NEWS revision 03b705cf
1Release 2.21.15 (2013-08-21)
2============================
3Some build fixes for the BSDs and alternate compiler, and conflicting
4configure options and a critcial fix for some PRIME setups.
5
6 * Don't fail to configure if DRI1 is not available and the user asks
7   for both DRI and KMS-only
8   [Regression from 2.21.14]
9
10 * Lots of miscellaneous fixes for older gcc, other compilers and BSD.
11
12 * Initial framework support for hosted X.
13
14 * Improve transition from fbcon to X with multiple outputs and extended
15   desktops.
16
17 * Ensure the framebuffer exists before checking it against the CRTC
18   constraints.
19   https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/121234
20   [Regression from 2.21.10]
21
22 * Add RGB overlay support for Ironlake and later.
23
24 * Don't release the overlay buffer too early, i.e. before the client
25   disables the Xv image.
26
27 * Tweak the ring switching logic to reduce use of semaphores on
28   Sandybridge and later - a delicate balancing act between trying to
29   use the faster engine for the task and preventing concurrent use of
30   multiple rings.
31
32
33Release 2.21.14 (2013-08-04)
34============================
35A few updates for the top-end Haswell systems (notably GT3 and GT3e
36systems), in particular, setting the appropriate thread counts for the
37larger GPU and handling a few scaling issues with the emedded DRAM. More
38important than the performance tuning though is a fix for a critical
39issue encountered by a few people on gen4/gen5 systems.
40
41 * Initial performance tuning for HSW:GT3e
42
43 * Fix a crash with Planetary Annihilation
44
45 * Disable triple buffering for compositors
46
47 * Reserve space in the exec buffer array for deferred VBOs.
48   https://bugs.freedesktop.org/show_bug.cgi?id=67504
49
50
51Release 2.21.13 (2013-07-27)
52============================
53A minor release to repair the build for non-Linux systems, and to undo a
54few more regressions.
55
56 * Avoid potential memory corruption with allocations of very small
57   depth 1 bitmaps.
58
59 * Fix source clipping whilst computing copy extents and exposures.
60   [Regression from 2.21.12]
61   https://bugs.freedesktop.org/show_bug.cgi?id=66970
62
63 * Use /proc/cpuinfo if cpuid4 is not available (old hardware or old
64   gcc), and repair the build on systems without cpuid.
65   [Regression from 2.21.12]
66
67 * Fix performance regression on Ironlake from inadvertently flushing
68   after every operation.
69   [Regression from 2.21.11]
70   https://bugs.freedesktop.org/show_bug.cgi?id=67157
71
72 * Fix conflict handling when probing initial KMS configuration for
73   ZaphodHeads.
74   [Regression from 2.21.11]
75   https://bugs.freedesktop.org/show_bug.cgi?id=67176
76
77 * Stop being overly restrictive and rejecting stale DRI2 buffers.
78   (DRI2 is inherently racy in that the client may have completed and
79   submitted rendering to buffers that are now invalid on the server.
80   This race is magnified by bugs within the Xserver where it forgets
81   to notify the DRI2 clients of certain invalidation events.)
82   https://bugs.freedesktop.org/show_bug.cgi?id=67210
83
84 * Handle a failure to use a GTT mmap for a pixmap upload and try
85   an alternative method before giving up.
86
87
88Release 2.21.12 (2013-07-14)
89============================
90In this release, we clear up the teething troubles from preserving the KMS
91configuration, notably external connections on Haswell and plugging in new
92outputs after startup were broken. Besides these regression fixes, there
93are a couple of fixes for some long standing issues, such as incorrect
94rendering on gen2, an infinite loop with very, very large pixmaps and a
95slight improvement to the tempermental gen4.
96
97 * Allow untiled scanouts again (required for large extended desktops on
98   gen2 and gen3).
99   [Regression from 2.21.11]
100
101 * Use the correct count of the number of dirty damage boxes for the quick
102   check on whether the existing damage contains the requested area.
103   The danger is that we may get a false result and skip migration and so
104   cause pixmap corruption (in the unlikely event that the application
105   frequently causes fallbacks).
106   [Regression from 2.21.11]
107   https://bugs.freedesktop.org/show_bug.cgi?id=66430
108
109 * Fix initial connection probing for multi-function encoders, such as the
110   external connections on Haswell.
111   [Regression from 2.21.11, initial connection probing]
112   https://bugs.freedesktop.org/show_bug.cgi?id=66488
113
114 * Fix gen2 rendercopy into a8 surfaces, for example, glyph uploads into
115   the glyph cache.
116
117 * Fix detection of user overrides for initial connection configuration.
118   The code used the xorg-server-1.15 values, having missed the
119   introduction of ZoomModes into that release.
120   [Regression from 2.21.11, initial connection probing]
121
122 * Always initialise the gamma ramp, even on unconnected CRTCs.
123   https://bugs.freedesktop.org/show_bug.cgi?id=66563
124   [Regression from 2.21.11, initial connection probing]
125
126 * Some more tuning of the gen4 vertex corruption workaround. The root
127   cause behind the GPU using incorrect texture coordinates is still
128   not solved, but by reducing the maximum number or rectangles in
129   flight through the GPU we reduce the likelihood of corruption.
130
131 * Fix compilation with gcc-4.5
132   [Regression from 2.21.11]
133
134 * Avoid integer overflow when performing tiled uploads and operations
135   on very large (>28k pixels wide or tall pixmaps)
136   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1200766
137
138
139Release 2.21.11 (2013-06-30)
140============================
141An eventful week. What started with a regression with some builds of
142firefox on some machines lead ultimately to the discovery of an older
143kernel bug. Aside from the work to fix the image bug and a few other
144older bugs that were reported and resolved this week, there is also a
145(hopefully) subtle change to the initial configuration of displays. In
146the absence of user overrides in xorg.conf, the DDX will try to preserve
147the same display configuration as used by the kernel, which hopefully
148will be the same configuration as setup by the BIOS. The result should
149be a boot sequence that does not resize at all (aka fastboot) - until
150the display manager takes over and loads a completely different
151configuration!
152
153 * Add reference counting of drmMaster for ZaphodHeads
154   https://bugs.freedesktop.org/show_bug.cgi?id=66041
155
156 * Add a GPU flush before changing blend modes on Ironlake
157   https://bugs.freedesktop.org/show_bug.cgi?id=51422
158
159 * Fix occasional missing images for inplace uploads
160   [regression from 2.21.10]
161   https://bugs.freedesktop.org/show_bug.cgi?id=66059
162
163 * Add missing utility files to the tarball and remove a few unused ones
164
165 * Initialise PolyPoint operand state before calling miWideDash
166   https://bugs.freedesktop.org/show_bug.cgi?id=66104
167
168 * Fix redirection handling for rendering into large surfaces
169   https://bugs.freedesktop.org/show_bug.cgi?id=66168
170   https://bugs.freedesktop.org/show_bug.cgi?id=66249
171
172 * Fix compilation of UXA with xorg-xserver < 1.10
173   [regression from 2.20.0]
174
175 * Fix consideration of gradients for deciding when to migrate render
176   operations
177   [performance regression from 2.21.10, the bug itself is older]
178   https://bugs.freedesktop.org/show_bug.cgi?id=66297
179
180Also fixed this week was:
181
182commit 22fd5ca947b58901927d100d2b1aa0f1672b3435
183Author: Chris Wilson <chris@chris-wilson.co.uk>
184Date:   Fri Jun 28 16:54:08 2013 +0100
185
186    drm/i915: Only clear write-domains after a successful wait-seqno
187
188which affects kernels 3.7 - 3.10, coming to a stable kernel near you soon.
189
190
191Release 2.21.10 (2013-06-22)
192============================
193Fixes missing support for Xv (with the textured video adaptor) on
194Haswell, and an old bug with wide monitors on various generation. Along
195with a few other regression fixes and performance tweaks for various
196corner cases.
197
198 * Do not lose track of fast pageflips across mode changes
199   [regression from 2.20.8]
200
201 * Fix listing of Visuals for Xv
202   [regression from 2.21.8]
203   https://bugs.freedesktop.org/show_bug.cgi?id=65479
204
205 * Improve coherency of concurrent CPU accesses to a pixmap
206   https://bugs.freedesktop.org/show_bug.cgi?id=61628
207
208 * Set sampler swizzling for textured video on Haswell
209   https://bugs.freedesktop.org/show_bug.cgi?id=65699
210
211 * Apply scanout stride limits
212   https://bugs.freedesktop.org/show_bug.cgi?id=65099
213
214 * Undo the self-copy for cloned pixmaps for loimpress's animations
215   [regression from 2.21.7]
216   https://bugs.freedesktop.org/show_bug.cgi?id=65665
217
218
219Release 2.21.9 (2013-06-06)
220===========================
221Consolidating the copy-on-write support, hopefully cleaning up the last of
222the regressions.
223
224 * Restore vsync on textured videos.
225   [regression from 2.21.8]
226   https://bugs.freedesktop.org/show_bug.cgi?id=65048
227
228 * Fix incorrect ordering of possible_clones with certain outputs, which
229   can lead to attempting to incorrectly clone 2 outputs and failing to
230   light them up.
231   [regression from 2.20.10]
232
233 * Fix performance regression from not promoting large fills to the GPU
234   [regression from 2.21.7]
235
236 * Undo the pixmap clone before performing a DRI2CopyRegion
237   [regression from 2.21.7]
238   https://bugs.freedesktop.org/show_bug.cgi?id=65250
239
240
241Release 2.21.8 (2013-05-27)
242===========================
243A quick release to cleanup a few regressions from the introduction of
244copy-on-write support, notably hitting wine applications and a memory
245leak for firefox.
246
247 * Only mark a PolyFillRect operation as replacing if it is unclipped
248   https://bugs.freedesktop.org/show_bug.cgi?id=64841
249
250 * Prevent potential NULL dereference of damage when checking COW support
251
252 * Fix invalidation of clone after dirtying the pixmap via the CPU
253
254 * Prevent discarding an operation before requiring it for a fallback
255
256 * Fix memory leak from replacing the clone under certain circumstances
257   https://bugs.freedesktop.org/show_bug.cgi?id=64978
258
259
260Release 2.21.7 (2013-05-21)
261===========================
262A couple of weeks turned into a month and a couple of weeks... Amidst
263the usual bug fixes, we have added the complete set of Haswell PCI IDs -
264hopefully future proofing ourselves against being surprised by new
265products. We can also now use the correct term for the top of the range
266Haswell variants, GT3.
267
268 * Fix several assertion failures hit by Jiri Slaby.
269
270 * Allow XvMC to also target overlay/sprite planes.
271
272 * Throw in a paranoid MI_FLUSH between BLT and RENDER operations on
273   Ironlake.
274   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1168066
275
276 * Prevent reuse of old framebuffers after a resize.
277   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1157678
278
279 * Fix compilation with --enable-valgrind and no --enable-debug
280
281 * Improve partial migration of render sources.
