aboutsummaryrefslogtreecommitdiffstats
path: root/src/mesa/state_tracker
Commit message (Collapse)AuthorAgeFilesLines
* nir: Distinguish between normal uniforms and UBOsJason Ekstrand2019-01-081-2/+1
| | | | | | | | | | | | | | | | Previously, NIR had a single nir_var_uniform mode used for atomic counters, UBOs, samplers, images, and normal uniforms. This commit splits this into nir_var_uniform and nir_var_ubo where nir_var_uniform is still a bit of a catch-all but the nir_var_ubo is specific to UBOs. While we're at it, we also rename shader_storage to ssbo to follow the convention. We need this so that we can distinguish between normal uniforms and UBO access at the deref level without going all the way back variable and seeing if it has an interface type. Reviewed-by: Alejandro Piñeiro <[email protected]> Reviewed-by: Caio Marcelo de Oliveira Filho <[email protected]>
* st/glsl: refactor st_link_nir()Timothy Arceri2019-01-071-36/+16
| | | | | | | | | | | | | | | | | | The functional change here is moving the nir_lower_io_to_scalar_early() calls inside st_nir_link_shaders() and moving the st_nir_opts() call after the call to nir_lower_io_arrays_to_elements(). This fixes a bug with the following piglit test due to the current code not cleaning up dead code after we lower arrays. This was causing an assert in the new duplicate varyings link time opt introduced in 70be9afccb23. tests/spec/glsl-1.10/execution/vsfs-unused-array-member.shader_test Moving the nir_lower_io_to_scalar_early() calls also allows us to tidy up the code a little and merge some loops. Reviewed-by: Eric Anholt <[email protected]>
* nir: rename nir_link_constant_varyings() nir_link_opt_varyings()Timothy Arceri2019-01-021-1/+1
| | | | | | | | | | The following patches will add support for an additional optimisation so this function will no longer just optimise varying constants. Tested-by: Dieter Nützel <[email protected]> Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Eric Anholt <[email protected]>
* st/glsl_to_nir: call nir_lower_load_const_to_scalar() in the stTimothy Arceri2019-01-021-1/+3
| | | | | | | | | This will help the new opt introduced in the following patches allowing us to remove extra duplicate varyings. Tested-by: Dieter Nützel <[email protected]> Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Eric Anholt <[email protected]>
* st/mesa: expose GL_NV_shader_atomic_float when ATOMFADD is supportedIlia Mirkin2018-12-261-0/+1
| | | | | Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* st/mesa: select ATOMFADD when source type is floatIlia Mirkin2018-12-262-0/+3
| | | | | Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* st/mesa: allow glDrawElements to work with GL_SELECT feedbackIlia Mirkin2018-12-262-23/+42
| | | | | | | | | | | Not sure if this ever worked, but the current logic for setting the min/max index is definitely wrong for indexed draws. While we're at it, bring in all the usual logic from the non-indirect drawing path. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=109086 Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Brian Paul <[email protected]> Reviewed-by: Roland Scheidegger <[email protected]>
* st/nir: Drop unused gl_program parameter in VS input handling helper.Kenneth Graunke2018-12-211-2/+2
| | | | | | | Nobody uses this, so let's drop it. This makes the helper callable from places without a gl_program. Reviewed-by: Marek Olšák <[email protected]>
* st/nir: Gather info after applying lowering FS variant featuresKenneth Graunke2018-12-211-0/+4
| | | | | | | | DrawPixels lowering, for example, adds new varyings that need to be accounted for in inputs_read. The earlier info gathering at link time cannot account for this. Reviewed-by: Marek Olšák <[email protected]>
* st/mesa: Combine the DrawPixels and Bitmap passthrough VS programs.Kenneth Graunke2018-12-214-33/+22
| | | | | | They're now identical, so we can just compile it once. Reviewed-by: Marek Olšák <[email protected]>
* st/mesa: Don't open code the drawpixels vertex shader.Kenneth Graunke2018-12-211-27/+11
| | | | | | Now that we always copy color, we can just use the util function. Reviewed-by: Marek Olšák <[email protected]>
