aboutsummaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/panfrost
Commit message (Collapse)AuthorAgeFilesLines
* pan/midgard: Use dummy tag for empty shadersAlyssa Rosenzweig2020-02-161-2/+2
| | | | | | | Fixes INSTR_INVALID_ENC in dEQP-GLES31.functional.compute.basic.empty Signed-off-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3835>
* panfrost: Allocate RAM backing of shared memoryAlyssa Rosenzweig2020-02-165-2/+43
| | | | | | | | | | | | Unlike other GPUs, Mali does not have dedicated shared memory for compute workloads. Instead, we allocate shared memory (backed to RAM), and the general memory access functions have modes to access shared memory (essentially, think of these modes as adding this allocates base + workgroupid * stride in harder). So let's allocate enough memory based on the shared_size parameter and supply it. Signed-off-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3835>
* panfrost: Rename bifrost_framebuffer->mali_framebufferAlyssa Rosenzweig2020-02-162-19/+19
| | | | | | | | | | | | (And bifrost_fb_extra to mali_framebuffer_extra, bifrost_render_target to mali_render_target) These structures are the norm on midgard t760+, drop the bifrost names, it's silly... unrelated to the rest of the series but while I'm messing with pandecode and cleaning up bifrost abstractions, might as well. Signed-off-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3835>
* panfrost: Unify bifrost_scratchpad with mali_shared_memoryAlyssa Rosenzweig2020-02-162-3/+6
| | | | | | | | | | It looks like these are the same structure, so this allows us to reuse mali_shared_memory across architectures, and dispels with the Bifrost-specific mystery of the scratchpads... nothing so mysterious after all, just stack. Signed-off-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3835>
* panfrost: Identify mali_shared_memory structureAlyssa Rosenzweig2020-02-163-18/+14
| | | | | | | | | This small structure is used to configure shared memory and stack for compute shaders, and is also present at the beginning of framebuffer descriptors. Let's factor it out. Signed-off-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3835>
* panfrost: Ensure compute shader_meta is zeroedAlyssa Rosenzweig2020-02-161-1/+2
| | | | | | | In theory the hardware doesn't care but it'll make for easier traces. Signed-off-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3835>
* panfrost: Simplify swizzle translationAlyssa Rosenzweig2020-02-141-32/+5
| | | | | | | | | It lines up anyway, and Gallium shouldn't change this. (And if it does, we'll deal with that then since CI would start failing.) Signed-off-by: Alyssa Rosenzweig <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3824> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3824>
* panfrost: Inline panfrost_get_default_swizzleIcecream952020-02-145-20/+23
| | | | | | | | | | | This commit replaces panfrost_get_default_swizzle with an inlined implementation where the returned values can be determined at compile time. According to perf, this previously used about 2% CPU for Openarena. Reviewed-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3824>
* panfrost: Set the MALI_WRITES_{Z,S} flags when neededBoris Brezillon2020-02-054-1/+17
| | | | | | | | | | | | | In order to make Z/S writes from fragment shaders effective, we need to set the MALI_WRITES_{Z,S} flags when the shader has a FRAG_RESULT_{DEPTH,STENCIL} output variable. Now that shaders can change the S value, we can expose the STENCIL_EXPORT cap. Signed-off-by: Boris Brezillon <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3697>
* panfrost: Add the MALI_WRITES_{Z,S} flagsBoris Brezillon2020-02-052-7/+7
| | | | | | | | | | | | | | We discovered 2 new shader flags used when a fragment shader updates the depth/stencil value through a ZS writeout. If those flags are not set, the depth/stencil value stored in the depth/stencil tilebuffer remain unchanged. While at it, rename unknown2 into flags_hi and rename flags into flags_lo. Signed-off-by: Boris Brezillon <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3697>
* panfrost: Z24 variants should be sampled as R32UIBoris Brezillon2020-02-051-1/+9
| | | | | | | | | | | Midgard has no dedicated samplers for Z24S8 and Z24X8 formats, and the GPU expects the depth to be encoded in an IEEE 32-bit float. Turn all Z24_UNORM variants into R32UI and let the shader do the conversion using bfe+fmul instructions. Signed-off-by: Boris Brezillon <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3697>
* panfrost: Use size0 when calculating the offset to a depth levelAlyssa Rosenzweig2020-02-042-7/+12
| | | | | | | | | | Previously, we were using cubemap_stride and sometimes overwriting other BOs such as shaders. Signed-off-by: Alyssa Rosenzweig <[email protected]> Signed-off-by: Tomeu Vizoso <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3692>
* panfrost: Only clamp the LOD to disable mipmapping when neededTomeu Vizoso2020-02-041-6/+7
| | | | | | | | | Otherwise, we may be incrementing max_lod over the limit, causing a DATA_INVALID_FAULT. Signed-off-by: Tomeu Vizoso <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3692>