282
283 * Fix origin of trapezoids.
284   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1178020
285
286 * Introduce copy-on-write support for cloning pixmaps. The ultimate
287   goal here is to efficiently support the TearFree mode of operation,
288   but this provides immediate benefits with firefox - most importantly
289   because of the inefficient way it now implements scrolling.
290
291
292Release 2.21.6 (2013-04-06)
293===========================
294A surprising highlight of this release is a little refresh to the KMS
295support for OpenBSD. OpenBSD now has its own KMS implementation which is
296mostly compatible with the interface in Linux, with one or two tweaks
297supplied by Mark Kettenis. This release continues to cleanup behaviour
298for Haswell.
299
300 * Workaround a failure by the xserver to invalidate DRI buffers
301   following a pixmap change for XComposite redirection.
302   https://bugs.freedesktop.org/show_bug.cgi?id=62614
303
304 * Fix computation of clip extents for stippling
305   https://bugs.freedesktop.org/show_bug.cgi?id=62618
306
307 * Support KMS on OpenBSD, by Mark Kettenis
308
309 * Clean up sockets upon CloseScreen (making ourselves better behaved
310   for muxed setups).
311
312 * Fix the tests for AVX/AVX2 support in CPUID and remember to check for
313   OS support as well.
314
315 * Report a monotonic UST value for undisplayed drawables rather than 0
316   by Daniel Kurtz
317
318 * Fix video playback on gen4 through a complex clip (more gen4 GPU woes)
319   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1162046
320
321
322Release 2.21.5 (2013-03-21)
323===========================
324Haswell reintroduces a command to load the scanline window from the
325command stream and so requires its own specialised wait-for-vsync routine
326- failure to do so was then causing hangs when trying to do tearfree video
327or use a compositor.
328
329 * Prevent buffer leak if a non-fullscreen Window is closed with multiple
330   pending swap events.
331
332 * Fix offset transformation for fallback gradient paths.
333   https://bugs.freedesktop.org/show_bug.cgi?id=62198
334
335 * Prevent Glamor from crashing if misconfigured.
336   Thanks to Michel Dänzer.
337
338 * Prevent UXA from crashing if torn down during PreInit.
339   Thanks to Aaron Plattner.
340
341 * Prevent miscompilation with different functional units having different
342   compiler flags. Some functions were expected to be inlined and so
343   recompiled with the current target. However, some compilers were
344   choosing to emit subroutine calls instead without noticing that the
345   ABI was different between the caller and callee - causing corruption.
346   https://bugs.freedesktop.org/show_bug.cgi?id=62198
347
348 * Fix rendering of CompositeTriFan with recent Xorg.
349
350 * Apply the video src-offset fix highlighted in the last release!
351   A typo prevented the fix from working for gen4+.
352   https://bugs.freedesktop.org/show_bug.cgi?id=62343
353
354 * Fix rendering of multiple glyphs to very large destination surfaces
355   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1156387
356
357 * Fix scanline waits for Haswell
358   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1156679
359
360
361Release 2.21.4 (2013-03-11)
362===========================
363More bugs, more fixes, more releases. A minor new feature being introduced
364is the runtime detection of CPU instructions sets along with specialised
365paths to take advantage of the available CPU.
366
367 * Honour LinearFramebuffer for clumsy PowerXpress integration
368
369 * Disable read-read optimisations of mappings whilst the root
370   cause of a corruption issue remains elusive.
371   https://bugs.freedesktop.org/show_bug.cgi?id=61628
372
373 * Disable 8-bpp framebuffers in UXA as the regressions therein
374   remain unfixed.
375
376 * Restart vertex checks after lock contention
377   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1124576
378
379 * Handle Screen Pixmap recreation whilst exported via PRIME.
380
381 * Correct application of scale factors to video source offsets.
382   https://bugs.freedesktop.org/show_bug.cgi?id=61610
383
384 * Chain up CloseScreen so that resources are actually freed across regen
385   https://bugs.freedesktop.org/show_bug.cgi?id=56608
386
387 * Fix Haswell CRW PCI-IDs
388
389 * Handle batch submission failure during DRI copies
390   https://bugs.freedesktop.org/show_bug.cgi?id=61708
391
392 * Probe for kernel support of requested Screen depth
393   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1135403
394
395 * Correct GPU limits for early gen2 and gen3 architectures and prevent
396   an infinite recursion for particular image sizes.
397
398
399Release 2.21.3 (2013-02-20)
400===========================
401A few minor bugfixes, another point release.
402
403 * Fix tracking of DRI pixmaps and their backing bo across reparenting. If
404   we tried to execute a SwapBuffers after a Window was reparented, but
405   before the DRI client has updated its references, then we would end up
406   manipulating an exported pixmap without a flush flag set. In the worst
407   case, this would culminate in a segfault in the driver.
408   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1127497
409
410 * Restore the gen4 workarounds for flickering rendering - a few cases still
411   remain, as the root cause persists.
412   https://bugs.freedesktop.org/show_bug.cgi?id=60402
413
414 * Double check that the device has KMS enabled before claiming. This allows
415   X to gracefully fallback to VESA/fbdev rather than bailing out.
416   https://bugs.freedesktop.org/show_bug.cgi?id=60987
417
418 * Fix the UXA render programs for projective transforms on Ivybridge.
419
420
421Release 2.21.2 (2013-02-10)
422===========================
423Pass the brown paper bags, I need half a dozen or so. That seemingly
424innocuous build fix with xorg-1.13 happened to have the little side-effect
425of breaking glyph rendering with xorg-1.12 and older on 64-bit machines.
426
427
428Release 2.21.1 (2013-02-10)
429===========================
430A fix for a potential GPU hang on 945gm (GMA950) and earlier chipsets,
431along with backporting SNA to the packages found in stable distributions
432like Debian 6.0 (Squeeze).
433
434 * Cleanup compilation warnings from deblint, thanks to Paul Menzel
435
436 * Minor build improvements by Damien Lespiau.
437
438 * Disable generating span geometry for non-rectilinear spans on gen4
439   in order to work around and prevent one class of render corruption.
440
441 * Prevent cache thrashing and severe performance degradation on LLC
442   machines for streaming texture updates. However, note the effect was
443   only observed on just one particular laptop.
444
445 * Fix alignment of subsurface proxies for old chipsets.
446   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1120108
447
448 * Repair build against Xserver-1.6 and contemporary packages.
449
450
451Release 2.21.0 (2013-02-01)
452===========================
453A few new features:
454
455 * Enable render acceleration for Haswell GT1/GT2.
456
457 * Enable multi-threaded rasterisation of trapezoids and fallback composition
458
459 * Utilise a new kernel interface (v3.9) for processing relocations
460
461along with a few older features from the 2.20.x series:
462
463 * PRIME support for hotplug GPUs and hybrid systems
464
465 * Support for IvyBridge GT1 machines, aka HD2500 graphics.
466
467 * Stable 830gm/845g support, at last!
468
469As usual we have a large number of bug fixes since the last release:
470
471 * Prevent a stray relocation being left after a buffer is removed from
472   a batch, leading to GPU hangs.
473
474 * Make the driver more robust against its own failures to submit batches
475   by falling back to software rendering.
476
477 * Fix emission of scanline waits for secondary pipes on gen6/7. Otherwise
478   you may encounter GPU hangs in MI_WAIT_FOR_EVENT.
479
480 * Fix a missing corner pixel when drawing rectangles with PolyLines
481   https://bugs.freedesktop.org/show_bug.cgi?id=55484
482
483 * Don't try to use Y-tiling colour buffers with mesa/i915c as mesa
484   doesn't support them and wil fallback to software rendering
485
486 * Ensure that any cached mmaps are invalidated for a SwapBuffers
487   https://bugs.freedesktop.org/show_bug.cgi?id=60042
488
489 * Correctly handle the composition of rotated displays too large for the
490   3D pipeline
491   https://bugs.freedesktop.org/show_bug.cgi?id=60124
492
493 * Fix the computation of the planar video frame size
494   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1104180
495
496
497Release 2.20.19 (2013-01-20)
498============================
499A quick release as the last broke USB DisplayLink slave outputs badly. The
500performance of those displays was unusable due to an inadvertent change that
501caused us to flush the entire scanout over the USB for every drawing
502operation.
503
504 * Implement the GNOME Build API. A couple of minor changes to make
505   integrators and distributors lives a little easier, or at least more
506   consistent.
507
508 * Correctly offset inplace trapezoids for subwindows, such as the GTK+
509   close button after it has a background image uploaded.
510
511 * Explicitly prevent ring-switching for synchronized rendering to
512   scanouts (for vsync).
513
514 * Clip dirty region to slave pixmaps (otherwise UDL is nigh unusuable)
515   https://bugs.freedesktop.org/show_bug.cgi?id=59539
516
517
518Release 2.20.18 (2013-01-16)
519============================
520A bunch of miscellaneous fixes for assertion failures and various
521performance regressions when mixing new methods for offloads, along with
522a couple of improvements for rendering with gen4.
523
524 * Remove use of packed unnormalized texture coordinates on gen4/5 as
525   these GPUs do not support unnormalized coordinates in the sampler.
526
527 * Remove dependency upon x86 asm for cross-building to unsupported
528   architectures.
529   https://bugs.gentoo.org/show_bug.cgi?id=448570
530
531 * Apply damage around PRIME updates in the correct order.
532
533 * Correctly read the initial backlight level for when the user
534   overrides UXA's choice of backlight controller.
535
536 * Throttle UXA and prevent it queuing work much faster than the GPU can
537   complete it. This realised itself in impossible performance figures and
538   the entire display freezing for several seconds whlist the GPU caught
539   up. One side effect is that it also caused the DDX to consume more
540   memory than was required as it could not recycle buffers quick
541   enough, and in some cases this produces a marked improvement in
542   performance. Also note on gen2/3 this requires a new libdrm [2.4.41]
543   in order to prevent a bug causing the DDX to fallback to swrast.
544
545Release 2.20.17 (2012-12-26)
546============================
547A minor update to prepare for co-operating with the kernel over managing
548stability on 830gm/845g. On this pair of chipsets, the kernel will perform
549an extra copy of the batchbuffer into reserved memory, which prevents them
550from randomly dying. However, that extra copy does have a noticeable
551impact upon throughput, so we also have a mechanism for userspace to
552opt-out of the kernel workaround and take responsibility for ensuring its
553batches are coherent.
554
555 * Build fixes against xorg-1.14
556   https://bugs.freedesktop.org/show_bug.cgi?id=58552
557   https://bugs.freedesktop.org/show_bug.cgi?id=58406
558
559 * Fixed the origin of cropped (textured) video windows (Xv and XvMC)
560   https://bugs.freedesktop.org/show_bug.cgi?id=23033
561
562 * Fix potential corruption when using images larger than ~1GiB
563
564
565Release 2.20.16 (2012-12-15)
566============================
567Rejoice! We have found a trick to make 830gm/845g stable at long last.
