.10 compared to 2.9
2371------------------------------------
2372 * New driver debugging options available in xorg.conf
2373   (DebugFlushBatches, DebugFlushCaches, DebugWait). See "man intel"
2374   for more details.
2375
2376 * Video overlay support with KMS. This currently requires Linux
2377   2.6.33, but a backport to 2.6.32 is available here:
2378
2379   http://gitorious.org/daniel-s-linux-stuff/linux-kernel/commits/intel-kms-overlay-for-2.6.32
2380
2381Notable fixes in 2.10 compared to 2.9
2382-----------------------------------
2383 * Fix crash when XRenderComposite is called with a -1 value for
2384   width/height, (crash at login when using compiz).
2385
2386   http://bugs.freedesktop.org/show_bug.cgi?id=24724
2387
2388 * Fix to support larger extended desktops, (such as a 2560x1600 plus a
2389   1920x1200 monitor).
2390
2391
2392 * Fix glyph corruption due to insufficient cache flushing:
2393
2394   https://bugs.freedesktop.org/show_bug.cgi?id=24315
2395
2396Several other correctness and performance fixes are also included.
2397
2398Note: The driver code of the 2.10.0 release is identical to that of
2399the 2.9.99.901 release candidate. The only changes in 2.10.0 compared
2400to 2.9.99.902 are a minor change to the configure script, and some
2401small updates to the documentation.
2402
2403Snapshot 2.9.99.902 (2009-12-10)
2404================================
2405This is the second release candidate in preparation for the upcoming
24062.10.0 release. We will appreciate any feedback we can get from
2407testing of this snapshot to improve the 2.10.0 release.
2408
2409Compared to the previous snapshot, this snapshot now correctly
2410verifies that libdrm 2.4.16 (or newer) is available. The previous
2411snapshot also required libdrm 2.4.16 for correct functionality, but
2412neglected to notify the user of this at configure, compile, or install
2413time and would instead simply fail to run.
2414
2415Snapshot 2.9.99.901 (2009-11-30)
2416================================
2417This is the first release candidate in preparation for the upcoming
24182.10.0 release. We will appreciate any feedback we can get from
2419testing of this snapshot to improve the 2.10.0 release.
2420
2421With this snapshot, the xf86-video-intel driver requires kernel
2422modesetting (KMS) as all of the user modesetting (UMS) support has
2423been removed from the driver.
2424
2425Release 2.9.0 (2009-09-28)
2426==========================
2427We are pleased to announce the 2.9.0 release of the xf86-video-intel
2428driver. (If you have been following along at home, this release is
2429identical to the 2.8.99.902 release candidate other than the version
2430number).
2431
2432New features in 2.9 compared to 2.8
2433-----------------------------------
2434* Support for the B43 chipset.
2435
2436Major fixes in 2.9.0 compared to 2.8.0
2437--------------------------------------
2438* Multiple fixes to make the driver stable for 8xx chipsets, (855GM,
2439  865G, etc.). The 2.8 driver series was extremely unstable with many
2440  of these chipsets.
2441
2442  https://bugs.freedesktop.org/show_bug.cgi?id=22904 (and many duplicates)
2443  https://bugs.freedesktop.org/show_bug.cgi?id=22947
2444
2445* Add support for BACKLIGHT property when using kernel modesetting
2446  (KMS). This allows backlight adjustment with programs such as
2447  "xbacklight -set <percentage>" or "xrandr --set BACKLIGHT <value>".
2448
2449  https://bugs.freedesktop.org/show_bug.cgi?id=20963
2450
2451* Fix so that "xrandr --scale" works when using KMS.
2452
2453  https://bugs.freedesktop.org/show_bug.cgi?id=22893
2454
2455* Fix segfaults of X server when logging out.
2456
2457  https://bugs.freedesktop.org/show_bug.cgi?id=20516
2458
2459* Avoid falling back to software for 1-bit alpha-only masks.
2460
2461  https://bugs.freedesktop.org/show_bug.cgi?id=23184
2462
2463* Fix pixel sampling position for 8xx chipsets, (fixes some cairo-test
2464  suite failures).
2465
2466Snapshot 2.8.99.901 (2009-09-09)
2467================================
2468This is the first release candidate in preparation for the upcoming
24692.9.0 release. We will appreciate any feedback we can get from testing
2470of this snapshot to improve the 2.9.0 release.
2471
2472See the commit log for changes in this release, and developers, please
2473provide me summaries for any major features in this release that you
2474think should appear in NEWS and in the 2.9.0 release notes.
2475
2476Note: We are still investigating failures with 865G chipsets, (these
2477failures existed in the 2.8 series as well). We are hoping to fix
2478these failures before the 2.9.0 release.
2479
2480Release 2.8.0 (2009-07-20)
2481==========================
2482We are pleased to present this major release of the xf86-video-intel
2483driver. This release is very similar to the 2.7.99.902 release
2484candidate but includes a couple of additional bug fixes, (for bugs
2485#22760 and @22483).
