summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* anv/entrypoints: Stop using the C preprocessorJason Ekstrand2016-06-071-12/+2
| | | | | | | | Now that we emit guards for everything, we can just generate the files and trust build flags to keep us safe. This should also fix the tarball problems. Reviewed-by: Emil Velikov <[email protected]>
* anv/entrypoints: Emit #if guards for all platformsJason Ekstrand2016-06-071-10/+22
| | | | Reviewed-by: Emil Velikov <[email protected]>
* platform_android: prevent deadlock in droid_swap_buffersHaixia Shi2016-06-071-5/+12
| | | | | | | | | | | | | To avoid blocking other EGL calls, release the display mutex before we enqueue buffer to android frameworks and re-acquire the mutex upon return. v2: moved lock/unlock inside droid_window_enqueue_buffer(). TEST=verify pinch zoom in Photos app no longer causes hangs Signed-off-by: Haixia Shi <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* mesa: automake: distclean git_sha1.h when building OOTEmil Velikov2016-06-071-0/+6
| | | | | | | | | | | | | | | In the case of out-of-tree (OOT) builds, in particular when building from tarball, we'll end up with the file in both srcdir and builddir. We want the former to remain intact (since we need it on rebuild) while the latter should be removed otherwise `make distclean' gets angry at us. Ideally there'll be a solution that feels a bit less of a hack. Until then this does the job exactly as expected. Cc: <[email protected]> Signed-off-by: Emil Velikov <[email protected]>
* mesa: automake: ensure that git_sha1.h.tmp has the right attributesEmil Velikov2016-06-071-0/+1
| | | | | | | | | | ... when copied from git_sha1.h. As the latter file can we lacking the write attribute, one should set it explicitly. Otherwise we'll get a warning/failure at cleanup stage. Cc: <[email protected]> Signed-off-by: Emil Velikov <[email protected]>
* mesa: automake: add directory prefix for git_sha1.hEmil Velikov2016-06-071-1/+1
| | | | | | | | | | Otherwise the build will assume that we've talking about builddir, which is not the case in the else statement. Here the file is already generated and is part of the tarball. Cc: <[email protected]> Signed-off-by: Emil Velikov <[email protected]>
* egl: android: don't add the image loader extension for !render_nodeEmil Velikov2016-06-071-8/+7
| | | | | | | | | | | | | | | | | | | | With earlier commit we introduced support for render_node devices, which was couples with the use of the image loader extension. As the work was inspired by egl/wayland we (erroneously) added the extension for the !render_node path as well. That works for wayland, as the implementations of the DRI2 and IMAGE loader extensions converge behind the scenes. As that is not yet the case for Android we shouldn't expose the extension. Fixes: 34ddef39cef ("egl: android: add dma-buf fd support") Cc: <[email protected]> Reported-by: Mauro Rossi <[email protected]> Tested-by: Mauro Rossi <[email protected]> Acked-by: Rob Herring <[email protected]> Signed-off-by: Emil Velikov <[email protected]>
* gallium/radeon: add support for sharing textures with DCC between processesMarek Olšák2016-06-073-4/+51
| | | | | | v2: use a function for calculating WORD1 of bo metadata Reviewed-by: Nicolai Hähnle <[email protected]>
* gallium/radeon: don't discard DCC if an external user can write to itMarek Olšák2016-06-073-12/+31
| | | | | | | | | We don't import textures with DCC now, but soon we will. v2: if we can't disable DCC for image writes, at least decompress DCC at bind time Reviewed-by: Nicolai Hähnle <[email protected]>
* i915: fix typo CAP.Dave Airlie2016-06-071-1/+1
| | | | Signed-off-by: Dave Airlie <[email protected]>
* glsl: initialise pointer to NULLJakob Sinclair2016-06-071-0/+1
| | | | | | | | | | | Could cause issues if you tried to read from an uninitialised pointer. This just initalises the pointer to null to avoid that being a problem. Discovered by Coverity. CID: 1343616 Signed-off-by: Jakob Sinclair <[email protected]> Reviewed-by: Iago Toral Quiroga <[email protected]>
* i965/gen8: fix cull distance emission for tessellation shaders.Dave Airlie2016-06-071-3/+5
| | | | | | | | | | This fixes some cases of: GL45-CTS.cull_distance.functional on Skylake. Reviewed-by: Chris Forbes <[email protected]> Cc: "12.0" <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
