summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* glsl: Convert function inlining to the util hash tableThomas Helland2016-09-121-3/+3
| | | | | Signed-off-by: Thomas Helland <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* mesa: Convert string_to_uint_map to the util hash tableThomas Helland2016-09-121-44/+22
| | | | | | | | | | | And remove the now unused hash_table_replace. V2: Actually do the equivalent thing, and don't leak memory V3: fix minor typo in comment (Timothy Arceri) Signed-off-by: Thomas Helland <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* util: Move hash_table_call_foreach to util hash tableThomas Helland2016-09-122-14/+13
| | | | | | | | | | It is included through the util/hash_table include in the program hash_table, so this should be safe. This will be needed when we start converting each use of the program_hash_table, as some places need this function. Signed-off-by: Thomas Helland <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* mesa: Remove prog_hash_table.cThomas Helland2016-09-126-85/+23
| | | | | | | | | | | | | Here we make the prog_hash_table functionally equivalent to the one in util by wrapping the remaing functions that differ. We also move the functions to the header so we can remove the c file. This enables us to do a step-by-step replacement of the table. Signed-off-by: Thomas Helland <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* mesa: Remove unused hash table includesThomas Helland2016-09-124-4/+0
| | | | | | | This should prevent us from rebuilding the world. Signed-off-by: Thomas Helland <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* freedreno/a3xx: disable filtering for texture buffers and int texturesIlia Mirkin2016-09-111-0/+2
| | | | Signed-off-by: Ilia Mirkin <[email protected]>
* st/clover: Define __OPENCL_VERSION__ on the device sideNiels Ole Salscheider2016-09-101-0/+3
| | | | | | | | This is required by the OpenCL standard. Signed-off-by: Niels Ole Salscheider <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]> Reviewed-by: Vedran Miletić <[email protected]>
* gm107/ir: allow indirect inputs to be loaded by frag shaderIlia Mirkin2016-09-102-5/+21
| | | | | | | | | Looks like the GM107 IPA op does not allow a separate offset when using an indirect register. Instead we must use AL2P like we do for indirect vertex operations on Kepler+. Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Samuel Pitoiset <[email protected]>
* gm107/ir: AL2P writes to a predicate registerIlia Mirkin2016-09-101-0/+1
| | | | | | | | | | | We have to force it to write to predicate 7 (aka PT) in order for it not to mess up another predicate. Unclear what would be returned in the predicate, perhaps an error code for out-of-bounds requests. Blob doesn't seem to check it. Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Samuel Pitoiset <[email protected]> Cc: [email protected]
* i965: Fix calculation of the image height at start levelAntia Puentes2016-09-101-6/+21
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - Fixes CTS tests: * GL44-CTS.shader_image_size.advanced-nonMS-cs-float * GL44-CTS.shader_image_size.advanced-nonMS-cs-int * GL44-CTS.shader_image_size.advanced-nonMS-cs-uint * GL44-CTS.shader_image_size.advanced-nonMS-gs-float * GL44-CTS.shader_image_size.advanced-nonMS-gs-int * GL44-CTS.shader_image_size.advanced-nonMS-gs-uint * GL44-CTS.shader_image_size.advanced-nonMS-tes-float * GL44-CTS.shader_image_size.advanced-nonMS-tes-int * GL44-CTS.shader_image_size.advanced-nonMS-tes-uint * GL44-CTS.shader_image_size.advanced-nonMS-vs-float * GL44-CTS.shader_image_size.advanced-nonMS-vs-int * GL44-CTS.shader_image_size.advanced-nonMS-vs-uint v1: (written by Dave Airlie) Always shift height images for levels. Fixed the CTS test. v2: Only shift height if the texture is not an 1D_ARRAY, it fixes assertion in GL44-CTS.texture_view.gettexparameter due to the original patch (Antia). v3: Remove the loop. Do not shift height either for 1D textures. Use an explicit switch and add an assertion (levels == 0) for multisampled textures (Jason). v4: Rectangle textures can not have levels either (Ilia Mirkin). Signed-off-by: Dave Airlie <[email protected]> Signed-off-by: Antia Puentes <[email protected]> Reviewed-by: Anuj Phogat <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* radeonsi: flush TC L2 before using a compute indirect bufferMarek Olšák2016-09-091-2/+10