2486
2487Compared to the 2.7 releases this driver contains several new features
2488and many bug fixes. The driver now depends on X server 1.6 or later,
2489and eliminates several obsolete code paths, (XAA and EXA removed in
2490favor of UXA, DRI1 support eliminated). The driver certainly will work
2491best with an i915 module from a recent kernel (2.6.31) and with
2492kernel-modesetting (KMS, specified by loading i915 with the option
2493"modeset=1").
2494
2495New features in 2.8.0 compared to 2.7
2496-------------------------------------
2497* Remove XAA/EXA/DRI1 and NoAccel, and only support UXA/DRI2 now
2498  (requiring xserver 1.6).
2499
2500* VLD support added into XvMC for Mpeg2, which decreases CPU usage
2501  when playing video.
2502
2503Major fixes in 2.8.0 compared to 2.7
2504------------------------------------
2505* Fix broken front-buffer rendering
2506  https://bugs.freedesktop.org/show_bug.cgi?id=19174
2507
2508* Fix disabling of XvMC disabled
2509  https://bugs.freedesktop.org/show_bug.cgi?id=20790
2510
2511* Fix broken VT switch on some machines
2512  https://bugs.freedesktop.org/show_bug.cgi?id=19578
2513
2514* Improve performance of trapezoid rendering for UXA
2515  https://bugs.freedesktop.org/show_bug.cgi?id=21376
2516
2517* Vblank sync'd GL buffer swap (fixing video tearing under composite)
2518  https://bugs.freedesktop.org/show_bug.cgi?id=20664
2519
2520* Fix G41 DRM support
2521  https://bugs.freedesktop.org/show_bug.cgi?id=21095
2522
2523* Fix crash on some 8xx
2524  https://bugs.freedesktop.org/show_bug.cgi?id=18974
2525
2526Various fixes for GPU hangs, and more...
2527
2528Snapshot 2.7.99.902 (2009-07-13)
2529--------------------------------
2530This is the first release candidate in preparation for the upcoming
25312.8.0 release. Most major and critical bugs should be fixed, but some
2532minor bugs may still be present. We will appreciate any feedback we
2533can get from testing of this snapshot to improve the 2.8.0 release.
2534
2535Snapshot 2.7.99.901 (2009-06-10)
2536--------------------------------
2537This is the first release candidate in preparation for the upcoming
25382.8.0 release. Most critical bugs should be fixed, but some major
2539bugs may still be present. We will appreciate any feedback we can
2540get from testing of this snapshot to improve the 2.8.0 release.
2541
2542Some of the major fixes in this snapshot include:
2543
2544	* Several fixes for intermittent GPU hangs/crashes, (some of
2545	  which already appeared in the 2.7.1 release)
2546
2547	* Fix to avoid X crash when viewing large images in browser
2548
2549	* Fix X server failure when running old (non-GEM) kernel
2550
2551	* Fixes for SDVO LVDS mode detection
2552
2553	* Fix major performance regression of trapezoid rendering
2554          compared to XAA/EXA
2555
2556	* New support for tear-free video when using KMS
2557
2558	* New support for tear-free DRI2 CopyRegion
2559
2560Snapshot 2.7.99.1 (2009-04-28)
2561------------------------------
2562This is a development snapshot very early in the process toward
2563developing 2.8. There have been some big changes to the code, and
2564we're anxious to get feedback on these changes as early as possible.
2565
2566Here is a summary of the biggest changes:
2567
2568  * Driver now depends on X server 1.6 or later
2569
2570  * Eliminate XAA and EXA support (in favor of UXA)
2571
2572  * Eliminate DRI1 support
2573
2574  * Fixes for running without DRI at all
2575
2576These code removals represent a deletion of a substantial amount of
2577code, (and hopefully piles of bugs), as well as reduce the maintenance
2578effort going forward as the number of combinatorial configurations for
2579the driver are greatly reduced. This means that users are much more
2580likely to be running code that has actually been tested, and it will
2581be much easy for developers to replicate bugs that users experience.
2582
2583One of the things that would be most useful in testing this release is
2584to revisit any outstanding bugs that you have previously reported. If
2585the buggy behavior is gone, (or the bug is no longer relevant---such
2586as a bug that's specific to XAA only), please feel free to indicate so
2587in bugzilla or even just close the bug.
2588
2589If you confirm that the bug is still present, please indicate so in
2590the bug report. (I was going to ask that you select a 1.7.99 version,
2591but it looks like bugzilla only has versions for products not
2592compoenents, while we use a "xorg" product and a "driver/intel"
2593component.) We definitely want to make any such confirmed bugs a
2594priority, so it would be nice to have a consistent mechanism to search
2595for these bugs. Suggestions are welcome on the best approach.
2596
2597Thanks in advance for any testing or feedback on this snapshot.
