summaryrefslogtreecommitdiffstats
path: root/src/mesa/drivers/dri/i965
Commit message (Collapse)AuthorAgeFilesLines
* i965: Relax restriction on scheduling last instruction.Matt Turner2016-03-301-20/+3
| | | | | | | | | | | | | | | | | | | | | | I think when this code was written, basic blocks were always ended by a control flow instruction or an end-of-thread message. That's no longer the case, and removing this restriction actually helps things: instructions in affected programs: 7267 -> 7244 (-0.32%) helped: 4 total cycles in shared programs: 66559580 -> 66431900 (-0.19%) cycles in affected programs: 28310152 -> 28182472 (-0.45%) helped: 9577 HURT: 879 GAINED: 2 The addition of the is_control_flow() checks is not a functional change, since the add_insts_from_block() does not put them in the list of instructions to schedule. I plan to change this in a later patch. Reviewed-by: Francisco Jerez <[email protected]>
* i965/vec4/tcs: Set conditional mod on TCS_OPCODE_SRC0_010_IS_ZERO.Matt Turner2016-03-302-2/+3
| | | | | | | | | | | | | | | | | | Missing this causes an assertion failure in the scheduler with the next patch. Additionally, this gives cmod propagation enough information to optimize code better. total instructions in shared programs: 7112991 -> 7112852 (-0.00%) instructions in affected programs: 25704 -> 25565 (-0.54%) helped: 139 total cycles in shared programs: 64812898 -> 64810674 (-0.00%) cycles in affected programs: 127224 -> 125000 (-1.75%) helped: 139 Acked-by: Francisco Jerez <[email protected]>
* Revert "i965: Don't add barrier deps for FB write messages."Matt Turner2016-03-301-4/+3
| | | | | | | | | | | | | | This reverts commit d0e1d6b7e27bf5f05436e47080d326d7daa63af2. The change in the vec4 code is a mistake -- there's never an FS_OPCODE_FB_WRITE in vec4 code. The change in the fs code had the (harmless) effect of not recognizing an FB_WRITE as a scheduling barrier even if it was marked EOT -- harmless because the scheduler marked the last instruction of a block as a barrier, something I'm changing in the following patches. This will be reimplemented later in the series.
* i965: Simplify full scheduling-barrier conditions.Matt Turner2016-03-301-27/+8
| | | | | | | All of these were simply code for "architecture register file" (and in the case of destinations, "not the null register"). Reviewed-by: Francisco Jerez <[email protected]>
* i965: Remove incorrect cycle estimates.Matt Turner2016-03-301-10/+0
| | | | | | | | These printed the cycle count the last basic block (sched.time is set per basic block!). We have accurate, full program, data printed elsewhere. Reviewed-by: Francisco Jerez <[email protected]>
* glsl: add transform feedback buffers to resource listTimothy Arceri2016-03-311-1/+1
| | | | Reviewed-by: Dave Airlie <[email protected]>
* mesa: split transform feedback buffer into its own structTimothy Arceri2016-03-313-7/+7
| | | | | | | This will be used in a following patch to implement interface query support for TRANSFORM_FEEDBACK_BUFFER. Reviewed-by: Dave Airlie <[email protected]>
* glsl: use bitmask of active xfb buffer indicesTimothy Arceri2016-03-311-1/+1
| | | | | | | | | | | | | This allows us to print the correct binding point when not all buffers declared in the shader are bound. For example if we use a single buffer: layout(xfb_buffer=2, offset=0) out vec4 v; We now print '2' when the buffer is not bound rather than '0'. Reviewed-by: Dave Airlie <[email protected]>
* i965: Don't inline intel_batchbuffer_require_space().Matt Turner2016-03-302-26/+28
| | | | | | | | | | | | It's called by the inline intel_batchbuffer_begin() function which itself is used in BEGIN_BATCH. So in sequence of code emitting multiple packets, we have inlined this ~200 byte function multiple times. Making it an out-of-line function presumably improved icache usage. Improves performance of Gl32Batch7 by 3.39898% +/- 0.358674% (n=155) on Ivybridge. Reviewed-by: Abdiel Janulgue <[email protected]>
* i965: Don't use CUBE wrap modes for integer formats on IVB/BYT.Kenneth Graunke2016-03-291-1/+5
| | | | | | | | | | | | | | | | There is no linear filtering for integer formats, so we should always be using CLAMP_TO_EDGE mode. Fixes 46 dEQP cases on Ivybridge (which were likely broken by commit 0faf26e6a0a34c3544644852802484f2404cc83e). This workaround doesn't appear to be necessary on any other hardware; I haven't found any documentation mentioning errata in this area. v2: Only apply on Ivybridge/Baytrail to avoid regressing GLES3.1 tests. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> [v1]
* Revert "i965: Set address rounding bits for GL_NEAREST filtering as well."Kenneth Graunke2016-03-291-6/+3
| | | | | | | This reverts commit 60d6a8989ab44cf47accee6bc692ba6fb98f6a9f. It's pretty sketchy, and apparently regressed a bunch of dEQP tests on Sandybridge.