| | | | | | | There is no known test for this. Cc: 12.0 <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: fix the VGT performance tweak for small instancesMarek Olšák2016-09-091-5/+6
| | | | | | | | Based on the VGT spec. The Vulkan driver doesn't do it optimally and they plan to fix it. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: remove the cache_flush atomMarek Olšák2016-09-097-12/+9
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* winsys/amdgpu: replace OUT_CS with radeon_emitMarek Olšák2016-09-091-14/+12
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* winsys/radeon: replace OUT_CS with radeon_emitMarek Olšák2016-09-091-7/+5
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* doc: document GALLIUM_DRIVERChristoph Haag2016-09-091-0/+2
| | | | | | v2: Add dot at end of sentence Reviewed-by: Nicolai Hähnle <[email protected]>
* egl/android: Set EGL_MAX_PBUFFER_WIDTH and EGL_MAX_PBUFFER_HEIGHTHaixia Shi2016-09-091-0/+2
| | | | | | | | | | | | Set config attributes EGL_MAX_PBUFFER_WIDTH and EGL_MAX_PBUFFER_HEIGHT to hard-coded non-zero values. These two attributes are required on Android. v2: use _EGL_MAX_PBUFFER_WIDTH/HEIGHT from egldefines.h (based on discussion on the first version) Signed-off-by: Tomasz Figa <[email protected]> Signed-off-by: Tapani Pälli <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* android: depend on libmesa_genxml from i965 Android.gen.mkTapani Pälli2016-09-091-0/+2
| | | | | | | | Static library dependency is required to pull the generated XML headers into the generated C file. Signed-off-by: Tapani Pälli <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* i965: release GLSL IR in LinkShader after it's not neededTapani Pälli2016-09-091-0/+11
| | | | | Signed-off-by: Tapani Pälli <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* glsl: use hash instead of exec_list in copy propagationTapani Pälli2016-09-091-51/+41
| | | | | | | | | | | | This change makes copy propagation pass faster. Complete link time spent in test case attached to bug 94477 goes down to ~400 secs from over 500 secs on my HSW machine. Does not fix the actual issue but brings down the total. No regressions seen in CI. v2: do not leak hash_table structure Signed-off-by: Tapani Pälli <[email protected]> Reviewed-by: Eric Anholt <[email protected]>
* i965/fs: Fail the shader compile instead of asserting when we can't spillJason Ekstrand2016-09-081-2/+3
| | | | | | | | | | | Blorp doesn't handle spilling so we set allow_spilling to false in that case. The blorp 16x MSAA resolve shader spills in 16-wide but not 8-wide. This commit makes it so that we fail the 16-wide compile and successfully fall back to 8-wide instead of just assert-failing when trying to compile the 16-wide shader. Signed-off-by: Jason Ekstrand <[email protected]> Reviewed-by: Francisco Jerez <[email protected]>
* nir/gcm: Add global value numbering supportJason Ekstrand2016-09-082-6/+25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Unlike the current CSE pass, global value numbering is capable of detecting common values even if one does not dominate the other. For instance, in you have if (...) { ssa_1 = ssa_0 + 7; /* use ssa_1 */ } else { ssa_2 = ssa_0 + 7; /* use ssa_2 */ } Global value numbering doesn't care about dominance relationships so it figures out that ssa_1 and ssa_2 are the same and converts this to if (...) { ssa_1 = ssa_0 + 7; /* use ssa_1 */ } else { /* use ssa_1 */ } Obviously, we just broke SSA form which is bad. Global code motion, however, will repair this for us by turning this into ssa_1 = ssa_0 + 7; if (...) { /* use ssa_1 */ } else { /* use ssa_1 */ } This intended to eventually mostly replace CSE. However, conventional CSE may still be useful because it's less of a scorched-earth approach and doesn't require GCM. This makes it a bit more appropriate for use as a clean-up in a late optimization run. Signed-off-by: Jason Ekstrand <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* nir/gcm: Call nir_metadata_preserveJason Ekstrand2016-09-081-0/+3
