kc3-lang/angle/src/libANGLE/renderer/vulkan/vk_helpers.cpp

src/libANGLE/renderer/vulkan/vk_helpers.cpp

Log

Author Commit Date CI Message
Charlie Lao b6bd039c 2020-10-06T13:34:53 Vulkan: Fine-grained depth stencil layout transition barrier Gfxbench Manhattan uses depth attachment as texture in the same render pass, which creates a feedback loop. The layout we are using for this has kAllShadersPipelineStageFlags even though it only used by fragment shader. This creates an unnecessary vertex/fragment dependency that hurts performance on some mobile GPUs. This CL split DepthStencilReadOnly to DSAttachmentReadAndFragmentShaderRead, DSAttachmentReadAndAllShadersRead, and DepthStencilAttachmentReadOnly, so that if it is used by fragment shader, we only barrier against fragment stage. If not used by any shader, we don't block any shader stages. To make names consistent, this CL also renames DepthStencilAttachmentAndFragmentShaderRead to DSAttachmentWriteAndFragmentShaderRead, renames DepthStencilAttachmentAndAllShadersRead to DSAttachmentWriteAndAllShadersRead. Bug: b/169891513 Change-Id: Iede497e8af028861f9166a32d122f26dc7aa864a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2453599 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Charlie Lao d151af7b 2021-05-20T10:52:46 Vulkan: Improve generateMipmapsWithBlit barrier usage ImageHelper::generateMipmapsWithBlit() implements glGenerateMipmap using vkCmdBlitImage. For each level it changes layout to VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL and issue a blit. In order to keep all mipmap levels' layout consistent, it issues a vkCmdPipelineBarrier call for the the very last mipmap level to change it to VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL as well. This creates an unnecessary barrier for the next operation, that now it sees a barrier against transfer stage. This CL changes all levels to VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL so that the whole image will have consistent layout. After glGenerateMipmap call, the next reasonable usage is use it in fragment shader, so the barrier is not wasted. Bug: angleproject:5986 Change-Id: I3690ae8417d97bc504fcf036d016f289e6370526 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2909757 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Mohan Maiya 81db1a46 2021-05-20T13:45:26 Vulkan: TextureVk inherits a few attributes from ImageHelper Cache VkImageCreateFlags in ImageHelper. This allows texture target siblings to inherit a few VkImage create attributes thus avoiding image respecification in certain code paths. Bug: angleproject:2514 Bug: angleproject:5281 Test: PbufferTest.ClearAndBindTexImageSrgbSkipDecode*Vulkan Change-Id: Ic7397fabdce185264e06488355ca47f809338519 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2910473 Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi cf3af29a 2021-05-14T16:49:36 Vulkan: Optimize respecifying an image When recreating a texture image to change a usage flag or the base/max levels, this change directly places the levels of the previous image as updates to the new image. This is instead of copying the image to a temp one and staging that. Bug: angleproject:4835 Change-Id: Ibc210b9ff0e8d11cba10b1cd9ab262c8f706cea5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2898417 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com>
Shahbaz Youssefi 8bd3d7d5 2021-05-17T13:45:33 Vulkan: Fix a bug releasing DynamicBuffer-owned buffer There was one instance of BufferVk releasing a buffer it had allocated from a DynamicBuffer. This shouldn't have happened as the DynamicBuffer owns the buffers. Bug: angleproject:5720 Change-Id: I435512f4bb099130126bf3efb48a238fcd9f3ddb Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2896168 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 9c262ad0 2021-05-14T12:26:48 Vulkan: Cleanup texture image respecify Prior to this change, respecifying a texture image due to usage, base or max level changes incurred a copy of every level and layer to a temporary buffer which was then staged as an update to the new image. This code was somewhat messy (for example with respect to depth/stencil images), error prone (e.g. previously had bugs with compressed textures) and disallowed further optimizations such as in anglebug.com/4835. This change does the following: - ImageHelper::SubresourceUpdate now takes ref-counted images, instead of image pointers. This allows the same image to be staged for multiple updates. - Respecifying an image is still done through a copy, but to an identical (temp) image instead of buffer, and each level of the image is staged as an update. * Further optimization is to stage the old image itself directly as updates to the new image Bug: angleproject:4835 Change-Id: I4a3ef2d616c9ab459ff65f918b0fb6d9a2161b73 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2897537 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com>
Shahbaz Youssefi e354ff1a 2021-03-05T04:07:21 Vulkan: Allow DynamicBuffer suballocation in BufferVk When allocations are made from DynamicBuffer, they suballocate from a possibly larger BufferHelper. In BufferVk, the offset of the suballocation was discarded, which limited the use of DynamicBuffer to a pool of small buffers. This change applies any such offset that may arise from suballocations everywhere, and makes BufferVk use a larger buffer size when the GL_DYNAMIC_* buffer usage hints are provided. Bug: angleproject:5719 Change-Id: I3df3317f7acff1b1b06a5e3e2bb707616a7d0512 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2738650 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Charlie Lao 46a139ad 2021-05-06T11:20:51 Vulkan: set DS layout before using it in the endRenderPass In CommandBufferHelper::endRenderPass(), we are checking depth stencil's initialLayout to change storeOp to None if the layout is read only. But the layout was set after that check, which essentially voids the optimization. This CL moves the finalizeDepthStencilImageLayout() call before the layout is used. This CL also moves the depth stencil loadOp/storeOp to a new function finalizeDepthStencilLoadStoreOp(). When depthImage gets deleted before renderpass ends, we could also apply the same load/store optimization just like we did at endRenderPass() time. Bug: b/187425444 Change-Id: I89814274352f09cbf1f7b58a91bbaf131b983fb1 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2877933 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Commit-Queue: Charlie Lao <cclao@google.com>
Mohan Maiya 84d22197 2021-04-26T17:11:27 Vulkan: Add device local fallback in findCompatibleMemoryIndex(...) When finding a compatible memory index for an external buffer, we need to have a fallback path that requires just the VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT property. Bug: angleproject:5909 Change-Id: Id3faffb3e357be2e5e6f5fce9e64f0f2889d199a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2862561 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Jamie Madill ccc0fbaa 2021-03-24T17:56:38 Vulkan: Related fixes for buffer descriptor set cache. Includes some stats counter gathering and a few related refactors and cleanups. Also includes a new overlay widget. Bug: angleproject:5736 Change-Id: Ida8d2cd815c5b598c6a442dd9bbfdf51e9c05180 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2785431 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Tim Van Patten <timvp@google.com>
Mohan Maiya 84b31c03 2021-04-25T08:30:14 Vulkan: Fix bug in VkImageFormatListCreateInfoKHR On some vendors VVL throws the following error, snippet only - [ VUID-VkImageViewCreateInfo-pNext-01585 ] Validation Error ... image was created with a VkImageFormatListCreateInfo in pNext of vkImageCreateInfo, but none of the formats match the VkImageViewCreateInfo::format ... The Vulkan spec states: If a VkImageFormatListCreateInfo structure was included in the pNext chain of the VkImageCreateInfo structure used when creating image and VkImageFormatListCreateInfo::viewFormatCount is not zero then format must be one of the formats in VkImageFormatListCreateInfo::pViewFormats. It looks like VkImageFormatListCreateInfoKHR::pViewFormats needs to contain all formats including the format of the VkImage itself. Bug: angleproject:5281 Change-Id: I93c6900d99791ef4f9f116cb114f068e0a318bf4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2849566 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Charlie Lao fb288312 2021-04-12T12:38:03 Vulkan: Consolidate mReadOnlyDepthStencilMode to mRenderPassUsageFlags Since now ImageHelper object has a mRenderPassUsageFlags tracking how it is been used by current RenderPassCommands, we can consolidate mReadOnlyDepthStencilMode into a bit in the mRenderPassUsageFlags. The read only mode is just a special attachment mode, so this makes code more consistent and able to get rid of the extra dword for mReadOnlyDepthStencilMode. Bug: b/181797383 Change-Id: Ie9b7be1c18c392e5b2712bdae6ab6506de8d6d34 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2821942 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Charlie Lao 5eac3169 2021-04-09T15:54:11 Vulkan: Support sample/render to depth texture with different LOD Previously we supported read only depth attachment that attachment and sample are using the same texture. It was always assuming it will be read only if the same texture is attached to FBO and same time bound to texture unit. But we never supported if it actually writing to depth texture, but at the different level with sampling. This is supported OpenGL operation as long as the levels render to is outside [base_level, max_level]. Later on, we added support of the above operation, but limited to color buffers. This CL extends the same support of render and sample to the same depth texture with non-overlapping levels. Android game black desert mobile is running into this usage case. Bug: b/181797383 Change-Id: I7ee1d52f27603f933102ad1b098684309449c406 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2819487 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Ian Elliott <ianelliott@google.com>
