Permalink
Commits on Jul 9, 2016
  1. nvc0: initial support for GP100 GPUs

    Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
    committed Jul 9, 2016
Commits on Jul 8, 2016
  1. glx: Fix for commit 2c86668.

    Ian suggested these changes in his review and I made them, but I pushed
    the old version of the patch.
    mattst88 committed Jun 23, 2016
  2. docs: add news item and link release notes for 12.0.0/12.0.1

    Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
    Emil Velikov committed with evelikov Jul 8, 2016
  3. docs: add sha256 checksums for 12.0.1

    Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
    (cherry picked from commit edfc17a19a17e8dae41a703a9cb4b9eb3c255c34)
    Emil Velikov committed with evelikov Jul 8, 2016
  4. docs: add release notes for 12.0.1

    Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
    (cherry picked from commit 04277f058d00238937e664cf546c43b16cea7b2b)
    Emil Velikov committed with evelikov Jul 8, 2016
  5. docs: add sha256 checksums for 12.0.0

    Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
    (cherry picked from commit 3a146a789c5aa45e0de52bff764c2b371dd253aa)
    Emil Velikov committed with evelikov Jul 8, 2016
  6. docs: Update 12.0.0 release notes

    Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
    (cherry picked from commit 8b06176f310f65628ce136b90a99005278ba5e0d)
    Emil Velikov committed with evelikov Jul 7, 2016
  7. glx: Undo memory allocation checking damage.

    This partially reverts commit d41f539.
    
    That untested commit broke the tex-skipped-unit piglit test and the
    arbvparray Mesa demo when run with indirect GLX.
    
    state->array_state is used during initialization, so its assignment cannot be
    moved to the end of the function.
    
    The backtrace looked like:
    
    Program received signal SIGSEGV, Segmentation fault.
    0x00007ffff77c7a5c in __glXGetActiveTextureUnit (state=0x6270e0) at indirect_vertex_array.c:1952
    1952           return state->array_state->active_texture_unit;
    (gdb) bt
    0  0x00007ffff77c7a5c in __glXGetActiveTextureUnit (state=0x6270e0) at indirect_vertex_array.c:1952
    1  0x00007ffff77cbf62 in get_client_data (gc=0x626f50, cap=34018, data=0x7fffffffd7a0) at single2.c:159
    2  0x00007ffff77cce51 in __indirect_glGetIntegerv (val=34018, i=0x7fffffffd830) at single2.c:498
    3  0x00007ffff77c4340 in __glXInitVertexArrayState (gc=0x626f50) at indirect_vertex_array.c:193
    
    Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
    mattst88 committed Jun 23, 2016
  8. glx: Fix indirect multi-texture GL_DOUBLE coordinate arrays.

    There is no draw arrays protocol support for multi-texture coordinate
    arrays, so it is implemented by sending batches of immediate mode
    commands from emit_element_none in indirect_vertex_array.c.  This sends
    the target texture unit (which has been previously setup in the
    array_state header field), followed by the texture coordinates.  But for
    GL_DOUBLE coordinates the texture unit must be sent *after* the texture
    coordinates. This is documented in the glx protocol description, and can
    also be seen in the indirect.c immediate mode commands generated from
    gl_API.xml. Sending the target texture unit in the wrong place can crash
    the remote X server.
    
    To fix this required some more extensive changes to
    indirect_vertex_array.c and indirect_vertex_array_priv.h, in order to
    remove the texture unit value out of the array_state "header" field, and
    send it separately.
    
    Reviewed-by: Matt Turner <mattst88@gmail.com>
    Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=61907
    Colin McDonald committed with mattst88 Jun 23, 2016
  9. glx: Correct opcode typos in __indirect_glTexCoordPointer.

    At the same time, replace opcode numbers with names in
    __indirect_glVertexAttribPointer.
    
    Reviewed-by: Matt Turner <mattst88@gmail.com>
    Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=61907
    Colin McDonald committed with mattst88 Jun 23, 2016
  10. glx: Call __glXInitVertexArrayState() with a usable gc.

    For each indirect context the indirect vertex array state must be initialised
    by __glXInitVertexArrayState in indirect_vertex_array.c.  As noted in the
    routine header it requires that the glx context has been setup prior to the
    call, in order to test the server version and extensions.
    
