[ANNOUNCE] xf86-video-intel 2.99.917

Chris Wilson chris at chris-wilson.co.uk
Sun Dec 21 06:47:58 PST 2014

Snapshot 2.99.917 (2014-12-21)
3 months drifted by whilst I looked elsewhere for bugs.. The highlight of
bugs fixed here are a couple of workarounds required for Broadwell and
making sure that the rasterisation code is symmetric under inversions.
However, as a couple of crashers slipped through into 2.99.916 (though
not actual regressions in 2.99.916 per se) and 3 months have passed, we
should make one more snapshot before an imminent release.

 * Beware of recomputing the clear hint in the middle of MI recursion

 * Fix crash from rendering an empty Glyph string under PRIME as the
   secondary driver

 * Restore application of default monitor options to the first output,
   a regression in 2.99.915 and the MST support

 * Finally fix ZaphodHeads blocking on waiting for flip completion events.
   Before kernel 3.19, O_NONBLOCK support is broken and so we must avoid
   reading if we are not expecting an event.

 * Backwards compatibilty fix for fake triple buffering with PRIME and

 * Fix a rendering issue with output rotation and software fallbacks.

 * Enable MST discovery even without udev support

 * Fix TearFree operation after the GPU is wedged

 * Fix projective sampling on gen6+.

 * Fix rendering and corruption with gen8.

 * Fix crash when using Xinerama.