* nvc0: add support for VOTE tgsi opcodesIlia Mirkin2016-06-066-25/+78
| | | | | Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Samuel Pitoiset <[email protected]>
* st/mesa: expose GL_ARB_shader_group_vote when supported by backendIlia Mirkin2016-06-062-4/+11
| | | | | Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Dave Airlie <[email protected]>
* gallium: add PIPE_CAP_TGSI_VOTE for when the VOTE ops are allowedIlia Mirkin2016-06-0617-0/+17
| | | | | Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Dave Airlie <[email protected]>
* gallium: add VOTE_* opcodes to implement GL_ARB_shader_group_voteIlia Mirkin2016-06-063-1/+26
| | | | | Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Dave Airlie <[email protected]>
* mesa: hook up core bits of GL_ARB_shader_group_voteIlia Mirkin2016-06-0612-1/+67
| | | | | | Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Ian Romanick <[email protected]> Reviewed-by: Dave Airlie <[email protected]>
* glsl: Make opt_copy_propagation_elements actually propagate into loops.Kenneth Graunke2016-06-061-2/+17
| | | | | | | | | | | | | | | | | | | | | | We've had a FINISHME here since Eric originally wrote the code in 2011. This patch implements his suggested approach, which makes us actually able to copy propagate into the loops, at the unfortunate cost of making this pass even more expensive. The shader-db statistics are basically a wash: No change in instruction counts. total cycles in shared programs: 78685980 -> 78680730 (-0.01%) cycles in affected programs: 2102646 -> 2097396 (-0.25%) helped: 48 HURT: 83 I figured if we're going to do this for one copy propagation pass, we may as well do it in both. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* glsl: Make opt_copy_propagation actually propagate into loops.Kenneth Graunke2016-06-061-6/+24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We've had a FINISHME here since Eric originally wrote the code in 2010. This patch implements his suggested approach, which makes us actually able to copy propagate into the loops, at the unfortunate cost of making this pass even more expensive. The shader-db statistics are not terribly impressive: total instructions in shared programs: 9008589 -> 9008613 (0.00%) instructions in affected programs: 4293 -> 4317 (0.56%) helped: 0 HURT: 10 total cycles in shared programs: 78550978 -> 78575760 (0.03%) cycles in affected programs: 655426 -> 680208 (3.78%) helped: 75 HURT: 88 GAINED: 2 Most of the "regressions" appear to be us successfully copy propagating uniforms, which i965 is loading as pull constants instead of push, so we occasionally have two pulls instead of one. That doesn't seem like this pass's job - it's propagating correctly, and we should be smarter about pull loads in the backend. This patch is also useful for a couple of reasons: 1. It can clean up copies created by varying packing (previously, we couldn't if the uses were inside a loop). This fixes a bug when interpolateAt*() is used on a packed varying inside a loop: glsl_to_nir struggles to see through the extra copy and mistakenly believed the variable was not an input. 2. It will help propagate uniform array access created by lower_const_array_to_uniforms(). Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* nv50/ir: use round toward 0 when converting doubles to integersSamuel Pitoiset2016-06-061-1/+3
| | | | | | | | | | | | Like floats, we should use the round toward 0 mode instead of the nearest one (which is the default) for doubles to integers. This fixes all arb_gpu_shader_fp64 piglits which convert doubles to integers (16 tests). Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Ilia Mirkin <[email protected]> Cc: "11.2 12.0" <[email protected]>
* gallium/radeon: don't re-set BO metadata after CMASK deallocationMarek Olšák2016-06-061-1/+0
| | | | | | CMASK has no effect on metadata, because it's not sharable. Reviewed-by: Nicolai Hähnle <[email protected]>
* st/mesa: change SQRT lowering to fix the game RisenMarek Olšák2016-06-061-7/+9
| | | | | | | Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94627 (against nouveau) Acked-by: Ilia Mirkin <[email protected]>
