|
2611f98b
|
2020-06-02T09:52:27
|
|
Vulkan: Add IMAGE_USE_STORAGE bit only if necessary.
The VK_IMAGE_USAGE_STORAGE_BIT has negative performance impact. Right
now we needed for overlay widgets. This CL will only add the bit if we
actually have widget enabled.
Bug: b/157774833
Change-Id: I3027df886c9b34ccfd667152fa4fb090dfadb45b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2225810
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Charlie Lao <cclao@google.com>
|
|
9047ce5b
|
2020-05-26T11:23:35
|
|
Vulkan: Fix bad delete of swapchain.
We could end up deleting a dirty pointer if the driver set an
output value during a call to vkCreateSwapchainKHR that generates
an error. In practice this turned up in some edge cases with
SwiftShader. Fix this by using a temporary variable.
Bug: chromium:1079475
Change-Id: Ia40e1fac94cbd56fd3bb800a179bb38b7f11fa26
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2216292
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
6de7ee52
|
2020-05-25T17:00:01
|
|
Clean up overlay RenderPass count reporting.
This fixes the trace perf test to accurately report how many
RPs in each frame. Instead of counting the RPs on a flush we now
count only on a swap call. This won't work for offscreen surfaces
which is fine - the overlay doesn't really have the same use for
offscreen rendering. Also ignores the first frame in graph data so
we can ignore the first setup frame in the trace tests.
Also skips the redundant extra "flush" call that would generate an
empty space in the RP graph.
Gives a cleaner measurement for optimizing the XFB RP count.
Bug: angleproject:4622
Change-Id: I5762c500cdb216700247095984ae62b4f8741602
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2215309
Reviewed-by: Tim Van Patten <timvp@google.com>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
9823d86a
|
2020-05-11T13:13:41
|
|
Fix EGLSurfaceTest.ResizeWindow on Fuchsia
A VkSurfaceKHR may not have an intrinsic size as is the case on Fuchsia.
This is broken since 718ae5088 ("Vulkan: Always query EGL_WIDTH and
EGL_HEIGHT"); queries for EGL_WIDTH or EGL_HEIGHT now always return -1.
Switch back to the old behavior in this case.
Bug: angleproject:4624
Change-Id: I7e7bf569db9aec9890b2cb184056be5a6031bd98
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2191173
Reviewed-by: Ian Elliott <ianelliott@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Michael Spang <spang@chromium.org>
|
|
|
d657e1d7
|
2020-04-24T13:13:18
|
|
Vulkan: Defer framebuffer clears.
This works by storing the deferred clears in the ImageHelper's staging
buffers. We apply the deferred clears onto the RenderPass right before
we begin to draw. Storing the clears in the ImageHelper solves problems
where we clear GL Textures in a Framebuffer and then unbind the
Textures and sample from them. Or do other commands like CopyTexImage.
Note that because the staging buffer clears only handle full-image
clears we need to immediately apply some scissored clears where before
we would use the RP. This should be a pretty rare occurrence and it is
possible to optimize that in the future.
Reduces the RenderPass count in the Manhattan "frame 10" trace from max
22 to max 20. May improve perf slightly on Android or may have effects
too small to measure. Should not regress performance.
Bug: angleproject:4517
Change-Id: I02150d531022afb903f1058f070937ec6337bd88
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2142711
Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
718ae508
|
2020-04-28T08:42:07
|
|
Vulkan: Always query EGL_WIDTH and EGL_HEIGHT
This fixes the failures of the dEQP EGL resize tests on Android.
Those tests don't actually resize the window (which would allow the
tests to pass), but do change internal Android Surface values that
vkGetPhysicalDeviceSurfaceCapabilitiesKHR queries. Therefore, by
freshly querying these values the test passes.
Bug: b/153329980
Change-Id: Ie966f221bfaa14988c1503c7f5503f0b21476bcb
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2165639
Reviewed-by: Ian Elliott <ianelliott@google.com>
Commit-Queue: Ian Elliott <ianelliott@google.com>
|
|
|
1fc354fa
|
2020-04-22T11:49:00
|
|
Re-land: "Vulkan: Forward RenderBuffer/Surface dirty messages."
Re-land fixes ASSERT on Android GLES.
Previously we were only forwarding Texture staging changes to the
Framebuffer for re-sync. We could miss out on clears that were staged
in Surfaces and Renderbuffer attachments. Complete the chain by making
the SurfaceVk and RenderbufferVk forward notifications from the
VkImage.
Bug: angleproject:4517
Change-Id: I91d31213b115a58546b574caa574eebbb098266c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2159294
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
623d0443
|
2020-04-21T15:15:03
|
|
Revert "Vulkan: Forward RenderBuffer/Surface dirty messages."
This reverts commit 3fb33ade6ec8f286797bff5645304df580b1a963.
Reason for revert: Causing WebGL test failures on GLES/ANGLE/Android.
Bug: chromium:1072739
Original change's description:
> Vulkan: Forward RenderBuffer/Surface dirty messages.
>
> Previously we were only forwarding Texture staging changes to the
> Framebuffer for re-sync. We could miss out on clears that were staged
> in Surfaces and Renderbuffer attachments. Complete the chain by making
> the SurfaceVk and RenderbufferVk forward notifications from the
> VkImage.