* i965: Set address rounding bits for GL_NEAREST filtering as well.Kenneth Graunke2016-03-281-3/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Yuanhan Liu decided these were useful for linear filtering in commit 76669381 (circa 2011). Prior to that, we never set them; it seems he tried to preserve that behavior for nearest filtering. It turns out they're useful for nearest filtering, too: setting these fixes the following dEQP-GLES3 tests: functional.fbo.blit.rect.nearest_consistency_mag functional.fbo.blit.rect.nearest_consistency_mag_reverse_src_x functional.fbo.blit.rect.nearest_consistency_mag_reverse_src_y functional.fbo.blit.rect.nearest_consistency_mag_reverse_dst_x functional.fbo.blit.rect.nearest_consistency_mag_reverse_dst_y functional.fbo.blit.rect.nearest_consistency_mag_reverse_src_dst_x functional.fbo.blit.rect.nearest_consistency_mag_reverse_src_dst_y functional.fbo.blit.rect.nearest_consistency_min functional.fbo.blit.rect.nearest_consistency_min_reverse_src_x functional.fbo.blit.rect.nearest_consistency_min_reverse_src_y functional.fbo.blit.rect.nearest_consistency_min_reverse_dst_x functional.fbo.blit.rect.nearest_consistency_min_reverse_dst_y functional.fbo.blit.rect.nearest_consistency_min_reverse_src_dst_x functional.fbo.blit.rect.nearest_consistency_min_reverse_src_dst_y functional.fbo.blit.rect.nearest_consistency_out_of_bounds_mag functional.fbo.blit.rect.nearest_consistency_out_of_bounds_mag_reverse_src_x functional.fbo.blit.rect.nearest_consistency_out_of_bounds_mag_reverse_src_y functional.fbo.blit.rect.nearest_consistency_out_of_bounds_mag_reverse_dst_x functional.fbo.blit.rect.nearest_consistency_out_of_bounds_mag_reverse_dst_y functional.fbo.blit.rect.nearest_consistency_out_of_bounds_mag_reverse_src_dst_x functional.fbo.blit.rect.nearest_consistency_out_of_bounds_mag_reverse_src_dst_y functional.fbo.blit.rect.nearest_consistency_out_of_bounds_min functional.fbo.blit.rect.nearest_consistency_out_of_bounds_min_reverse_src_x functional.fbo.blit.rect.nearest_consistency_out_of_bounds_min_reverse_src_y functional.fbo.blit.rect.nearest_consistency_out_of_bounds_min_reverse_dst_x functional.fbo.blit.rect.nearest_consistency_out_of_bounds_min_reverse_dst_y functional.fbo.blit.rect.nearest_consistency_out_of_bounds_min_reverse_src_dst_x functional.fbo.blit.rect.nearest_consistency_out_of_bounds_min_reverse_src_dst_y Apparently, BLORP has always set these bits unconditionally. However, setting them unconditionally appears to regress tests using texture projection, 3D samplers, integer formats, and vertex shaders, all in combination, such as: functional.shaders.texture_functions.textureprojlod.isampler3d_vertex Setting them on Gen4-5 appears to regress Piglit's tests/spec/arb_sampler_objects/framebufferblit. Honestly, it looks like the real problem here is a lack of precision. I'm just hacking around problems here (as embarassing as it is). Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jordan Justen <[email protected]>
* i965: Always use BRW_TEXCOORDMODE_CUBE when seamless filtering.Kenneth Graunke2016-03-281-3/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | When using seamless cube map mode and NEAREST filtering, we explicitly overrode the wrap modes to CLAMP_TO_EDGE. This was to implement the following spec text: "If NEAREST filtering is done within a miplevel, always apply apply wrap mode CLAMP_TO_EDGE." However, textureGather() ignores the sampler's filtering mode, and instead returns the four pixels that would be blended by LINEAR filtering. This implies that we should do proper seamless filtering, and include pixels from adjacent cube faces. It turns out that we can simply delete the NEAREST -> CLAMP_TO_EDGE overrides. Normal cube map sampling works by first selecting the face, and then nearest filtering fetches the closest texel. If the nearest texel was on a different face, then that face would have been chosen. So it should always be within the face anyway, which effectively performs CLAMP_TO_EDGE. Fixes 86 dEQP-GLES31.texture.gather.basic.cube.