568Ever since the switch to GEM and dynamic video memory, those early
569second generation chipsets have been plagued by instability. The lack of
570flushing cachelines from the CPU to GMCH was eventually solved by using
571an undocmented bit, but 830/845 were still hanging under memory pressure.
572These deaths were all due to garbage finding its way into the command
573streamer, and they go away if we take a leaf out of the original driver
574and never reuse those pages for anything else. So for the first time
575ever, I have been able to complete running the test suite on an 845g,
576even whilst thrashing the page and buffer caches!
577
578 * Run the SF stage as single-threaded on gen4 to workaround a few issues
579   https://bugs.freedesktop.org/show_bug.cgi?id=57410
580
581 * Keep the scanout SURFACE_STATE separate to avoid overriding its
582   memory access control on gen6/7 (i.e. writes to the scanout need to
583   be kept out of the render cache)
584
585 * Tune batch flushing after an operation to an exported surface under a
586   compositor.
587
588 * Make sure the source is on the CPU for inplace composition of trapezoids
589   using the CPU
590   https://bugs.freedesktop.org/show_bug.cgi?id=56825
591
592 * Immediately flush in the block hander after a split batch to reduce
593   latency between the two halves of an operation.
594   https://bugs.freedesktop.org/show_bug.cgi?id=51718
595
596 * Install a fallback config if we fail to install the desired config
597   at VT switch (i.e. booting, after resume with 3 incompatible pipes on
598   Ivybridge)
599
600 * Pin batches to avoid CS incoherence on 830/845
601   https://bugs.freedesktop.org/show_bug.cgi?id=26345
602
603
604Release 2.20.15 (2012-12-03)
605============================
606And lo, enabling more of the common acceleration paths for gen4 revealed
607another lurking bug - something is wrong with how we prepare Y-tiling
608surfaces for rendering. For the time being, we can surreptiously disable
609them for gen4 and avoid hitting GPU hangs.
610
611 * Avoid clobbering the render state after failing to convert the
612   operation to use the blitter.
613   https://bugs.freedesktop.org/show_bug.cgi?id=57601
614
615 * Disable shadow tracking upon server regeneration, and so fix a crash
616   if you restart the server whilst a RandR transform (e.g. rotation) is
617   in effect.
618   https://bugs.freedesktop.org/show_bug.cgi?id=52255
619   https://bugs.freedesktop.org/show_bug.cgi?id=56608
620
621
622Release 2.20.14 (2012-11-26)
623============================
624The highlight of this release is gen4, from 965g to gm45. Quite an old
625bug surfaced in the shader assembly, sparking a chance to review a few
626design choices within that backend and experiment on fresh ways to
627workaround the remaining issues.
628
629 * Avoid using inplace XOR'ed uploads for very large buffers
630   https://bugs.freedesktop.org/show_bug.cgi?id=57031
631
632 * Fix the gen4/5 opacity shader
633   https://bugs.freedesktop.org/show_bug.cgi?id=57054
634
635 * Queue a pending vblank request after flip completion
636   https://bugs.freedesktop.org/show_bug.cgi?id=56423
637
638 * Avoid migrating an uninitialised pixmap for use as a render source
639   https://bugs.freedesktop.org/show_bug.cgi?id=47597
640
641 * Improve handing of texture fallbacks for 830/845.
642   https://bugs.freedesktop.org/show_bug.cgi?id=57392
643
644
645Release 2.20.13 (2012-11-11)
646============================
647Nothing but bug fixes. Many thanks to everyone who took the time to
648report their issues, and for their help in improving the driver.
649
650 * Sanity check the platform probe points to our expected i915 device
651   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1069031
652
653 * Prevent 16-bit overflow for computing the sample area to upload of
654   sources for render operations
655   https://bugs.freedesktop.org/show_bug.cgi?id=56324
656
657 * Clamp the drawable box for migration to prevent 16-bit overflow
658   https://bugs.freedesktop.org/show_bug.cgi?id=56591
659
660 * Disable RandR hotplug events if Xinerama is enabled and thereby prevent
661   a crash upon hotplug
662   https://bugs.freedesktop.org/show_bug.cgi?id=55260
663
664 * Call ValidatePicture before attempting to flatten the alphamaps
665   https://bugs.freedesktop.org/show_bug.cgi?id=56367
666
667 * Clip the trapezoid correctly if it ends on the boundary pixel
668   https://bugs.freedesktop.org/show_bug.cgi?id=56395
669
670 * Make sure the pipeline choice is propagated to the scanline wait
671   across a batch flush
672   https://bugs.freedesktop.org/show_bug.cgi?id=47597
673
674 * Set the valid drawable box when choosing placement of BLT composite ops
675   https://bugs.freedesktop.org/show_bug.cgi?id=47597
676
677 * Prevent use-after-free when promoting a partial-GPU bo to a full-GPU bo
678   https://bugs.freedesktop.org/show_bug.cgi?id=56591
679
680 * gen4 opacity spans require the per-rectangle workaround
681   https://bugs.freedesktop.org/show_bug.cgi?id=55500
682
683 * Prevent use of invalid damage pointers when redirecting rendering
684   https://bugs.freedesktop.org/show_bug.cgi?id=56785
685
686
687Release 2.20.12 (2012-10-20)
688============================
689More bug reports, more bug fixes! Perhaps the headline feature is
690that with a secure batches, coming to a 3.8 kernel near you, we may
691finally have the ability to perform updates to the scanout synchronized
692to the refresh rate on later SandyBridge and IvyBridge chipsets. It comes
693at quite a power cost as we need to keep the GPU out of its power saving
694modes, but it should allow legacy vsync to function at last. But this
695should allow us to address a longstanding issue with tearing on
696SandyBridge+.
697
698 * Fix component-alpha rendering on IvyBridge, for example subpixel
699   antialiased glyphs.
700   https://bugs.freedesktop.org/show_bug.cgi?id=56037
701
702 * Flush before some "pipelined" state changes on gen4. The evidence is
703   that the same flushes as required on gen5+ are also required for gen4.
704   https://bugs.freedesktop.org/show_bug.cgi?id=55627
705
706 * Prevent a potential crash when forcing a stall on a busy CPU bo
707   https://bugs.freedesktop.org/show_bug.cgi?id=56180
708
709[Release 2.20.11 contained a typo causing UXA to fail immediately.]
710
711Release 2.20.10 (2012-10-14)
712============================
713The last couple of weeks have been fairly retrospective, a dive into
714prehistory tidying up the earlier generations which lay languishing as
715the core progressed and lead to a number of annoying core bugs being
716fixed.
717
718 * Release DRM master earlier during shutdown so switching between
719   multiple X servers works automatically.
720   https://bugs.freedesktop.org/show_bug.cgi?id=55446
721
722 * Suppress error propagation from DRI2GetMSC and behave as if the pipe
723   was simply off to avoid unexpected errors in the clients
724   https://bugs.freedesktop.org/show_bug.cgi?id=55395
725
726 * A few fixes to i8xx batch emission, ensuring that the GPU is always
727   in a valid state.
728   https://bugs.freedesktop.org/show_bug.cgi?id=55455
729
730 * Prevent a use-after-free during UXA shutdown due to inspecting
731   the glamor flags after the glamor interface had been freed.
732
733 * Prevent a crash combining TearFree and rotations.
734   https://bugs.freedesktop.org/show_bug.cgi?id=55527
735
736 * Correct a missing damage upload along PutImage after using the CPU bo
737   as a source for the GPU.
738   https://bugs.freedesktop.org/show_bug.cgi?id=55508
739
740 * Fix compilation for older glibc without O_CLOEXEC
741   https://bugs.freedesktop.org/show_bug.cgi?id=55577
742
743 * Fix out-of-tree builds failing to recompile the gen4 assemblies
744   https://bugs.freedesktop.org/show_bug.cgi?id=55645
745
746 * Fix non-standard build host configuration handling for intel-gen4asm
747   https://bugs.freedesktop.org/show_bug.cgi?id=55646
748
749 * Fix a potential batch buffer overflow when replacing the last BLT fill
750   operation with a copy
751   https://bugs.freedesktop.org/show_bug.cgi?id=55700
752
753 * Flush the render pipeline more frequently on Ironlake as not all
754   pipelined state changes are.
755   https://bugs.freedesktop.org/show_bug.cgi?id=51422
756
757 * Detect when we need to read the destination for the background raster
758   op during fallbacks.
759   https://bugs.freedesktop.org/show_bug.cgi?id=55810
760
761 * Avoid a potential deference of an invalid CPU mmap after doing an
762   inplace tiled upload.
763   https://bugs.freedesktop.org/show_bug.cgi?id=55812
764
765 * Prevent sign extension when packing the upload data for CopyPlane
766   https://bugs.freedesktop.org/show_bug.cgi?id=55823
767
768 * Fix some render corruption with a UDL slave output and pageflipping
769
770
771Release 2.20.9 (2012-09-29)
772===========================
773And so it came to pass that a critical bug was uncovered in UXA. The
774kernel does not like to pageflip when the pipe is off, yet due to the
775delayed nature of a pageflip and the relaxed checking performed by UXA,
776we could request a pageflip after turning off the display (DPMS). The
777kernel rejected that pageflip and the error handling path failed to
778restore sanity, and when the screen came back it was stuck on the image
779seen before it went to sleep. (Note that there are also some related
780kernel bugs, but this update should prevent the most conspicious of the
781freezes.) Many thanks to Timo Aaltonen for his efforts in tracking down
782the issue.
783
784In other news:
785
786  * Prepare for xorg-1.14, the api is being tweaked again.
787
788  * Handle early FreeScreen in UXA.
789    https://bugs.freedesktop.org/show_bug.cgi?id=55346
790
791  * Reenable XvMC support
792
793  * Do not replace the GPU bo when uploading into the shadow/CPU copy
794    https://bugs.freedesktop.org/show_bug.cgi?id=54978
795
796  * Fix use of an uninitialised GC when drawing glyphs to a depth=1 pixmap
797
798
799Release 2.20.8 (2012-09-16)
800===========================
801Another new small feature, another new release. And a few more bugs
802fixed as well! But what is this new feature, do I hear you ask? Why, it
803is nothing less than enabling the ValleyView SDV! The lucky person to
804have their hands on one will now be able to enjoy X in full TechniColor.
805
806For the rest of us, a few more bugs were fixed with interesting
807combinations of software and rendering patterns.
808
809 * Add an extra layer of defence against trying to use a non-GEM device
810   with UXA. This should already be taken care of with the new probe,
811   but the extra sanity check already existed in the code but was doing
812   nothing.
813   https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/962892
814
815 * Fix computation of valid CRTCs bitmask for ZaphodHead
816   Some systems can only handle certain outputs on certain pipes
817   (Screens), and this information was not being propagated through to X
818   and randr correctly, breaking valid configurations and not detecting
819   invalid configurations correctly.