>
> Bug: angleproject:4517
> Change-Id: I15f36c225517e915f656d0c2464a99631ef9f945
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2150699
> Reviewed-by: Cody Northrop <cnorthrop@google.com>
> Commit-Queue: Jamie Madill <jmadill@chromium.org>
TBR=cnorthrop@google.com,timvp@google.com,jmadill@chromium.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: angleproject:4517
Change-Id: I4fd4d79dc11da0ba86065d3a80975de4c791a2a0
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2159288
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
f13689b6
|
2020-04-17T14:43:28
|
|
Vulkan: Use renderpass' finalLayout to transit to ImageLayout::Present
This detects the last renderpass is actually rendering to the image we
are going to present and this is the very last renderpass to that image.
We use finalLayout of the renderpass to do layout transition to
VK_IMAGE_LAYOUT_PRESENT_SRC_KHR instead of a standalone barrier call.
Bug: b/153885625
Change-Id: Ifc17c77a1ba12f22c9d6038775f3e254c35655ca
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2155250
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
3fb33ade
|
2020-04-09T15:27:12
|
|
Vulkan: Forward RenderBuffer/Surface dirty messages.
Previously we were only forwarding Texture staging changes to the
Framebuffer for re-sync. We could miss out on clears that were staged
in Surfaces and Renderbuffer attachments. Complete the chain by making
the SurfaceVk and RenderbufferVk forward notifications from the
VkImage.
Bug: angleproject:4517
Change-Id: I15f36c225517e915f656d0c2464a99631ef9f945
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2150699
Reviewed-by: Cody Northrop <cnorthrop@google.com>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
2a0c3596
|
2020-03-31T15:36:45
|
|
Vulkan: Clean up robust and emulated image clears.
This change consolidates image clears in multiple places into a single
site in ImageHelper initialization. It adds support for appending clear
image commands as well as prepend (the default). We prepend clears
because image initialization happens after data upload.
The Vulkan robust clear path now works like the other back-ends. The
change flushed out a bug where partially uninitialized CopyTexImage was
not correctly initializing a texture before triggering a full resource
clear. Texture::copyImage now uses a workaround where we first init the
image before clearing it. After the init we upload the new data.
We'll use the appending clears path when implementing deferred clears.
Bug: angleproject:4517
Change-Id: If9212f3b8cdd0fc8b7e729d364530801a644e164
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2130627
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
|
|
|
a30c6f35
|
2020-04-07T11:25:20
|
|
Vulkan: Pass RendererVk to BufferHelper::destroy.
Bug: angleproject:2162
Change-Id: I7c0090d48b0a5bac222b1edf4c2db280d5b32930
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2139993
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Mohan Maiya <m.maiya@samsung.com>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
|
|
b07816d6
|
2020-01-02T15:13:37
|
|
Vulkan: Add robust init for NULL texture image.
In gl::ReadPixels(), ValidateReadPixels() will try to
flush staged updates before readPixels. In the case
where a texture was initialized with null RGBA data,
no Framebuffer dirty bits are set, and thus the staged
clear would never be flushed from the staged updates.
1. Add robust init in TextureVk::initImage to ensure
image is initialized.
Test: ./angle_end2end_tests --gtest_filter=\
RobustResourceInitTest*_Vulkan_AllocateNonZeroMemory
2. Update stageSubresourceRobustClear() to
kEmulatedInitColorValue in the case where robust resource
was initialized without full RGBA format to update
init value of robust resource.
Test: texture-attachment-formats.html in
webgl_conformance_vulkan_passthrough_tests.
3. Revert "Suppress Vulkan RobustResourceInit tests."
Revert commit a8e6a463121b9920564fb6efd7f2c857d7023fe3.
Reason for revert: re-enable related robust tests.
Bug: angleproject:4255
Change-Id: I79f20e0c02c2f1b1cd68ab590f0f765229f9e780
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1985503
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
e488d8b8
|
2020-02-26T16:11:29
|
|
Vulkan: Implement Android pre-rotation
As an Android GLES driver on top of Vulkan, ANGLE must pre-rotate
rendering on behalf of the application. This involves modifying the
vertex shader to multiply gl_Position with a mat2 "rotation matrix".
Not doing so means that SurfaceFlinger (SF) will perform a costly
rotation blit before presenting every image. Setting
WindowSurfaceVk::mPreTransform to mCurrentTransform tells SF to not do
the blit.
When the surface is rotated 90 or 270 degrees, the width and height
must be swapped for:
- The swapchain images, and for any depth, stencil, and/or multisample
attachments used with the swapchain image.
- The viewport, scissor, and render area.
Because the Metal back-end shares the TranslatorVulkan, it will define
the same preRotation (mat2) DriverUniform that is used for Vulkan.
Bug: angleproject:3502
Change-Id: I968dbe8869ba0f50de18dd41f1195e847c06b545
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2038272
Commit-Queue: Ian Elliott <ianelliott@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
aa09ca69
|
2020-02-24T13:33:30
|
|
Vulkan: Update ContextVk command buffer access methods.
Use the following names:
- flushAndBeginRenderPass
- endRenderPassAndGetCommandBuffer
- flushAndGetPrimaryCommandBuffer
This clarifies that a flush is part of the method.