* panfrost: Fix the damage box clamping logicBoris Brezillon2020-02-031-0/+2
| | | | | | | | | | | | When the rendering are is not covering the whole FBO, and the biggest damage rect is empty, we can have damage.max{x,y} > damage.min{x,y}, which leads to invalid reload boxes. Fixes: 65ae86b85422 ("panfrost: Add support for KHR_partial_update()") Signed-off-by: Boris Brezillon <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3676> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3676>
* panfrost: Fix non-debug buildsIcecream952020-02-021-0/+1
| | | | | | | | | | | | | For non-debug builds, where assertions are compiled out, GCC complains about the end of the non-void function panfrost_translate_channel_width being reached. Fixes: 226c1efe9a8 ("panfrost: Add more info to some assertions") Reported-by: Piotr Oniszczuk Suggested-by: Boris Brezillon <[email protected]> Reviewed-by: Boris Brezillon <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3665> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3665>
* panfrost: Use DBG macro to avoid noise in the consoleTomeu Vizoso2020-01-316-11/+10
| | | | | | | | | | It pollutes the output of programs that use Panfrost and can confuse its callers, such as test runners. Signed-off-by: Tomeu Vizoso <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3625> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3625>
* panfrost: Add more info to some assertionsTomeu Vizoso2020-01-311-2/+4
| | | | | | Signed-off-by: Tomeu Vizoso <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3625>
* panfrost: Don't copy uniforms when the size is zeroIcecream952020-01-271-1/+1
| | | | | | | | | This fixes a crash when using Gallium HUD with QuakeSpasm when gamma correction shaders (a QuakeSpasm feature, not part of Mesa) are used. Reviewd-by: Alyssa Rosenzweig <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3549> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3549>
* panfrost: Fix 32-bit warning for `indices`Alyssa Rosenzweig2020-01-241-1/+1
| | | | | | | | | | | | ../src/gallium/drivers/panfrost/pan_context.c: In function ‘panfrost_draw_vbo’: ../src/gallium/drivers/panfrost/pan_context.c:1551:70: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast] ctx->payloads[PIPE_SHADER_FRAGMENT].prefix.indices = (u64) NULL; ^ Signed-off-by: Alyssa Rosenzweig <[email protected]> Reported-by: Icecream95 <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3543> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3543>
* pan/decode: Rotate trace filesIcecream952020-01-231-0/+4
| | | | | | Reviewed-by: Alyssa Rosenzweig <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3525> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3525>
* panfrost: Drop mysterious zero=0xFFFF fieldAlyssa Rosenzweig2020-01-231-13/+0
| | | | | | | | | | It doesn't seem to affect any results and it's not at all clear if/why the blob sometimes(?) sets it? So let's clean this up since this solution isn't correct anyway. Signed-off-by: Alyssa Rosenzweig <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3513> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3513>
* panfrost: Fix crash in compute variant allocationAlyssa Rosenzweig2020-01-221-1/+2
| | | | | | | Signed-off-by: Alyssa Rosenzweig <[email protected]> Fixes: d8a3501f1b2 ("panfrost: Dynamically allocate shader variants") Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3515> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3515>
* panfrost: Add ASTC texture formatsIcecream952020-01-214-1/+28
| | | | | | Acked-by: Daniel Stone <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3414>
* panfrost: Add ETC1/ETC2 texture formatsIcecream952020-01-213-4/+34
| | | | | | Acked-by: Daniel Stone <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3414>
* panfrost: Rework linear<--->tiled conversionsAlyssa Rosenzweig2020-01-211-3/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | There's a lot going on here (it's a ton of commits squashed together since otherwise this would be impossible to review...) 1. We have a fast path for linear->tiled for whole (aligned) tiles, but we have to use a slow path for unaligned accesses. We can get a pretty major win for partial updates by using this slow path simply on the borders of the update region, and then hit the fast path for the tile-aligned interior. This does require some shuffling. 2. Mark the LUTs constant, which allows the compiler to inline them, which pairs well with loop unrolling (eliminating the memory accesses and just becoming some immediates.. which are not as immediate on aarch64 as I'd like..) 3. Add fast path for bpp1/2/8/16. These use the same algorithm and we have native types for them, so may as well get the fast path. 4. Drop generic path for bpp != 1/2/8/16, since these formats are generally awful and there's no way to tile them efficienctly and honestly there's not a good reason too either. Lima doesn't support any of these formats; Panfrost can make the opinionated choice to make them linear. 5. Specialize the unaligned routines. They don't have to be fully generic, they just can't assume alignment. So now they should be nearly as fast as the aligned versions (which get some extra tricks to be even faster but the difference might be neglible on some workloads). 6. Specialize also for the size of the tile, to allow 4x4 tiling as well as 16x16 tiling. This allows compressed textures to be efficiently tiled with the same routines (so we add support for tiling ASTC/ETC textures while we're at it) Signed-off-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Vasily Khoruzhick <[email protected]> Tested-by: Vasily Khoruzhick <[email protected]> #lima on Mali400 Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3414>