* radeonsi: add a performance tweak for 4 SE partsMarek Olšák2016-06-061-0/+11
| | | | | | Ported from Vulkan. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: simplify PRIMGROUP_SIZE computation for tessellationMarek Olšák2016-06-061-9/+1
| | | | | | | | Ported from Vulkan. v2: keep the comment Reviewed-by: Nicolai Hähnle <[email protected]>
* r600g: use hw MSAA resolve for non-trivial resolvesMarek Olšák2016-06-061-9/+53
| | | | | | This improves MSAA resolve performance. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: use hw MSAA resolve for non-trivial resolvesMarek Olšák2016-06-061-10/+54
| | | | | | This improves MSAA resolve performance. Reviewed-by: Nicolai Hähnle <[email protected]>
* mesa/program_resource: return -1 for index if no location.Dave Airlie2016-06-071-3/+16
| | | | | | | | | | | | | | | | | The GL4.5 spec quote seems clear on this: "The value -1 will be returned by either command if an error occurs, if name does not identify an active variable on programInterface, or if name identifies an active variable that does not have a valid location assigned, as described above." This fixes: GL45-CTS.program_interface_query.output-built-in [airlied: use _mesa_program_resource_location_index as suggested by Eduardo] Reviewed-by: Eduardo Lima Mitev <[email protected]> Cc: "12.0" <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
* radeonsi: set descriptor dirty mask on shader buffer unbindNicolai Hähnle2016-06-061-0/+1
| | | | | | | | Found randomly while skimming the code. This might have caused VM faults in robustness tests. Cc: 12.0 <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* st/mesa: fix resource leak in try_pbo_readpixelsNicolai Hähnle2016-06-061-1/+1
| | | | | | | Found by inspection after seeing https://bugs.freedesktop.org/show_bug.cgi?id=96343 Reviewed-by: Marek Olšák <[email protected]>
* tgsi: fix mixed data type comparison in tgsi_point_sprite.cCharmaine Lee2016-06-061-3/+3
| | | | | | | | | | | | | | | | Cast the unsigned semantic index to integer datatype before comparing to max_generic, otherwise, max_generic which is initialized to -1 will be converted to unsigned int before the comparison, causing a wrong semantic index to be assigned to a shader output. Fixes the assert running TurboCAD_gl.trace. (VMware bug 1667265) Also tested with glretrace, mesa demos pointblast, spriteblast and pointcoord. v2: use the original max_generic variable but add the (int) cast to the semantic index, as suggested by Brian. Reviewed-by: Brian Paul <[email protected]>
* svga: print shader linkage info when tgsi debug bit is onCharmaine Lee2016-06-061-2/+5
| | | | | | | | When TGSI debug flag is enabled, print the shader linkage info as well. Tested with mesa demos with SVGA_DEBUG=tgsi Reviewed-by: Brian Paul <[email protected]>
* st/mesa: check shader image format support before using PBO downloadIlia Mirkin2016-06-061-0/+6
| | | | | | | | | | | ARB_shader_image_load_store only requires a very fixed list of formats to be supported, while textures may be in all kinds of formats, like BGRA which are presently not supported on at least Kepler. Signed-off-by: Ilia Mirkin <[email protected]> Tested-by: Samuel Pitoiset <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* tgsi: use truncf in micro_truncLars Hamre2016-06-061-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | | Switches to using truncf in micro_trunc. Fixes the following piglit tests (for softpipe): /spec/glsl-1.30/execution/built-in-functions/... fs-trunc-float fs-trunc-vec2 fs-trunc-vec3 fs-trunc-vec4 vs-trunc-float vs-trunc-vec2 vs-trunc-vec3 vs-trunc-vec4 /spec/glsl-1.50/execution/built-in-functions/... gs-trunc-float gs-trunc-vec2 gs-trunc-vec3 gs-trunc-vec4 Signed-off-by: Lars Hamre <[email protected]> Reviewed-by: Roland Scheidegger <[email protected]>
* i965/gs/scalar: Fix load input for doublesSamuel Iglesias Gonsálvez2016-06-061-18/+54
| | | | | | Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]> Reviewed-by: Timothy Arceri <[email protected]> Cc: "12.0" <[email protected]>
* i965/fs: fix offset when loading double vector input varyingsSamuel Iglesias Gonsálvez2016-06-061-1/+21
