Mesa 11.1.3 Release Notes / April 17, 2016

Mesa 11.1.3 is a bug fix release which fixes bugs found since the 11.1.2 release.

Mesa 11.1.3 implements the OpenGL 4.1 API, but the version reported by glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) / glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used. Some drivers don't support all the features required in OpenGL 4.1. OpenGL 4.1 is only available if requested at context creation because compatibility contexts are not supported.

SHA256 checksums

9e86c72b6b2e8adb53c1c4a0002ab267b45094d753eb9404b1db34f81ce94ccf  mesa-11.1.3.tar.gz
51f6658a214d75e4d9f05207586d7ed56ebba75c6b10841176fb6675efa310ac  mesa-11.1.3.tar.xz

New features

None

Bug fixes

This list is likely incomplete.

• Bug 27512 - Illegal instruction _mesa_x86_64_transform_points4_general
• Bug 91526 - World of Warcraft (on Wine) has UI corruption with nouveau
• Bug 92193 - [SKL] ES2-CTS.gtf.GL2ExtensionTests.compressed_astc_texture.compressed_astc_texture fails
• Bug 93358 - [HSW] Unreal Elemental demo - assertion error in copy_image_with_blitter
• Bug 93418 - Geometry Shaders output wrong vertices on Sandy Bridge
• Bug 93524 - Clover doesn't build
• Bug 93667 - Crash in eglCreateImageKHR with huge texture size
• Bug 93813 - Incorrect viewport range when GL_CLIP_ORIGIN is GL_UPPER_LEFT
• Bug 94050 - test_vec4_register_coalesce regression
• Bug 94073 - Miscompilation of abs_vec3_vert_xvary_ref.vert in WebGL conformance
• Bug 94088 - [llvmpipe] SIGFPE pthread_barrier_destroy.c:40
• Bug 94193 - [llvmpipe] Line antialiasing looks different when GL_LINE_STIPPLE is enabled with pattern 0xffff
• Bug 94195 - [llvmpipe] Does not build with LLVM 3.7.x on Windows
• Bug 94388 - r600_blit.c:281: r600_decompress_depth_textures: Assertion `tex->is_depth && !tex->is_flushing_texture' failed.
• Bug 94412 - Trine 3 misrender
• Bug 94481 - softpipe - access violation in img_filter_2d_nearest
• Bug 94595 - [Mesa AMD&swrast] Texture views attached as framebuffers return their viewed tecture's color encoding and render incorrectly
• Bug 94954 - test_vec4_copy_propagation fails in `make check`

Changes

Anuj Phogat (1):

• i965: Fix assert conditions for src/dst x/y offsets

Ben Widawsky (2):

• i965: Make sure we blit a full compressed block
• i965/skl: Add two missing device IDs

Brian Paul (1):

• mesa: fix incorrect viewport position when GL_CLIP_ORIGIN = GL_LOWER_LEFT

Chris Forbes (1):

• i965/blorp: Fix hiz ops on MSAA surfaces

Christian König (1):

• radeon/uvd: disable MPEG1

Christian Schmidbauer (1):

• st/nine: specify WINAPI only for i386 and amd64

Daniel Czarnowski (3):

• egl_dri2: NULL check for xcb_dri2_get_buffers_reply()
• egl_dri2: set correct error code if swapbuffers fails
• egl: support EGL_LARGEST_PBUFFER in eglCreatePbufferSurface(...)

Dave Airlie (1):

• mesa/fbobject: propogate Layered when reusing attachments.

Derek Foreman (1):

• egl/wayland: Try to use wl_surface.damage_buffer for SwapBuffersWithDamage

Dongwon Kim (1):

• egl: move Null check to eglGetSyncAttribKHR to prevent Segfault

Emil Velikov (10):

• docs: add sha256 checksums for 11.1.2
• get-pick-list.sh: Require explicit "11.1" for nominating stable patches
• cherry-ignore: do not pick nv50/ir commit
• automake: add nine to make distcheck
• install-gallium-links: port changes from install-lib-links
• automake: add more missing options for make distcheck
• mesa; add get-extra-pick-list.sh script into bin/
• egl/x11: check the return value of xcb_dri2_get_buffers_reply()
• nvc/ir: remove duplicate variable declaration
• Update version to 11.1.3

Francisco Jerez (4):

• i965: Reupload push and pull constants when we get new shader image unit state.
• i965/fs: Add missing analysis invalidation in opt_sampler_eot().
• i965/fs: Add missing analysis invalidation in fixup_3src_null_dest().
• i965/vec4: Consider removal of no-op MOVs as progress during register coalesce.