Bug: angleproject:4029
Change-Id: I6c870761339ea7aa39c83142200781ba39ad6a4b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2068129
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Tim Van Patten <timvp@google.com>
|
|
|
d3aad0f4
|
2020-02-20T08:59:58
|
|
Vulkan: Remove command graph size overlay widget.
No longer applicable without the command graph.
Bug: angleproject:4029
Change-Id: If75e3aca56eb18567074e14ceb23b3fd7a0afb3d
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2065919
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
c58458e6
|
2020-02-19T14:51:41
|
|
Vulkan: Remove CommandGraph code.
Also updates relevant comments to no longer refer to a graph.
Bug: angleproject:4029
Change-Id: Ic29716e9ae4926870f902947d49d8fee7af98662
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2057804
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
|
|
|
ce4918f1
|
2020-02-19T09:39:44
|
|
Vulkan: Sanitize Images & Buffers with non-zero values.
Only enabled for specific tests at the moment. This CL allows our tests
to sanitizes memory for the robust resource access extension. It is
quite slow so should not be enabled by default.
Only works for 1 level 2D color textures and buffers. Makes several
flaky robust resource initialization tests consistently fail.
Controlled via an angle::Feature in FeaturesVk.
It works by initializing memory to an abitrary non-zero value:
- if newly allocated memory is mappable, we map it in init and set it
- if a buffer or texture can be a transfer destination, we use a
staging resource
- otherwise we don't attempt to initialize the resource.
Bug: angleproject:4384
Change-Id: I9b4f347bfcddf3096f491ed0243bef86837feaa0
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2043271
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
|
|
|
20b1259a
|
2020-02-05T17:08:05
|
|
Vulkan: Command graph linearization (Step 2).
ES 2.0 is feature complete.
Passes all of the angle_end2end_tests with the new linear command
recording method. Also runs the T-Rex benchmark without any obvious
glitches. Likely has issues with creating too many RenderPasses. ES3
is mostly untouched.
Bug: angleproject:4029
Change-Id: Ic5acf3d768495fbffd07b07bf0a6f2b5787c51f8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2012900
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
806ba566
|
2019-12-12T13:02:01
|
|
Extend ANGLE_iosurface_client_buffer to Vulkan backend for Swangle
Implement an IOSurface-backed pBuffer surface for the Vulkan backend
on Mac, through SwANGLE. ANGLE will pass a raw pointer to Swiftshader
and handle locking/unlocking the IOSurface.
Bug: chromium:1015454
Change-Id: Ia3ead55334736003d405b54ba8dcc7701706fbb2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1965434
Commit-Queue: Jonah Ryan-Davis <jonahr@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
1a1a1427
|
2020-01-23T13:57:21
|
|
Expose eglGetMscRateCHROMIUM from EGL_CHROMIUM_sync_control
When ANGLE is using the GL backend on GLX, we can expose
eglGetMscRateCHROMIUM via glXGetMscRateOML. Otherwise, this function
should return false.
Bug: chromium:1042393
Change-Id: Id9b308c2217e07ee9860e2869be0e23b7a0c7411
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2017048
Commit-Queue: Jonah Ryan-Davis <jonahr@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
564eb6f2
|
2019-11-22T16:46:02
|
|
Implement EGL_IMG_context_priority
Change RendererVk to have 3 VkQueues instead of one.
Each queue has a priority. To match extension: Low, Med, High.
gl::Context contains priority.
ContextVk contains a reference to one of the queues.
Every call to vulkan that uses queue, uses the associated context queue.
Bug: angleproject:3962
Change-Id: Ibd913a07a81c77bd975921d6dbae6a222842e88b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1978154
Commit-Queue: Jeff Vigil <j.vigil@samsung.com>
Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
3712b2e4
|
2020-01-24T17:08:30
|
|
Vulkan: Command graph linearization (Step 1).
This initial prototype introduces a new feature to the Vulkan back-end
that disables the deferred command recording. The intent is to have a
lower CPU overhead during submission calls which currently walk a DAG.
The feature is not complete. Currently it only passes the ANGLE
SimpleOperationTests. Moreover it is extremely simple and only allows
use of one command buffer at a time. In the future we'll allow open
command buffers for recording outside and inside render pass commands
at the same time. We'll also support collapsing RenderPasses together
for some use cases.
Currently the prototype only passes "SimpleOperationTest". There are
quite a few unimplemented features like queries, XFB, etc.
Bug: angleproject:4029
Change-Id: I82760986683f55e37ac4ea559de6f4cffb6ef84e
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1953485
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Tobin Ehlis <tobine@google.com>
|
|
942d9152
|
2019-12-23T15:31:52
|
|
EGL: Add support for EGL_KHR_gl_colorspace extension
If the Vulkan backend supports VK_EXT_swapchain_colorspace extension,
enable the option for applications to create surfaces with non-linear
formats. Not all formats have non-linear versions and is platform
specific.
Tests: dEQP-EGL.functional.wide_color.*
angle_deqp_egl_tests --use-angle=vulkan --deqp-gl-context-type=egl --deqp-case=dEQP-EGL.functional.wide_color.*
Bug: angleproject:2514
Change-Id: I441ee797cceef92c84473bfa18605c4fd8180de1
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1951963
Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
|
|
5fd73782
|
2019-08-09T11:46:46
|
|
Vulkan: Use volk to load vk* func ptrs
Thanks to Jamie Madill for some fixes to get all CI test passing w/ volk.