820
821 * Disable global glyph caching with ZaphodHeads.
822   The glyph privates need to be Screen private, but at the moment are
823   global leading to conflicts and invalid rendering with multiple heads.
824   https://bugs.freedesktop.org/show_bug.cgi?id=54707
825
826 * Prevent direct read back of unmappable buffers
827   https://bugs.freedesktop.org/show_bug.cgi?id=54808
828
829 * Tile large uploads whilst replacing the alpha channel
830   https://bugs.freedesktop.org/show_bug.cgi?id=54808
831
832 * Correct the source offset when converting a RENDER composite
833   operation into a BLT composite operation.
834   https://bugs.freedesktop.org/show_bug.cgi?id=54868
835
836 * Correct a minor typo in flattening alphamaps which caused the
837   replacement pixmap to often end up with a height of zero.
838   https://bugs.gentoo.org/show_bug.cgi?id=434860
839
840 * Don't discard the CPU damage if only part of is being replaced.
841   This is most evident when combining software renders like Opera with
842   accelerated rendering.
843   https://bugs.freedesktop.org/show_bug.cgi?id=54937
844
845Release 2.20.7 (2012-09-08)
846===========================
847Continuing the flurry of releases, this week saw the release of
848xorg-1.13 bringing with it the first steps towards hotpluggable gpu
849support and the ability to offload DRI rendering onto other GPUs. In
850light of that advance, this release includes the support code by Dave
851Airlie to integrate PRIME into xf86-video-intel and make those new
852features available.
853
854Aside from landing PRIME, a few other bugs were fixed since 2.20.6:
855
856 * Fix framebuffer leak on server regeneration
857
858 * Fix texture cache flushing on IvyBridge and Kwin with plastique
859   https://bugs.freedesktop.org/show_bug.cgi?id=54488
860
861 * Redirect large solid fills on SandyBridge+. By large I mean greater
862   than 8192 or 16384 pixels on SandyBridge and IvyBridge respectively.
863   https://bugs.freedesktop.org/show_bug.cgi?id=54134
864
865 * Fix up backlight option handling in the manpage and uxa.
866   https://bugs.freedesktop.org/show_bug.cgi?id=54397
867
868 * Unbreak ZaphodHeads.
869   https://bugs.freedesktop.org/show_bug.cgi?id=52438
870
871Release 2.20.6 (2012-09-02)
872===========================
873A serious bug that caused a crash on SandyBridge and IvyBridge when
874mixing CPU and GPU operations on the same buffer, and an annoyance from
875bad scheduling of windowed swapbuffer updates causing low framerates and
876jitter. Plus the usual smattering of assertion fixes and a long standing
877issue with incoherent page access to a streaming buffer.
878
879 * Low frame rates in Blobby Valley when "fullscreen"
880   https://bugs.freedesktop.org/show_bug.cgi?id=54274
881
882 * Incoherent concurrent access with the CPU and GPU
883   https://bugs.freedesktop.org/show_bug.cgi?id=51422
884   https://bugs.freedesktop.org/show_bug.cgi?id=52299
885
886 * Add Option "Backlight" to override automatic selection of the
887   backlight interface.
888
889 * Avoid overwriting the composite operation info when testing if we
890   can transfer the operation to the BLT.
891
892Release 2.20.5 (2012-08-26)
893===========================
894Another silly bug found, another small bugfix release. The goal was for
895the driver to bind to all Intel devices supported by the kernel.
896Unfortunately we were too successful and started claiming Pouslbo,
897Medfield and Cedarview devices which are still encumbered by propietary
898IP and not supported by this driver.
899
900Bugs fixed since 2.20.4:
901
902 * Only bind to Intel devices using the i915 kernel module
903
904 * Regression in the bitmap-to-region code, e.g. icewm window buttons
905   https://bugs.freedesktop.org/show_bug.cgi?id=53699
906
907Release 2.20.4 (2012-08-18)
908===========================
909Continuing the small bugfix releases, the only real feature is initial
910enabling for Haswell for the purpose of rendering verification and
911validation - by no means is it complete!
912
913Bugs fixed since 2.20.3:
914
915  * Some potential errors along failure paths found by a static analyser
916  with the help of Zdenek Kablac.
917
918  * Eliminate zero-sized rectangles from PolyFillRectangles as the code
919  assumes that they did not exist and so caused corruption.
920
921  * Remove the UXA warning for failing to tile the front buffer if it is
922  disallowed by hardware, and so expected.
923
924  * Fix the validation of the XV pipe parameter.
925
926  * Fix 8x8 tiled pattern fills
927  https://bugs.freedesktop.org/show_bug.cgi?id=53353
928
929  * Fix compile failure when using --with-builderstring
930
931  * Restore w/a flush for gen4 fill/copy/video, fortunately rare
932  operations as at least for fill/copy we prefer to use the BLT.
933  https://bugs.freedesktop.org/show_bug.cgi?id=53119
934
935  * Restore preferred use of the RENDER ring for SNB+ DRI copies.
936
937Release 2.20.3 (2012-08-04)
938===========================
939Just a minor bugfix for gen4 chipsets (965gm, gm45 and friends) that
940crept into 2.20.2. As an added bonus, the pessimistic workaround for a
941GPU hang on gen4 has been relaxed and the shaders have been overhauled
942which should pave the way to eliminating the last of the uncommon CPU
943operations, along with immediately realising a small perforamnce
944improvement.
945
946Bugs fixed since 2.20.2:
947
948 * Update DPMS bookkeeping after modeset
949   https://bugs.freedesktop.org/show_bug.cgi?id=52142
950
951 * Avoid overlapping gpu/cpu damage after ignoring cpu damage in the
952   consideration of placement for the operation.
953
954 * Enable acceleration by default on 830gm/845g. The GMCH on this pair
955   of chipsets is notoriously incoherent, so the GPU is almost certainly
956   going to hang at some point, though unlikely to hang the system and
957   should automatically disable acceleration (and thence behave
958   identically as if the acceleration was disabled from the start).
959   Option "NoAccel" can be used to disable all 2D acceleration and
960   Option "DRI" can be used to disable all 3D acceleration.
961   https://bugs.freedesktop.org/show_bug.cgi?id=52624
962
963 * Fix vertex bookkeeping for gen4 that was causing corruption in the
964   command stream.
965
966Release 2.20.2 (2012-07-27)
967===========================
968For the last 9 months, since 2.16.901, we have been shipping a driver that
969does not work on IvyBridge GT1 systems (HD2500 graphics); we were telling
970the GPU to use an invalid number of threads for the pixel shader and this
971in turned caused the GPU to hang.
972
973Also fixed since the last release just a few days ago:
974
975* Support for the gmux backlight controller on Apple laptops
976  https://bugs.freedesktop.org/show_bug.cgi?id=52423
977
978* Fix X -configure not to list this driver as matching any Intel device,
979  just the VGA class devices will do!
980
981* A crash in SNA when repeatedly switching xrandr rotations
982
983* Corruption in SNA observed in kwin on IvyBridge
984  https://bugs.freedesktop.org/show_bug.cgi?id=52473
985
986Release 2.20.1 (2012-07-22)
987===========================
988A week in, grab the brown paper bags, for it is time to reveal a couple
989of critical bugs that spoilt the 2.20.0 release.
990
991Firstly we have the restoration of DRI for i810. I am sure that the
992solitary user will be overjoyed in a couple of years when a new xserver
993is forced upon him. That enjoyment will be short-lived when as no actual
994acceleration remains, not even shadow, for the chipset.
995
996Perhaps a little more wildly felt, I hope!, will be that the SNA
997fallbacks were broken on 64-bit machines if they required clipping. One
998little misplaced cast of a pointer, and the screen is filled with
999corruption.
1000
1001Among the other tweaks this week:
1002
1003* A bug affecting gen4 handling of trapezoids was fixed, and CPU
1004  overhead reduced.
1005  https://bugs.freedesktop.org/show_bug.cgi?id=52158
1006
1007* A fix for a bug causing corruption of a DRI2 unredirected client
1008  window that was resized whilst under a compositor.
1009
1010* Support for snoopable buffers on non-LLC architectures, coming to
1011  a future kernel. The aim to accelerate transfers between the CPU
1012  and the GPU, in particular to dramatically improve readback
1013  performance, and to further minimise clflushes.
1014
1015* Improvement to the composite performance on GT2 SandyBridge and
1016  IvyBridge devices, in particular the render copy is significantly
1017  improved.
1018
1019* Improved handling for when acceleration is disabled, including
1020  permitting DRI2 to remain supported even if the X server believes
1021  the GPU wedged.
1022
1023* Shadow support was dropped from UXA as it was neither complete nor
1024  correct, use SNA instead.
1025
1026Release 2.20.0 (2012-07-15)
1027===========================
1028First the big news, a new acceleration method that aims to be faster and
1029consume far less CPU than UXA is now available for selection at runtime.
1030This snazzy new architecture can be selected through use of
1031  Option "AccelMethod" "sna"
1032in your xorg.conf. Whilst it has been under development for some time, it
1033has not yet had the same degree of widespread testing of UXA, so tread
1034lightly. Try it and if you spot anything that can be improved, please do
1035report a bug.
1036
1037Otherwise we have the usual smattering of bug fixes for UXA:
1038
1039 * Use a white source whilst adding glyphs to the glyph mask
1040   (This fixes blank glyphs if using a font that mixes ARGB and A glyphs.)
1041
1042 * Avoid fallbacks for glyph-to-dst in e.g. gnome-terminal
1043   https://bugs.freedesktop.org/show_bug.cgi?id=50508
1044
1045 * Force unused outputs off when VT switching
1046   https://bugs.freedesktop.org/show_bug.cgi?id=50772
1047
1048 * Copy the fbcon across to the Screen pixmap at startup.
1049   (This patch has been kicking around in the distributions for years.)
1050
1051 * Many missed malloc failures checks and forgotten frees found by a static
1052   analyzer. Thanks Zdenek Kabelac!
1053
1054 * Leak of the back buffer when terminating an application after pageflipping
1055   https://bugs.freedesktop.org/show_bug.cgi?id=50670
1056
1057 * Double check that the pipe is on before emitting a WAIT_ON_EVENT.
1058   In conjunction with an uptodate kernel, this should eliminate any
1059   hangs when changing resolutions or adding/removing displays.
1060   https://bugs.freedesktop.org/show_bug.cgi?id=50668
1061
1062 * Update to new Xorg APIs. Future proofing for the next generation of
1063   hotplug Xorg display servers.
1064
1065Many thanks to everyone who has reported a bug and otherwise helped to
1066improve the driver.
1067
1068Release 2.19.0 (2012-04-29)
1069===========================
1070More stability fixes for UXA and support for another variant of IvyBridge.
1071Given the severity of the stability fixes, I strongly recommend everybody
1072to upgrade to 2.19.0.
1073
1074* Prevent waiting on scanlines whilst not in control of the VT and therefore
1075  whilst referencing foreign CRTC configurations.