* tests. Signed-off-by: Kenneth Graunke <[email protected]> Suggested-by: Ian Romanick <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* i965: Fix brw_render_cache_set_check_flush's PIPE_CONTROLs.Kenneth Graunke2016-03-282-3/+22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Our driver uses the brw_render_cache mechanism to track buffers we've rendered to and are about to sample from. Previously, we did a single PIPE_CONTROL with the following bits set: - Render Target Flush - Depth Cache Flush - Texture Cache Invalidate - VF Cache Invalidate - Instruction Cache Invalidate - CS Stall This combined both "top of pipe" invalidations and "bottom of pipe" flushes, which isn't how the hardware is intended to be programmed. The "top of pipe" invalidations may happen right away, without any guarantees that rendering using those caches has completed. That rendering may continue altering the caches. The "bottom of pipe" flushes do wait for the rendering to complete. The CS stall also prevents further work from happening until data is flushed out. What we wanted to do was wait for rendering complete, flush the new data out of the render and depth caches, wait, then invalidate any stale data in read-only caches. We can accomplish this by doing the "bottom of pipe" flushes with a CS stall, then the "top of pipe" flushes as a second PIPE_CONTROL. The flushes will wait until the rendering is complete, and the CS stall will prevent the second PIPE_CONTROL with the invalidations from executing until the first is done. Fixes dEQP-GLES3.functional.texture.specification.teximage2d_pbo subtests on Braswell and Skylake. These tests hit the meta PBO texture upload path, which binds the PBO as a texture and samples from it, while rendering to the destination texture. The tests then sample from the texture. For now, we leave Gen4-5 alone. It probably needs work too, but apparently it hasn't even been setting the (G45+) TC invalidation bit at all... v2: Add Sandybridge post-sync non-zero workaround, for safety. Cc: [email protected] Suggested-by: Francisco Jerez <[email protected]> Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Francisco Jerez <[email protected]>
* i965: Whack UAV bit when FS discards and there are no color writes.Kenneth Graunke2016-03-281-2/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | dEQP-GLES31.functional.fbo.no_attachments.* draws a quad with no framebuffer attachments, using a shader that discards based on gl_FragCoord. It uses occlusion queries to inspect whether pixels are rendered or not. Unfortunately, the hardware is not dispatching any pixel shaders, so discards never happen, and the full quad of pixels increments PS_DEPTH_COUNT, making the occlusion query results bogus. To understand why, we have to delve into the WM_INT internal signalling mechanism's formulas. The "WM_INT::Pixel Shader Kill Pixel" signal is defined as: 3DSTATE_WM::ForceKillPixel == ON || (3DSTATE_WM::ForceKillPixel != Off && !WM_INT::WM_HZ_OP && 3DSTATE_WM::EDSC_Mode != PREPS && (WM_INT::Depth Write Enable || WM_INT::Stencil Write Enable) && ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ (3DSTATE_PS_EXTRA::PixelShaderKillsPixels || 3DSTATE_PS_EXTRA:: oMask Present to RenderTarget || 3DSTATE_PS_BLEND::AlphaToCoverageEnable || 3DSTATE_PS_BLEND::AlphaTestEnable || 3DSTATE_WM_CHROMAKEY::ChromaKeyKillEnable)) Because there is no depth or stencil buffer, writes to those buffers are disabled. So the highlighted condition is false, making the whole "Kill Pixel" condition false. This then feeds into the following "WM_INT::ThreadDispatchEnable" condition: 3DSTATE_WM::ForceThreadDispatch != OFF && !WM_INT::WM_HZ_OP && 3DSTATE_PS_EXTRA::PixelShaderValid && (3DSTATE_PS_EXTRA::PixelShaderHasUAV || WM_INT::Pixel Shader Kill Pixel || WM_INT::RTIndependentRasterizationEnable || (!3DSTATE_PS_EXTRA::PixelShaderDoesNotWriteRT && 3DSTATE_PS_BLEND::HasWriteableRT) || (WM_INT::Pixel Shader Computed Depth Mode != PSCDEPTH_OFF && (WM_INT::Depth Test Enable || WM_INT::Depth Write Enable)) || (3DSTATE_PS_EXTRA::Computed Stencil && WM_INT::Stencil Test Enable) || (3DSTATE_WM::EDSC_Mode == 1 && (WM_INT::Depth Test Enable || WM_INT::Depth Write Enable || WM_INT::Stencil Test Enable))) Given that there's no depth/stencil testing, no writeable render target, and the hardware thinks kill pixel doesn't happen, all of these conditions are false. We have to whack some bit to make PS invocations happen. There are many options. Curro suggested using the UAV bit. There's some precedence in doing that - we set it for fragment shaders that do SSBO/image/atomic writes when no color buffer writes are enabled. We can simply include discard here too. Fixes 64 dEQP-GLES31.functional.fbo.no_attachments.* tests. v2: Add a comment suggested and written by Jason Ekstrand. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Francisco Jerez <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* mesa: replace gl_context->Multisample._Enabled with ↵Bas Nieuwenhuizen2016-03-246-12/+14
| | | | | | | | | | | | _mesa_is_multisample_enabled. This removes any dependency on driver validation of the number of framebuffer samples. Signed-off-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]> Reviewed-by: Marek Olšák <[email protected]> Tested-by: Brian Paul <[email protected]>
* i965/peephole_ffma: Don't fuse exact addsJason Ekstrand2016-03-231-1/+3
| | | | Reviewed-by: Francisco Jerez <[email protected]>
* i965/fs: Don't constant-fold RCPJason Ekstrand2016-03-221-15/+0
| | | | | | No shader-db changes on Broadwell Reviewed-by: Matt Turner <[email protected]>
* i965: Remove the RCP+RSQ algebraic optimizationsJason Ekstrand2016-03-222-22/+0
| | | | | | | | | NIR already has this optimization and it can do much better than the little peephole in the backend. No shader-db change on Haswell or Broadwell. Reviewed-by: Matt Turner <[email protected]>
* i965: Have NIR lower flrp on pre-GEN6 vec4 backendIan Romanick2016-03-221-2/+26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously we were doing the lowering by hand in vec4_visitor::emit_lrp. By doing it in NIR, we have the opportunity for NIR to do additional optimization of the expanded code. This also enables optimizations added by the next commit. shader-db results: G4X / Ironlake total instructions in shared programs: 4024401 -> 4016538 (-0.20%) instructions in affected programs: 447686 -> 439823 (-1.76%) helped: 2623 HURT: 0 total cycles in shared programs: 84375846 -> 84328296 (-0.06%) cycles in affected programs: 16964960 -> 16917410 (-0.28%) helped: 2556 HURT: 41 Unsurprisingly, no changes on later platforms. v2: Formatting and comment changes suggested by Matt. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* i965: fix invalid memory writeMarc-André Lureau2016-03-211-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | I noticed some heap corruption running virgl tests, and valgrind helped me to track it down to the following error: ==29272== Invalid write of size 4 ==29272== at 0x90283D4: push_loop_stack (brw_eu_emit.c:1307) ==29272== by 0x9029A7D: brw_DO (brw_eu_emit.c:1750) ==29272== by 0x90554B0: fs_generator::generate_code(cfg_t const*, int) (brw_fs_generator.cpp:1999) ==29272== by 0x904491F: brw_compile_fs (brw_fs.cpp:5685) ==29272== by 0x8FC5DC5: brw_codegen_wm_prog (brw_wm.c:137) ==29272== by 0x8FC7663: brw_fs_precompile (brw_wm.c:638) ==29272== by 0x8FA4040: brw_shader_precompile(gl_context*, gl_shader_program*) (brw_link.cpp:51) ==29272== by 0x8FA4A9A: brw_link_shader (brw_link.cpp:260) ==29272== by 0x8DEF751: _mesa_glsl_link_shader (ir_to_mesa.cpp:3006) ==29272== by 0x8C84325: _mesa_link_program (shaderapi.c:1042) ==29272== by 0x8C851D7: _mesa_LinkProgram (shaderapi.c:1515) ==29272== by 0x4E4B8E8: add_shader_program (vrend_renderer.c:880) ==29272== Address 0xf2f3cb0 is 0 bytes after a block of size 112 alloc'd ==29272== at 0x4C2AA98: calloc (vg_replace_malloc.c:711) ==29272== by 0x8ED11F7: ralloc_size (ralloc.c:113) ==29272== by 0x8ED1282: rzalloc_size (ralloc.c:134) ==29272== by 0x8ED14C0: rzalloc_array_size (ralloc.c:196) ==29272== by 0x9019C7B: brw_init_codegen (brw_eu.c:291) ==29272== by 0x904F565: fs_generator::fs_generator(brw_compiler const*, void*, void*, void const*, brw_stage_prog_data*, unsigned int, bool, gl_shader_stage) (brw_fs_generator.cpp:124) ==29272== by 0x9044883: brw_compile_fs (brw_fs.cpp:5675) ==29272== by 0x8FC5DC5: brw_codegen_wm_prog (brw_wm.c:137) ==29272== by 0x8FC7663: brw_fs_precompile (brw_wm.c:638) ==29272== by 0x8FA4040: brw_shader_precompile(gl_context*, gl_shader_program*) (brw_link.cpp:51) ==29272== by 0x8FA4A9A: brw_link_shader (brw_link.cpp:260) ==29272== by 0x8DEF751: _mesa_glsl_link_shader (ir_to_mesa.cpp:3006) if_depth_in_loop is an array of size p->loop_stack_array_size, and push_loop_stack() will access if_depth_in_loop[p->loop_stack_depth+1], thus the condition to grow the array should be p->loop_stack_array_size <= (p->loop_stack_depth + 1) (it's currently off by 2...) This can be reproduced by running the following test with virgl test server: LIBGL_ALWAYS_SOFTWARE=y GALLIUM_DRIVER=virpipe bin/shader_runner ./tests/shaders/glsl-fs-unroll-explosion.shader_test -auto Signed-off-by: Marc-André Lureau <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* i965: Fix assert conditions for src/dst x/y offsetsAnuj Phogat2016-03-211-3/+3
| | | | | | Cc: [email protected] Signed-off-by: Anuj Phogat <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* i965/blorp: Make BlitFramebuffer() do sRGB encoding in ES 3.x.Kenneth Graunke2016-03-211-1/+4
| | | | | | | | | | | | | | | | | | According to the ES 3.0 and GL 4.4 specifications, glBlitFramebuffer is supposed to perform sRGB decoding and encoding whenever sRGB formats are in use. The ES 3.0 specification is completely clear, and has always stated this. However, the GL specification has changed behavior in 4.1, 4.2, and 4.4. The original behavior stated that no sRGB encoding should occur. The 4.4 behavior matches ES 3.0's wording. However, implementing the new behavior appears to break applications such as Left 4 Dead 2. This patch changes Meta to apply the ES 3.x rules in ES 3.x, but leaves OpenGL alone for now, to avoid breaking applications. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Topi Pohjolainen <[email protected]>
* i965/blorp: Refactor sRGB encoding/decoding.Kenneth Graunke2016-03-214-11/+23
| | | | | | | | | | | | Because the rules for sRGB are so insane, we change brw_blorp_miptrees to take decode_srgb and encode_srgb flags, which control linearization of the source and destination separately. This should make it easy to implement whatever crazy combination of rules people throw at us. For now, it should be equivalent. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Topi Pohjolainen <[email protected]>
* i965: Stop XY clipping point and line primitives.Kenneth Graunke2016-03-181-1/+7
| | | | | | | | | | | | | | | | | | | | | | | Wide points and lines are not supposed to be clipped by the viewport. Rather, they should be rendered, and any fragments outside of the viewport should be discarded. The traditional use case for this behavior is rendering moving wide point particles. When the center of the point approaches the viewport edge, clipping would make it pop out of view early. Fixes: - dEQP-GLES2.functional.clipping.point.wide_point_clip - dEQP-GLES3.functional.clipping.point.wide_point_clip - dEQP-GLES3.functional.clipping.point.wide_point_clip_viewport_center - dEQP-GLES3.functional.clipping.point.wide_point_clip_viewport_corner - dEQP-GLES3.functional.clipping.line.wide_line_clip_viewport_center - dEQP-GLES3.functional.clipping.line.wide_line_clip_viewport_corner Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94453 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94454 Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jordan Justen <[email protected]>
* i965: Scissor to the viewport when rendering points/lines.Kenneth Graunke2016-03-182-5/+8