* st/mesa: Drop !passColor optimization in drawpixels shaders.Kenneth Graunke2018-12-212-28/+20
| | | | | | | | | | | | | | | | | | | | The glDrawPixels passthrough vertex shader copies position and texcoord vertex attributes to varying outputs. It also optionally copies a third gl_Color attribute, which sometimes is unnecessary. Until now, we've compiled separate variants of the shader, one of which does this extra copy, and the other of which doesn't. We have done this since 2007. But, the vertex shader runs for a whopping four vertices, and so the cost of a copying a single input to output is likely inconsequential. In theory, we could bind one fewer vertex element - but we always bind all three regardless. So, we don't even get that savings. This patch unifies the two, so we always copy the optional color, and save having to compile the variant. It also makes the VS input interface match up with the vertex element state without any dead (unused) input attributes. Reviewed-by: Marek Olšák <[email protected]>
* st/mesa: Drop dead 'passthrough_fs' field.Kenneth Graunke2018-12-211-2/+0
| | | | | | | Dead since 2015 (commit 5142564734bd68f165b02e29e384ebbcf91cce38). Reviewed-by: Ilia Mirkin <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* st/mesa: remove sampler associated with buffer texture in pbo logicIlia Mirkin2018-12-201-5/+0
| | | | | | | | | | | | | | A long time ago, when this was first implemented, not having a sampler bound would cause problems on Fermi. I didn't work out the reasons, but the solution was simple -- just put the samplers back in. Since then, regular texturing paths appear to have lost their associated samplers which required a fuller investigation and fix in nouveau. Now that this is done, this code should no longer need a sampler state for fetching texels from a buffer texture. Signed-off-by: Ilia Mirkin <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* nir/opt_peephole_select: Don't peephole_select expensive math instructionsIan Romanick2018-12-171-1/+1
| | | | | | | | | | | | | | | | On some GPUs, especially older Intel GPUs, some math instructions are very expensive. On those architectures, don't reduce flow control to a csel if one of the branches contains one of these expensive math instructions. This prevents a bunch of cycle count regressions on pre-Gen6 platforms with a later patch (intel/compiler: More peephole select for pre-Gen6). v2: Remove stray #if block. Noticed by Thomas. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Thomas Helland <[email protected]> Reviewed-by: Lionel Landwerlin <[email protected]>
* nir/opt_peephole_select: Don't try to remove flow control around indirect loadsIan Romanick2018-12-171-1/+1
| | | | | | | | | | | | | | | | | | | That flow control may be trying to avoid invalid loads. On at least some platforms, those loads can also be expensive. No shader-db changes on any Intel platform (even with the later patch "intel/compiler: More peephole select"). v2: Add a 'indirect_load_ok' flag to nir_opt_peephole_select. Suggested by Rob. See also the big comment in src/intel/compiler/brw_nir.c. v3: Use nir_deref_instr_has_indirect instead of deref_has_indirect (from nir_lower_io_arrays_to_elements.c). v4: Fix inverted condition in brw_nir.c. Noticed by Lionel. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Lionel Landwerlin <[email protected]>
* st/nir: Use nir_src_as_uint for tokensJason Ekstrand2018-12-161-5/+1
| | | | Reviewed-by: Marek Olšák <[email protected]>
* mesa/st: Expose compute shaders when NIR support is advertised.Eric Anholt2018-12-131-7/+12
| | | | | | | | We have a NIR path, and V3D doesn't have TGSI input for compute (only what TTN can handle for the various gallium-internal shaders). Reviewed-by: Kenneth Graunke <[email protected]> Reviewed-by: Jordan Justen <[email protected]>
* mesa/st/nir: fix missing nir_compact_varyingsRob Clark2018-12-111-1/+2
| | | | | | | | | LinkedTransformFeedback is normally populated, which had nerf'd varying packing since the check was introduced. Fixes: dbd52585fa9 st/nir: Disable varying packing when doing transform feedback. Signed-off-by: Rob Clark <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* st/mesa: Add support for EXT_multisampled_render_to_textureKristian H. Kristensen2018-12-063-0/+5