 * Fix rendering of right-to-left or bottom-to-top PolySegments in UXA

Complete list of changes since 2.99.916

Chris Wilson (165):
      sna: Avoid u16 underflow when computing reserved batch space
      sna: Initialise and check for batch space
      sna: Add some DBG spam for BLT boxes
      sna/glyphs: Prevent NULL vfunc deref with glyphs
      sna: Fix use of GETBLOB ioctl for grabbing the MST path
      sna: Relax early failure to set KMS modes
      sna: Emit assertions with FatalError
      sna: Do apply damage twice for miSpans.PolyFillRect
      sna: Prune damage that covers the entire target Pixmap
      backlight: Expose interface to switch backlight on/off entirely
      sna: Request the backlight to be disabled along with DPMS off
      sna: Do not mark the pixmap as cleared in the middle of a miSpans decomposition
      sna/gen8: Re-enable userptr
      sna: Use default monitor options on the first output
      sna: DBG compile fix
      sna/blt: Fix computation of remainaing boxes on gen8+
      sna: Initialise remaining batch space
      sna: Avoid stalls when promoting to the GPU before an operation
      sna: Mark the CRTC as disabled when hiding the plane
      intel-virtual-output: Initialise width/height for a disabled output
      sna: Fix include guard
      sna: Remvoe unused code
      configure: Require the presence of xorg/glamor.h for --enable-glamor
      uxa: Stub out intel_sync_init|fini when not compiled in
      sna: Update the list of outputs after a failed modeset
      sna: Last ditch attempt to make extra large batches fit
      sna/io: Initialise return code to catch early segfaults
      test: Add DrawSegments
      uxa: Fix conversion of Segments into Rectangles
      sna: Constify argument to box_from_seg()
      uxa/glamor: Silence compiler warnings
      sna/gen8: Update DBG output of surface offset
      sna/trapezoids: Thread tristrip rasterisation
      sna/dri2: Decouple the Client event link on Drawable free
      sna: Skip present unflip if the output is rotated
      sna/present: Make DBG statements uniform before present_event_notify()
      sna: Initialise the per-CRTC shadow bo on first use
      sna/trapezoids: Flesh out alternate rasterisers for tristrips
      sna: Reuse the same buffer when panning large CRTCs
      sna: Tweak partial Picture extraction to prefer limiting to max 3D coordinates
      sna: Relax constant numCrtcs assertion when ZaphodHeads is active
      sna/gen4+: Assert that the offset coordinate is within range if negative
      sna: Use shadow CRTC bo if we fail to obtain an framebuffer for the Pixmap
      sna: Check for scanout pitch restrictions on linear GPU bo
      sna: Check scanout Pixmaps are the correct pitch and convert if necessary
      sna: Pass scanout hints along to move-to-gpu
      Add Skylake PCI IDs
      sna: Allow move-to-gpu to fail during ScreenCreateResources
      man: Document the extended values supported by Option "DRI"
      man: Mention Option "AccelMethod" "none"
      sna: Use move-to-gpu for explicit handling of damage during fbcon copy
      sna; Markup that we are ignoring errors from early ScreenCreateResources
      sna: Clear padding in rotation properties for valgrind
      sna: Fix SHOW_BATCH_AFTER debugging aide
      sna/trapezoids: Fix loss of precision through projection onto sample grid
      sna/trapezoids: Consistently use NO_GPU_THREADS to disable threading for DBG
      sna/trapezoids: Fix comparison on right edge of trapezoids
      sna/dri2: Disable tracking stale back buffers without reuse-notifications
      sna: Update vblank interval after enabling outputs
      sna: Only use copy redisplay fast path if we have a source GPU bo
      sna: Double check for no residual CPU damage along redisplay fast path
      sna/trapezoids: Improve sample locations for imprecise rasterisation
      sna/trapezoids: Remove redundant coverage scaling
      sna: Avoid the intermediate allocation for PRIME pixmaps
      sna: Add pixmap id to DBG hints inside rende migration
      sna/trapezoids: Remove stale assert
      sna: Wrap rrGetInfo so that we can probe when udev is not active
      sna: Retrieve private pointer from vblank cookie
      sna: Fix the TearFree flip handler for the change in argument order
      intel: Store pointer to struct intel_device
      compat: Hide changes in miHandleExposures() interface
      sna: Log if we disable the backlight after failure to switch it on
      sna/trapezoids: Treat mono-edges within a single column as vertical
      sna/trapezoids: Sacrifice precision to avoid 64bit overflow
      backlight: Open bl_power using WRONLY for writing.
      sna: Drain all udev events before checking output status
      sna: Prevent recursion during last-gasp disabling of outputs
      sna: Replace GPU render operations with CPU callbacks after wedged
      sna: Fixup a failed per-CRTC TearFree flip
      sna: Allow TearFree updates to continue even when the GPU is wedged
      sna/dri2: Exchange pitches as well
      sna/gen6: Apply gen7 flushing
      sna: Prefer the GPU for wide lines and arcs
      sna: Avoid pwriting large batches
      sna/trapezoids: Prefer GPU for AddTraps()
      sna/dri2: Fix interoperation with PRIME and older Xorg
      sna: Fix double application of batch shrinking to relocation targets
      sna/gen8: Annotate more 64bit pointer locations
      sna/gen8: Clamp URB allocations for GT3
      sna/gen8: Override the URB output from the VS
      sna/gen8: Load the sampler result directly into the fb write message
      sna: Fix use of uninitialized value when handling migration for PRIME
      sna: Adapt to changes in internal render API in Xorg-1.17
      sna: Update the reloc->delta for the shrunk batch
      sna/trapezoids: Difference between two 32-bit quantities is 33-bits in size
      sna/trapezoids: Prevent overflow of edge gradient in mono rasteriser
      sna: Fix build for GC API changes in Xorg-
      sna: Reuse the same fence until retired
      sna: Retire all requests when a bo is found to be not busy
      sna: Remove extraneous tab whitespace before parameter name
      sna/brw: Fix EU tests to select the right gen
      sna/gen6+: Fix projective sample coordinates
      sna: Modicum of extra DBG for transformed redisplay
      sna: Fix forced fallbacks from init
      sna: Treat external bo carefully when marking as idle
      backlight: bl_power uses 0 for on and 4 for off
      sna: Use v2 of get_tiling ioctl to check for bit17 swizzling
      sna: Reuse the queried aperture size for videoRam determination
      sna: Correct units for videoRam
      sna/gen8: Clear instancing enabled bit between batches
      sna: Wrap prefaulting with sigtrap()
      sna: Experimental support for write-combining mmaps
      sna: Fix debug mmapping of active + released aux buffers
      Remove defunct glamor support
      sna/xv: Give the illusion of hardware progress
      sna: Check picture format on destination before beginning a GPU operation
      sna: Check after filtering points before submitting boxes to be drawn
      sna: Only reshow unhidden cursors after modesetting
      sna/gen7: Be paranoid and force a pipecontrol stall at least every 4 flushes
      sna: Reorder PARAMs based on upstream inclusion
      sna: Unroll pwrite using our cached mmappings
      Disable DRI3 by default
      sna/dri2: Remove unused function
      sna: Add more checks and asserts for BLT capable bo
      sna: Tweak alignment constraints on gen8 to allow BLT
      sna: DBG compile fix
      sna/dri2: Add more DBG to explain invalid frames
      sna/dri2: Improve precision of completion event for no-op swaps
      sna/transform: Correctly check for imprecise fractional translations
      sna: Reject attempts to set an invalid mode
      sna: Report when we disable TearFree due to an error in operation
      sna: Clean up DBG compile warnings
      sna: Do not post shadow updates in the middle of output reconfiguration
      sna: Add a few more asserts to sna_crtc_flip()
      sna: Reject modesets with 0 outputs
      sna: Cancel pending mode change notification
      sna: Generate a backtrace on assertion failures
      sna: Explicitly disable unused pipes after an output reconfiguration
      sna: Clamp downsampling tile size
      sna: Validate that the reg/insn size is well defined
      sna: Update the fence marker after submitting a new batch
      sna: Clear the read flags for the shared drm fd after use
      sna: Inline check for pending events before reading from drm fd
      sna/gen6: Poke PSMI control around WAIT_FOR_EVENT to prevent idling
      sna/dri3: Mesa relies upon implicit fences for X/Compositor synchronisation
      uxa/dri: Fix triple buffer swapping
      uxa/dri: Remove defunct intel->back_pixmap
      sna: Catch NULL RandR screen private
      sna: Add another Xinerama vs RandR guard
      sna/trapezoids: Check the pixmap is wrapped before deferencing it
      sna: Fix TearFree multi-monitor operation whilst wedged
      sna: Tune handling of kernel error detection during intial setup
      sna: Make error debug tracing available at normal debug levels
      intel-virtual-output: Add version number to verbose output
      intel-virtual-output: Do not attempt to tidy up the local display if it fails
      sna/dri2: Decouple Window cache on Pixmap changes
      sna/dri2: Prevent NULL pointer dereference of Window private
      sna/dri2: Look for preferred CRTC for visible portion of the window
      sna/dri2: Report changes in pitch when swapping bo
      sna/dri2: Preserve the shadow CRTCs when copying to other areas
      sna: Skip allocation of new TearFree buffer if not actually flipping
      sna/dri2: Check for implicit flushing after updating a Pixmap buffer
      sna: Fix Drawable offsets prior to performing shadow discard
      2.99.917 snapshot