| | | | | | | | | | | | | | We're about to start allowing wide points/lines whose vertices are outside the viewport past the clipper. This scissoring hack ensures that any fragments generated are still restricted to the viewport. It is not necessary on Gen8+ as those platforms already discard fragments which are outside the viewport. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94453 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94454 Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jordan Justen <[email protected]>
* i965: Include the viewport in the scissor rectangle.Kenneth Graunke2016-03-181-4/+4
| | | | | | | | | | We'll need to use scissoring to restrict fragments to the viewport soon. It seems harmless to include it generally, so let's do that. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94453 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94454 Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jordan Justen <[email protected]>
* i965: Introduce an is_drawing_lines() helper.Kenneth Graunke2016-03-181-0/+30
| | | | | | | | | Similar to is_drawing_points(). v2: Account for isoline tessellation output topology. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jordan Justen <[email protected]>
* i965: Move is_drawing_points to brw_state.h.Kenneth Graunke2016-03-182-24/+24
| | | | | | | | | I need to use this in multiple source files. v2: Rebase on TES output domain fix. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jordan Justen <[email protected]>
* i965: Fix gl_TessLevelOuter[] for isolines.Kenneth Graunke2016-03-182-6/+22
| | | | | | | | | | | | | | | | | | | | | | | Thanks to James Legg for finding this! From the ARB_tessellation_shader spec: "The number of isolines generated is derived from the first outer tessellation level; the number of segments in each isoline is derived from the second outer tessellation level." According to the PRM, "TF.LineDensity determines # lines" while "TF.LineDetail determines # segments". Line Density is stored at DWord 6, while Line Detail is at DWord 7. So, they're not reversed like they are for triangles and quads. Fixes Piglit's spec/arb_tessellation_shader/execution/isoline, and about 24 dEQP isoline tests (with GL_EXT_tessellation_shader hacked on - it's not normally enabled). Cc: [email protected] Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94524 Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jordan Justen <[email protected]>
* i965: Decode non-normalized coordinates bit in SAMPLER_STATE.Kenneth Graunke2016-03-181-2/+3
| | | | | | | We weren't printing this for some reason. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Eduardo Lima Mitev <[email protected]>
* i965: Account for TES in is_drawing_points().Kenneth Graunke2016-03-181-1/+8
| | | | | | | | | | | | | Now that we implement tessellation shaders, the TES might be the last stage enabled. If it's outputting points, then the primitive type reaching the SF is points. We need to account for this. Caught by Ilia Mirkin. v2: Update dirty bit comment above caller (caught by Iago) Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Iago Toral Quiroga <[email protected]>
* nir: add a bit_size parameter to nir_ssa_dest_initConnor Abbott2016-03-171-2/+5
| | | | | | | | | | | | | | | | | | | | | | v2: Squash multiple commits addressing the new parameter in different files so we don't break the build (Iago) v3: Fix tgsi (Samuel) v4: Fix nir_clone.c (Samuel) v5: Fix vc4 and freedreno (Iago) v6 (Sam) - Fix build errors in nir_lower_indirect_derefs - Use helper to get type size from nir_alu_type. Signed-off-by: Iago Toral Quiroga <[email protected]> Signed-off-by: Samuel Iglesias Gonsalvez <[email protected]> Tested-by: Rob Clark <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]> Reviewed-by: Iago Toral Quiroga <[email protected]>
* nir: rename nir_const_value fields to include bitsize informationIago Toral Quiroga2016-03-176-63/+63
| | | | | Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]>