This change updates all ANGLE targets that use Vulkan to dyanmically
link all of the VK entrypoints using the volk OSS library from
https://github.com/zeux/volk.
It's only two source files so baking them directly into ANGLE repo.
Also it's used in both the tests and libANGLE trees so added to
src/common/third_party/volk dir.
Updated volk and the renderer to track latest instance and device
that were loaded and renderer will refresh vk* function pointers if
the current and previous device and/or instance don't match. This
prevents errors in the test framework as we transition between
backends, especially between VK HW & SwiftShader ICDs.
This change rolls the Vulkan Loader forward to use the latest loader
version which no longer allows static linking but requires dynamic
linking.
Bug: angleproject:3740
Bug: angleproject:4092
Bug: angleproject:4162
Bug: angleproject:4210
Bug: angleproject:4225
Change-Id: I8a0b7d24c9545bbfdfaa4b9357a9bfe6793e0140
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1965640
Commit-Queue: Tobin Ehlis <tobine@google.com>
Reviewed-by: Tobin Ehlis <tobine@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
c898ec1a
|
2019-11-04T15:20:18
|
|
Add EGL GGP extensions.
This CL adds two new extensions:
* EGL_ANGLE_ggp_stream_descriptor:
Introduces a new attribute to CreateWindowSurface. Allows the app
to pass in a stream descriptor to VkCreateSurfaceKHR.
Mirrors VK_GGP_stream_descriptor_surface.
* EGL_ANGLE_swap_with_frame_token:
Introduces a new function 'eglSwapBuffersWithFrameTokenANGLE'. This
allows the app to pass a GGP frame token down to vkQueuePresentKHR.
Mirrors VK_GGP_frame_token.
Bug: angleproject:4078
Change-Id: I4313ac4c264e68999905049f661bc64b44f72fab
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1897315
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
|
|
|
5883a44b
|
2019-11-02T19:34:47
|
|
Vulkan: Don't pass width/height to Surface constructors.
These values are accessible already from the Surface state. Sometimes
they don't apply (e.g. GGP). Remove them to keep the code clearer.
Updates the offscreen surface constructor to pull width and height from
the attributes map instead of using parameters. Otherwise they weren't
used.
Refactoring change only.
Bug: angleproject:4078
Change-Id: I9e49eadc7116562f62bd8d11342d6b8835376719
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1895762
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
68591eff
|
2019-10-13T15:05:23
|
|
Vulkan: Store ImageView access in the graph.
This will ensure we don't destroy the image views when they are still
in use by other Contexts.
Bug: angleproject:2464
Change-Id: I1d3ba2ad241250e31ea32873446c4cb23971750d
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1843236
Reviewed-by: Tim Van Patten <timvp@google.com>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
beacd8c8
|
2019-10-17T14:34:02
|
|
Vulkan: Rename format fields.
Renames 'angleFormat' to 'intendedFormat'. Also renames 'bufferFormat'
and 'imageFormat' to 'actualImageFormat' and 'actualBufferFormat'. This
renaming should make it clearer to the reader what the meaning of the
different format fields are. Intended format is the front-end format
and the actual formats are the formats we pass to Vulkan. Also updates
the documentation.
Bug: angleproject:4009
Change-Id: If61bf7250e88f7ed3d452718574c963d718e27b2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1866077
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
a14555a7
|
2019-10-13T15:05:22
|
|
Vulkan: Refactor RenderTargetVk ImageViews.
Instead of storing a pointer to an ImageView, store a
pointer to an ImageViewHelper. The level and layer index
are all that's needed to pull the right image view from
the ImageViewHelper.
Will make it easier to store ImageView references in the
graph to track lifetime.
Bug: angleproject:2464
Change-Id: I773fa79a89ceefc99ff70e4eed4985f9f01fe26a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1857549
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Ian Elliott <ianelliott@google.com>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
|
|
|
0b9ebe58
|
2019-10-15T11:15:42
|
|
Vulkan: Add "ImageViewHelper".
This allows views to track a different lifetime than vk::ImageHelper.
This in turn will fix the race condition on ContextVk destruction when
releasing ImageViews owned by TextureVk and RenderbufferVk. For now
this is a refactoring change only.
Bug: angleproject:2464
Change-Id: I9581975bd5d4913233bbed8439dd4a632cc78a2a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1843231
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Ian Elliott <ianelliott@google.com>
|
|
|
88bc4d3d
|
2019-10-04T20:41:24
|
|
Vulkan: Get rid of "fetch" level/layer image views.
Instead always use 2D array views when rendering to cube maps. We need
to keep the "fetch" image view for the main image for emulating
seamless cube map textures on GL 2.0. We also use it for some texture
copying init. These uses could potentially be replaced with sample
instructions if we wanted to remove this last "fetch" view.
Bug: angleproject:3975
Change-Id: I90f9be9fc6542d8ec27f6722132c1b27acf72176
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1842226
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
f10bf6bf
|
2019-09-26T10:27:18
|
|
Vulkan: Implement multi-threaded GL.