1076
1077* Pixmap (and bo leak) during fallback glyph composition
1078
1079* Remove broken acceleration for rendering glyphs directly upon the
1080  destination pixmap, exposed by cairo-1.12.0 (and coincidentally fix
1081  another Pixmap leak upon fallback handling).
1082
1083* Add support for Ivy Bridge GT2 Server chipset [PCI id 0x016a]
1084
1085* Remove broken damage flushing with CompositeRectangles
1086  https://bugs.freedesktop.org/show_bug.cgi?id=32547
1087
1088* Fix crash upon server start with multiple monitors
1089  https://bugs.freedesktop.org/show_bug.cgi?id=47395
1090
1091* Fix composition issues resulting from overly aggressive Pixmap reuse
1092  https://bugs.freedesktop.org/show_bug.cgi?id=47345
1093
1094Release 2.18.0 (2012-02-24)
1095===========================
1096Time passes, a few more bugs have crept out of the woodwork that are a
1097compelling reason to update.
1098
1099Bugs fixed in this release (compared to 2.17.0)
1100-----------------------------------------------
1101
1102* Limit maximum object size so that all of the source, mask and
1103  destination can be mapped into the aperture simultaneously by basing the
1104  limit on the mappable aperture size rather than the size of the total
1105  GATT.
1106
1107* Incorrect clipping of polygons
1108  https://bugs.freedesktop.org/show_bug.cgi?id=43649
1109  Regression from 2.15.901
1110
1111* Limit number of VMA cached to avoid hitting the per-process VMA limit
1112  There still is a residual bug in that we seem to have so many objects
1113  floating around in the first place and that still leads to exhaustion
1114  of system limits.
1115  https://bugs.freedesktop.org/show_bug.cgi?id=43075
1116  https://bugs.freedesktop.org/show_bug.cgi?id=40066
1117
1118* Latency in processing user-input during continuous rendering
1119
1120Release 2.17.0 (2011-11-16)
1121==============================
1122A few months have passed, and we have accumulated a surprising number of
1123bug fixes. Oops! We would strongly encourage everyone to upgrade.
1124
1125Bugs fixed in this snapshot (compared to 2.16.902)
1126--------------------------------------------------
1127
1128* Video clobbering composite batch state
1129  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=635953
1130
1131Snapshot 2.16.902 (2011-11-13)
1132==============================
1133This is the second release candidate in preparation for the upcoming
11342.17.0 release. We will appreciate any feedback we can get from
1135testing of this snapshot, and thanks to everyone who tested and reported
1136upon issues in the last snapshot.
1137
1138A very old bug demonstrating that nobody has ever used alpha maps, and a
1139couple of more recent mistakes were found.
1140
1141Bugs fixed in this snapshot (compared to 2.16.901)
1142--------------------------------------------------
1143
1144 * Incorrect reuse of surface bindings within a batch for multiple formats
1145   https://bugs.freedesktop.org/show_bug.cgi?id=40926
1146
1147 * Nothing was rendered for text with procedural sources
1148   https://bugs.freedesktop.org/show_bug.cgi?id=31819
1149
1150 * Handle fallbacks involving alpha maps
1151
1152Snapshot 2.16.901 (2011-10-30)
1153==============================
1154This is the first release candidate in preparation for the upcoming
11552.17.0 release. We will appreciate any feedback we can get from
1156testing of this snapshot.
1157
1158The basis of this release are to push out a couple of critical stability
1159fixes for SandyBridge and IvyBridge. There is very little else to see
1160here. Those 200+ other commits do not exist...
1161
1162Bugs fixed in this snapshot (compared to 2.16.0)
1163--------------------------------------------------
1164
1165 * Workaround blitter hang on SandyBridge and IvyBridge
1166   https://bugzilla.kernel.org/show_bug.cgi?id=27892
1167   https://bugs.freedesktop.org/show_bug.cgi
1168
1169 * Workaround pipe control issues on SandyBridge
1170
1171 * Use correct maximum PS thread count on IvyBridge
1172
1173 * Protect against failed pixmap allocation for XV
1174   https://bugs.freedesktop.org/show_bug.cgi?id=40439
1175
1176Release 2.16.0 (2011-08-09)
1177==============================
1178A new quarter, a new release! The key feature of this release, looking past
1179the bug fixes, is the enabling of IvyBridge acceleration. We have also fixed
1180many bugs and graphical glitches and would encourage everyone to upgrade.
1181
1182Bugs fixed in this snapshot (compared to 2.15.901)
1183--------------------------------------------------
1184
1185* Build fix for xserver-1.7.7
1186
1187Snapshot 2.15.901 (2011-07-30)
1188==============================
1189This is the first release candidate in preparation for the upcoming
11902.16.0 release. We will appreciate any feedback we can get from
1191testing of this snapshot.
1192
1193The highlight of this snapshot is the full enabling of IvyBridge with
1194acceleration for Render and Xv, along with handling of the shared render
1195buffer allocations required for Mesa. Also of note is that deep-color
1196support is enabled (for all chipsets), which allows you to drive your
119730-bit monitor at its native colour depth.
1198
1199Bugs fixed in this snapshot (compared to 2.15.0)
1200--------------------------------------------------
1201
1202* Misuse of the Resource database causing crashes after DRI clients close
1203  https://bugs.freedesktop.org/show_bug.cgi?id=37700
1204
1205* Crash on large strings
1206  https://bugs.freedesktop.org/show_bug.cgi?id=36860
1207
1208* Incorrect rendering for some core drawing operations
1209  http://bugs.freedesktop.org/show_bug.cgi?id=28768
1210  http://bugs.freedesktop.org/show_bug.cgi?id=28798
1211  http://bugs.freedesktop.org/show_bug.cgi?id=28908
1212  http://bugs.freedesktop.org/show_bug.cgi?id=29401
1213
1214* Crash in Xv due to insufficient checks on batch space
1215  https://bugs.freedesktop.org/show_bug.cgi?id=36319
1216
1217Release 2.15.0 (2011-04-14)
1218==============================
1219We are pleased to announce this major release of the xf86-video-intel
1220driver, roughly on schedule at 3 months since 2.14.0. With the many bug
1221fixes in this release, we encourage everyone to upgrade to 2.15.
1222
1223The priority for this quarter has been simply to be unexciting and stabilise
1224the driver further, seeking to capitalise upon the improvements elsewhere
1225in the stack.
1226
1227Bugs fixed in this snapshot (compared to 2.14.903)
1228--------------------------------------------------
1229
1230* Turn off relaxed fencing by default for older chipsets
1231  This was continuing to destabilize those system, so for the release
1232  we disabled the feature. If you wish to help us debug this, you can
1233  re-enable the optimisation with Option "RelaxedFencing" "True".
1234  Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=36147
1235
1236* Build fix for xserver-1.7.7
1237
1238* KDE glitches on SNB
1239  [Technically fixed in the previous snapshot, but I'm really pleased
1240   that this got fixed in time for the release!]
1241  https://bugs.freedesktop.org/show_bug.cgi?id=35808
1242
1243
1244Snapshot 2.14.903 (2011-04-11)
1245==============================
1246This is the third release candidate in preparation for the upcoming
12472.15.0 release. We will appreciate any feedback we can get from
1248testing of this snapshot.
1249
1250There was a bit of churn since 2.14.902 as a potential fix for a
1251performance regression was tried but had to reverted when it was found to
1252cause glitches running Compiz on SandyBridge. Otherwise, there were just a
1253couple of fixes for building against old xservers and running on an
1254obscure chipset.
1255
1256Bugs fixed in this snapshot (compared to 2.14.902)
1257--------------------------------------------------
1258
1259* Prevent issuing an invalid scanline wait command
1260  https://bugs.freedesktop.org/show_bug.cgi?id=35576
1261
1262* The 946GZ in not a 945, but a 965.
1263  https://bugs.freedesktop.org/show_bug.cgi?id=35854
1264
1265* Fix tile sizes for gen2 (finally).
1266
1267* Allow building of recent dri2 changes against old xservers.
1268
1269
1270Snapshot 2.14.902 (2011-03-29)
1271==============================
1272This is the second release candidate in preparation for the upcoming
12732.15.0 release. We will appreciate any feedback we can get from
1274testing of this snapshot.
1275
1276As befits testing of release candidates, no major regression was found and
1277a couple more bugs have been fixed.
1278
1279Bugs fixed in this snapshot (compared to 2.14.901)
1280--------------------------------------------------
1281
1282* Clients disappearing with pending swaps
1283
1284* Incorrect clipping of Xv output on i915 across extended desktops
1285  https://bugs.freedesktop.org/show_bug.cgi?id=35346
1286
1287* Introduction of a LinearFramebuffer option. (Defaults to tiled for
1288  performance and power saving.)
1289
1290
1291Snapshot 2.14.901 (2011-03-02)
1292==============================
1293This is the first release candidate in preparation for the upcoming
12942.15.0 release. We will appreciate any feedback we can get from
1295testing of this snapshot.
1296
1297Still no further along my grandiose plans to improve Render performance,
1298aside from the performance tuning lower in the stack, instead we have had
1299a steady stream of bug fixes.
1300
1301Bugs fixed in this snapshot (compared to 2.14.0)
1302------------------------------------------------
1303
1304* Green pixels within partially off-screen video playback
1305  https://bugs.freedesktop.org/show_bug.cgi?id=24767
1306
1307* Defer creation of the glyph cache to generation startup
1308  https://bugs.freedesktop.org/show_bug.cgi?id=33412
1309
1310* Incorrect maximum addresses for video decoder state
1311  https://bugs.freedesktop.org/show_bug.cgi?id=34017
1312
1313* Failure to handle oversized temporary surfaces
1314  https://bugs.freedesktop.org/show_bug.cgi?id=34399
1315
1316* Relaxed tiling corruption on gen2
1317
1318* Crash when destroying a foreign DRI drawable
1319  https://bugs.freedesktop.org/show_bug.cgi?id=34787
1320
1321
1322Release 2.14.0 (2011-01-07)
1323===========================
1324We are pleased to announce this major release of the xf86-video-intel
1325driver, on schedule at 3 months since 2.13.0. With the many bug fixes
1326in this release, we encourage everyone to upgrade to 2.14.
1327
1328The most notable feature of this release is the inclusion of acceleration
1329support, both RENDER and XVIDEO, for Sandy Bridge; note that this requires
1330a 2.6.37 kernel. In improving the acceleration code paths to support Sandy
1331Bridge, we were also able to improve the performance of the general i965+
1332RENDER acceleration. For example on i3, glyph (aa10 and rgb10) performance
1333is about 3x faster, as is performing small copies between windows, with the
1334improvement becoming negligible once you hit around 100 pixels square.