* nir: update opcode definitions for different bit sizesConnor Abbott2016-03-171-0/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Some opcodes need explicit bitsizes, and sometimes we need to use the double version when constant folding. v2: fix output type for u2f (Iago) v3: do not change vecN opcodes to be float. The next commit will add infrastructure to enable 64-bit integer constant folding so this is isn't really necessary. Also, that created problems with source modifiers in some cases (Iago) v4 (Jason): - do not change bcsel to work in terms of floats - leave ldexp generic Squashed changes to handle different bit sizes when constant folding since otherwise we would break the build. v2: - Use the bit-size information from the opcode information if defined (Iago) - Use helpers to get type size and base type of nir_alu_type enum (Sam) - Do not fallback to sized types to guess bit-size information. (Jason) Squashed changes in i965 and gallium/nir drivers to support sized types. These functions should only see sized types, but we can't make that change until we make sure that nir uses the sized versions in all the relevant places. A later commit will address this. Signed-off-by: Iago Toral Quiroga <[email protected]> Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]> Reviewed-by: Iago Toral Quiroga <[email protected]>
* i965/nir: fix check to resolve booleans to work with sized nir_alu_typeSamuel Iglesias Gonsálvez2016-03-171-1/+1
| | | | | | | | | | As nir_alu_type has now embedded the data size, the check for the instruction's output type (to see if a boolean resolve is required) should ignore the data size part. Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Iago Toral Quiroga <[email protected]>
* i965/nir: Lower nir compute shader shared variablesJordan Justen2016-03-173-0/+11
| | | | | Signed-off-by: Jordan Justen <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* i965: Skip execution size adjustment for instructions of width 4Iago Toral Quiroga2016-03-171-1/+13
| | | | | | | | | | | | | | | | | | | | This code in brw_set_dest adjusts the execution size of any instruction with a dst.width < 8. However, we don't want to do this with instructions operating on doubles, since these will have a width of 4, but still need an execution size of 8 (for SIMD8). Unfortunately, we can't just check the size of the operands involved to detect if we are doing an operation on doubles, because we can have instructions that do operations on double operands interpreted as UD, operating on any of its 2 32-bit components. Previous commits have made it so we never emit instructions with a horizontal width of 4 that don't have the correct execution size set for gen6+, so we can skip it in this case, avoiding the conflicts with fp64 requirements. Expanding the same fix to other hardware generations requires many more changes but since we are not targetting fp64 support on them wer don't really care for now. Reviewed-by: Topi Pohjolainen <[email protected]>
* i965/vec4/gen6: fix exec_size for MOV with a width of 4 in generate_gs_ff_sync()Samuel Iglesias Gonsalvez2016-03-171-1/+3
| | | | | Signed-off-by: Samuel Iglesias Gonsalvez <[email protected]> Reviewed-by: Topi Pohjolainen <[email protected]>
* i965/vec4/gen6: fix exec_size for instructions with destination width of 4Samuel Iglesias Gonsalvez2016-03-171-0/+6
| | | | | Signed-off-by: Samuel Iglesias Gonsalvez <[email protected]> Reviewed-by: Topi Pohjolainen <[email protected]>
* i965/vec4/gen6: fix exec_size for instructions with width of 4 in ↵Samuel Iglesias Gonsalvez2016-03-171-0/+3
| | | | | | | generate_gs_svb_write() Signed-off-by: Samuel Iglesias Gonsalvez <[email protected]> Reviewed-by: Topi Pohjolainen <[email protected]>
* i965/gs/gen6: fix execsize for instructions with width of 4 in ↵Samuel Iglesias Gonsalvez2016-03-171-1/+10
| | | | | | | | | | gen6_sol_program() v2: - Add assert (Topi). Signed-off-by: Samuel Iglesias Gonsalvez <[email protected]> Reviewed-by: Topi Pohjolainen <[email protected]>