    Currently __glXInitVertexArrayState is called from indirect_bind_context in
    indirect_glx.c, as follows:
    
       state = gc->client_state_private;
       if (state->array_state == NULL) {
          glGetString(GL_EXTENSIONS);
          glGetString(GL_VERSION);
          __glXInitVertexArrayState(gc);
       }
    
    But, the gc context is not yet usable at this stage, so the server queries
    fail, and __glXInitVertexArrayState is called without the server version and
    extension information it needs.  This breaks multi-texturing as
    glXInitVertexArrayState doesn't get GL_MAX_TEXTURE_UNITS.  It probably also
    breaks setup of other arrays: fog, secondary colour, vertex attributes.
    
    To fix this I have moved the call to __glXInitVertexArrayState to the end of
    MakeContextCurrent in glxcurrent.c, where the glx context is usable.
    
    Fixes a regression caused by commit 4fbdde8. Fixes ARB_vertex_program
    usage in the arbvparray Mesa demo when run with indirect GLX and also
    the tex-skipped-unit piglit test when run with indirect GLX.
    
    Reviewed-by: Matt Turner <mattst88@gmail.com>
    Acked-by: Ian Romanick <ian.d.romanick@intel.com>
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=61907
    Colin McDonald committed with mattst88 Jun 23, 2016
  11. radeon/uvd: simplify sending context buffer message

    Just send it whenever it is allocated.
    
    Signed-off-by: Christian König <christian.koenig@amd.com>
    Reviewed-by: Leo Liu <leo.liu@amd.com>
    Christian König committed Jul 7, 2016
  12. radeon/uvd: fix contex buffer destruction in the error path

    Destroying a not allocated buffer is harmless.
    
    Signed-off-by: Christian König <christian.koenig@amd.com>
    Reviewed-by: Leo Liu <leo.liu@amd.com>
    Christian König committed Jul 7, 2016
  13. radeon/uvd: move polaris fw check into radeon_video.c v2

    It's actually not very clever to claim to support H.264
    and then fail to create a decoder.
    
    v2: prefix FW macro with UVD_.
    
    Signed-off-by: Christian König <christian.koenig@amd.com>
    Reviewed-by: Leo Liu <leo.liu@amd.com>
    Christian König committed Jul 7, 2016
  14. radeon/video: fix coding style in radeon_video.c v2

    v2: fix other tabs as well.
    
    Signed-off-by: Christian König <christian.koenig@amd.com>
    Reviewed-by: Leo Liu <leo.liu@amd.com>
    Christian König committed Jul 7, 2016
  15. svga: simplify/fix 1D/2D array resource copies

    Fixes the one of the piglit arb_copy_image-targets tests for 1D arrays.
    Previously, we were applying the 1D array z/face adjustment twice.
    
    Also simplify the copy_region_vgpu10() function.  It never has to copy
    multiple array layers/slices.  The Mesa code for glCopyImageSubData does
    the loop over slices/faces.
    
    Reviewed-by: Charmaine Lee <charmainel@vmware.com>
    Brian Paul committed Jul 6, 2016
  16. mesa: print number of samples in renderbuffer_storage error msg

    Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
    Brian Paul committed Jul 6, 2016
  17. svga: remove unused variable

    Reviewed-by: Charmaine Lee <charmainel@vmware.com>
    Brian Paul committed Jul 6, 2016
  18. svga: add dumping for more device commands

    Signed-off-by: Brian Paul <brianp@vmware.com>
    Brian Paul committed Jul 6, 2016
  19. svga: silence a couple unused variable warnings

    Signed-off-by: Brian Paul <brianp@vmware.com>
    Brian Paul committed Jul 6, 2016
  20. svga: rebind using render target surfaces in hw draw state

    Currently when we rebind framebuffer resources at the beginning of
    the command buffer, we use the color buffer surfaces saved in the context
    hw clear state. But the surfaces could be different from the actual
    emitted render target surfaces if any of the color buffer surfaces
    is also used for shader resource, in that case, we create
    a backed surface for the collided render target surface. So to rebind
    the framebuffer resources correctly, use the render target surfaces saved
    in the context hw draw state.
    
    Tested with Heaven, Lightsmark2008, MTT piglit, glretrace, conform.
    
    Reviewed-by: Brian Paul <brianp@vmware.com>
    Charmaine Lee committed with Brian Paul Jun 30, 2016
  21. svga: invalidate gb surface before it is reused

    With this patch, a guest-backed surface will be invalidated
    using the SVGA_3D_CMD_INVALIDATE_GB_SURFACE command before
    the surface is reused. This fixes the updating dirty image error
    from the device when a surface is reused.
    
    v2: Instead of invalidating the surface when it is reused,
        send the invalidate command before the surface is put into
        the recycle pool.
    
    v3: (1) surface invalidate is a noop operation in Linux winsys, since
            surface invalidation is not needed for DMA path.
        (2) Instead of invalidating the surface content in
            svga_screen_surface_destroy() when a surface is to be destroyed,
            it is done in svga_screen_cache_flush() when the surface is
            no longer referenced in a command buffer and is ready to
            be moved to the unused list. At this point, the surface will
            be moved to the invalidate list. When the surface invalidation
            is submitted, the surface will be moved to the unused list.
    