| | | | | | Signed-off-by: Jason Ekstrand <[email protected]> Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* r300g: Set R300_VAP_CNTL on RSxxx to avoid triangle flickeringMax Staudt2016-09-091-1/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On the RSxxx chip series, HW TCL is missing and r300_emit_vs_state() is never called. However, if R300_VAP_CNTL is never set, the hardware (at least the RS690 I tested this on) comes up with rendering artifacts, and parts that are uploaded before this "fix" remain broken in VRAM. This causes artifacts as in fdo#69076 ("triangle flickering"). It seems like this setup needs to happen at least once after power on for 3D rendering to work properly. In the DDX with EXA, this happens in RADEON_SWITCH_TO_3D() when processing an XRENDER Composite or an Xv request. So playing back a video or starting a GTK+2 application fixes 3D rendering for the rest of the session. However, this auto-fix doesn't happen when EXA is not used, such as with GLAMOR or Wayland. This patch ensures the register is configured even in absence of the DDX's EXA module. The register setting is taken from: xf86-video-ati -- RADEONInit3DEngineInternal() mesa/src/mesa/drivers/dri/r300 -- r300EmitClearState() Tested on RS690. CC: <[email protected]> Signed-off-by: Max Staudt <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
* gallium: remove PIPE_BIND_TRANSFER_READ/WRITEMarek Olšák2016-09-0829-95/+24
| | | | | | | | not used in any useful way Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]> Reviewed-by: Roland Scheidegger <[email protected]>
* radeonsi: unify si_set_optimal_micro_tile_mode call sitesMarek Olšák2016-09-081-8/+4
| | | | | | | There is nothing special happening in those code blocks. Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: fix texture reinterpretation after DCC fast clearMarek Olšák2016-09-081-12/+20
| | | | | | | | | | The problem is that TC-compatible DCC clear codes translate into different clear values when you change the format. I have a new piglit reproducing the issue. Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: enable DCC fast clear for 128-bit formatsMarek Olšák2016-09-081-13/+32
| | | | | Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: clamp integer clear color values for DCC fast clearMarek Olšák2016-09-081-4/+12
| | | | | | | It should be possible to get TC-compatible fast clear more often now. Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* Revert "radeonsi: enable SDMA on CIK"Marek Olšák2016-09-081-0/+4
| | | | | | | This reverts commit 0241d8300f66ee2c6c2c55fe64ac88d76440c591. It doesn't work with mobile Bonaire. It looks like the programming of tiling parameters is wrong on some chips.
* doc: fix typo of GALLIUM_HUD_TOGGLE_SIGNALChristoph Haag2016-09-081-1/+1
| | | | | | | In the original commit message in 56a1c10 it was wrongly used too: - env GALLIUM_HUD_SIGNAL_TOGGLE: toggle visibility via signal Reviewed-by: Nicolai Hähnle <[email protected]>
* nir/spirv: Refactor variable deocration handlingJason Ekstrand2016-09-081-15/+29
| | | | | | | | | | | Previously, we dind't apply variable decorations to the members of a split structure variable. This doesn't quite work, unfortunately, because things such as the "flat" qualifier may get applied to an entire structure instead of propagated to the members. This fixes 9 of the new CTS tests in the dEQP-VK.glsl.linkage.varying.struct.* group. Signed-off-by: Jason Ekstrand <[email protected]> Cc: "12.0" <[email protected]>
* nir/spirv: Break variable decoration handling into a helperJason Ekstrand2016-09-081-74/+81
| | | | | Signed-off-by: Jason Ekstrand <[email protected]> Cc: "12.0" <[email protected]>
* aubinator: only use program_invocation_short_name with glibc/cygwinJonathan Gray2016-09-081-1/+7
| | | | | | | | | | program_invocation_short_name is a gnu extension. Limit use of it to glibc and cygwin and otherwise use getprogname() which is available on BSD and OS X. Signed-off-by: Jonathan Gray <[email protected]> Reviewed-by: Kristian Høgsberg <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* aubinator: include libgen.h for basename(3)Jonathan Gray2016-09-081-0/+1