1335
1336
1337Snapshot 2.13.903 (2010-01-04)
1338==============================
1339This is the second release candidate in preparation for the upcoming
13402.14.0 release, and if no showstoppers turn up in wider testing will
1341become the next release. Besides the usual bug fixes, we've included
1342patches by Mario Kleiner to improve vblank handing which both improve the
1343timing accuracy and remove a few potential races -- provided you also
1344have the corresponding fixes in the kernel.
1345
1346Bug fixes
1347---------
1348
1349 * Crash on hot-unpluging with an active fullscreen GL application, e.g. a
1350   compositing WM or video player
1351   https://bugs.freedesktop.org/show_bug.cgi?id=32770
1352
1353 * Freeze upon maximising client-side software renderers, e.g. flash.
1354   https://bugs.freedesktop.org/show_bug.cgi?id=31367
1355
1356 * Fix confusion over G35 generation
1357   https://bugs.freedesktop.org/show_bug.cgi?id=32478
1358
1359
1360Snapshot 2.13.902 (2010-12-10)
1361==============================
1362This is the first release candidate in preparation for the upcoming
13632.14.0 release. Sandybridge support should be now be ready for use,
1364and we will appreciate any feedback we can get from this snapshot.
1365Improving the code base for Sandybridge should also have had knock
1366on effects for earlier chips as well, in particular Ironlake and to
1367a lesser extent i965.
1368
1369New requirements compared to 2.13
1370---------------------------------
1371 * Librdrm >= 2.4.23
1372
1373Bug fixes
1374---------
1375
1376 * Crash after failing to allocate memory
1377   https://bugs.freedesktop.org/show_bug.cgi?id=31487
1378
1379 * EFAULT during pwrite
1380   https://bugs.freedesktop.org/show_bug.cgi?id=29752
1381
1382 * Mark outputs as on and restore backlight after mode-set
1383
1384 * A hang with i965+ with newer kernels.
1385
1386
1387Snapshot 2.13.901 (2010-11-05)
1388==============================
1389This is an intermediate snapshot of ongoing driver development. The
1390primary purpose of this snapshot is to capture some recent
1391improvements, (particularly in Sandybridge support), for further
1392testing.
1393
1394Release 2.13.0 (2010-09-30)
1395===========================
1396We are pleased to announce this major release of the xf86-video-intel
1397driver, on schedule at 3 months since 2.12.0. With the many bug fixes
1398in this release, we encourage everyone using 2.12 to upgrade to 2.13.
1399
1400[This release is functionally identical to the earlier 2.12.902
1401release candidate.]
1402
1403New requirements compared to 2.12
1404---------------------------------
1405 * Librdrm >= 2.4.22
1406
1407Bug fixes
1408---------
1409 * Attempt to fix infinite MI_WAIT_FOR_EVENT while watching video
1410   https://bugs.freedesktop.org/show_bug.cgi?id=28964
1411
1412   [Various new checks have been added to the video code here, but the
1413   bug fix hasn't yet been verified by the original reporter.]
1414
1415 * Fix buffer-object leak
1416   https://bugs.freedesktop.org/show_bug.cgi?id=26946
1417
1418 * Fix memory leak on server reset
1419
1420 * Fix crash due to unchecked pixmap allocation
1421   https://bugs.freedesktop.org/show_bug.cgi?id=29187
1422
1423 * Fix for video artifacts when using dualscreen
1424   https://bugs.freedesktop.org/show_bug.cgi?id=29213
1425
1426 * Fix for incorrect characters in gnome-terminal when using compiz
1427   https://bugs.freedesktop.org/show_bug.cgi?id=28438
1428
1429 * Fix for hanging, full-screen applications, (flash, compiz, etc.)
1430   https://bugs.freedesktop.org/show_bug.cgi?id=29584
1431
1432 * Fix selection of backlight device on multi-GPU systems
1433   https://bugs.freedesktop.org/show_bug.cgi?id=29273
1434
1435 * Fix to avoid crash with extremely large glyphs
1436   https://bugs.freedesktop.org/show_bug.cgi?id=29430
1437
1438 * Fix for eDP panels incorrectly being given only a single, valid mode
1439   https://bugs.freedesktop.org/show_bug.cgi?id=30069
1440
1441 * Fix GPU hang involving clipped SRC copies
1442   https://bugs.freedesktop.org/show_bug.cgi?id=30120
1443
1444 * Fix to compile for 1.6 series X server.
1445
1446 * Fix to retry framebuffer allocation after an initial failure.
1447
1448 * Fix to disable dri2 after fallbacks are forced on.
1449
1450Snapshot 2.12.902 (2010-09-28)
1451==============================
1452This is the second release candidate in preparation for the upcoming
14532.13.0 release. The comes exactly one week after 2.12.902 and includes
1454only a handful of changes.
1455
1456One significant change is that the drive now requires libdrm 2.4.22
1457(or newer), as several people reported compilation failures of
1458xf86-video-intel 2.12.901 with libdrm 2.4.21.
1459
1460Other changes include a fix to compile for the 1.6 series X server, a
1461fix to retry framebuffer allocation after an initial failure, and a
1462fix to disable dri2 after fallbacks are forced on.
1463
1464Snapshot 2.12.901 (2010-09-21)
1465==============================
1466This is the first release candidate in preparation for the upcoming
14672.13.0 release. We will appreciate any feedback we can get from
1468testing of this snapshot.
1469
1470Bugs fixed in this snapshot (compared to 2.12.0)
1471------------------------------------------------
1472 * Attempt to fix infinite MI_WAIT_FOR_EVENT while watching video
1473   https://bugs.freedesktop.org/show_bug.cgi?id=28964
1474
1475   [Various new checks have been added to the video code here, but the
1476   bug fix hasn't yet been verified by the original reporter.]
1477
1478 * Fix buffer-object leak
1479   https://bugs.freedesktop.org/show_bug.cgi?id=26946
1480
1481 * Fix memory leak on server reset
1482
1483 * Fix crash due to unchecked pixmap allocation
1484   https://bugs.freedesktop.org/show_bug.cgi?id=29187
1485
1486 * Fix for video artifacts when using dualscreen
1487   https://bugs.freedesktop.org/show_bug.cgi?id=29213
1488
1489 * Fix for incorrect characters in gnome-terminal when using compiz
1490   https://bugs.freedesktop.org/show_bug.cgi?id=28438
1491
1492 * Fix for hanging, full-screen applications, (flash, compiz, etc.)
1493   https://bugs.freedesktop.org/show_bug.cgi?id=29584
1494
1495 * Fix selection of backlight device on multi-GPU systems
1496   https://bugs.freedesktop.org/show_bug.cgi?id=29273
1497
1498 * Fix to avoid crash with extremely large glyphs
1499   https://bugs.freedesktop.org/show_bug.cgi?id=29430
1500
1501 * Fix for eDP panels incorrectly being given only a single, valid mode
1502   https://bugs.freedesktop.org/show_bug.cgi?id=30069
1503
1504 * Fix GPU hang involving clipped SRC copies
1505   https://bugs.freedesktop.org/show_bug.cgi?id=30120
1506
1507Release 2.12.0 (2010-06-24)
1508===========================
1509We are pleased to announce this major release of the xf86-video-intel
15102D driver. It contains many correctness and performance improvements
1511compared to the 2.11 releases. We encourage everyone using 2.11 to
1512upgrade to 2.12.
1513
1514[This release contains several bug fixes since the earlier 2.11.901
1515release candidate---see below for details.]
1516
1517New requirements compared to 2.11
1518---------------------------------
1519 * Libdrm >= 2.4.21
1520
1521Major performance improvements
1522------------------------------
1523There have been many performance improvements targeting the
1524915G/945G/Pineview graphics chipsets (thank to Chris Wilson!). Areas
1525improved include:
1526
1527 * Major improvements to glyph rendering
1528
1529 * Avoid software fallback when using extended desktops, (for example,
1530   video playback)
1531
1532 * Better performance when using large windows, (for example with
1533   firefox scrolling)
1534
1535 * Avoiding migration ping-pong with gigantic drawing
1536
1537The following speedups (measured with cairo-perf-trace/cairo-traces)
1538capture some of the improvements. These were measured on a Pineview
1539system with X server 1.8.1. This compares how this 2.12 release
1540improves performance compared to 2.11.0:
1541
1542	Trace			Speedup
1543	-----			-------
1544	ocitysmap		5.41x speedup
1545	firefox-talos-gfx       2.41x speedup
1546	gnome-terminal-vim      2.15x speedup
1547	poppler                 1.86x speedup
1548	firefox-planet-gnome	1.43x speedup
1549	midori-zoomed		1.39x speedup
1550	swfdec-giant-steps      1.36x speedup
1551	xfce4-terminal-a1	1.17x speedup
1552	evolution		1.09x speedup
1553	swfdec-youtube		1.06x speedup
1554
1555Other improvements
1556------------------
1557Daniel Vetter improved Xvmc and overlay, adding GEM support.
1558
1559Bug fixes
1560---------
1561There have been many conformance improvements targeting the
1562915G/945G/Pineview graphics chipsets. The rendercheck test suite
1563passes once again and there has been a significant reduction in the
1564number of errors detected by the cairo test suite (notably center
1565sampling and EXTEND_NONE behaviors are fixed).
1566
1567Other notable bug fixes include:
1568
1569 * Fix for bug 28446 Garbled fonts with Mathematica
1570   https://bugs.freedesktop.org/show_bug.cgi?id=28446
1571
1572 * Fixes for page-flipping and other DRI2-handling bugs, (Thanks Jesse
1573   Barnes, and Kristian Høgsberg)
1574
1575 * Avoid corruption when using extended desktops, (for example, video
1576   playback)
1577
1578(The bug fixes below are new since the 2.11.901 release candidate)
1579
1580 * Fix corrupted output when screen rotated
1581   https://bugs.freedesktop.org/show_bug.cgi?id=28461
1582
1583 * Fix corrupted rendering in KDE due to missing flush.
1584
1585 * Fixes i830_uxa_put_image for the following issues:
1586
1587      Bug 28569 - [i965] IGN's flash-based video player crashes X
1588      https://bugs.freedesktop.org/show_bug.cgi?id=28569
1589    
1590      Bug 28573 - [i965] Fullscreen flash and windowed SDL games fail to
1591                  update the screen
1592      https://bugs.freedesktop.org/show_bug.cgi?id=28573
1593
1594 * Fix visual corruption of scrollbar in Chromium
1595
1596 * Fix tiling limits to resolve this issue:
1597
1598      Bug 28497 - Graphics corruption after opening a specific website
1599      https://bugs.freedesktop.org/show_bug.cgi?id=28497
1600
1601Snapshot 2.11.901 (2010-06-14)
1602==============================
1603This is the first release candidate in preparation for the upcoming
16042.12.0 release. We will appreciate any feedback we can get from
1605testing of this snapshot to improve the 2.12.0 release.