    Tested with piglit, glretrace.
    
    Reviewed-by: Brian Paul <brianp@vmware.com>
    Reviewed-by: Sinclair Yeh <syeh@vmware.com>
    Charmaine Lee committed with Brian Paul Mar 8, 2016
  22. svga: fix use of provoking vertex control

    If the SVGA3D_DEVCAP_DX_PROVOKING_VERTEX query returns false, never
    define rasterizer state objects with provokingVertexLast set.  Despite
    what the device reports, it may interpret the provokingVertexLast flag
    anyway.  This fixes an issue when using capability clamping.
    
    Tested with piglit provoking-vertex and glsl-fs-flat-color tests.
    
    VMware bug 1550143.
    
    Reviewed-by: <charmainel@vmware.com>
    Brian Paul committed May 17, 2016
  23. vl: add half pixel to v_tex before adding offsets

    Since pixel center lies at 0.5, add half_pixel to vtex
    before adding offsets to it.
    
    Signed-off-by: Nayan Deshmukh <nayan26deshmukh@gmail.com>
    Reviewed-by: Christian König <christian.koenig@amd.com>
    ndesh26 committed with Christian König Jul 5, 2016
  24. nvc0/ir: remove unused resource info loading helpers

    Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
    Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
    hakzsam committed Jul 5, 2016
  25. nvc0/ir: refactor the surfaces info loading logic

    Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
    Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
    hakzsam committed Jul 5, 2016
  26. nvc0/ir: move the shift left op inside loadTexHandle()

    Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
    Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
    hakzsam committed Jul 5, 2016
  27. radeonsi: disable multi-threading when shader dumps are enabled

    Otherwise, shader dumps can become interleaved and unusable.
    
    Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
    Reviewed-by: Marek Olšák <marek.olsak@amd.com>
    Nicolai Hähnle committed Jul 6, 2016
  28. radeonsi: use multi-threaded compilation in debug contexts

    We only have to stay single-threaded when debug output must be synchronous.
    This yields better parallelism in shader-db runs for me.
    
    Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
    Reviewed-by: Marek Olšák <marek.olsak@amd.com>
    Nicolai Hähnle committed Jul 7, 2016
  29. st/mesa: set debug callback async flag

    Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
    Reviewed-by: Marek Olšák <marek.olsak@amd.com>
    Nicolai Hähnle committed Jul 7, 2016
  30. gallium: add async flag to pipe_debug_callback

    v2: fix typo db -> cb
    
    Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
    Reviewed-by: Marek Olšák <marek.olsak@amd.com>
    Nicolai Hähnle committed Jul 7, 2016
  31. radeonsi: catch a potential state tracker error with non-MSAA FBs

    At least st/mesa ensures this, so I'd rather not handle deviations in radeonsi.
    
    Reviewed-by: Marek Olšák <marek.olsak@amd.com>
    Nicolai Hähnle committed Jul 6, 2016
  32. radeonsi: explicitly choose center locations for 1xAA on Polaris

    Unlike SC, the small primitive filter does not automatically use center
    locations in 1xAA mode, so this is needed to avoid artifacts caused by
    the small primitive filter discarding triangles that it shouldn't.
    
    As a side effect of how the effective number of samples is now calculated,
    this patch also avoids submitting the sample locations for line/poly smoothing
    when they're not really needed.
    
    Cc: 12.0 <mesa-stable@lists.freedesktop.org>
    Reviewed-by: Marek Olšák <marek.olsak@amd.com>
    Nicolai Hähnle committed Jul 6, 2016
  33. r600g: call cayman_emit_msaa_sample_locs only when needed

    In the case of nr_samples <= 1, that function is (currently) a no-op anyway.
    
    Reviewed-by: Marek Olšák <marek.olsak@amd.com>
    Nicolai Hähnle committed Jul 6, 2016
  34. mesa: Mark R*32F formats as filterable when an extension is present.

    GL_OES_texture_float_linear marks R32F, RG32F, RGB32F, and RGBA32F
    as texture filterable.
    
    Fixes glGenerateMipmap GL errors when visiting a WebGL demo in Chromium:
    http://www.iamnop.com/particles
    
    Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
    Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
    kaydenl committed Jul 7, 2016