Charlie Lao 92d3912d 2021-04-01T10:49:14 Vulkan: Remove maxLevel and immutable from ImageHelper::initExternal Now that we have removed mBaseLevel and mMaxLevel from ImageHelper class in previous two CLs. The initExternal function can be modified to not pass in base/max level information. This CL also removes immutable from argument and let caller directly pass in firstLevel to it. Bug: b/181800403 Change-Id: I595c1faffabc7582af35a9578b2ecdf00c39207f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2800414 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Ian Elliott <ianelliott@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Charlie Lao afd32d8e 2021-03-30T15:50:57 Vulkan: Remove mMaxLevel from ImageHelper class This CL removes mMaxLevel from ImageHelper class. Instead, it now uses front end's max level information when possible. Bug: b/181800403 Change-Id: Ie0f6bd11e3ca0d4ddfc98f21261396c4d71b7140 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2796153 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Ian Elliott <ianelliott@google.com>
Charlie Lao 08b7c55c 2021-03-30T13:51:04 Vulkan: Remove mBaseLevel from ImageHelper class This removes mBaseLevel from ImageHelper class. With the mFirstAllocatedLevel tracking exactly which GL level has been allocated, this cached mBaseLevel is no longer needed. Bug: b/181800403 Change-Id: I99d66c93b0c8f1bd20a5811b51f512a27927201e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2795275 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Ian Elliott <ianelliott@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Charlie Lao b5a71140 2020-08-20T17:24:21 Vulkan: Make storage actually immutable for immutable textures The immutable textures are intended to be "immutable". Right now we are still re-allocating VkImage object based on base_level and max_level. This causes a problem for rendering to a level that is beyond [base_level, max_level], which is totally within OpenGL spec. This CL makes an immutable texture always allocate from 0 to max levels that are specified by a glTexStorage call. Changing base_level will not trigger re-allocation of VkImage object. Bug: b/181800403 Change-Id: I4b4ddea17b7f6f7bfd8f36bfe8bb3a35b5c180b6 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2368038 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Ian Elliott <ianelliott@google.com> Commit-Queue: Charlie Lao <cclao@google.com>
Charlie Lao fd3b8faf 2021-03-30T11:55:49 Vulkan: Split fragment shader from ColorAttachmentAndAllShadersRead Even though it is not common usage case that the same texture is used as attachment and texture, but aztec ruins is using it. And earlier version of gfxbench's T-Rex also running into this. So performance is still important. This CL splits the texture from fragment shader usage case out so we can have lighter barrier. Bug: angleproject:5780 Change-Id: Ifdcbfe24488fdac62826c5af6ecfbb05f87c0499 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2795269 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Charlie Lao 81432c24 2021-03-19T14:03:02 Vulkan: Rename onImageHelperRelease to finalizeImageLayout Make the API name reflect what it does. With recent changes, this get called from places that are not releasing ImageHelper object, but is try to reference the current layout. This new name makes more sense than onImageHelperRelease (at that time when this API was introduced, it was only called when ImageHelper is being released). Bug: b/175584609 Change-Id: Ie5f5fcdbd97436724d7eb016374d8b4178e7ba1d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2776261 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Charlie Lao 867411a4 2021-03-08T19:08:14 Vulkan: Track images used for both attachment and sampler This CL adds a new layout ImageLayout::ColorAttachmentAndShaderRead. We detect that the same ImageHelper object is used for both sampler and attachment and will pick this new layout instead of ColorAttachment. Bug: b/175584609 Change-Id: I2089f59d4dc1ad4de7edab0e067797adcfa9b020 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2749479 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Charlie Lao da437f26 2021-03-08T19:08:14 Vulkan: Defer color image layout changes at endRenderPass time Right now color render target's image's layout change are done at beginRenderPass time. The problem is that the layout also depends on whether texture is also being used as a sampler or not. That information is not known when renderpass starts. We did some special treatment for depth stencil attachment so that its layout determination is deferred until endRenderPass time. This CL expands that same mechanism to color attachment as well. Right now the color attachment will still pick the same ImageLayout::ColorAttachment layout since the logic to detect it is also used for texture sampling is not there yet. Bug: b/175584609 Change-Id: Id7486174d475f894461578b31d0d40fdd90e808a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2744121 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Charlie Lao <cclao@google.com>
Charlie Lao 378b3261 2021-03-08T18:50:53 Vulkan: Move CommandBufferHelper::reset() closer to constructor This is a mechanical change. Similar to constructor, the reset() function is doing initialization of data members. Move them next to each other helps adding a new data members which requires modifications to both functions. Bug: b/175584609 Change-Id: I11386677bc5c9838104025d6a4ecf6f2b58fa92c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2745842 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Commit-Queue: Charlie Lao <cclao@google.com>
Mohan Maiya 81dcf078 2021-03-08T11:21:31 Reland "Vulkan: Support EXT_sRGB_write_control" This is a reland of 6073af536cf627742696823edc82c9b0a481a8bc with 2 changes - 1. Don't enable the extension even in nonConformant mode 2. Don't enable VK_KHR_image_format_list for swiftshader Original change's description: > Vulkan: Support EXT_sRGB_write_control > > Implement support for EXT_sRGB_write_control. This extension > requires VK_KHR_image_format_list to be supported. > > The spec requires this functionality to work with glBlitFramebuffer > as well but support for that will be added in a follow up change. > As such, this extension is only exposed in non-conformant mode. > > Bug: angleproject:5075 > Tests: SRGBFramebufferTest.*Vulkan* > Change-Id: I59b38f6cd810a3d0d67ec29f4f19c25f65f70862 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2617243 > Commit-Queue: Mohan Maiya <m.maiya@samsung.com> > Reviewed-by: Jamie Madill <jmadill@chromium.org> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Bug: angleproject:5075 Change-Id: I8e149d196a39c3c4769bfa8690792f3c53831299 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2762647 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Peng Huang b27740f3 2021-03-09T16:15:15 Revert "Vulkan: Support EXT_sRGB_write_control" This reverts commit 6073af536cf627742696823edc82c9b0a481a8bc. Reason for revert: crbug.com/1186140 Original change's description: > Vulkan: Support EXT_sRGB_write_control > > Implement support for EXT_sRGB_write_control. This extension > requires VK_KHR_image_format_list to be supported. > > The spec requires this functionality to work with glBlitFramebuffer > as well but support for that will be added in a follow up change. > As such, this extension is only exposed in non-conformant mode. > > Bug: angleproject:5075 > Tests: SRGBFramebufferTest.*Vulkan* > Change-Id: I59b38f6cd810a3d0d67ec29f4f19c25f65f70862 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2617243 > Commit-Queue: Mohan Maiya <m.maiya@samsung.com> > Reviewed-by: Jamie Madill <jmadill@chromium.org> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Bug: angleproject:5075,chromium:1186140 Change-Id: Ib0d4d60fe7434fb950f99db2c210aab9af7d2d0e No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2743663 Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Commit-Queue: Peng Huang <penghuang@chromium.org>