Ilia Mirkin (21):

• nvc0/ir: fix converting between predicate and gpr
• nvc0: add some missing PUSH_SPACE's
• nvc0: avoid negatives in PUSH_SPACE argument
• glsl: make sure builtins are initialized before getting the shader
• glsl: return cloned signature, not the builtin one
• nv50/ir: fix quadop emission in the presence of predication
• st/mesa: fix up result_src.type when doing i2u/u2i conversions
• meta/copy_image: use precomputed dst_internal_format to avoid segfault
• st/mesa: force depth mode to GL_RED for sized depth/stencil formats
• glx: update to updated version of EXT_create_context_es2_profile
• nv50,nvc0: bump minimum texture buffer offset alignment
• nvc0: reset TFB bufctx when we no longer hold a reference to the buffers
• glsl: avoid stack smashing when there are too many attributes
• nvc0: fix blit triangle size to fully cover FB's > 8192x8192
• nv50: reset TFB bufctx when we no longer hold a reference to the buffers
• nv50/ir: force-enable derivatives on TXD ops
• st/mesa: only minify depth for 3d targets
• nv50/ir: fix indirect texturing for non-array textures on nvc0
• nvc0/ir: fix picking of coordinates from tex instruction for textureGrad
• nvc0: disable primitive restart and index bias during blits
• nv50/ir: we can't load local memory directly into an output

Jason Ekstrand (1):

• nir/lower_vec_to_movs: Better report channels handled by insert_mov

Kenneth Graunke (3):

• mesa: Make glGet queries initialize ctx->Debug when necessary.
• mesa: Allow Get*() of several forgotten IsEnabled() pnames.
• i965: Only magnify depth for 3D textures, not array textures.

Koop Mast (1):

• st/clover: Add libelf cflags to the build

Marc-André Lureau (1):

• virtio_gpu: Add virtio 1.0 PCI ID to driver map

Marek Olšák (3):

• radeonsi: fix Hyper-Z on Stoney
• gallium/radeon: don't use temporary buffers for persistent mappings
• radeonsi: fix Hyper-Z hangs on P2 configs

Matt Turner (3):

• i965/vec4: don't copy ATTR into 3src instructions with complex swizzles
• i965/fs: Don't CSE negated multiplies with saturation.
• i965/vec4: Update vec4 unit tests for commit 01dacc83ff.

Nanley Chery (2):

• mesa/image: Make _mesa_clip_readpixels() work with renderbuffers
• mesa/readpix: Clip ReadPixels() area to the ReadBuffer's

Nicolai Hähnle (2):

• r600g: clear compressed_depthtex/colortex_mask when binding buffer texture
• st/mesa: use the texture view's format for render-to-texture

Nishanth Peethambaran (2):

• st/omx: Remove trailing spaces
• st/omx/dec: Correct the timestamping

Oded Gabbay (8):

• gallium/radeon: Correctly translate colorswaps for big endian
• llvmpipe: use vpkswss when dst is signed
• gallium/radeon: return correct values for BE in r600_translate_colorswap
• gallium/radeon: remove separate BE path in r600_translate_colorswap
• gallium/r600: Don't let h/w do endian swap for colorformat
• gallium/radeon: disable evergreen_do_fast_color_clear for BE
• r600g: Do colorformat endian swap for PIPE_USAGE_STAGING
• radeonsi: Do colorformat endian swap for PIPE_USAGE_STAGING

Olivier Pena (1):

• scons: support for LLVM 3.7.

Patrick Baggett (1):

• mesa: Use SSE prefetch instructions rather than 3DNow instructions

Rob Herring (10):

• Android: remove dependence on .SECONDEXPANSION
• Android: glsl: fix dependence on YACC_HEADER_SUFFIX from build system
• Android: add -Wno-date-time flag for clang
• Android: remove headers from LOCAL_SRC_FILES
• Android: clean-up and fix DRI module path handling
• freedreno: drop unnecessary -Wno-packed-bitfield-compat
• gallium/radeon: Add space between string literal and identifier
• r600: Make enum alu_op_flags unsigned
• virtio_gpu: Add PCI ID to driver map
• Android: fix x86 gallium builds

Roland Scheidegger (2):

• softpipe: fix anisotropic filtering crash
• draw: fix line stippling

Samuel Pitoiset (1):

• nvc0: make sure to delete samplers used by compute shaders

Steinar H. Gunderson (1):

• mesa: Fix locking of GLsync objects.

Tamil velan (1):

• radeon/uvd: increase max height to 4096 for VI and newer

Thomas Hellstrom (2):

• winsys/svga: Fix an uninitialized return value
• winsys/svga: Increase the fence timeout

Vinson Lee (1):

• llvmpipe: Do not use barriers if not using threads.

xavier (1):

• r600/sb: Do not distribute neg in expr_handler::fold_assoc() when folding multiplications.