1606
1607New requirements compared to 2.10
1608---------------------------------
1609 * Libdrm >= 2.4.19
1610
1611Major performance improvements
1612------------------------------
1613There have been many performance improvements targeting the
1614915G/945G/Pineview graphics chipsets (thank to Chris Wilson!). Areas
1615improved include:
1616
1617 * Major improvements to glyph rendering
1618
1619 * Avoid software fallback when using extended desktops, (for example,
1620   video playback)
1621
1622 * Better performance when using large windows, (for example with
1623   firefox scrolling)
1624
1625 * Avoiding migration ping-pong with gigantic drawing
1626
1627The following speedups (measured with cairo-perf-trace/cairo-traces)
1628capture some of the improvements. These were measured on a Pineview
1629system with a very recent X server, (from git just before 1.8.1). This
1630compares how this 2.12 release candidate improves performance compared
1631to 2.11.0:
1632
1633	Trace			Speedup
1634	-----			-------
1635	ocitysmap		5.41x speedup
1636	firefox-talos-gfx       2.41x speedup
1637	gnome-terminal-vim      2.15x speedup
1638	poppler                 1.86x speedup
1639	firefox-planet-gnome	1.43x speedup
1640	midori-zoomed		1.39x speedup
1641	swfdec-giant-steps      1.36x speedup
1642	xfce4-terminal-a1	1.17x speedup
1643	evolution		1.09x speedup
1644	swfdec-youtube		1.06x speedup
1645
1646Other improvements
1647------------------
1648Daniel Vetter improved Xvmc and overlay, adding GEM support.
1649
1650Bug fixes
1651---------
1652There have been many conformance improvements targeting the
1653915G/945G/Pineview graphics chipsets. The rendercheck test suite
1654passes once again and there has been a significant reduction in the
1655number of errors detected by the cairo test suite (notably center
1656sampling and EXTEND_NONE behaviors are fixed).
1657
1658Other notable bug fixes include:
1659
1660 * Fix for bug 28446 Garbled fonts with Mathematica
1661   https://bugs.freedesktop.org/show_bug.cgi?id=28446
1662
1663 * Fixes for page-flipping and other DRI2-handling bugs, (Thanks Jesse
1664   Barnes, and Kristian Høgsberg)
1665
1666 * Avoid corruption when using extended desktops, (for example, video
1667   playback)
1668
1669Release 2.11.0 (2010-03-29)
1670===========================
1671New in 2.11: DRI2 and page flipping
1672-----------------------------------
1673The most significant new feature of this release is support for new
1674DRI2 APIs, allowing page flipping to occur for swaps that are
1675full-screen and not rotated.
1676
1677Performance improvements in 2.11 compared to 2.10
1678-------------------------------------------------
1679Dramatically improved performance of large pixmaps on
1680memory-constrained hardware, (such as 945), by using tiling.
1681Note that this fix is new since the most recent release
1682candidate for this release (2.10.903).
1683
1684    https://bugs.freedesktop.org/show_bug.cgi?id=25375
1685
1686Elimination of software fallback with alpha-only pixmap (a fix which
1687eliminated a full second from the Moblin boot time):
1688
1689    https://bugs.freedesktop.org/show_bug.cgi?id=26189
1690
1691Notable bug fixes in 2.11 compared to 2.10
1692------------------------------------------
1693Fix for undesired black borders on some images, (caued several
1694problems with firefox):
1695
1696    http://bugs.freedesktop.org/show_bug.cgi?id=17933
1697
1698And other fixes to provide more conformance with existing
1699software-rendering, (as measures by the cairo test suite), such as the
1700sampling location for nearest-neighbor sampling.
1701
1702Other changes
1703-------------
1704Eric Anholt and Daniel Vetter both removed large piles of old and
1705useless code now that the driver requires kernel modesetting
1706(KMS). This continues to reduce the mainteance burden of the driver,
1707making it easier to isolate and fix bugs.
1708
1709Snapshot 2.10.903 (2010-03-22)
1710==============================
1711Some notable bug fixes
1712----------------------
1713* Fix unpredictable results in page-flipping code due to access of an
1714  uninitialized variable.
1715
1716* Two fixes that address a number of long-outstanding failures in the
1717  cairo test suite:
1718
1719    Remove coord-adjust for nearest centre-sampling.
1720
1721    Fill alpha on xrgb images. Closes:
1722
1723	Bug 17933 - x8r8g8b8 doesn't sample alpha=0 outside surface bounds
1724	http://bugs.freedesktop.org/show_bug.cgi?id=17933
1725
1726* Fix a regression introduced in 2.10.902:
1727
1728    Fatal server error:
1729    i915_emit_composite_setup: ADVANCE_BATCH:
1730    under-used allocation 100/104
1731
1732	Bug 21723 [i915 bisected] rendering makes X crash
1733	http://bugs.freedesktop.org/show_bug.cgi?id=27123
1734
1735Other minor fixes (avoid trying to print an error message with a freed
1736string, set a variable to NULL after freeing), and documentation
1737updates are also included.
1738
1739Snapshot 2.10.902 (2010-03-15)
1740==============================
1741Some significant bug fixes
1742--------------------------
1743Fix video color problems due to YUV plane ordering issue. This fixed a
1744regression that showed up recently (only in these release candidates).
1745
1746Fix recent regression leading to corruption under compiz:
1747
1748  https://bugs.freedesktop.org/show_bug.cgi?id=26814
1749
1750Further DRI2 improvements/fixes from Mario Kleiner and Jesse Barnes,
1751(including support for SwapBuffers request with divisor != 0).
1752
1753Fix build against X server 1.6 branch.
1754
1755Avoid flooding logs after an error occurs.
1756
1757And dramatic code cleanups
1758--------------------------
1759Eric Anholt removed several cases of old, useless code, (conditions
1760that could not occur now that we require KMS, etc.).
1761
1762Daniel Vetter eliminated piles of code by modernizing the XvMC
1763memory-management and completely removing the old i830_memory
1764allocator (i830_allocate_memory, i830_free_memory), etc.
1765
1766Snapshot 2.10.901 (2010-02-26)
1767==============================
1768The most significant new feature of this release is support for new
1769DRI2 APIs, allowing page flipping to occur for swaps that are
1770full-screen and not rotated.
1771
1772Some significant bug fixes
1773--------------------------
1774* Avoid fallback due to failing to extract pixel value from an
1775  alpha-only solid, (this fix eliminated 1 second from Mobilin boot
1776  time).
1777
1778  https://bugs.freedesktop.org/show_bug.cgi?id=26189
1779
1780* Handle full-height blits without tearing in CopyRegion
1781
1782  https://bugs.freedesktop.org/show_bug.cgi?id=22475
1783
1784Several other fixes and cleanups are included as well, (including the
1785removal of stale utility code that is now maintained in the separate
1786intel-gpu-tools module).
1787
1788Release 2.10.0 (2010-01-04)
1789===========================
1790New requirements of 2.10 compared to 2.9
1791----------------------------------------
1792 * Linux kernel with kernel-modesetting (KMS)
1793
1794   All user-modesetting code has now been removed from the driver.
1795   We highly recommend kernel version 2.6.32 or later.
1796
1797 * Libdrm >= 2.4.16
1798
1799New features in 2.10 compared to 2.9
1800------------------------------------
1801 * New driver debugging options available in xorg.conf
1802   (DebugFlushBatches, DebugFlushCaches, DebugWait). See "man intel"
1803   for more details.
1804
1805 * Video overlay support with KMS. This currently requires Linux
1806   2.6.33, but a backport to 2.6.32 is available here:
1807
1808   http://gitorious.org/daniel-s-linux-stuff/linux-kernel/commits/intel-kms-overlay-for-2.6.32
1809
1810Notable fixes in 2.10 compared to 2.9
1811-----------------------------------
1812 * Fix crash when XRenderComposite is called with a -1 value for
1813   width/height, (crash at login when using compiz).
1814
1815   http://bugs.freedesktop.org/show_bug.cgi?id=24724
1816
1817 * Fix to support larger extended desktops, (such as a 2560x1600 plus a
1818   1920x1200 monitor).
1819
1820
1821 * Fix glyph corruption due to insufficient cache flushing:
1822
1823   https://bugs.freedesktop.org/show_bug.cgi?id=24315
1824
1825Several other correctness and performance fixes are also included.
1826
1827Note: The driver code of the 2.10.0 release is identical to that of
1828the 2.9.99.901 release candidate. The only changes in 2.10.0 compared
1829to 2.9.99.902 are a minor change to the configure script, and some
1830small updates to the documentation.
1831
1832Snapshot 2.9.99.902 (2009-12-10)
1833================================
1834This is the second release candidate in preparation for the upcoming
18352.10.0 release. We will appreciate any feedback we can get from
1836testing of this snapshot to improve the 2.10.0 release.
1837
1838Compared to the previous snapshot, this snapshot now correctly
1839verifies that libdrm 2.4.16 (or newer) is available. The previous
1840snapshot also required libdrm 2.4.16 for correct functionality, but
1841neglected to notify the user of this at configure, compile, or install
1842time and would instead simply fail to run.
1843
1844Snapshot 2.9.99.901 (2009-11-30)
1845================================
1846This is the first release candidate in preparation for the upcoming
18472.10.0 release. We will appreciate any feedback we can get from
1848testing of this snapshot to improve the 2.10.0 release.
1849
1850With this snapshot, the xf86-video-intel driver requires kernel
1851modesetting (KMS) as all of the user modesetting (UMS) support has
1852been removed from the driver.
1853
1854Release 2.9.0 (2009-09-28)
1855==========================
1856We are pleased to announce the 2.9.0 release of the xf86-video-intel
1857driver. (If you have been following along at home, this release is
1858identical to the 2.8.99.902 release candidate other than the version
1859number).
1860
1861New features in 2.9 compared to 2.8
1862-----------------------------------
1863* Support for the B43 chipset.
1864
1865Major fixes in 2.9.0 compared to 2.8.0
1866--------------------------------------
1867* Multiple fixes to make the driver stable for 8xx chipsets, (855GM,
1868  865G, etc.). The 2.8 driver series was extremely unstable with many
1869  of these chipsets.
1870
1871  https://bugs.freedesktop.org/show_bug.cgi?id=22904 (and many duplicates)
1872  https://bugs.freedesktop.org/show_bug.cgi?id=22947
1873
1874* Add support for BACKLIGHT property when using kernel modesetting
1875  (KMS). This allows backlight adjustment with programs such as
1876  "xbacklight -set <percentage>" or "xrandr --set BACKLIGHT <value>".
1877
1878  https://bugs.freedesktop.org/show_bug.cgi?id=20963
1879
1880* Fix so that "xrandr --scale" works when using KMS.
1881
1882  https://bugs.freedesktop.org/show_bug.cgi?id=22893
1883
1884* Fix segfaults of X server when logging out.
1885
1886  https://bugs.freedesktop.org/show_bug.cgi?id=20516
1887
1888* Avoid falling back to software for 1-bit alpha-only masks.
1889
1890  https://bugs.freedesktop.org/show_bug.cgi?id=23184
1891
1892* Fix pixel sampling position for 8xx chipsets, (fixes some cairo-test
1893  suite failures).
