aboutsummaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/radeonsi
Commit message (Collapse)AuthorAgeFilesLines
* radeonsi: use enums in si_shader.hMarek Olšák2016-04-181-93/+119
| | | | | Reviewed-by: Edward O'Callaghan <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: always use PFP_SYNC_ME when doing flushes and waitsMarek Olšák2016-04-182-1/+10
| | | | | | This is typically used by the closed driver before SURFACE_SYNC. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: don't do VS/PS partial flushes if SURFACE_SYNC waits tooMarek Olšák2016-04-181-11/+14
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: add safety assertions for meta cache flushesMarek Olšák2016-04-181-0/+4
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: don't use ACQUIRE_MEM on the graphics ringMarek Olšák2016-04-181-18/+8
| | | | | | | | | It's only required on the compute ring. This matches the closed driver. The compute flag is removed to prevent confusion and Bas's compute shader patches remove it in the whole function. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: remove TODO and correct a comment in si_emit_cache_flushMarek Olšák2016-04-181-2/+1
| | | | | | Yes, that flag is really needed. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: don't flush CB/DB caches for performance countersMarek Olšák2016-04-181-3/+6
| | | | | | | | I'm not sure about this. This will make the engines go idle, but the caches will be unflushed. This should match app behavior without performance counters, which can be a good thing. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: mark ssbo and images descriptor pointers dirty at beginning of CSNicolai Hähnle2016-04-141-0/+2
| | | | | | Without this, we were getting non-deterministic VM faults under high pressure. Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: don't overwrite the scratch offset in shader prologsMarek Olšák2016-04-141-3/+4
| | | | | | Prologs only look at num_input_sgprs. Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radeonsi: fold num_user_sgprs where it is possibleMarek Olšák2016-04-141-16/+4
| | | | Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radeonsi: fix SGPRS calculation once moreMarek Olšák2016-04-142-55/+23
| | | | | | | | | | | | | This fixes GS piglit failures after adding SI_PARAM_SHADER_BUFFERS, which bumped NUM_USER_SGPRS and uncovered this bug on SI. If this was fixed in LLVM, these workarounds wouldn't be needed. LLVM would have to look at the calling convention to know how many SGPR inputs are declared, and add VCC and the scratch wave offset (which is enabled even if we spill SGPRs but not VGPRs, oh well). Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radeonsi: disable hw ETC2 on PolarisMarek Olšák2016-04-141-1/+1
| | | | | | not supported by hw directly, but it's still fully supported by the driver Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: gate PIPE_CAP_SHADER_BUFFER_OFFSET_ALIGNMENT by LLVM versionNicolai Hähnle2016-04-131-1/+2
| | | | | | | | | Otherwise we incorrectly claim ARB_ssbo support even with older LLVM versions. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94917 Reviewed-by: Michel Dänzer <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* gallium/radeon: add an env variable to force a level of aniso filteringMarek Olšák2016-04-131-3/+7
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: enable GLSL 4.20 and therefore OpenGL 4.2Nicolai Hähnle2016-04-121-1/+2
| | | | | | | This is the last necessary bit for OpenGL 4.2 support. All driver-specific functionality has already been implemented as part of extensions. Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: fix a critical SI hang since PIPELINESTAT_START/STOP was addedMarek Olšák2016-04-132-9/+2
| | | | | For some reason unknown to me, SI hangs if the event is written after CONTEXT_CONTROL.
* radeonsi: fix bounds check in si_create_vertex_elementsNicolai Hähnle2016-04-121-1/+1
| | | | | | | | | This was triggered by dEQP-GLES3.functional.vertex_array_objects.all_attributes Cc: "11.1 11.2" <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: enable shader buffer pipe capsNicolai Hähnle2016-04-121-2/+2
| | | | | Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: add shader buffer support to TGSI_OPCODE_RESQNicolai Hähnle2016-04-121-6/+11
| | | | | Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: add shader buffer support to TGSI_OPCODE_STORENicolai Hähnle2016-04-121-18/+111
| | | | | Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: add shader buffer support to TGSI_OPCODE_LOADNicolai Hähnle2016-04-121-19/+70
| | | | | Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: add shader buffer support to TGSI_OPCODE_ATOM*Nicolai Hähnle2016-04-121-15/+46
| | | | | Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: add offset parameter to buffer_append_argsNicolai Hähnle2016-04-121-5/+10
| | | | | Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: adjust buffer_append_args to take a 128 bit resourceNicolai Hähnle2016-04-121-7/+23
| | | | | | | Move the buffer resource extraction code out into its own function. Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: preload shader buffers in shadersNicolai Hähnle2016-04-121-0/+17
| | | | | Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: implement set_shader_buffersNicolai Hähnle2016-04-125-58/+125
| | | | | Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: move resetting of constant buffers into a separate functionNicolai Hähnle2016-04-121-15/+26