| | | | | | | | | | | | | | | | | When we are not packing a double input varying, we might need to read its data in a non-aligned to 64-bit offset, so we read the wrong data. This is happening when using explicit locations in varyings because Mesa disables packing varying for that case. const_index is in 32-bit size units but offset() is multiplying it by destination type size units. When operating with double input varyings, const_index value could be not aligned to 64 bits. To fix it, we load the double vector as if it was a float based vector with twice the number of components. Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]> Reviewed-by: Timothy Arceri <[email protected]> Cc: "12.0" <[email protected]>
* i965/fs: fix FS_OPCODE_CINTERP for unpacked double input varyingsSamuel Iglesias Gonsálvez2016-06-061-1/+12
| | | | | | | | | | | Data starts at suboffet 3 in 32-bit units (12 bytes), so it is not 64-bit aligned and the current implementation fails to read the data properly. Instead, when there is is a double input varying, read it as vector of floats with twice the number of components. Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]> Reviewed-by: Timothy Arceri <[email protected]> Cc: "12.0" <[email protected]>
* glsl: geom shader max_vertices layout must match.Dave Airlie2016-06-063-9/+13
| | | | | | | | | | | | | From GLSL 4.5 spec, "4.4.2.3 Geometry Outputs". "all geometry shader output vertex count declarations in a program must declare the same count." Fixes: GL45-CTS.geometry_shader.output.conflicted_output_vertices_max Reviewed-by: Alejandro Piñeiro <[email protected]> Cc: "11.2 12.0" <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
* anv/pipeline: Add support for caching the push constant mapJason Ekstrand2016-06-061-4/+29
| | | | | Signed-off-by: Jason Ekstrand <[email protected]> Reviewed-by: Kristian Høgsberg <[email protected]>
* glsl: use enum glsl_interface_packing in more places. (v2)Dave Airlie2016-06-0611-36/+46
| | | | | | | | | | | | | | | Although the glsl_types.h stores this in a bitfield, we should hide that from everyone else. Hide the cast in an accessor method and use the enum everywhere. This makes things a bit nicer in gdb, and improves type safety. v2: fix a few pieces of interface I missed that caused some piglit regressions. Signed-off-by: Dave Airlie <[email protected]> Reviewed-by: Ilia Mirkin <[email protected]>
* i965: don't use NumLayers for 3D textures.Dave Airlie2016-06-061-1/+1
| | | | | | | | | | | | For 3D textures we shouldn't be using NumLayers, we need to get it from the depth. This fixes: GL45-CTS.geometry_shader.layered_framebuffer.clear_call_support Reviewed-by: Eduardo Lima Mitev <[email protected]> Cc: "12.0" <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
* glsl: for anonymous struct matching use without_array() (v3)Dave Airlie2016-06-061-3/+20
| | | | | | | | | | | | | | | | | With tessellation shaders we can have cases where we have arrays of anon structs, so make sure we match using without_array(). Fixes: GL45-CTS.tessellation_shader.tessellation_control_to_tessellation_evaluation.gl_in v2: test lengths match as well (Ilia) v3: descend array lengths to check for matches as well (Ilia) Reviewed-by: Ilia Mirkin <[email protected]> Cc: "12.0" <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
* glsl/ast: don't crash when func_name is NULLDave Airlie2016-06-061-0/+4
| | | | | | | | | | | | This fixes a crash in GL43-CTS.shader_subroutine.subroutines_not_allowed_as_variables_constructors_and_argument_or_return_types If we can't find the func_name in one of these paths, we have emitted an earlier error so just return here. Reviewed-by: Timothy Arceri <[email protected]> Cc: "11.2 12.0" <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
* glsl: handle ast_aggregate in has_sequence_subexpression. (v2)Dave Airlie2016-06-061-1/+1
| | | | | | | | | | | | | | | GL43-CTS.compute_shader.work-group-size does uniform uint g_uniform[gl_WorkGroupSize.z + 20] = { 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24 }; The initializer triggers the GLSL 4.30/GLES3 tests for constant sequence subexpressions, so it doesn't happen unless you are using those, so just return false as this path is now reachable. v2: update commit msg with diagnosis Acked-by: Timothy Arceri <[email protected]> Cc: "11.2 12.0" <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