The main component of this change is to make vk::BufferHelper,
vk::ImageHelper and vk::SyncHelper use a common path. We introduce a
new "vk::SharedGarbage" helper class that stores small lists of garbage
from individual objects like an ImageHelper or BufferHelper. The
SharedGarbage is stored in the RendererVk with the ResourceUse of the
helper object. The ResourceUse tells RendererVk when it is safe to
destroy the GarbageObjects.
New "onGraphAccess" commands are added in a few places to enable the
common garbage collection path. A couple Context-only resources like
default attributes now are referenced where they were not before.
Also reorganizes some functions so we can add a few helpful ASSERTs
to our graph dependencies. Added "updateCurrentAccessNodes" for this.
Also adds a "RendererScoped" helper to replace many uses of
"ContextScoped".
The multithreading EGL tests mostly pass but have some remaining
flakiness so cannot yet be enabled.
Bug: angleproject:2464
Change-Id: Ia3e3ae8848d731abf3f21ebe04c33e381e130be0
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1808444
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
cb16fb5f
|
2019-08-29T16:53:55
|
|
Vulkan: Support texture base and max levels
The Vulkan backend uses a vkImage that matches the number
of effective levels in the GL texture. This is due to the fact
that GL textures can have really strange layouts that only make
sense when base level and max level are applied.
For instance, take the following layout with disjoint mip levels:
Level 0: 4x4 RGBA
Level 1: 2x2 RGBA
Level 2: 10x10 RGB
If base level is set to zero and max level is set to 1, the image is
still considered mip-complete:
Level 0: 4x4 RGBA ==> Base Level 0 ==> Level 0: 4x4 RGBA
Level 1: 2x2 RGBA ==> Max Level 1 ==> Level 1: 2x2 RGBA
Level 2: 10x10 RGB
If base and max level are then both set to 2, the texture is still
considered complete, but of a different size and format:
Level 0: 4x4 RGBA
Level 1: 2x2 RGBA
Level 2: 10x10 RGB ==> Base/Max Level 2 ==> Level 2: 10x10 RGB
When the base or max level is changed, we must recreate the vkImage to
match the new level count.
To support that, we:
- Stage updates from the current image to the new image
- Only stage updates if there aren't already staged updates for a level
- Free the current image and so it can be recreated at the next draw
This CL does the following:
- Refactors TextureVk::copyImageDataToBuffer to support staging updates
without flush
- Adds TextureVk::copyImageDataToBufferAndGetData to support previous
use model
- Adds TextureVk::changeLevels, triggered during syncState, which stages
updates and releases the current image.
- Updates ImageHelper::flushStagedUpdates to understand base/max levels
- Updates TextureVk::ensureImageInitialized and TextureVk::generateMipmap
to account for base/max level
- Tracks base and max levels in ImageHelper
- Adds ImageHelper::stageSubresourceUpdateFromBuffer to support
this use case
- Adds ImageHelper::isUpdateStaged to determine if changeLevels
should propagate data
- Makes gl::TextureTypeToTarget available for use outside of ImageIndex
- Enables several deqp and end2end tests
Bug: angleproject:3148
Test: dEQP-GLES3.functional.texture.mipmap.*base_level*
Test: dEQP-GLES3.functional.texture.mipmap.*max_level*
Change-Id: I14ca071c9c62eb310dfed7ef9290dc65fc3ff696
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1776933
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Commit-Queue: Cody Northrop <cnorthrop@google.com>
|
|
|
33ffed01
|
2019-09-26T10:19:35
|
|
Vulkan: Clean up garbage APIs.
Instead of dumping resources to a context, we use the release APIs
consistently. Refactoring/cleanup change only. Should have very litte
impact on runtime behaviour.
Bug: angleproject:2464
Change-Id: I2dc7f8316c466f7ccfad50a7b792ba0ee7bc2e49
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1804883
Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
b540da89
|
2019-09-19T14:19:12
|
|
Vulkan: Use ResourceUse to track object lifetime.
With the new resource tracking scheme the CommandGraph, tracking a
"Context serial" aka "current" serial is no longer necessary for
CommandGraphResources. Serial tracking has been moved to the shared
ResourceUse struct that gets updated on a command submission. Thus
we don't need to store the serial as a current separate piece of info
in BufferHelper/ImageHelper.
Will lead to further redesign for the multi-threading support for
Vulkan.
Bug: angleproject:2464
Change-Id: I1ae4bcc27fcfb93422b4ab4c9682a458e482f295
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1785990
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
d7f28aae
|
2019-09-19T14:19:10
|
|
Vulkan: Pass CommandGraph when updating serials.
This is in preparation for storing a pointer to a shared resource use
structure.
Bug: angleproject:2464
Change-Id: I8f4ba1c71de6ad6a27ac06fc8012a0e94267cc16
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1785988
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
7e44ec26
|
2019-08-26T15:59:48
|
|
GL_EXT_multisampled_render_to_texture extension. Part 1.
Adding new parameters for extension without adding any real code change. Since no new code paths were added, we expect all tests to pass as before.