1894
1895Snapshot 2.8.99.901 (2009-09-09)
1896================================
1897This is the first release candidate in preparation for the upcoming
18982.9.0 release. We will appreciate any feedback we can get from testing
1899of this snapshot to improve the 2.9.0 release.
1900
1901See the commit log for changes in this release, and developers, please
1902provide me summaries for any major features in this release that you
1903think should appear in NEWS and in the 2.9.0 release notes.
1904
1905Note: We are still investigating failures with 865G chipsets, (these
1906failures existed in the 2.8 series as well). We are hoping to fix
1907these failures before the 2.9.0 release.
1908
1909Release 2.8.0 (2009-07-20)
1910==========================
1911We are pleased to present this major release of the xf86-video-intel
1912driver. This release is very similar to the 2.7.99.902 release
1913candidate but includes a couple of additional bug fixes, (for bugs
1914#22760 and @22483).
1915
1916Compared to the 2.7 releases this driver contains several new features
1917and many bug fixes. The driver now depends on X server 1.6 or later,
1918and eliminates several obsolete code paths, (XAA and EXA removed in
1919favor of UXA, DRI1 support eliminated). The driver certainly will work
1920best with an i915 module from a recent kernel (2.6.31) and with
1921kernel-modesetting (KMS, specified by loading i915 with the option
1922"modeset=1").
1923
1924New features in 2.8.0 compared to 2.7
1925-------------------------------------
1926* Remove XAA/EXA/DRI1 and NoAccel, and only support UXA/DRI2 now
1927  (requiring xserver 1.6).
1928
1929* VLD support added into XvMC for Mpeg2, which decreases CPU usage
1930  when playing video.
1931
1932Major fixes in 2.8.0 compared to 2.7
1933------------------------------------
1934* Fix broken front-buffer rendering
1935  https://bugs.freedesktop.org/show_bug.cgi?id=19174
1936
1937* Fix disabling of XvMC disabled
1938  https://bugs.freedesktop.org/show_bug.cgi?id=20790
1939
1940* Fix broken VT switch on some machines
1941  https://bugs.freedesktop.org/show_bug.cgi?id=19578
1942
1943* Improve performance of trapezoid rendering for UXA
1944  https://bugs.freedesktop.org/show_bug.cgi?id=21376
1945
1946* Vblank sync'd GL buffer swap (fixing video tearing under composite)
1947  https://bugs.freedesktop.org/show_bug.cgi?id=20664
1948
1949* Fix G41 DRM support
1950  https://bugs.freedesktop.org/show_bug.cgi?id=21095
1951
1952* Fix crash on some 8xx
1953  https://bugs.freedesktop.org/show_bug.cgi?id=18974
1954
1955Various fixes for GPU hangs, and more...
1956
1957Snapshot 2.7.99.902 (2009-07-13)
1958--------------------------------
1959This is the first release candidate in preparation for the upcoming
19602.8.0 release. Most major and critical bugs should be fixed, but some
1961minor bugs may still be present. We will appreciate any feedback we
1962can get from testing of this snapshot to improve the 2.8.0 release.
1963
1964Snapshot 2.7.99.901 (2009-06-10)
1965--------------------------------
1966This is the first release candidate in preparation for the upcoming
19672.8.0 release. Most critical bugs should be fixed, but some major
1968bugs may still be present. We will appreciate any feedback we can
1969get from testing of this snapshot to improve the 2.8.0 release.
1970
1971Some of the major fixes in this snapshot include:
1972
1973	* Several fixes for intermittent GPU hangs/crashes, (some of
1974	  which already appeared in the 2.7.1 release)
1975
1976	* Fix to avoid X crash when viewing large images in browser
1977
1978	* Fix X server failure when running old (non-GEM) kernel
1979
1980	* Fixes for SDVO LVDS mode detection
1981
1982	* Fix major performance regression of trapezoid rendering
1983          compared to XAA/EXA
1984
1985	* New support for tear-free video when using KMS
1986
1987	* New support for tear-free DRI2 CopyRegion
1988
1989Snapshot 2.7.99.1 (2009-04-28)
1990------------------------------
1991This is a development snapshot very early in the process toward
1992developing 2.8. There have been some big changes to the code, and
1993we're anxious to get feedback on these changes as early as possible.
1994
1995Here is a summary of the biggest changes:
1996
1997  * Driver now depends on X server 1.6 or later
1998
1999  * Eliminate XAA and EXA support (in favor of UXA)
2000
2001  * Eliminate DRI1 support
2002
2003  * Fixes for running without DRI at all
2004
2005These code removals represent a deletion of a substantial amount of
2006code, (and hopefully piles of bugs), as well as reduce the maintenance
2007effort going forward as the number of combinatorial configurations for
2008the driver are greatly reduced. This means that users are much more
2009likely to be running code that has actually been tested, and it will
2010be much easy for developers to replicate bugs that users experience.
2011
2012One of the things that would be most useful in testing this release is
2013to revisit any outstanding bugs that you have previously reported. If
2014the buggy behavior is gone, (or the bug is no longer relevant---such
2015as a bug that's specific to XAA only), please feel free to indicate so
2016in bugzilla or even just close the bug.
2017
2018If you confirm that the bug is still present, please indicate so in
2019the bug report. (I was going to ask that you select a 1.7.99 version,
2020but it looks like bugzilla only has versions for products not
2021compoenents, while we use a "xorg" product and a "driver/intel"
2022component.) We definitely want to make any such confirmed bugs a
2023priority, so it would be nice to have a consistent mechanism to search
2024for these bugs. Suggestions are welcome on the best approach.
2025
2026Thanks in advance for any testing or feedback on this snapshot.
2027
2028Release 2.7.0 (2009-04-15)
2029--------------------------
2030Compared to the 2.6 series, 2.7.0 has a large number of bug fixes, but
2031also a few significant features, such as:
2032
2033        SDVO-TV support, available on ADD2 card (bug#9992) and
2034        D945GCLF2 board (bug#17776).
2035
2036        Basic SDVO-LVDS support
2037
2038        XV video display without tearing
2039	[Though this isn't working for all users yet, see
2040	https://bugs.freedesktop.org/show_bug.cgi?id=21076 ]
2041
2042	Various fixes for UXA, DRI2, and Kernel modesetting.
2043
2044We encourage users to use kernel modesetting and UXA acceleration with
2045this release, which should give the best performance and
2046robustness. When KMS is available, UXA is the default acceleration
2047used by the driver, (EXA is the default otherwise).
2048
2049Known issue:
2050
2051	Some Linux kernel versions (such as 2.6.29) are known to have
2052	broken PAT code that causes recent versions of this driver to
2053	fail, (which can manifest as the X server simply not
2054	starting). This can be verified by adding the "nopat" option
2055	to the kernel command-line and seeing the failure go away. We
2056	hope that newer kernels in the 2.6.29.x as well as 2.6.30 and
2057	above will have working PAT code.
2058
2059Some of the most notable bugs fixed in 2.7.0 include:
2060
2061	[GM45 965GM] bad htotal causes panel startup failure
2062	https://bugs.freedesktop.org/show_bug.cgi?id=17292
2063
2064	[xrandr TV] need TV output property control
2065	https://bugs.freedesktop.org/show_bug.cgi?id=12763
2066
2067	[TV] "xrandr --set TV_FORMAT" gets BadMatch error
2068	https://bugs.freedesktop.org/show_bug.cgi?id=16566
2069
2070	[945 tiling] Low performance due to no A17 workaround
2071	https://bugs.freedesktop.org/show_bug.cgi?id=16835
2072
2073	[TV]Flicker when launching applications in the 2.4-branch
2074	https://bugs.freedesktop.org/show_bug.cgi?id=17405
2075
2076	[945GM FBC] FBC causes underruns & flicker
2077	https://bugs.freedesktop.org/show_bug.cgi?id=18651
2078
2079	[xv] Textured video suffers from tearing
2080	https://bugs.freedesktop.org/show_bug.cgi?id=19635
2081
2082	[G45] Random hangs with UXA
2083	https://bugs.freedesktop.org/show_bug.cgi?id=19734
2084
2085	[945GM] Any 3D app is slow in resolution higher than 800x600
2086	with UXA+DRI2, due to tiling
2087	https://bugs.freedesktop.org/show_bug.cgi?id=19738
2088
2089	[i915 UXA,EXA] rotation messes display with tiling on
2090	https://bugs.freedesktop.org/show_bug.cgi?id=20265
2091
2092	[G45] DRI2/UXA gives solid white instead of transparency
2093	https://bugs.freedesktop.org/show_bug.cgi?id=20321
2094
2095	LVDS output not detected
2096	https://bugs.freedesktop.org/show_bug.cgi?id=20517
2097
2098	xf86-video-intel-2.6.3: Xv crashes X server
2099	https://bugs.freedesktop.org/show_bug.cgi?id=20525
2100
2101	[G965 non-GEM] systray in KDE 4 completely broken
2102	https://bugs.freedesktop.org/show_bug.cgi?id=20527
2103
2104	[SDVO-TV]the desktop is teared in four sections on the screen
2105	https://bugs.freedesktop.org/show_bug.cgi?id=20550
2106
2107	Intel video driver 2.6.3 crashes with XVideo
2108	https://bugs.freedesktop.org/show_bug.cgi?id=20563
2109
2110	[855GM] Xv crash with non-KMS
2111	https://bugs.freedesktop.org/show_bug.cgi?id=20585
2112
2113	2.6.99.902 breaks native 1680x1050 mode on TMDS -- EDID miss
2114	https://bugs.freedesktop.org/show_bug.cgi?id=20594
2115
2116	[945GM TV] 2.6.99.902 sets a too high CONTRAST-value
2117	https://bugs.freedesktop.org/show_bug.cgi?id=20670
2118
2119	[915GM] fail to detect LVDS with new VBT code
2120	https://bugs.freedesktop.org/show_bug.cgi?id=20752
2121
2122	[regression i965]tiled output when start X
2123	https://bugs.freedesktop.org/show_bug.cgi?id=20803
2124
2125	2.6.99.902: LVDS wrongly detected as disconnected
2126	https://bugs.freedesktop.org/show_bug.cgi?id=20826
2127
2128	vt switching fails and crashes X
2129	https://bugs.freedesktop.org/show_bug.cgi?id=20863
2130
2131	[HDMI] The screen will flicker when some application runs on
2132	G45-64
2133	https://bugs.freedesktop.org/show_bug.cgi?id=20875
2134
2135	TexturedVideo is offsetted wrongly on dualhead
2136	https://bugs.freedesktop.org/show_bug.cgi?id=20980
2137
2138	[EXA] xvideo hang X
2139	https://bugs.freedesktop.org/show_bug.cgi?id=21027
2140
2141	[EXA] x11perf performance regression
2142	https://bugs.freedesktop.org/show_bug.cgi?id=21029
2143
2144	And many others...
2145