2598
2599Release 2.7.0 (2009-04-15)
2600--------------------------
2601Compared to the 2.6 series, 2.7.0 has a large number of bug fixes, but
2602also a few significant features, such as:
2603
2604        SDVO-TV support, available on ADD2 card (bug#9992) and
2605        D945GCLF2 board (bug#17776).
2606
2607        Basic SDVO-LVDS support
2608
2609        XV video display without tearing
2610	[Though this isn't working for all users yet, see
2611	https://bugs.freedesktop.org/show_bug.cgi?id=21076 ]
2612
2613	Various fixes for UXA, DRI2, and Kernel modesetting.
2614
2615We encourage users to use kernel modesetting and UXA acceleration with
2616this release, which should give the best performance and
2617robustness. When KMS is available, UXA is the default acceleration
2618used by the driver, (EXA is the default otherwise).
2619
2620Known issue:
2621
2622	Some Linux kernel versions (such as 2.6.29) are known to have
2623	broken PAT code that causes recent versions of this driver to
2624	fail, (which can manifest as the X server simply not
2625	starting). This can be verified by adding the "nopat" option
2626	to the kernel command-line and seeing the failure go away. We
2627	hope that newer kernels in the 2.6.29.x as well as 2.6.30 and
2628	above will have working PAT code.
2629
2630Some of the most notable bugs fixed in 2.7.0 include:
2631
2632	[GM45 965GM] bad htotal causes panel startup failure
2633	https://bugs.freedesktop.org/show_bug.cgi?id=17292
2634
2635	[xrandr TV] need TV output property control
2636	https://bugs.freedesktop.org/show_bug.cgi?id=12763
2637
2638	[TV] "xrandr --set TV_FORMAT" gets BadMatch error
2639	https://bugs.freedesktop.org/show_bug.cgi?id=16566
2640
2641	[945 tiling] Low performance due to no A17 workaround
2642	https://bugs.freedesktop.org/show_bug.cgi?id=16835
2643
2644	[TV]Flicker when launching applications in the 2.4-branch
2645	https://bugs.freedesktop.org/show_bug.cgi?id=17405
2646
2647	[945GM FBC] FBC causes underruns & flicker
2648	https://bugs.freedesktop.org/show_bug.cgi?id=18651
2649
2650	[xv] Textured video suffers from tearing
2651	https://bugs.freedesktop.org/show_bug.cgi?id=19635
2652
2653	[G45] Random hangs with UXA
2654	https://bugs.freedesktop.org/show_bug.cgi?id=19734
2655
2656	[945GM] Any 3D app is slow in resolution higher than 800x600
2657	with UXA+DRI2, due to tiling
2658	https://bugs.freedesktop.org/show_bug.cgi?id=19738
2659
2660	[i915 UXA,EXA] rotation messes display with tiling on
2661	https://bugs.freedesktop.org/show_bug.cgi?id=20265
2662
2663	[G45] DRI2/UXA gives solid white instead of transparency
2664	https://bugs.freedesktop.org/show_bug.cgi?id=20321
2665
2666	LVDS output not detected
2667	https://bugs.freedesktop.org/show_bug.cgi?id=20517
2668
2669	xf86-video-intel-2.6.3: Xv crashes X server
2670	https://bugs.freedesktop.org/show_bug.cgi?id=20525
2671
2672	[G965 non-GEM] systray in KDE 4 completely broken
2673	https://bugs.freedesktop.org/show_bug.cgi?id=20527
2674
2675	[SDVO-TV]the desktop is teared in four sections on the screen
2676	https://bugs.freedesktop.org/show_bug.cgi?id=20550
2677
2678	Intel video driver 2.6.3 crashes with XVideo
2679	https://bugs.freedesktop.org/show_bug.cgi?id=20563
2680
2681	[855GM] Xv crash with non-KMS
2682	https://bugs.freedesktop.org/show_bug.cgi?id=20585
2683
2684	2.6.99.902 breaks native 1680x1050 mode on TMDS -- EDID miss
2685	https://bugs.freedesktop.org/show_bug.cgi?id=20594
2686
2687	[945GM TV] 2.6.99.902 sets a too high CONTRAST-value
2688	https://bugs.freedesktop.org/show_bug.cgi?id=20670
2689
2690	[915GM] fail to detect LVDS with new VBT code
2691	https://bugs.freedesktop.org/show_bug.cgi?id=20752
2692
2693	[regression i965]tiled output when start X
2694	https://bugs.freedesktop.org/show_bug.cgi?id=20803
2695
2696	2.6.99.902: LVDS wrongly detected as disconnected
2697	https://bugs.freedesktop.org/show_bug.cgi?id=20826
2698
2699	vt switching fails and crashes X
2700	https://bugs.freedesktop.org/show_bug.cgi?id=20863
2701
2702	[HDMI] The screen will flicker when some application runs on
2703	G45-64
2704	https://bugs.freedesktop.org/show_bug.cgi?id=20875
2705
2706	TexturedVideo is offsetted wrongly on dualhead
2707	https://bugs.freedesktop.org/show_bug.cgi?id=20980
2708
2709	[EXA] xvideo hang X
2710	https://bugs.freedesktop.org/show_bug.cgi?id=21027
2711
2712	[EXA] x11perf performance regression
2713	https://bugs.freedesktop.org/show_bug.cgi?id=21029
2714
2715	And many others...
2716