Mohan Maiya 6073af53 2021-03-08T11:21:31 Vulkan: Support EXT_sRGB_write_control Implement support for EXT_sRGB_write_control. This extension requires VK_KHR_image_format_list to be supported. The spec requires this functionality to work with glBlitFramebuffer as well but support for that will be added in a follow up change. As such, this extension is only exposed in non-conformant mode. Bug: angleproject:5075 Tests: SRGBFramebufferTest.*Vulkan* Change-Id: I59b38f6cd810a3d0d67ec29f4f19c25f65f70862 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2617243 Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi e366e2c3 2021-02-27T01:00:02 Vulkan: Keep dynamic buffer's free list trimmed ContextVk's staging buffer never gets a chance to free its free buffer list. During application load time, a large amount of memory may be allocated from this buffer to stage texture updates and they would remain throughout the life of the application. This change ensures that the free buffer list doesn't grow unbounded. In the Manhattan trace, this saves >1GB of memory on Linux. There are now three policies for vk::DynamicBuffer: - Always reuse buffers: This is useful for dynamic buffers that make frequent small allocations, such as default uniforms, driver uniforms, default vertex attributes and UBO updates. - Never reuse buffers: This is for situations where the buffer is unlikely to be used after some initial usage, such as texture data upload or vertex format emulation (as the conversion result is cached, so it's never redone). - Limited reuse of buffers: For the staging buffer in the context which is shared by all immutable texture data uploads, it's useful to keep a limited number of buffers (1 in this change) to support future texture streaming while allowing a large number of buffers allocated in a burst to be discarded. Bug: angleproject:5690 Change-Id: Ic39ce61e6beb3165dbce4b668e1d3984a2b35986 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2725499 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Mohan Maiya 58c35d42 2021-03-02T09:00:55 Vulkan: Unmap buffer when adding to mBufferFreeList The commit b5af8bde13 moved the unmap call on a buffer to when it is placed in mBufferFreeList. There was one place that was missed, rectifying that oversight here. Bug: angleproject:5689 Change-Id: Iaed2916de374d55ab59276e2679c23ed7ef9c4e8 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2733476 Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Stephen White d8557cc8 2021-03-03T10:17:46 Fix for readPixels() to PBO fast path. Include the "pixels" (data) pointer as offset. Bug: angleproject:5702 Change-Id: Idc9f3ee4c0dbb1cc9308393bbf877d6137598486 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2732789 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Stephen White <senorblanco@chromium.org>
Shahbaz Youssefi 07025aa3 2021-03-02T00:28:32 Overlay: Widget for vk::DynamicBuffer allocations Bug: angleproject:5690 Change-Id: Idfa591903627bbebffe306b387e95cbec1195338 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2725767 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Mohan Maiya b5af8bde 2021-02-24T09:47:43 Vulkan: No need to unmap buffer when adding to mInFlightBuffers When we move a BufferHelper to the mInFlightBuffers list don't try to unmap it there, instead do that when it gets added to mBufferFreeList This removes 0.75% CPU overhead from a Manhattan30 offscreen run Bug: angleproject:5689 Change-Id: I01a52cdf4803fb8364140b01bedbd3d0e6cc61fd Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2723492 Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 3013f27c 2021-02-23T22:37:42 Vulkan: Fix missing visibility barrier for glReadPixels glReadPixels uses a temporary buffer for readback and was missing an onBufferTransferWrite() call. Bug: angleproject:5669 Change-Id: I3c6daba176e332711fa1ab51e279526717486dd0 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2717005 Reviewed-by: Jie A Chen <jie.a.chen@intel.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 4a24786e 2021-02-23T23:32:48 Vulkan: Fix PBO read pixels with unaligned output pitch Vulkan takes the output pitch in texels instead of bytes in VkBufferImageCopy::bufferRowLength. This means that it's impossible to use vkCmdCopyImageToBuffer when the pitch is not a multiple of texel size. This change makes sure the fallback path is taken. Bug: angleproject:5667 Change-Id: I3f2ef312bb8288de8ca3c6730d85a3c14858812e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2717006 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi e96d1744 2021-02-12T14:14:02 Vulkan: Optimize glMemoryBarrier Previous to this change, glMemoryBarrier was processed as it is issued. This made it impossible to know whether a draw call would follow or a dispatch call, and what resources it would use. The render pass was conservatively broken due to this limitation. To address this limitation, handling of glMemoryBarrier is deferred until the next draw or dispatch call. Note that glMemoryBarrier acts as two barriers: - An execution+memory barrier: shader writes are made visible to subsequent accesses - Another execution barrier: shader accesses are finished before subsequent writes An important observation is that for most resources, ANGLE actually necessarily has to issue memory barriers automatically to conform with Vulkan. In terms of memory barrier thus, ANGLE already does the right thing except for when there's no binding change. This means WaW hazards (i.e. storage buffer and image writes) with no binding change require a memory barrier as a result of glMemoryBarrier. In all other cases, it's enough for glMemoryBarrier to break the render pass if necessary and ensure that corresponding bindings are marked dirty (for the execution or memory barriers to happen automatically later). Bug: angleproject:5070 Change-Id: Ide359c43362f8a78805ecf797a91de7aa79221f9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2693473 Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Mohan Maiya 4afd0f27 2021-02-17T13:35:57 Vulkan: Add VK_KHR_image_format_list support to initExternal The code for initializing a VkImage with VK_KHR_image_format_list extension support is now internal to initExternal. Also provides a flag to inform callers when VK_KHR_image_format_list support is enabled. Bug: angleproject:5520 Change-Id: Ie2ade93d1403ab3f9cbcf7d80684bbca201d5d8d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2702159 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Shahbaz Youssefi e8c0aa81 2021-01-26T23:40:36 Vulkan: Clean up transform feedback extension pause/resume 1. The xfb counter buffer barrier issued was wrong, following a typo in the spec. This barrier is now correctly issued using the usual barrier APIs. 2. A mechanism was added to automatically pause/resume transform feedback when a program pipeline needs to be rebound. This is incorrect as it misses the xfb counter buffer barrier. The render pass is broken instead if transform feedback is active/unpaused and the program pipeline is changed. 3. The transform feedback counter buffers are now disposed of when transform feedback is ended. This avoids an unnecessary barrier that this change would have otherwise incurred (and hence render pass break) in Manhattan which repurposes the same transform feedback object. Bug: angleproject:5528 Change-Id: I1ffe8b4b8975645ba43afd70e9cdbb0765529da5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2651647 Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Ian Elliott 60b03e62 2021-02-16T14:31:19 Create getRotatedExtents() and related methods Created: - WindowsSurfaceVk::getRotatedWidth() - WindowsSurfaceVk::getRotatedHeight() - RenderTarget::getRotatedExtents() - ImageHelper::getRotatedExtents() - ImageHelper::getRotatedLevelExtents2D() Note: The FramebufferVk class doesn't use any of these methods Bug: b/175793022 Change-Id: I64395688bfdb172d32853763743fc5f266a6b792 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2686102 Commit-Queue: Ian Elliott <ianelliott@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Mohan Maiya 2b69462e 2021-02-12T15:55:00 Vulkan: Consolidate image format list initialization code HardwareBufferImageSiblingVkAndroid, RenderbufferVk, SurfaceVk and TextureVk all had similar code to initialize VkImage with VK_KHR_image_format_list extension support. Consolidate all of them into a helper. Bug: angleproject:5520 Change-Id: Ib7ac582a35a26c0da15879e688bcaf0c2a92ea87 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2692641 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Shahbaz Youssefi 79ae52dd 2021-02-11T14:18:41 Vulkan: Fix missing visibility barrier for host-visible buffer writes See https://chromium-review.googlesource.com/c/angle/angle/+/1661252 for context. This was accidentally broken during the command graph rework. This will eventually be validated by syncval. See https://github.com/KhronosGroup/Vulkan-ValidationLayers/issues/2329. Bug: angleproject:5070 Change-Id: Ic16fa900e554d46e54b42fc3fbe0f96d5327fa0f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2689379 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 031dd5c5 2021-02-11T14:56:43 Vulkan: Add fast path for pack buffer copies When vkCmdCopyImageToBuffer is possible, it is used to implement glReadPixels with PBO instead of a CPU readback. A few tests are added to verify synchronization between pack/unpack and compute shader writes. Bug: angleproject:5070 Change-Id: Ia5da0feb31348f6b52d164b973ff95e2a0510582 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2691466 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Ian Elliott <ianelliott@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Ian Elliott ce540d8c 2021-02-01T17:38:10 Create MSAA-swapchain windows' VkImages differently The actual VkImage needs to have rotated extents, but the ImageHelper needs to have non-rotated extents in order for the rest of ANGLE's pre-rotation to work. Bug: b/175793022 Change-Id: I6fa25ab8c636886787ac50b194e566111308f30b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2666514 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Cody Northrop <cnorthrop@google.com> Commit-Queue: Ian Elliott <ianelliott@google.com>