* i965: set correct execsize for MOVS with a width of 4 in brw_find_live_channelIago Toral Quiroga2016-03-171-0/+3
| | | | Reviewed-by: Topi Pohjolainen <[email protected]>
* i965/eu: set execution size for SEND message in brw_send_indirect_messageIago Toral Quiroga2016-03-171-0/+3
| | | | Reviewed-by: Topi Pohjolainen <[email protected]>
* i965/fs: Set exec size for gen7 pull const loadsIago Toral Quiroga2016-03-171-0/+1
| | | | | | | | | | v2 (Topi): - No need to set the execsize for the indirect send message, the next patch will handle that. - Set the execution size explicitly instead of taking it from the width of the dst that we set before. Reviewed-by: Topi Pohjolainen <[email protected]>
* i965/eu: set correct execution size in brw_NOPIago Toral Quiroga2016-03-171-2/+3
| | | | | | | v2: NOP should have an execsize of 1 (Matt) Reviewed-by: Matt Turner <[email protected]> Reviewed-by: Topi Pohjolainen <[email protected]>
* meta: Don't use integer handles for shaders or programs.Kenneth Graunke2016-03-163-33/+39
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, we gave our internal clear/blit shaders actual GL handles and stored them in the shader/program hash table. We used ordinary GL API entrypoints to work with them. We thought this shouldn't be a problem because GL doesn't allow applications to invent their own names for shaders or programs. GL allocates all names via glCreateShader and glCreateProgram. However, having them in the hash table is a bit risky: if a broken application guesses the name of our shaders or programs, it could alter them, potentially screwing up future meta operations. Also, test cases can observe the programs in the hash table. Running a single dEQP process that executes the following test list: dEQP-GLES3.functional.negative_api.buffer.clear dEQP-GLES3.functional.negative_api.shader.compile_shader dEQP-GLES3.functional.negative_api.shader.delete_shader would result in the last two tests breaking. The compile_shader test calls glCompileShader(9) straight away, and since it hasn't even created any shaders or programs, it expects to get a GL_INVALID_VALUE error because there's no such name. However, because the clear test ran first, it created Meta programs, so an object named "9" did exist. This patch reworks Meta to work with gl_shader and gl_shader_program pointers directly. These internal programs have bogus names, and are never stored in the hash tables, so they're invisible to applications. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94485 Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Topi Pohjolainen <[email protected]>
* meta: Use the _mesa_meta_compile_and_link_program helper more places.Kenneth Graunke2016-03-161-11/+3
| | | | | | | Less boilerplate. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Topi Pohjolainen <[email protected]>
* meta: Use ARB_explicit_attrib_location in the rest of the meta shaders.Kenneth Graunke2016-03-161-2/+2
| | | | | | | | | | | | | | | | | This is cleaner than using glBindAttribLocation(). Not all drivers support the extension, but I don't think those drivers use GLSL in the first place. Apparently some Meta shaders already use GL_ARB_explicit_attrib_location, so I think it should be okay. Honestly, I'm not sure how the old code worked anyway - we bound the attribute location for "texcoords", while all the shaders capitalized or spelled it differently. v2: Convert another instance in brw_meta_fast_clear.c. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* i965/fs: Restrict inequality that can only hold equal in saturate propagation.Francisco Jerez2016-03-141-1/+1
| | | | | | | | | | Should have no functional change. The IP value of an instruction that reads src_var cannot possibly be after the end of the live interval of the variable it's reading from, by the definition of live interval. Might save future readers a momentary WTF while trying to understand this code. Reviewed-by: Matt Turner <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-22 20:10:22 +0000