| | | | | | | | | | Include libgen.h for basename as required by posix. The definition is not found on at least OpenBSD otherwise. Signed-off-by: Jonathan Gray <[email protected]> Reviewed-by: Kristian Høgsberg <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* aubinator: stop using non portable error() functionJonathan Gray2016-09-081-16/+30
| | | | | | | | | | | | | error() is a gnu extension and is not present on OpenBSD and likely other systems. Convert use of error to fprintf/strerror/exit. Signed-off-by: Jonathan Gray <[email protected]> Reviewed-by: Kristian Høgsberg <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* egl: Fix up indentation on previous commitAdam Jackson2016-09-081-2/+2
| | | | | | This was requested in review but I pushed the wrong version. Signed-off-by: Adam Jackson <[email protected]>
* egl: Document why EGL_OPENGL{, _ES}_API are mostly identicalAdam Jackson2016-09-081-0/+10
| | | | | Signed-off-by: Adam Jackson <[email protected]> Reviewed-by: Alejandro Piñeiro <[email protected]>
* anv: Link to libX11-xcb only when unneededChad Versace2016-09-081-3/+5
| | | | | | | | | | | | | The Makefile unconditionally linked libX11-xcb into libvulkan_intel.so. But it's needed only if HAVE_PLATFORM_X11. Fixes build of libvulkan_intel.so on Chromium OS, which has no X11 libraries. Fixes: 71258e9462c ("anv/x11: Add support for Xlib platform") Cc: Kevin Strasser <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* swr: fixes for format mapping and texture sizingTim Rowley2016-09-081-2/+16
| | | | Signed-off-by: Tim Rowley <[email protected]>
* intel/blorp: Allow single slice converter to suppress number of layersTopi Pohjolainen2016-09-081-3/+0
| | | | | Signed-off-by: Topi Pohjolainen <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* spirv/nir: Implement OpAtomicLoad/Store for shared variablesLionel Landwerlin2016-09-071-1/+34
| | | | | | | Missing bits from 2afb950161f847d9b0a7be03dfb62cacc5ea51ba. Signed-off-by: Lionel Landwerlin <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir/spirv: Remove an erroneous "fall through" commentJason Ekstrand2016-09-071-1/+0
|
* EGL: Combine the GL and GLES current contexts (v2)Kyle Brenneman2016-09-074-107/+22
| | | | | | | | | | | | | | | | | | | | | | Only keep track of a single current context, instead of separate contexts for GL and GLES. In EGL 1.4 (and 1.5), EGL_OPENGL_API and EGL_OPENGL_ES_API are supposed to be interchangeable for all purposes except for eglCreateContext. The _EGLThreadInfo::CurrentContexts array is now a single pointer to the current context, which may be a GL or GLES context. In addition, it now keeps track of the current API as an enum instead of an index. eglMakeCurrent will now replace the current context, regardless of which client API is used for for the current and new contexts. It no longer checks for a conflicting context. In addition, calling eglMakeCurrent with EGL_NO_CONTEXT will now release the current context regardless of the current API. v2: Rebased against master (Adam Jackson) Reviewed-by: Adam Jackson <[email protected]>
* gbm: wire up fence extensionRob Clark2016-09-074-1/+6
| | | | | | | | v2: make fence extension optional to not break non-i965 classic drivers, and move __DRI2_FENCE into core extensions, based on comments from Emil Signed-off-by: Rob Clark <[email protected]>
* freedreno: reject imports with bogus pitchRob Clark2016-09-073-8/+7
| | | | Signed-off-by: Rob Clark <[email protected]>
* gbm: add missing R8 and GR88 formatsRob Clark2016-09-072-1/+13
| | | | | Reviewed-by: Eric Engestrom <[email protected]> Signed-off-by: Rob Clark <[email protected]>
* spirv/nir: Add support for OpAtomicLoad/StoreLionel Landwerlin2016-09-071-11/+113
| | | | | | | | | | | | Fixes new CTS tests : dEQP-VK.spirv_assembly.instruction.compute.opatomic.load dEQP-VK.spirv_assembly.instruction.compute.opatomic.store v2: don't handle images like ssbo/ubo (Jason) Signed-off-by: Lionel Landwerlin <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* radeonsi: skip redundant INDEX_TYPE writesMarek Olšák2016-09-073-20/+32
| | | | | | Ported from Vulkan. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: add more unlikely() uses into si_draw_vboMarek Olšák2016-09-071-5/+5
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-26 00:55:35 +0000