Peng Huang c6a8bc4a 2021-01-30T14:58:22 Fix a crash in DynamicBuffer::~DynamicBuffer() The crash is because mBufferFreeList is not moved from the rhs value for DynamicBuffer's move constructor. Bug: chromium:1170339 Change-Id: I7d41b775ab56af8fa9412369d5a4008f59da3053 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2661615 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi eae262e7 2021-01-27T13:56:49 Vulkan: Fix image layout barriers for tessellation shaders Also fixes a bug where invalid stages may be specified for example if AllGraphicsReadOnly or DepthStencilReadOnly layouts are used and geometry or tessellation shaders are not supported by the implementation. Bug: angleproject:5557 Change-Id: Ia25a6aec8138c67701c63da65783263d8a7bda27 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2653911 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Peng Huang 122a1cc5 2021-01-26T20:06:24 Fix several vulkan backend problem on Android. * Load AHardwarebuffer API dynamically, so vulkan backend can be built with old NDK, but can work with newer android releases. * Do not link with libvulkan on android. * Expose EGL_ANDROID_get_native_client_buffer extension with vulkan backend. Bug: chromium:1170339 Change-Id: Idf7f6867a86ae40ba6d57a86e419c610ba404ba8 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2653506 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Peng Huang <penghuang@chromium.org>
Jamie Madill e357cc6f 2021-01-14T14:47:02 Revert "Vulkan: Prefer host cached memory for Buffers" This reverts commit 326f128b941d54c0277178f87e784da8ce2190eb. Reason for revert: Fails of Mali platform. See bug: Bug: angleproject:5544 Original change's description: > Vulkan: Prefer host cached memory for Buffers > > This improves perf because vertex data upload uses host visible memory, > and uncached memory can be slow. > > Test: > > Android on Windows with 1080Ti: > glmark2 [build] vbo=false 265 -> 3900 fps > > Bug: angleproject:5515 > Change-Id: Ib75582bbdb4ca4a22d3a121c7b477a23a8a809ef > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2615742 > Commit-Queue: Lingfeng Yang <lfy@google.com> > Reviewed-by: Jamie Madill <jmadill@chromium.org> > Reviewed-by: Tim Van Patten <timvp@google.com> TBR=geofflang@chromium.org,lfy@google.com,doughorn@google.com,timvp@google.com,jmadill@chromium.org # Not skipping CQ checks because original CL landed > 1 day ago. Bug: angleproject:5515 Change-Id: I59695b1981b43471dbd309d4e859bae7a4821969 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2627604 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 8db7915e 2021-01-12T10:56:47 Vulkan: Increment buffer view serials only when used This avoids the texel buffer view serials (shared with image view serials) recklessly getting large. Bug: angleproject:5185 Change-Id: I0debe73771094b395f6526475aa3c22a91f9e05e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2623811 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill 2e9706d8 2021-01-08T17:29:42 Vulkan: Use angle::FormatID instead of VkFormat. This change switches the internal enums we pass around from VkFormat to FormatID. The end goal of the refactor is to allow the Vulkan back-end to store packed tables indexed by FormatID. Because VkFormat has large gaps in its enum space we'd otherwise need to use unordered data structures like unordered_map. The change removes the redundant VkFormat storage from vk::Format and uses a new table query to return the VkFormat that 1:1 matches an angle::FormatID. We also include a reverse mapping for use with native Vulkan get functions for Android. Also moves sRGB conversion functions into renderer_utils. A couple sRGB formats that don't exist in GL are no longer handled by the sRGB conversion functions. These formats should be extremely rare. Bug: angleproject:5438 Change-Id: Id8b49773ca0c556f9f5a6a10fcf0d9762b93bbea Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2618204 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Ian Elliott <ianelliott@google.com>
Lingfeng Yang 326f128b 2021-01-07T14:25:44 Vulkan: Prefer host cached memory for Buffers This improves perf because vertex data upload uses host visible memory, and uncached memory can be slow. Test: Android on Windows with 1080Ti: glmark2 [build] vbo=false 265 -> 3900 fps Bug: angleproject:5515 Change-Id: Ib75582bbdb4ca4a22d3a121c7b477a23a8a809ef Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2615742 Commit-Queue: Lingfeng Yang <lfy@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Shahbaz Youssefi f99ccb08 2021-01-06T23:29:21 Vulkan: Add a perf warning when dropping superseded updates If an update to an image subresource is entirely superseded by a following update, the former update is dropped. This change adds a perf warning for this. Bug: angleproject:3461 Change-Id: Iaf4984ba52f3dfc86d5cc87f8be82123badebd04 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2613203 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi a7158eb5 2021-01-06T21:33:24 Vulkan: Fix image leak in staged updates in ImageHelper When removing superseding updates, the superseded update was not released, causing a memory leak. This change also makes SubresourceUpdate non-copyable and correctly implements the move assignment operator such that swap between different update types are correct. As a result, the destructor can now ASSERT that the image is not leaked. Bug: chromium:1146516 Bug: chromium:1163354 Change-Id: I7531c91d8559c23b2e09159118fe645d12fc601f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2613201 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill e91097bf 2020-12-29T14:05:56 Vulkan: Clean up "actual"/"intended" naming. Clarifies that the GL internal format is an "intended" format and the Vulkan formats are "actual" formats. This makes all the format fields use the same consistent naming pattern. Bug: angleproject:5438 Change-Id: I935a49895109e9e06eae5ef98d5614dfd1128ff8 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2605728 Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Mohan Maiya 98b56e60 2020-12-12T16:28:21 Vulkan: Accumulate internal cache stats in renderer The CacheStats of all internal caches are accumulated by the renderer. In order to see the hit ratios of all caches, the following GN args must be enabled: is_debug = true angle_enable_perf_counter_output = true Bug: angleproject:5447 Test: Manual verification with angle_end2end_tests Change-Id: Iaca3249192e9e4e130d8291b7759c459d79b06ee Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2588430 Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 56330564 2020-12-10T00:46:04 Vulkan: Support layered framebuffers This feature is introduced by geometry shaders, where all the layers of a texture can be attached to a framebuffer. The geometry shader would use gl_Layer to decide which layer the primitive should be rendered to. Bug: angleproject:3571 Change-Id: Ib2ae8e227b226295f9e2f62f6b230839070bc95c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2582711 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Charlie Lao 067a12f0 2020-11-19T16:47:25 Reland "Vulkan: Use specialization constant for halfRenderAreaWidth" This is a reland of ff38106d62fca2e36ec2db1cd580b546fd53ae43 Original change's description: > Vulkan: Use specialization constant for halfRenderAreaWidth > > halfRenderArea is used by fragment shader to adjust gl_FragCoord for > rotation and yflip compensation. This CL bakes halfRenderAreaWidth into > shader via specialization constant, thus allow compiler to consolidate > the calculation into one MAD instruction. > > Bug: b/173800146 > Change-Id: Id66301278e3389e2582369b695825e632bccecee > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2551541 > Commit-Queue: Charlie Lao <cclao@google.com> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Reviewed-by: Tim Van Patten <timvp@google.com> > Reviewed-by: Jamie Madill <jmadill@chromium.org> Bug: b/173800146 Change-Id: Ifc36086f4c5d6a44fb5456aa6e31bc8a783bba71 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2579648 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi f691b3b5 2020-12-02T13:11:54 Vulkan: Support PrimitivesGenerated query This query uses the Vulkan transform feedback extension. In GL, GL_TRANSFORM_FEEDBACK_PRIMITIVES_WRITTEN and GL_PRIMITIVES_GENERATED queries can be independently begun/ended. However, Vulkan requires that queries from pools of the same type can only be active one at a time. This forbids the two GL queries from being handled by two VK queries when they are simultaneously begun. This change makes these queries share their QueryHelper objects. The Vulkan transform feedback queries unconditionally retrieve both results anyway, so this is just a matter of making sure the two GL queries are merged as one when they are simultaneously used. The change fixes a number of issues as collateral: - TransformFeedbackPrimitivesWritten queries when !emulated were not released - Stashed queries were never released - If no render pass is open when a query ends, then getResult(no_wait) ended up waiting Bug: angleproject:5404 Change-Id: I8ce13ea76ffd31b3152ded7c713c6466d0315504 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2573580 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Courtney Goeltzenleuchter aafcb504 2020-12-08T09:32:10 Vulkan: Add ensureSubmission to queueSubmitOneOff Some callers of queueSubmitOneOff require that the command being queued to have been sent to the GPU. The new ensureSubmission parameter indicates that behavior and when running with threaded worker will wait for the worker queue to empty before returning. Bug: b/170312581 Change-Id: Ib620fb37f4b9b4431451ccbd10807c0dff1842af Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2579041 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Cody Northrop <cnorthrop@google.com> Commit-Queue: Courtney Goeltzenleuchter <courtneygo@google.com>