Bug: angleproject:980428
Change-Id: I551b46a66f422eabd357fd021e00cf266a991efb
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1772377
Commit-Queue: Rafael Cintron <rafael.cintron@microsoft.com>
Reviewed-by: Rafael Cintron <rafael.cintron@microsoft.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
050b124d
|
2019-06-30T03:26:18
|
|
Reland "Vulkan: Debug overlay"
This is a reland of e54d0f90d1a165404236fd7abd1b05ddd041a686
This was reverted due to a build failure as a result of a missing
virtual destructor in the widget base class.
Original change's description:
> Vulkan: Debug overlay
>
> A debug overlay system for the Vulkan backend designed with efficiency
> and runtime configurability in mind. Overlay widgets are of two
> fundamental types:
>
> - Text widgets: A single line of text with small, medium or large font.
> - Graph widgets: A bar graph of data.
>
> Built on these, various overlay widget types are defined that gather
> statistics. Five such types are defined with one widget per type as
> example:
>
> - Count: A widget that counts something. VulkanValidationMessageCount
> is an overlay widget of this type that shows the number of validation
> messages received from the validation layers.
> - Text: A generic text. VulkanLastValidationMessage is an overlay
> widget of this type that shows the last validation message.
> - PerSecond: A value that gets reset every second automatically. FPS is
> an overlay widget of this type that simply gets incremented on every
> swap().
> - RunningGraph: A graph of last N values. VulkanCommandGraphSize is an
> overlay of this type. On every vkQueueSubmit, the number of nodes in
> the command graph is accumulated. On every present(), the value is
> taken as the number of nodes for the whole duration of the frame.
> - RunningHistogram: A histogram of last N values. Input values are in
> the [0, 1] range and they are ranked to N buckets for histogram
> calculation. VulkanSecondaryCommandBufferPoolWaste is an overlay
> widget of this type. On vkQueueSubmit, the memory waste from command
> buffer pool allocations is recorded in the histogram.
>
> Overlay font is placed in libANGLE/overlay/ which gen_overlay_fonts.py
> processes to create an array of bits, which is processed at runtime to
> create the actual font image (an image with 3 layers).
>
> The overlay widget layout is defined in overlay_widgets.json which
> gen_overlay_widgets.py processes to generate an array of widgetss, each
> of its respective type, and sets their properties, such as color and
> bounding box. The json file allows widgets to align against other
> widgets as well as against the framebuffer edges.
>
> Two compute shaders are implemented to efficiently render the UI:
>
> - OverlayCull: This shader creates a bitset of Text and Graph widgets
> whose bounding boxes intersect a corresponding subgroup processed by
> OverlayDraw. This is done only when the enabled overlay widgets are
> changed (a feature that is not yet implemented) or the surface is
> resized.
> - OverlayDraw: Using the bitsets generated by OverlayCull, values that
> are uniform for each workgroup (set to be equal to hardware subgroup
> size), this shader loops over enabled widgets that can possibly
> intersect the pixel being processed and renders and blends in texts
> and graphs. This is done once per frame on present().
>
> Currently, to enable overlay widgets an environment variable is used.
> For example:
>
> $ export ANGLE_OVERLAY=FPS:VulkanSecondaryCommandBufferPoolWaste
> $ ./hello_triangle --use-angle=vulkan
>
> Possible future work:
>
> - On Android, add settings in developer options and enable widgets based
> on those.
> - Spawn a small server in ANGLE and write an application that sends
> enable/disable commands remotely.
> - Implement overlay for other backends.
>
> Bug: angleproject:3757
> Change-Id: If9c6974d1935c18f460ec569e79b41188bd7afcc
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1729440
> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Bug: angleproject:3757
Change-Id: I47915d88b37b6f882c686c2de13fca309a10b572
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1780897
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
fc58af47
|
2019-09-02T07:46:44
|
|
Revert "Vulkan: Debug overlay"
This reverts commit e54d0f90d1a165404236fd7abd1b05ddd041a686.
Reason for revert: causes compile failure on Linux CFI bot.
Sample build: https://ci.chromium.org/p/chromium/builders/ci/Linux%20CFI/14810
Sample log: https://logs.chromium.org/logs/chromium/buildbucket/cr-buildbucket.appspot.com/8903575125463586160/+/steps/compile/0/stdout?format=raw
Original change's description:
> Vulkan: Debug overlay
>
> A debug overlay system for the Vulkan backend designed with efficiency
> and runtime configurability in mind. Overlay widgets are of two
> fundamental types:
>
> - Text widgets: A single line of text with small, medium or large font.
> - Graph widgets: A bar graph of data.
>
> Built on these, various overlay widget types are defined that gather
> statistics. Five such types are defined with one widget per type as
> example:
>
> - Count: A widget that counts something. VulkanValidationMessageCount
> is an overlay widget of this type that shows the number of validation
> messages received from the validation layers.
> - Text: A generic text. VulkanLastValidationMessage is an overlay
> widget of this type that shows the last validation message.
> - PerSecond: A value that gets reset every second automatically. FPS is
> an overlay widget of this type that simply gets incremented on every
> swap().
> - RunningGraph: A graph of last N values. VulkanCommandGraphSize is an
> overlay of this type. On every vkQueueSubmit, the number of nodes in
> the command graph is accumulated. On every present(), the value is
> taken as the number of nodes for the whole duration of the frame.