| | | | | | | | | In gallium, we model the attachment sample count as a new nr_samples field in pipe_surface. A driver can indicate support for the extension using the new pipe cap, PIPE_CAP_MULTISAMPLED_RENDER_TO_TEXTURE. Reviewed-by: Eric Anholt <[email protected]> Signed-off-by: Kristian H. Kristensen <[email protected]>
* st/mesa: expose EXT_render_snorm on GLESMarek Olšák2018-12-041-0/+8
| | | | Reviewed-by: Ilia Mirkin <[email protected]>
* st/mesa: expose GL_OES_texture_viewMarek Olšák2018-12-041-0/+1
| | | | | | | | | | For format fallbacks like ETC and ASTC, switching between sRGB and linear decoding is undefined, or at least is not bit-exact. Same as EXT_texture_sRGB_decode on GLES. There are no piglit or dEQP regresssions. Reviewed-by: Ilia Mirkin <[email protected]>
* Revert "st/mesa: silenced unhanded enum warning in st_glsl_to_tgsi.cpp"Matt Turner2018-12-031-1/+0
| | | | | | | This reverts commit 198c50f4873758e9f64d89eea262af5dd1644df9. This needs to be reverted after commit 017199d2d2e4 ("mesa: Revert INTEL_fragment_shader_ordering support")
* mesa/st: do not probe for the same texture-formats twiceErik Faye-Lund2018-12-031-4/+4
| | | | | | | This should be equalent of what we did before. Signed-off-by: Erik Faye-Lund <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* android: st/mesa: fix building error due to sched_getcpu()Mauro Rossi2018-12-011-1/+1
| | | | | | | | | | Android has cpufeatures library but pinning of threads is not supported PIPE_OS_LINUX code path causes build error due to sched_getcpu() unavailable thus we need to avoid setting HAVE_SCHED_GETCPU for Android Fixes: 48f2160 ("st/mesa: regularly re-pin driver threads to the CCX where the app thread is") Signed-off-by: Mauro Rossi <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* mesa/st: better colormask check for clear fallbackRob Clark2018-11-271-2/+8
| | | | | | | | | For RGB surfaces (for example) we don't really care that the colormask is 0x7 instead of 0xf. This should not trigger clear_with_quad() slowpath. Signed-off-by: Rob Clark <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* mesa/st: swap order of clear() and clear_with_quad()Rob Clark2018-11-271-3/+3
| | | | | | | | | | | If we can't clear all the buffers with pctx->clear() (say, for example, because of ColorMask), push the buffers we *can* clear with pctx->clear() first. Tilers want to see clears coming before draws to enable fast- paths, and clearing one of the attachments with a quad-draw first confuses that logic. Signed-off-by: Rob Clark <[email protected]> Reviewed-by: Eric Anholt <[email protected]>
* mesa/st: Make st_pipe_vertex_format static.Mathias Fröhlich2018-11-262-4/+1
| | | | | Reviewed-by: Marek Olšák <[email protected]> Signed-off-by: Mathias Fröhlich <[email protected]>
* mesa/st: Use binding information from the VAO in feedback rendering.Mathias Fröhlich2018-11-263-62/+68
| | | | | | | | | | | | Use VAO binding information in feedback rendering. In theory it should reduce the amount of buffer objects scheduled for rendering. Feedback rendering is implemented in a crude way anyhow, so I do not expect much gain here. But for the sake of code reuse we should use the same code for the same task. And finally if feeback rendering may get improved the array setup is already well done there. Reviewed-by: Marek Olšák <[email protected]> Signed-off-by: Mathias Fröhlich <[email protected]>
* mesa/st: Avoid extra references in the feedback draw function scope.Mathias Fröhlich2018-11-261-2/+1
| | | | | | | | | | | | The change removes the reference that is held on the entries of the vbuffers[] array. The new code does not do that anymore as following the code into draw_set_vertex_buffers() the draw context holds an other reference as long as it is reset down the function again. So it should be already by that argument save to remove that additional reference count. Reviewed-by: Marek Olšák <[email protected]> Signed-off-by: Mathias Fröhlich <[email protected]>