Courtney Goeltzenleuchter bcdd4e58 2020-11-03T13:33:44 Vulkan: Fix validation errors with YUV Had several validation errors with YUV ImageTests using Vulkan back-end. This corrects those issues and tests now pass. Initializing YUV AHardwareBuffers requires Android API 29. The CI bots are not yet to Android 29 so this must be tested locally by compiling with the appropriate api level in args.gn: android32_ndk_api_level = 29 android64_ndk_api_level = 29 Test: ImageTestES3.SourceYUVAHBTargetExternalRGBSampleInitData Note: Some tests can run at api 26: ImageTestES3.ClearYUVAHB ImageTestES3.YUVValidation Bug: angleproject:4852 Bug: b/172649538 Bug: b/175021871 Change-Id: I383d24faa9911f77a71bd9b764232ad519b54e9a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2530454 Commit-Queue: Courtney Goeltzenleuchter <courtneygo@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Cody Northrop <cnorthrop@google.com>
Ian Elliott 1ad5791d 2020-12-03T22:05:51 Revert "Vulkan: Use specialization constant for halfRenderAreaWidth" This reverts commit ff38106d62fca2e36ec2db1cd580b546fd53ae43. Reason for revert: Suspect that this CL may be causing win-angle-rel-32 bot failures Original change's description: > Vulkan: Use specialization constant for halfRenderAreaWidth > > halfRenderArea is used by fragment shader to adjust gl_FragCoord for > rotation and yflip compensation. This CL bakes halfRenderAreaWidth into > shader via specialization constant, thus allow compiler to consolidate > the calculation into one MAD instruction. > > Bug: b/173800146 > Change-Id: Id66301278e3389e2582369b695825e632bccecee > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2551541 > Commit-Queue: Charlie Lao <cclao@google.com> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Reviewed-by: Tim Van Patten <timvp@google.com> > Reviewed-by: Jamie Madill <jmadill@chromium.org> TBR=timvp@google.com,syoussefi@chromium.org,jmadill@chromium.org,cclao@google.com # Not skipping CQ checks because original CL landed > 1 day ago. Bug: b/173800146 Change-Id: I7bc9966ebbe949f3999d7927afa629dd2e3a4187 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2572801 Reviewed-by: Ian Elliott <ianelliott@google.com> Reviewed-by: Yuly Novikov <ynovikov@chromium.org> Commit-Queue: Ian Elliott <ianelliott@google.com>
Shahbaz Youssefi 2c8bc1b7 2020-11-26T16:06:24 Vulkan: Clean up render pass query code In preparation for support of primitives generated queries. Bug: angleproject:5404 Change-Id: Ic123aaff33f8903994ff8fcc9158954ac023ec13 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2562126 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com>
Jamie Madill 720b85e6 2020-12-01T20:45:29 Reland "Vulkan: Use specialization constant for halfRenderAreaWidth" This reverts commit b1917fe1a314709cd9f7f53b2923d6be0e8bbf6c. Reason for revert: Not the root cause. See bug. Original change's description: > Revert "Vulkan: Use specialization constant for halfRenderAreaWidth" > > This reverts commit ff38106d62fca2e36ec2db1cd580b546fd53ae43. > > Reason for revert: Speculative revert to clear up AMD. See bug. > > Bug: chromium:1154356 > > Original change's description: > > Vulkan: Use specialization constant for halfRenderAreaWidth > > > > halfRenderArea is used by fragment shader to adjust gl_FragCoord for > > rotation and yflip compensation. This CL bakes halfRenderAreaWidth into > > shader via specialization constant, thus allow compiler to consolidate > > the calculation into one MAD instruction. > > > > Bug: b/173800146 > > Change-Id: Id66301278e3389e2582369b695825e632bccecee > > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2551541 > > Commit-Queue: Charlie Lao <cclao@google.com> > > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > > Reviewed-by: Tim Van Patten <timvp@google.com> > > Reviewed-by: Jamie Madill <jmadill@chromium.org> > > TBR=timvp@google.com,syoussefi@chromium.org,jmadill@chromium.org,cclao@google.com > > Change-Id: Idab612bcad4f179857c41ffc4bf8fa2355fae946 > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Bug: b/173800146 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2568244 > Reviewed-by: Jamie Madill <jmadill@chromium.org> > Commit-Queue: Jamie Madill <jmadill@chromium.org> TBR=timvp@google.com,syoussefi@chromium.org,jmadill@chromium.org,cclao@google.com # Not skipping CQ checks because this is a reland. Bug: chromium:1154356 Bug: b/173800146 Change-Id: I5b86bbee8600c6597620f31d8d7b08ca8b3d2d57 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2568246 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Jamie Madill b1917fe1 2020-12-01T20:42:36 Revert "Vulkan: Use specialization constant for halfRenderAreaWidth" This reverts commit ff38106d62fca2e36ec2db1cd580b546fd53ae43. Reason for revert: Speculative revert to clear up AMD. See bug. Bug: chromium:1154356 Original change's description: > Vulkan: Use specialization constant for halfRenderAreaWidth > > halfRenderArea is used by fragment shader to adjust gl_FragCoord for > rotation and yflip compensation. This CL bakes halfRenderAreaWidth into > shader via specialization constant, thus allow compiler to consolidate > the calculation into one MAD instruction. > > Bug: b/173800146 > Change-Id: Id66301278e3389e2582369b695825e632bccecee > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2551541 > Commit-Queue: Charlie Lao <cclao@google.com> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Reviewed-by: Tim Van Patten <timvp@google.com> > Reviewed-by: Jamie Madill <jmadill@chromium.org> TBR=timvp@google.com,syoussefi@chromium.org,jmadill@chromium.org,cclao@google.com Change-Id: Idab612bcad4f179857c41ffc4bf8fa2355fae946 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: b/173800146 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2568244 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Jamie Madill e56fd224 2020-12-01T09:44:38 Vulkan: Fix BufferHelper leaks in DynamicBuffer. We were leaking BufferHelper pointers in a couple cases. Fix these systematically by using a unique_ptr template. This leak was detected running ANGLE with LSAN enabled. Bug: angleproject:5377 Change-Id: I30ab235105cf74fae2cfc62a22dde534e2d07e81 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2567641 Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Alexis Hétu <sugoi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Charlie Lao ff38106d 2020-11-19T16:47:25 Vulkan: Use specialization constant for halfRenderAreaWidth halfRenderArea is used by fragment shader to adjust gl_FragCoord for rotation and yflip compensation. This CL bakes halfRenderAreaWidth into shader via specialization constant, thus allow compiler to consolidate the calculation into one MAD instruction. Bug: b/173800146 Change-Id: Id66301278e3389e2582369b695825e632bccecee Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2551541 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill bccaaffd 2020-11-27T16:33:09 Vulkan: Use XFB queries with the XFB extension. This will enable accurate XFB primitive counts when using tessellation and geometry shaders. Adds new vk::QueryResult and gl::QueryTypeMap helper classes. Based on contributions by Mohan Maiya (m.maiya@samsung.com). Bug: angleproject:3572 Change-Id: Ie3f496deda887c13bb4ad7ab430e31d615849bfd Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2564002 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Mohan Maiya <m.maiya@samsung.com>
Shahbaz Youssefi a5da17cd 2020-11-26T17:09:37 Vulkan: Remove vk:: inside namespace vk Bug: angleproject:5404 Change-Id: I51df1ad69a65b17621a3cbe4e5d55621cc9ae6ec Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2562683 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 5850c748 2020-11-20T22:14:27 Vulkan: Emulated RGB copies in compute The copy between emulated RGB formats can take a number of paths: - Sample from src (reinterpreted as UINT), output to dst - Sample from src, output to temp buffer, copy to dst - Copy src to temp buffer, output to dst - Copy src to temp buffer, convert to another temp buffer, copy to dst While directly sampling from src and outputting to dst is more efficient, these are not always possible. The former may not have SAMPLED_IMAGE usage bit for the reinterpreted UINT format, and the latter may not have STORAGE_IMAGE usage at all. This change takes the universal approach of using two temp buffers. The ConvertVertex shader is used to transform between RGB and RGBA when copying from the first temp buffer to the second. Bug: angleproject:5278 Change-Id: I63d916cfdb4c389f5b817d89cd7348fdea703ce5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2556467 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Brandon Schade <b.schade@samsung.com>