> - RunningHistogram: A histogram of last N values. Input values are in
> the [0, 1] range and they are ranked to N buckets for histogram
> calculation. VulkanSecondaryCommandBufferPoolWaste is an overlay
> widget of this type. On vkQueueSubmit, the memory waste from command
> buffer pool allocations is recorded in the histogram.
>
> Overlay font is placed in libANGLE/overlay/ which gen_overlay_fonts.py
> processes to create an array of bits, which is processed at runtime to
> create the actual font image (an image with 3 layers).
>
> The overlay widget layout is defined in overlay_widgets.json which
> gen_overlay_widgets.py processes to generate an array of widgetss, each
> of its respective type, and sets their properties, such as color and
> bounding box. The json file allows widgets to align against other
> widgets as well as against the framebuffer edges.
>
> Two compute shaders are implemented to efficiently render the UI:
>
> - OverlayCull: This shader creates a bitset of Text and Graph widgets
> whose bounding boxes intersect a corresponding subgroup processed by
> OverlayDraw. This is done only when the enabled overlay widgets are
> changed (a feature that is not yet implemented) or the surface is
> resized.
> - OverlayDraw: Using the bitsets generated by OverlayCull, values that
> are uniform for each workgroup (set to be equal to hardware subgroup
> size), this shader loops over enabled widgets that can possibly
> intersect the pixel being processed and renders and blends in texts
> and graphs. This is done once per frame on present().
>
> Currently, to enable overlay widgets an environment variable is used.
> For example:
>
> $ export ANGLE_OVERLAY=FPS:VulkanSecondaryCommandBufferPoolWaste
> $ ./hello_triangle --use-angle=vulkan
>
> Possible future work:
>
> - On Android, add settings in developer options and enable widgets based
> on those.
> - Spawn a small server in ANGLE and write an application that sends
> enable/disable commands remotely.
> - Implement overlay for other backends.
>
> Bug: angleproject:3757
> Change-Id: If9c6974d1935c18f460ec569e79b41188bd7afcc
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1729440
> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
> Reviewed-by: Jamie Madill <jmadill@chromium.org>
TBR=geofflang@chromium.org,syoussefi@chromium.org,jmadill@chromium.org
Bug: angleproject:3757
Change-Id: Ib08e2e7b1a9449ca097673acb11655df5d2bbf31
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1778862
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
|
|
|
e54d0f90
|
2019-06-30T03:26:18
|
|
Vulkan: Debug overlay
A debug overlay system for the Vulkan backend designed with efficiency
and runtime configurability in mind. Overlay widgets are of two
fundamental types:
- Text widgets: A single line of text with small, medium or large font.
- Graph widgets: A bar graph of data.
Built on these, various overlay widget types are defined that gather
statistics. Five such types are defined with one widget per type as
example:
- Count: A widget that counts something. VulkanValidationMessageCount
is an overlay widget of this type that shows the number of validation
messages received from the validation layers.
- Text: A generic text. VulkanLastValidationMessage is an overlay
widget of this type that shows the last validation message.
- PerSecond: A value that gets reset every second automatically. FPS is
an overlay widget of this type that simply gets incremented on every
swap().
- RunningGraph: A graph of last N values. VulkanCommandGraphSize is an
overlay of this type. On every vkQueueSubmit, the number of nodes in
the command graph is accumulated. On every present(), the value is
taken as the number of nodes for the whole duration of the frame.
- RunningHistogram: A histogram of last N values. Input values are in
the [0, 1] range and they are ranked to N buckets for histogram
calculation. VulkanSecondaryCommandBufferPoolWaste is an overlay
widget of this type. On vkQueueSubmit, the memory waste from command
buffer pool allocations is recorded in the histogram.
Overlay font is placed in libANGLE/overlay/ which gen_overlay_fonts.py
processes to create an array of bits, which is processed at runtime to
create the actual font image (an image with 3 layers).
The overlay widget layout is defined in overlay_widgets.json which
gen_overlay_widgets.py processes to generate an array of widgetss, each
of its respective type, and sets their properties, such as color and
bounding box. The json file allows widgets to align against other
widgets as well as against the framebuffer edges.
Two compute shaders are implemented to efficiently render the UI:
- OverlayCull: This shader creates a bitset of Text and Graph widgets
whose bounding boxes intersect a corresponding subgroup processed by
OverlayDraw. This is done only when the enabled overlay widgets are
changed (a feature that is not yet implemented) or the surface is
resized.
- OverlayDraw: Using the bitsets generated by OverlayCull, values that
are uniform for each workgroup (set to be equal to hardware subgroup
size), this shader loops over enabled widgets that can possibly
intersect the pixel being processed and renders and blends in texts
and graphs. This is done once per frame on present().
Currently, to enable overlay widgets an environment variable is used.
For example:
$ export ANGLE_OVERLAY=FPS:VulkanSecondaryCommandBufferPoolWaste
$ ./hello_triangle --use-angle=vulkan
Possible future work:
- On Android, add settings in developer options and enable widgets based
on those.
- Spawn a small server in ANGLE and write an application that sends
enable/disable commands remotely.
- Implement overlay for other backends.