* mesa/st: Factor out array and buffer setup from st_atom_array.c.Mathias Fröhlich2018-11-262-22/+74
| | | | | | | | | Factor out vertex array setup routines from the array state atom. The factored functions will be used in feedback rendering in the next change. Reviewed-by: Marek Olšák <[email protected]> Signed-off-by: Mathias Fröhlich <[email protected]>
* mesa/st: Only unmap the uploader that was actually used.Mathias Fröhlich2018-11-261-1/+2
| | | | | | | | In st_atom_array, we only need to unmap the upload buffer that was actually used. Reviewed-by: Marek Olšák <[email protected]> Signed-off-by: Mathias Fröhlich <[email protected]>
* mesa/st: Only care about the uploader if it was used.Mathias Fröhlich2018-11-261-3/+3
| | | | | | | | In st_atom_array, we only need to care for unmapping the upload buffer if we actually used it. Reviewed-by: Marek Olšák <[email protected]> Signed-off-by: Mathias Fröhlich <[email protected]>
* mesa: Factor out struct gl_vertex_format.Mathias Fröhlich2018-11-213-16/+16
| | | | | | | | | | | | | | | | Factor out struct gl_vertex_format from array attributes. The data type is supposed to describe the type of a vertex element. At this current stage the data type is only used with the VAO, but actually is useful in various other places. Due to the bitfields being used, special care needs to be taken for the glGet code paths. v2: Change unsigned char -> GLubyte. Use struct assignment for struct gl_vertex_format. Reviewed-by: Brian Paul <[email protected]> Reviewed-by: Marek Olšák <[email protected]> Signed-off-by: Mathias Fröhlich <[email protected]>
* st/mesa: pin driver threads to a fixed CCX when glthread is enabledMarek Olšák2018-11-201-0/+11
| | | | | | | | radeonsi has 3 driver threads (glthread, gallium, winsys), other drivers may have 2 (glthread, gallium), so it makes sense to pin them to a random CCX and keep that irrespective of the app thread. Reviewed-by: Dave Airlie <[email protected]>
* st/mesa: regularly re-pin driver threads to the CCX where the app thread isMarek Olšák2018-11-202-0/+34
| | | | | | | | | This is used when glthread is disabled. Mesa pretty much chases the app thread on the CPU. The performance is the same as pinning the app thread. Reviewed-by: Dave Airlie <[email protected]>
* meson: Add tests to suitesDylan Baker2018-11-201-3/+5
| | | | | | | | | | | | | | | | Meson test has a concepts of suites, which allow tests to be grouped together. This allows for a subtest of tests to be run only (say only the tests for nir). A test can be added to more than one suite, but for the most part I've only added a test to a single suite, though I've added a compiler group that includes nir, glsl, and glcpp tests. To use this you'll need to invoke meson test directly, instead of ninja test (which always runs all targets). it can be invoked as: `meson test -C builddir --suite $suitename` (meson test has addition options that are pretty useful). Tested-By: Gert Wollny <[email protected]> Acked-by: Eric Engestrom <[email protected]>
* st/mesa: disable L3 thread pinningMarek Olšák2018-11-121-9/+0
| | | | | | | This implementation can have massive drawbacks. Cc: 18.3 <[email protected]> Reviewed-by: Edmondo Tommasina <[email protected]>
* st/mesa: make use of nir_link_constant_varyings()Timothy Arceri2018-11-101-0/+3
| | | | | | | | | | | | | | | | | | Shader-db results radeonsi (VEGA): Totals from affected shaders: SGPRS: 161464 -> 161368 (-0.06 %) VGPRS: 86904 -> 86292 (-0.70 %) Spilled SGPRs: 296 -> 314 (6.08 %) Spilled VGPRs: 0 -> 0 (0.00 %) Private memory VGPRs: 0 -> 0 (0.00 %) Scratch size: 0 -> 0 (0.00 %) dwords per thread Code Size: 3618596 -> 3573852 (-1.24 %) bytes LDS: 0 -> 0 (0.00 %) blocks Max Waves: 26189 -> 26276 (0.33 %) Wait states: 0 -> 0 (0.00 %) Reviewed-by: Eric Anholt <[email protected]>
* gallium: add PIPE_CONTEXT_LOSE_CONTEXT_ON_RESETMarek Olšák2018-11-091-0/+3
| | | | Tested-by: Dieter Nützel <[email protected]>