Shahbaz Youssefi 46107d3e 2020-11-18T22:03:22 Vulkan: Delay defining D/S content to endRenderPass Take the following situation: 1. Start RP with D/S undefined: loadOp = DONT_CARE, storeOp = STORE * At this point, onDepthStencilWrite calls image->onWrite, setting depth/stencil contents defined. 2. At endRP, observe depth/stencil is not used: storeOp = DONT_CARE 3. Start another RP with D/S: loadOp = LOAD, storeOp = STORE Because the call to image->onWrite was done at startRP, the contents of the depth/stencil image is marked as defined, and the next render pass is loading these data. This change moves image->onWrite to endRenderPass, and only calls it if storeOp = STORE, taking advantage of all the opportunistic optimizations that try to set storeOp to another value. Bug: angleproject:4836 Change-Id: I9858e5caa6b1f67f841a5c6356e66927356ef469 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2548319 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi b6ec24ab 2020-11-14T00:24:52 Vulkan: Support format reinterpretation in imageBuffers This is done by creating multiple buffer views over the buffer based on the format specified by the attached shaders. Bug: angleproject:3573 Change-Id: I0372a988938050cc092d8a0959a59d1c893fc6f4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2536909 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Shahbaz Youssefi 61a117fa 2020-11-16T12:48:11 Vulkan: Fix aliasing format for non-layered image bindings This was correctly handled for layered bindings by the change anglebug.com/3885#c20. This change uses the same mechanism for non-layered bindings. Bug: angleproject:5347 Change-Id: Ida55b3589d0bdbe5d824c1614c9bf9f867a7f652 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2541182 Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Mohan Maiya f197ebac 2020-11-16T15:16:56 Vulkan: Add EXT_external_buffer support Addition of buffer support for external memory Also adds new end2end tests for these usecases * SubData update * map/unmap buffer * dispatch compute with external buffer Bug: angleproject:5073 Test: ExternalBufferTestES31.*Vulkan Change-Id: Ib3cccaca77b76830effe49d3731782552e7424ec Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2525105 Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 86d7e4d8 2020-11-10T15:55:34 Vulkan: Support texture buffers Texture buffers are placed in the same descriptor set with the rest of the textures. However, the different code paths that handle textures have special cases for texture buffers as they create a different descriptor type (texel buffer instead of combined image sampler). Image view serials are used to track the buffer view serials as well so the texture descriptor cache can handle texture buffers as well. This CL is missing storage texel buffer support. Bug: angleproject:3573 Change-Id: Iff80ca22ff9b9957a0c9a3c7aaada1fa54b24ec8 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2532653 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Shahbaz Youssefi 60641abc 2020-11-16T16:12:31 Vulkan: Make ImageViewHelper a Resource Bug: angleproject:3573 Change-Id: I12e70418b3b971e802bc911409e170cbf8c61915 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2542223 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill cdf280c0 2020-11-13T10:14:58 Vulkan: Fix serial handling for queries Previous serial mechanism wasn't reliable. QueryHelper is now a vk::Resource and uses the vk::Resource lifecycle. Also added some utility methods to QueryVk to deal with stashed queries. Any question we want to ask about mQueryHelper we want to ask about the stashed queries as well. Bug: b/169055809 Bug: b/169788986 Bug: b/170312581 Change-Id: Ia34a7a433e61a2543cfb09491ebab55b054a26c9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2537718 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Courtney Goeltzenleuchter <courtneygo@google.com>
Shahbaz Youssefi 9e7f08fc 2020-11-06T16:55:51 Vulkan: Fix incorrect reordering of barriers Take the following situation, a simple copy from image to buffer: ANGLE_TRY(contextVk->onBufferTransferWrite(buffer)); ANGLE_TRY(contextVk->onImageTransferRead(aspectFlags, image)); CommandBuffer &commandBuffer = contextVk->getOutsideRenderPassCommandBuffer(); commandBuffer.copyImageToBuffer(imageHandle, layout, bufferHandle, 1, regions); Both `onBufferTransferWrite` and `onImageTransferRead` may flush either the outsideRP or insideRP command buffers. If buffer is not previously used, but image is used: - onBufferTransferWrite: buffer usage is recorded in outsideRP1 - onImageTransferREad: outsiderRP1 is flushed, outsideRP2 is started - copyImageToBuffer: recorded on outsideRP2, but buffer usage not recorded there - A following command that uses the buffer and requires barrier doesn't close outsideRP2 as it believes it was not used there Bug: angleproject:5319 Change-Id: Ib8994083fbc21969a538cda3784adee57b089415 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2523388 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Jamie Madill f19a4a20 2020-11-08T10:37:33 Vulkan: Move CommandBuffer management to RendererVk. This consolidates all relevant logic in a single place. We no longer need to interact with ContextVk in the worker thread. This switches the fixed pointer array size to a dynamically sized vector. Some of the EGL and ANGLE tests would use a large number of Contexts and we were consistently running out of available command buffers which would cause a deadlock situation. We can trust other parts of the code to throttle the application if it starts to get too far ahead of the device and dispense with the hard coded limit in the command buffer allocator itself. The resulting code is also quite a bit simpler and doesn't need a condition variable. Also fixes missing initialization in SecondaryCommandBuffer. Bug: b/172704839 Change-Id: Icc3a3daf5d6b272db556c0e4c93fb793583966a5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2525143 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Charlie Lao b22f8e8e 2020-10-29T15:50:39 Vulkan: Add specialization constants for surface rotation This plumbing through the specialization constant for surface rotation from ContextVk to pipeline program creation. It has not been used yet, so expecting no real functional change. This CL also converts lineRasterEmulation to use the same specialization constant path as surface rotation. Bug: b/171750979 Change-Id: Ic08c4f8bb576424d1752015e874d0977a58d78bb Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2508837 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Ian Elliott <ianelliott@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill c8a7c033 2020-11-07T09:12:47 Vulkan: Move primary command buffer to CommandQueue. This brings CommandQueue one step closer to the worker thread. Bug: b/172704839 Change-Id: I35225c5f302e34e3feb38a35c22b6d8e6f4ad10c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2524543 Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com> Reviewed-by: Tim Van Patten <timvp@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Stephan Hartmann eaab2e34 2020-11-09T11:24:09 Vulkan: add explicit template specialization for barrierImpl ImageHelper::barrierImpl with secondary command buffer is used in other translation units. Linking fails if compiler has optimized out symbol for secondary command buffer. Bug: angleproject:5324 Change-Id: I588d0d949a9b2eee04d8103effc485ab23d679c6 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2526082 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Mohan Maiya 18fcb4ad 2020-11-06T07:22:17 Vulkan: Enable a subset of ImageTest colorspace tests This patch does the following - 1. Expose EGL_KHR_image_gl_colorspace only if the underlying ICD supports VK_KHR_image_format_list extension 2. Create VkImageViews from EGLimage's format 3. Fix bugs in ImageTest colorspace tests Bug: angleproject:3756 Bug: angleproject:5281 Test: angle_end2end_tests.exe --gtest_filter=ImageTest*Colorspace*Vulkan Change-Id: I618ae009e83d0a6b27a0a50165f736784717e404 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2523670 Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Mohan Maiya 55934550 2020-11-05T07:13:39 Vulkan: Don't refresh image views unnecessarily If the underlying ICD supports VK_KHR_image_format_list extension we will have created both the linear and sRGB image views, reuse them instead of refreshing them. Bug: angleproject:3609 Bug: angleproject:4561 Bug: angleproject:5281 Change-Id: Ie34f08b3554a07f94911bebf925e38d702bc3c36 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2518387 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Shahbaz Youssefi d77e85a8 2020-11-03T22:23:18 Vulkan: Store image updates per level This optimization allows iterating only over updates of a certain level or range of levels, instead of having to iterate over every update and filter out the ones matching the desired level(s). Bug: angleproject:4891 Change-Id: Ied04f4b28f05d37b9add61c7f4d54cc328c0be86 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2519095 Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Brandon Schade 1e494cae 2020-10-07T19:26:01 Vulkan: Add support for EXT_copy_image Add support for GL_EXT_copy_image which allows image data transfer between image objects. This is implemented by using the vkCmdCopyImage API call. Bug: angleproject:3593 Test: dEQP-GLES31.functional.copy_image.