* mesa: Try to unbreak the MSVC build.Kenneth Graunke2016-06-052-0/+8
| | | | | | | PATH_MAX is apparently not a thing on Windows. Borrow the hack from pipe_loader.c to try and make this work. Signed-off-by: Kenneth Graunke <[email protected]>
* mesa: Add MESA_SHADER_CAPTURE_PATH for writing .shader_test files.Kenneth Graunke2016-06-053-0/+77
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This writes linked shader programs to .shader_test files to $MESA_SHADER_CAPTURE_PATH in the format used by shader-db (http://cgit.freedesktop.org/mesa/shader-db). It supports both GLSL shaders and ARB programs. All stages that are linked together are written in a single .shader_test file. This eliminates the need for shader-db's split-to-files.py, as Mesa produces the desired format directly. It's much more reliable than parsing stdout/stderr, as those may contain extraneous messages, or simply be closed by the application and unavailable. We have many similar features already, but this is a bit different: - MESA_GLSL=dump writes to stdout, not files. - MESA_GLSL=log writes each stage to separate files (rather than all linked shaders in one file), at draw time (not link time), with uniform data and state flag info. - Tapani's shader replacement mechanism (MESA_SHADER_DUMP_PATH and MESA_SHADER_READ_PATH) also uses separate files per shader stage, but allows reading in files to replace an app's shader code. v2: Dump ARB programs too, not just GLSL. v3: Don't dump bogus 0.shader_test file. v4: Add "GL_ARB_separate_shader_objects" to the [require] block. v5: Print "GLSL 4.00" instead of "GLSL 4.0" in the [require] block. v6: Don't hardcode /tmp/mesa. v7: Fix memoization of getenv(). v8: Also print "SSO ENABLED" (suggested by Timothy). v9: Also handle ES shaders (suggested by Ilia). v10: Guard against MESA_SHADER_CAPTURE_PATH being too long; add _mesa_warning calls on error handling (suggested by Ben). v11: Fix crash when variable is unset introduced in v10. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* nv50,nvc0: fix BGR10_A2UI vertex formatIlia Mirkin2016-06-051-1/+1
| | | | | | | | This is mostly academic as this is not reachable from GL, which only has the packed RGB10_A2UI vertex format. Signed-off-by: Ilia Mirkin <[email protected]> Cc: "12.0" <[email protected]>
* nvc0: do not clear surfaces bins in the validate functionSamuel Pitoiset2016-06-052-5/+2
| | | | | | | | | We should not call nouveau_bufctx_reset() inside a validate function. This only affects Fermi where images are aliased between 3D and CP. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Ilia Mirkin <[email protected]> Cc: "12.0" <[email protected]>
* nvc0: re-validate images after launching a grid on FermiSamuel Pitoiset2016-06-051-0/+3
| | | | | | | | | | | | | | | | | | | | | Images invalidation is a bit weird on Fermi and there is already a hack which forces invalidating all images when launching a computer shader to help in fixing 3D<->CP interaction. However, we need to re-validate images for compute because nvc0_compute_invalidate_surfaces() will destroy the previous binding. This is not really good for performance purposes but this might be improved later. This fixes the following piglits: - spec/arb_compute_shader/execution/basic-uniform-access - spec/arb_compute_shader/execution/mutiple-texture-reading - spec/arb_compute_shader/execution/multiple-workgroups - spec/glsl-4.30/execution/built-in-functions/cs-* (207 tests) Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Ilia Mirkin <[email protected]> Cc: "12.0" <[email protected]>
* radeonsi: fix images with level > 0Marek Olšák2016-06-051-1/+1
| | | | | | | | | | This should fix spec@arb_shader_image_load_store@level. Broken by: Commit: 95c5bbae66af3ca1f805d94f6fe8d8e4ba2c9c43 radeonsi: set some image descriptor fields at bind time Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* nvc0: reduce overhead from always marking images dirtyIlia Mirkin2016-06-041-9/+36
| | | | | | | | | | We would revalidate images when anything was touched at all. Which is unfortunate, since the state tracker does not use CSO's to reduce the workload. So instead implement a protocol to ensure that something has changed before revalidating all the images. Signed-off-by: Ilia Mirkin <[email protected]> Cc: "12.0" <[email protected]>