* panfrost,lima: De-Galliumize tiling routinesAlyssa Rosenzweig2020-01-211-2/+3
| | | | | | | | | | | | There's an implicit dependence on Gallium here that will add more complexity than needed when testing/optimizing out of driver as well as potentially Vulkanizing. We don't need a full pipe_box, just the x/y/w/h properties directly. Signed-off-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Vasily Khoruzhick <[email protected]> Tested-by: Vasily Khoruzhick <[email protected]> #lima on Mali400 Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3414>
* panfrost: Dynamically allocate shader variantsIcecream952020-01-182-3/+23
| | | | | | | | | This fixes a crash in LZDoom where over 16 shader variants are needed for a few shaders in some maps, and should also save a few kilobytes of RAM as most of the time only one or two variants of the 8 previously allocated are actually needed. Reviewed-by: Alyssa Rosenzweig <[email protected]>
* panfrost: Expose some functionality with dEQP flagAlyssa Rosenzweig2020-01-181-3/+3
| | | | | | | | These features are stable enough that they don't need to be hidden. Signed-off-by: Alyssa Rosenzweig <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3464> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3464>
* panfrost: Compact the bo_access readers arrayIcecream952020-01-181-4/+9
| | | | | | | | | | | | | | | | | | Previously, the array bo_access->readers was only cleared when there were no unsignaled fences, which in some situations never happened. That resulted in the array having thousands of NULL pointers, but only a handful of active readers. With this patch, all the unsignaled readers are moved to the front of the array, effectively building a new array only containing the active readers in-place. This results in the readers array usually only having a couple of elements. Reviewed-by: Boris Brezillon <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3419> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3419>
* android: Fix whitespace issueRobert Foss2020-01-151-1/+1
| | | | | Signed-off-by: Robert Foss <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]>
* panfrost: Fix linear depth texturesAlyssa Rosenzweig2020-01-144-20/+26
| | | | | | | | | | | | | | | | As pointed out by Boris, what we were calling PAN_LINEAR depth textures was in fact u-interleaved tiled (!), but we never noticed since we flipped the flag used for sampling, leading to all sorts of fun bugs when attempting to directly acess depth textures from the CPU. Which begs the question -- if what we called LINEAR was tiled, how do we actually render linear depth textures? It turns out the flags for AFBC form a mali_block_format 2-bit code just like their render-target counterparts, so we can render to any of the above. Signed-off-by: Alyssa Rosenzweig <[email protected]> Reported-by: Boris Brezillon <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3393> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3393>
* panfrost: Fix headers and gpu_headers memory leakDaniel Ogorchock2020-01-131-0/+3
| | | | | | | | | | The per-batch headers/gpu_headers dynarrays need to be freed during the batch cleanup to prevent leaking. Signed-off-by: Daniel Ogorchock <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3308> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3308>
* panfrost: Fix panfrost_bo_access memory leakDaniel Ogorchock2020-01-131-1/+3
| | | | | | | | | The bo access needs to be freed prior to removing it from its hash table. This prevents leaking them over time. Signed-off-by: Daniel Ogorchock <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3308>
* panfrost: Add negative lod bias supportIcecream952020-01-101-3/+3
| | | | Reviewed-by: Alyssa Rosenzweig <[email protected]>
* panfrost: Adjust for mismatch between hardware/Gallium in arrays/cubeAlyssa Rosenzweig2020-01-071-11/+33
| | | | | | | | | | | The hardware separates face selection and array indexing, it looks like, whereas Gallium smushes them together with some modulus fun. Let's fix it so mipmapped 2D arrays work without regressing cubemaps. Fixes dEQP-GLES3.functional.texture.filtering.2d_array.* among others. Signed-off-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Tomeu Vizoso <[email protected]>
* panfrost: Respect constant buffer_offsetAlyssa Rosenzweig2020-01-071-2/+4