* Change-Id: I30a34a8711b5d2e5834064d7453e03d6ec0df478 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2393955 Commit-Queue: Brandon Schade <b.schade@samsung.com> Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi a0e91016 2020-10-30T10:01:36 Vulkan: Don't break the render pass on scissor change Prior to this change, the render area was decided when the render pass was started, and remained fixed. If a small scissor was initially used, this created a render pass with a small area. If then the scissor region was expanded, the render pass was broken. This change instead expands the render area on scissor change to avoid breaking the render pass. If glInvalidateSubFramebuffer previously successfully resulted in storeOp=DONT_CARE, this optimization may need to undo that. As a result, the invalidate area is stored in the render pass and if the render area grows beyond that, invalidate is undone. Bug: angleproject:4988 Change-Id: I4e8039dec53a95a193a97cb40db3f71e397568d6 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2508983 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 4989667a 2020-11-03T16:28:37 Vulkan: Make ImageHelper::flushStagedUpdates take gl::LevelIndex More consistent with the rest of the public functions in this class. Bug: angleproject:4891 Change-Id: I72c7c3d21f8287cc1711772671c6e7ad524fd096 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2518179 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 8afe3f17 2020-11-03T10:55:31 Vulkan: Fix use of uninitialized data in staged clears When depth/stencil data are staged, only the depthStencil field of VkClearValue is initialized. However, when comparing staged clears, memcmp is used which also compares the extra bytes in the aliasing color field. Bug: chromium:1144491 Change-Id: Ic384ba792e9fd199d8e9c3e534ccdc6ea65ee9b8 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2517244 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Courtney Goeltzenleuchter 6bc362c4 2020-10-18T19:09:47 Allow single threaded CommandProcessor In replacing the legacy CommandQueue code with the threading capable CommandProcessor it would be good to be able to run the CommandProcessor in a single-thread environment. This CL changes the meaning of the feature flags for the commandProcessor and asynchronousCommandProcessing so that enabling commandProcess only changes the code paths to use the command processor but work it still done as part of the submitting thread (e.g. ContextVk). Enabling asynchronousCommandProcessing will cause a separate worker thread to be spawned which will asynchronously process the commands. This allows us to switch to the CommandProcessor without threading and then enable threading once performance issues are resolved. Bug: b/161912801 Bug: b/170329600 Change-Id: I534862b109a7e7708108190b7c3e894071d4c2ed Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2483580 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Commit-Queue: Courtney Goeltzenleuchter <courtneygo@google.com>
Mohan Maiya 3fcf17e6 2020-11-02T12:51:35 Vulkan: Bug fix in sRGBDecode logic When a sampler's GL_TEXTURE_SRGB_DECODE_EXT state was toggled between GL_DECODE_EXT and GL_SKIP_DECODE_EXT VkImageViews of the TextureVk object were not being updated. Add sRGB_decode state as part of ImageViewSubresourceSerial so we retrieve the correct VkImageView from the texture cache. Bug: angleproject:3609 Tests: angle_end2end_tests --gtest_filter=SRGBTextureTestES3.SRGBDecodeSamplerParameterToggle*Vulkan Change-Id: I897e461957d408b5a5b4f03fefc05f2e9684c7b7 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2514900 Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill 97843bda 2020-10-30T17:03:36 Vulkan: Fix EGL Surface robust init. The error here was related to using a single cache variable for the robust init setting for all the surfaces in a DisplayVk. Fix this by passing down the robust init setting from the SurfaceVk to image init. Bug: angleproject:5274 Change-Id: I9bc9c20990268d1d5166411fb53f8f2593fd1971 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2510694 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Tim Van Patten fcd35965 2020-09-29T14:15:51 Vulkan: Make DescriptorPoolHelper a Resource Descriptor pools need to live as long as the descriptor sets that are allocated from them. Using Serials while building a command to judge a pool's lifetime is prone to errors, since a command's Serial value isn't known until the command is submitted, leading to deleting pools too early relative to when the descriptor set is actually used. This CL updates DescriptorPoolHelper to inherit from Resource, so the descriptor pools can be retain()'ed. This allows the Resource's counter to indicate that a pool is in use until the command's Serial is known and can be recorded to indicate when the command completes. This prevents descriptor pools from being destroyed before the command completes (while the descriptor sets are still in use), or even before the command has been submitted. Destroying a descriptor pool resets all of the descriptors that were allocated from it, which can trigger a variety of VVL errors depending on when it's erroneously performed. This CL also adds the necessary retain() calls for the descriptor pools. In particular, the pools need to be retained each time a cached descriptor set that was allocated from it is re-used. This is relatively simple with the current design, since we always clear the descriptor set caches whenever a new pool is allocated, so the descriptor pool binding is always accurate. Bug: angleproject:5030 Test: VulkanMultithreadingTest::MultiContextDrawSmallDescriptorPools() Change-Id: Iac9e7efef338f169a6bf8ac3b2140e03dd326641 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2504457 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com> Commit-Queue: Tim Van Patten <timvp@google.com>
Jamie Madill 9e3eec54 2020-10-25T15:44:09 Revert "Vulkan: Make DescriptorPoolHelper a Resource" This reverts commit 5dcd29a6e532e4bd617af8767d488120b57f3b2c. Reason for revert: Breaking the ANGLE -> Chromium roller: https://chromium-review.googlesource.com/c/chromium/src/+/2496281 Original change's description: > Vulkan: Make DescriptorPoolHelper a Resource > > Descriptor pools need to live as long as the descriptor sets that are > allocated from them. Using Serials while building a command to judge a > pool's lifetime is prone to errors, since a command's Serial value > isn't known until the command is submitted, leading to deleting pools > too early relative to when the descriptor set is actually used. > > This CL updates DescriptorPoolHelper to inherit from Resource, so the > descriptor pools can be retain()'ed. This allows the Resource's counter > to indicate that a pool is in use until the command's Serial is known > and can be recorded to indicate when the command completes. This > prevents descriptor pools from being destroyed before the command > completes (while the descriptor sets are still in use), or even before > the command has been submitted. Destroying a descriptor pool resets all > of the descriptors that were allocated from it, which can trigger a > variety of VVL errors depending on when it's erroneously performed. > > This CL also adds the necessary retain() calls for the descriptor pools. > In particular, the pools need to be retained each time a cached > descriptor set that was allocated from it is re-used. This is relatively > simple with the current design, since we always clear the descriptor set > caches whenever a new pool is allocated, so the descriptor pool binding > is always accurate. > > Bug: angleproject:5030 > Test: MultithreadingTest::MultiContextDrawSmallDescriptorPools() > Change-Id: I5fdeeb46159448dfd679d7169e423048348be5ab > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2437609 > Commit-Queue: Tim Van Patten <timvp@google.com> > Reviewed-by: Jamie Madill <jmadill@chromium.org> > Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com> TBR=courtneygo@google.com,timvp@google.com,jmadill@chromium.org # Not skipping CQ checks because original CL landed > 1 day ago. Bug: angleproject:5030 Change-Id: I0fd6d9a0e1b0989b22368ef98652281288699deb Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2497222 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 33e30205 2020-10-17T19:03:45 Vulkan: sRGB cleanups. A few fixes to how we check for the sRGB override in the TextureVk class. In at least one instance there was a potential edge case where in syncState we might not create the Texture with the mutable bit the second time through the function. Bug: angleproject:5176 Change-Id: I4f1ca6e469b10514c3a0de3120be9ade62568084 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2482292 Reviewed-by: Mohan Maiya <m.maiya@samsung.