Egbert Eich (1):
      sna: Validate framebuffer tiling before creation

Keith Packard (17):
      Do not clear pending kernel events on mode switch
      Correct BO allocation alignment
      glamor: Add DRI3 support to glamor code paths
      Fix present debug output
      Remove redundant calls to drm_intel_bo_disable_reuse
      Declare variable holding option value as 'const char *'
      Don't use GetScratchPixmapHeader for shadow pixmaps
      Eliminate pixmap private 'stride' field, which duplicates devKind
      Do more checks for proposed flip pixmaps
      Move intel_alloc_framebuffer to intel_memory.c
      Rename uxa-specific functions and structs
      Add intel_flush to abstract flushing pending acceleration operations
      Move UXA hardware initialization into intel_uxa_init
      Move some UXA-specific code under #if USE_UXA.
      intel_present.c does not need to include uxa.h
      Use intel_uxa.h in all uxa-specific files
      glamor: Do version check for fd <-> pixmap functions

Mika Kuoppala (1):
      sna: gen8 BLT broken when address has bit 4 set

Rodrigo Vivi (3):
      sna: Fix URB offset shift on 3DSTATE_SBE
      sna: a bit of cleanup on gen8_render.h
      sna: Use VMask in 3DSTATE_PS

Sedat Dilek (1):
      sna: wc-mmap: Fix warning in configure

tobias.jakobi1 at uni-bielefeld.de (1):
      uxa/video: only call intel_xvmc_adaptor_init when xvmc is enabled

git tag: 2.99.917

MD5:  fa196a66e52c0c624fe5d350af7a5e7b  xf86-video-intel-2.99.917.tar.bz2
SHA1: 9af9ded7a29026c211e5eb50a547e3e33976301d  xf86-video-intel-2.99.917.tar.bz2
SHA256: 00b781eea055582820a123c47b62411bdf6aabf4f03dc0568faec55faf9667c9  xf86-video-intel-2.99.917.tar.bz2

MD5:  b35d49d8523baed52e4feb4836b1733b  xf86-video-intel-2.99.917.tar.gz
SHA1: 7a2a6039e0a7c6b2a4b8e685fb1288a2bbd69d06  xf86-video-intel-2.99.917.tar.gz
SHA256: 1888f2733f8dccf4ab24bc56ba62eeac41c231b98365027438702d6c55216576  xf86-video-intel-2.99.917.tar.gz

Chris Wilson, Intel Open Source Technology Centre
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: <http://lists.x.org/archives/xorg-announce/attachments/20141221/7e7d624e/attachment.sig>

More information about the xorg-announce mailing list