* st/mesa: Don't record garbage streamout information in the non-SSO case.Kenneth Graunke2018-11-033-31/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In the non-SSO case, where multiple shader stages are linked together, we were recording garbage pipe_stream_output_info structures for all but the last enabled geometry-processing stage. Specifically, we were using the gl_transform_feedback_info from shader_program->last_vert_prog (the stage whose outputs will be recorded)...but were pairing it with the output varying mappings from the current shader stage. For example, a program with a VS and GS, the VS's pipe_shader_state would have a pipe_stream_output_info based on the GS transform feedback info, but the VS output mapping. This generally worked out okay because only the pipe_stream_output_info for the last stage really matters - the others can be ignored. However, we'd like to avoid confusing the pipe driver. In particular, my new driver translates the stream out information to hardware packets at bind_{vs,tes,gs}_state() time...and was hitting asserts about garbage varyings that didn't exist. This patch changes st/mesa to record a blank pipe_stream_output_info with num_outputs = 0 for all stages prior to last_vert_prog. The last one is captured as normal. (In the fully-SSO case, nothing should change - each program contains a single shader stage, so last_vert_prog *is* the current shader.) Tested with llvmpipe (piglit's gpu profile), and freedreno (a3xx, gpu profile with -t transform.feedback). Fixes several hundred CTS tests on my new driver. Reviewed-by: Timothy Arceri <[email protected]>
* st/nir: Drop unused parameter from st_nir_assign_uniform_locations().Kenneth Graunke2018-11-031-2/+1
| | | | | | ARB programs won't have one of these, and we don't use it anyway. Reviewed-by: Rob Clark <[email protected]>
* st/mesa: Pull nir_lower_wpos_ytransform work into a helper function.Kenneth Graunke2018-11-032-29/+40
| | | | | | This will let me use it in the ARB program code as well. Reviewed-by: Rob Clark <[email protected]>
* mesa/st: Add support for EXT_texture_sRGB_R8Gert Wollny2018-11-022-0/+12
| | | | | | | | | | | | | | | | | | | | | | This only adds support on the Gallium core level, for the drivers it is likely that additional changes are needed to support the new texture format and thereby enabling the extension. Enables on softpipe and makes pass: dEQP-GLES31.functional.srgb_texture_decode.skip_decode.sr8.* v2: - add include for getting GL_SR8_EXT v4: - since the extension is not required don't bother providing a fallback (Ilia Mirkin) - split patch (2/2) to separate Gallium and mesa/st parts (Roland Scheidegger) - trim commit message to only contain the history of the patch relevant to this part v5: - don't include GLES headers (required enum has been added to glheader.h) (Ilia Mirkin) Signed-off-by: Gert Wollny <[email protected]> Reviewed-by: Ilia Mirkin <[email protected]>
* vbo: Remove set but not used weak field from _mesa_prim.Mathias Fröhlich2018-11-011-1/+0
| | | | | | | | The only reader of the weak field in _mesa_prim is pretty console printing. By that, remove the weak field from _mesa_prim. Reviewed-by: Brian Paul <[email protected]> Signed-off-by: Mathias Fröhlich <[email protected]>
* st/glsl_to_nir: fix next_stage gatheringTimothy Arceri2018-10-311-1/+1
| | | | | | | | ffs() just returns the bit that is set, we need to know what stage that bit represents so use u_bit_scan() instead. Fixes: 2ca5d9548fc4 ("st/glsl_to_nir: gather next_stage in shader_info") Reviewed-by: Marek Olšák <[email protected]>
* st/mesa: calculate buffer size correctly for packed uniformsTimothy Arceri2018-10-311-1/+1
| | | | | | Fixes: edded1237607 ("mesa: rework ParameterList to allow packing") Reviewed-by: Marek Olšák <[email protected]>
* glsl_to_tgsi: don't create 64-bit integer MAD/FMARhys Perry2018-10-301-0/+4
| | | | | | | | TGSI has no I64MAD/U64MAD opcode. Fixes: 278580729a5 ('st/glsl_to_tgsi: add support for 64-bit integers') Signed-off-by: Rhys Perry <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* gallium: rework PIPE_HANDLE_USAGE_* flagsMarek Olšák2018-10-301-1/+1
| | | | Only radeonsi uses them, so adjust them to match its needs.
* mesa: fix struct/class mismatchEric Engestrom2018-10-304-5/+5
| | | | | Signed-off-by: Eric Engestrom <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-07 14:02:32 +0000