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Tim Van Patten 5dcd29a6 2020-09-29T14:15:51 Vulkan: Make DescriptorPoolHelper a Resource Descriptor pools need to live as long as the descriptor sets that are allocated from them. Using Serials while building a command to judge a pool's lifetime is prone to errors, since a command's Serial value isn't known until the command is submitted, leading to deleting pools too early relative to when the descriptor set is actually used. This CL updates DescriptorPoolHelper to inherit from Resource, so the descriptor pools can be retain()'ed. This allows the Resource's counter to indicate that a pool is in use until the command's Serial is known and can be recorded to indicate when the command completes. This prevents descriptor pools from being destroyed before the command completes (while the descriptor sets are still in use), or even before the command has been submitted. Destroying a descriptor pool resets all of the descriptors that were allocated from it, which can trigger a variety of VVL errors depending on when it's erroneously performed. This CL also adds the necessary retain() calls for the descriptor pools. In particular, the pools need to be retained each time a cached descriptor set that was allocated from it is re-used. This is relatively simple with the current design, since we always clear the descriptor set caches whenever a new pool is allocated, so the descriptor pool binding is always accurate. Bug: angleproject:5030 Test: MultithreadingTest::MultiContextDrawSmallDescriptorPools() Change-Id: I5fdeeb46159448dfd679d7169e423048348be5ab Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2437609 Commit-Queue: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Courtney Goeltzenleuchter ed876984 2020-10-03T11:00:36 Vulkan: functionally complete worker thread Working on enhancing worker thread to completely own primary command buffers. This will include not only processing SCBs from main thread into a primary, but also submitting those command buffers to the queue. The CommandProcessor is a vk::Context so it can handle errors in the worker thread. When the main thread submits tasks to the worker thread it also syncs any outstanding errors from the worker. Include asynchronousCommandProcessing feature that will control whether the worker thread task does it's work in parallel or not. If false, we wait for the thread to complete it's work before letting the main thread continue. If true, the thread can execute in parallel with the main thread. Bug: b/154030730 Bug: b/161912801 Change-Id: I00f8f013d6cbb2af12a172c4f7927855db2f0ebf Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2328992 Commit-Queue: Courtney Goeltzenleuchter <courtneygo@google.com> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 2072ce50 2020-10-18T16:44:09 Rename "NonLinear" to "SRGB". In OpenGL parlance there are really only two color spaces: Linear and SRGB. Vulkan is much more complicated with linear and nonlinear SRGB, and multiple non-SRGB linear color spaces. Keep things simple by working with OpenGL nomenclature. Bug: angleproject:5176 Change-Id: Ia730a9a666a2b3325194b86dc588e7fd226c4183 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2483466 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Mohan Maiya 7bbe497a 2020-10-16T14:46:45 Vulkan: Implement EXT_texture_sRGB_decode Implement EXT_texture_sRGB_decode. This builds on the existing functionality from EXT_texture_sRGB_override, with 2 major edge cases: 1. sRGB_decode allows the texture state to be overridden by sampler state, which is implemented by forcing a a texture state sync during updateActiveTextures if a texture is bound to the same unit as a sampler with that state 2. texelFetch calls require us to reenable decoding, regardless of decode state. We add a new compiler pass (FlagSamplersWithTexelFetch) to mark samplers that are used with texelFetch in order to support this. This change also re-enables EXT_texture_sRGB_R8, which was disabled due to a dEQP bug that this change will bypass. Bug: angleproject:3609 Bug: angleproject:4503 Test: dEQP.GLES31/functional_srgb_texture_decode_skip_decode_* Test: GLES31/functional_state_query_texture_*_srgb_decode_* Test: GLES31/functional_state_query_sampler_*_srgb_decode_* Test: GLES31/functional_debug_negative_coverage_*_srgb_decode_* Test: GLES31/functional_android_extension_pack_extensions_ext_texture_srgb_decode Test: angle_end2end_tests --gtest_filter=SRGBTextureTest.*Vulkan* Change-Id: I4a67e487dc82e2f57c8c87d4bcd8ef442b6fe220 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2359481 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Shahbaz Youssefi 2aaeb81d 2020-10-06T17:20:29 Vulkan: loadOp=DONT_CARE + unused D/S => invalidate D/S If depth/stencil is not loaded, and it's not written to during the render pass, then treat it as if it was invalidated so storeOp can be set to DONT_CARE and its corresponding resolve attachment (if any) removed. This is especially useful for MSRTT as a resolve attachment is added at the start of render pass, and this optimization will give it a chance to undo that if depth/stencil was not actually used in the render pass. This situation can arise for example if a render pass is created for the sole purpose of clearing color. This change includes a bug fix for missing depth/stencil on*Access in the UtilsVk blit/resolve path. Bug: angleproject:4836 Change-Id: Ifc8eea3e6ffb3eb4bba19f03d1358f151ec69c44 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2453468 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Hyunchang Kim d9a88d63 2020-09-28T17:58:31 Vulkan: Support OES_texture_cube_map_array extension Add support for GL_OES_texture_cube_map_array to the Vulkan backend if the underlying ICD supports creation of of cube array ImageViews Bug: angleproject:3584 Test: dEQP-GLES31.functional.fbo.color.texcubearray.* Change-Id: I636cbf347d718abfc1600119b312578370b8e02e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2437989 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Charlie Lao df6b7298 2020-10-12T13:39:09 Vulkan: Use StoreOp_None_QCOM for read only depth stencil buffer For read only depth stencil buffers, there is no need to store depth or stencil value. But we can not use DontCare for storeOp because vulkan core spec says DontCare indicates data is undefined after this. VK_QCOM_render_pass_store_ops extension introduces a new store op that will leave data defined but skip the store. This CL utilize this if the extension is available. Bug: angleproject:5055 Change-Id: I104f3d01eb342a2d0cc900f342430e901bde1bff Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2462604 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 68bd685a 2020-10-10T22:58:41 Reland: "4 Vulkan content defined CLs." Reland "Vulkan: Avoid content restore by detecting no-op stencil" This relands commit 243d0f899e443cd931c78aba7489382dff79edbb. Reland "Vulkan: Restore at the end of RP if write-after-invalidate" This relands commit e5d52ac3b9a00656acdd912ee8cd62dd14784075. Reland "Vulkan: Invalidate/restore depth/stencil separately." This relands commit 61fa0878964a796f6d3b3c13bc3a3849403ecdbd. Reland "Vulkan: Move content-defined tracking to ImageHelper" This relands commit 2392e6b34c0ddfbfd7b4c3cb67323ba463e11a57. Reason for revert: Caused crashes in Fuchsia x64 and on ARM. Reland fixes content defined for external images. Original CL message: Content-defined tracking was done in render targets prior to this change. This had multiple drawbacks: - When a framebuffer attachment is changed (including the first time it's set), it's unknown whether the contents of the attachment is defined. - Invalidate takes effect at the end of render pass, at which point the render target objects may be gone. Attachment ImageHelpers are however correctly tracked. This change moves content-defined tracking to the ImageHelper itself, and tracks it per subresource. ImageHelper::onWrite() now receives the subresource that is being written, and marks it as having defined content. A future optimization can make use of this change to ImageHelper::onWrite to track "dirty" subresources. This can lead to the removal of unnecessary barriers when same-kind writes are done on different subresources of the image. See http://anglebug.com/3347#c15 Bug: b/167275320 Bug: angleproject:4836 Bug: angleproject:5159 Change-Id: If5c1ae7152657fd7c94db7d55bea4fb9ddf835ba Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2464825 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill d6b1c17b 2020-10-10T14:29:15 Revert 4 Vulkan content defined CLs. Revert "Vulkan: Avoid content restore by detecting no-op stencil" This reverts commit 243d0f899e443cd931c78aba7489382dff79edbb. Revert "Vulkan: Restore at the end of RP if write-after-invalidate" This reverts commit e5d52ac3b9a00656acdd912ee8cd62dd14784075. Revert "Vulkan: Invalidate/restore depth/stencil separately." This reverts commit 61fa0878964a796f6d3b3c13bc3a3849403ecdbd. Revert "Vulkan: Move content-defined tracking to ImageHelper" This reverts commit 2392e6b34c0ddfbfd7b4c3cb67323ba463e11a57. Causes crashes in Fuchsia x64 and on ARM. Original CL message: Content-defined tracking was done in render targets prior to this change. This had multiple drawbacks: - When a framebuffer attachment is changed (including the first time it's set), it's unknown whether the contents of the attachment is defined. - Invalidate takes effect at the end of render pass, at which point the render target objects may be gone. Attachment ImageHelpers are however correctly tracked. This change moves content-defined tracking to the ImageHelper itself, and tracks it per subresource. ImageHelper::onWrite() now receives the subresource that is being written, and marks it as having defined content. A future optimization can make use of this change to ImageHelper::onWrite to track "dirty" subresources. This can lead to the removal of unnecessary barriers when same-kind writes are done on different subresources of the image. See http://anglebug.com/3347#c15 Bug: b/167275320 Bug: angleproject:4836 Bug: angleproject:5159 Change-Id: I93d9dfe973caa7ce70aefa46b5b7d04a8637efb3 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2464822 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi e5d52ac3 2020-10-08T14:26:22 Vulkan: Restore at the end of RP if write-after-invalidate If a depth/stencil attachment is invalidated, but subsequently drawn to in the same render pass, undo the invalidate when the render pass is closed. Adapted from https://chromium-review.googlesource.com/c/angle/angle/+/2386478. Bug: b/167275320 Bug: angleproject:4836 Change-Id: I17a35bfd692ddc403ceaa6ec44b5c4f16ff9eed6 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2461464 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Ian Elliott <ianelliott@google.com>
kmx.io     mail     discord kmxgit v0.4.0