| | | | | | | | Fixes dEQP-GLES3.functional.ubo.multi_basic_types.single_buffer.* among others Signed-off-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Tomeu Vizoso <[email protected]>
* panfrost: Handle PIPE_FORMAT_R10G10B10A2_USCALEDAlyssa Rosenzweig2020-01-061-0/+2
| | | | | | | | | | Same format code as UINT... might be different in how it's fed into a shader but we'll deal with that when we get there. Fixes dEQP-GLES3.functional.vertex_arrays.single_attribute.output_types.usigned_int2_10_10_10.components4_vec2_quads1 Signed-off-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Tomeu Vizoso <[email protected]>
* panfrost: Report MSAA 4x supported for dEQPAlyssa Rosenzweig2020-01-061-1/+10
| | | | | | | | | Fixes dEQP-GLES3.functional.state_query.integers.max_samples_getinteger64 We'll have to actually implement multisampling next, but hey. Signed-off-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Tomeu Vizoso <[email protected]>
* panfrost: Cleanup tiling selection logicAlyssa Rosenzweig2020-01-061-13/+14
| | | | | | | | | | | | Make it a lot more obvious what we're doing and fix more than a few corner cases in the process. Fixes dEQP-GLES3.functional.buffer.map.write.render_as_index_array.pixel*, and likely others. Signed-off-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Tomeu Vizoso <[email protected]>
* panfrost: Implement sRGB blend shadersAlyssa Rosenzweig2020-01-062-8/+16
| | | | | | | | | | | | | We use the lowering in nir_format_convert. There are native ops for this so this is far from optimal and not remotely efficient but as with most blend shader things right now, it's hard enough to get it working, so let's focus on that for now. We'll make it fast later (once we have GLES3 stable, we can start optimizing these things). Fixes dEQP-GLES3.functional.fragment_ops.blend.fbo_srgb.* Signed-off-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Tomeu Vizoso <[email protected]>
* panfrost: Support rendering to non-zero Z/S layersAlyssa Rosenzweig2020-01-061-5/+5
| | | | | | | Fixes abort in STK's shadow implementation. Signed-off-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Tomeu Vizoso <[email protected]>
* panfrost: Texture from Z32F_S8 as R32FAlyssa Rosenzweig2020-01-061-0/+4
| | | | | | | | | Z32F_S8 becomes Z32F in texturing, which in turn just becomes R32F. Fixes dEQP-GLES3.functional.texture.format.sized.*.depth32f_stencil8* Signed-off-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Tomeu Vizoso <[email protected]>
* panfrost: Pack MRT blend shaders into a single BOAlyssa Rosenzweig2020-01-023-22/+38
| | | | | | | | | | Blend shader size and location in memory is considerably constrained, probably to facilitate optimizations (my guess is that blend shaders are run strictly out of i-cache). We need to pack the blend shaders for each RT of a single framebuffer together. The easiest way to do this is at draw time which is not terribly efficient but will hold us over for now. Signed-off-by: Alyssa Rosenzweig <[email protected]>
* panfrost: Handle RGB16F colour clearAlyssa Rosenzweig2020-01-021-0/+2
| | | | | | | | We don't handle this format yet, but we will soon, and the abort in pan_pack_color is possible even without exposing the format... Handling this gracefully might not be required by the spec but let's not crash. Signed-off-by: Alyssa Rosenzweig <[email protected]>
* panfrost: Store internal formatTomeu Vizoso2020-01-022-2/+7
| | | | | | | | It's needed by u_transfer_helper to know when the depth+stencil buffer has been split. Signed-off-by: Tomeu Vizoso <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]>
* panfrost: Map with size of first layer for 3D texturesTomeu Vizoso2020-01-022-2/+8
| | | | | | | As that's what Gallium expects in transfer.layer_stride. Signed-off-by: Tomeu Vizoso <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]>
* panfrost: Dynamically allocate array of texture pointersTomeu Vizoso2020-01-021-4/+22
| | | | | | | | With 3D textures we can have lots of layers, so better allocate it dynamically at runtime. Signed-off-by: Tomeu Vizoso <[email protected]> Reviewed-by: Alyssa Rosenzweig <[email protected]>
* panfrost: Respect glPointSize()Alyssa Rosenzweig2019-12-301-1/+5
| | | | | | We have native support for this somehow. Fixes the mesa demo `points` Signed-off-by: Alyssa Rosenzweig <[email protected]>
* panfrost: Remove MRT indirection in blend shadersAlyssa Rosenzweig2019-12-303-43/+39
| | | | | | | | | | Since we have a separate blend shader for each render target, let's simplify this structure and reduce the options memory footprint by 88% or something goofy like that. Should also enable separate blending per render target. Signed-off-by: Alyssa Rosenzweig <[email protected]>
* panfrost: Implement integer varyingsAlyssa Rosenzweig2019-12-301-1/+1
| | | | | | | | | We need to actually work out the varying format on demand, rather than assuming rgba32f. Fixes dEQP-GLES3.functional.fragment_out.basic.int.* Signed-off-by: Alyssa Rosenzweig <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-12 14:36:13 +0000