| | | | | | | This will be re-used for shader buffers. Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi: Mark ARB_robust_buffer_access_behavior as supported.Bas Nieuwenhuizen2016-04-121-1/+1
| | | | | Signed-off-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* gallium: Add capability for ARB_robust_buffer_access_behavior.Bas Nieuwenhuizen2016-04-121-0/+1
| | | | | | Signed-off-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Marek Olšák <[email protected]> Reviewed-by: Roland Scheidegger <[email protected]>
* radeonsi: move scissor and viewport states into gallium/radeonMarek Olšák2016-04-126-317/+14
| | | | | | Reviewed-by: Edward O'Callaghan <[email protected]> Reviewed-by: Grigori Goronzy <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: use guard band clippingMarek Olšák2016-04-121-4/+69
| | | | | | | | | | | | Guard band clipping speeds up rasterization for primitives that are partially off-screen. This change in particular results in small framerate improvements in a wide range of games. Started by Grigori Goronzy <[email protected]>. Reviewed-by: Edward O'Callaghan <[email protected]> Reviewed-by: Grigori Goronzy <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: compute scissor from viewport in set_viewport_statesMarek Olšák2016-04-122-30/+48
| | | | | | | | | and clamp it right before emitting. This is a prerequisite for computing the guard band. Reviewed-by: Edward O'Callaghan <[email protected]> Reviewed-by: Grigori Goronzy <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: disable aniso filtering for non-mipmap textures on SI-CIMarek Olšák2016-04-122-1/+45
| | | | | | | | | The closed driver does this, but it looks at base_level and last_level and uses a conditional assignment, which LLVM can't generate on SGPRs. That led me to invent this solution that abuses the image descriptor. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: clean up aniso state translationMarek Olšák2016-04-122-15/+8
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: enable some sampler fields to match the closed driverMarek Olšák2016-04-122-2/+10
| | | | | | copied from the Vulkan driver Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: allow clearing buffers >= 4 GBMarek Olšák2016-04-121-3/+3
| | | | | | | Only CMASK and DCC clears can use this, because only textures can be so large. Reviewed-by: Nicolai Hähnle <[email protected]>
* gallium/radeon: move pipeline stat context flags to common codeMarek Olšák2016-04-123-9/+6
| | | | | Reviewed-by: Edward O'Callaghan <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* gallium/radeon: unify checking streamout enable stateMarek Olšák2016-04-121-2/+1
| | | | | Reviewed-by: Edward O'Callaghan <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: fix mask checking when emitting scissors and viewportsMarek Olšák2016-04-121-4/+8
| | | | | Reviewed-by: Nicolai Hähnle <[email protected]> Reviewed-by: Grigori Goronzy <[email protected]>
* radeonsi: implement and rely on set_active_query_stateMarek Olšák2016-04-124-4/+45
| | | | | Reviewed-by: Edward O'Callaghan <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: Synchronize a streamout write after read hazard.Bas Nieuwenhuizen2016-04-121-0/+6
| | | | | | Signed-off-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: do per-pixel clipping based on viewport statesMarek Olšák2016-04-082-11/+85
| | | | | | | | | In other words, vport scissors are derived from viewport states. If the scissor test is enabled, the intersection of both is used. The guard band will disable clipping, so we have to clip per-pixel. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: raise number of samplers per shader to 32Nicolai Hähnle2016-04-071-3/+3
| | | | | Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94835 Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: expand the compressed color and depth texture masks to 64 bitsNicolai Hähnle2016-04-073-18/+18
| | | | | | | | | | | This is in preparation of raising the number of exposed sampler views to 32 bits, which will raise the total number of sampler views to 33 for the polygon stipple texture. That texture should never be compressed (and it's certainly not a depth texture), but this approach seems cleaner to me than special-casing the last slot in all affected code paths. Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: replace magic 16 by SI_NUM_USER_SAMPLERSNicolai Hähnle2016-04-071-1/+1
| | | | | Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: disable perfect ZPASS counts for PIPE_QUERY_OCCLUSION_PREDICATEMarek Olšák2016-04-071-2/+4
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: don't use the real barrier instruction in tess ctrl shadersMarek Olšák2016-04-071-0/+8
| | | | | Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: Enable ARB_framebuffer_no_attachmentsEdward O'Callaghan2016-04-071-1/+1
| | | | | Signed-off-by: Edward O'Callaghan <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: Improve assert info out of si_set_framebuffer_state()Edward O'Callaghan2016-04-071-0/+2
| | | | | | | | Lets give the developer a little hand if we are going to assert on a zero literal at the end of a branch. Signed-off-by: Edward O'Callaghan <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: Allow 16 samples MSAA mode for PIPE_FORMAT_NONEEdward O'Callaghan2016-04-071-0/+5
| | | | | | | | | For ARB_framebuffer_no_attachment; A is_format_supported() query with 'PIPE_FORMAT_NONE' passed implies a query of the number of samples supported from the framebuffer with no attachment. Signed-off-by: Edward O'Callaghan <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-12 02:04:41 +0000