Bug: angleproject:3757
Change-Id: If9c6974d1935c18f460ec569e79b41188bd7afcc
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1729440
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
c3f57231
|
2019-08-28T15:00:46
|
|
Vulkan: revamp present semaphore management
See doc/PresentSemaphores.md for details.
Bug: angleproject:3450
Bug: angleproject:3670
Change-Id: I52d5bd13a4af25f224d386c9584525c182af6f17
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1776880
Reviewed-by: Tim Van Patten <timvp@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
57ad1e12
|
2019-08-23T14:53:26
|
|
Vulkan: Introduce ContextScoped
Scoped is changed to DeviceScoped and a ContextScoped class is added.
The former destroys objects at the end of scope, and the latter releases
them.
Bug: angleproject:3757
Change-Id: Ia0bbded53e94af9c8c72c77d07306d6724c6c409
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1769060
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
7e48c9eb
|
2019-08-06T17:17:19
|
|
Add explicit integer casts
WebKit uses the -Wshorten-64-to-32 flag which warns on these cases.
Bug: 3439
Change-Id: I8c1de60da0f173ca2036e2120e79b857f5f2775f
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1740866
Commit-Queue: James Darpinian <jdarpinian@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
|
|
|
c327370e
|
2019-07-23T12:54:12
|
|
Vulkan: Pass VkExtent3D to TextureHelper::init.
Bug: angleproject:3189
Change-Id: I4b95240bb32fbc2b3d0c8f097e0552d0fe23417d
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1713094
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
|
|
|
deb43a31
|
2019-07-22T09:56:27
|
|
Vulkan: Store VkExtents3D in ImageHelper.
This makes the distinction between a gl::Extents (includes a depth
value for 2D array texture layer count) and a Vulkan extents (2D array
textures have a "1" for depth) clearer. Preparation refactor patch.
Bug: angleproject:3189
Change-Id: I9a13379c421e7f3c7856ac15b7a73013258ab9fe
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1709754
Reviewed-by: Jonah Ryan-Davis <jonahr@google.com>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
a1754dc8
|
2019-06-26T13:36:43
|
|
Vulkan: Recycle vkFence
This CL introduces a Recycler to reuse unreferenced vk::Fence, reducing
CPU time spent on vk::Fence init&destroy. Save around 15% of CPU time
for most glmark2 tests.
Bug: angleproject:3556
Change-Id: Ice5054305321c466c5be3bc368d04091f074729c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1679239
Commit-Queue: Tobin Ehlis <tobine@google.com>
Reviewed-by: Tobin Ehlis <tobine@google.com>
|
|
|
0966f3f8
|
2019-07-08T13:18:03
|
|
Vulkan: Remove flush semaphore chain
This avoids using an unbounded number of semaphores in between calls to
swapbuffers. Using two semaphores should be sufficient to synchronize
swaps.
In addition, fix tracking of VkPipelineStageFlags by creating a 2nd
vector parallel to the semaphores vector. The last fix assumed there
could only be 2 wait semaphores, but that bound only applied to signal
semaphores. After this change, there can only be one signal semaphore,
but there's still no bound to wait semaphores.
Bug: angleproject:3637
Change-Id: I7fbba67fa4bbdf62b9e9d530a924acd5236705d3
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1688435
Commit-Queue: Michael Spang <spang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
66b5ff58
|
2019-06-28T14:34:22
|
|
texture3D: Implement functionality and enable for Vulkan
Also update test expectations for texture3D.
Bug: angleproject:3188
Change-Id: If8a8e0a83a86c48c2afb0c36534c1e9d4120fe47
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1682782
Reviewed-by: Ian Elliott <ianelliott@google.com>
Reviewed-by: Lingfeng Yang <lfy@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Cody Northrop <cnorthrop@google.com>
|
|
|
f116aa9c
|
2019-06-17T15:08:11
|
|
Vulkan: detect swapchain recreation passively
Currently ANGLE vulkan query current window size every frame to detect
whether a swapchain recreation is needed, however profiling shows some
bottle neck on window size query on X11. So here removing the per frame
window size query and only depend the result of vkAcquireImage and
VkQueuePresent to detect a need for swapchain recreation.
Bug: angleproject:3535
Change-Id: I4ddf70b973d78dfcd5f8fab28e29e802edad1b2d
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1662759
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Ian Elliott <ianelliott@google.com>
Commit-Queue: Ian Elliott <ianelliott@google.com>
|
|
|
d5cef305
|
2019-06-19T14:21:33
|
|
Vulkan: Add more trace events.
This captures a bit more information about where the call time goes
during a flush. It can show at least on desktop NVIDIA that we spend a
fair amount of time in vkQueuePresentKHR.
Bug: angleproject:3117
Change-Id: I2d0195b9338bcac80e8dd8cfb550402271f286f9
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1650787
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
3ea463bf
|
2019-06-19T14:21:33
|
|
Move event tracer back into common.
Requires that we update the TRACE_EVENT macros to accept a platform as
an argument. The refactor isn't complete. In order to finish we'd need
to ensure we have the Display's PlatformMethods available at all sites.
Unblocks adding trace events directly in the perf tests.
Bug: angleproject:1892
Bug: angleproject:3117
Change-Id: Iee0ca086ccfe23acab3fc186fb042f018711a94c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1664794
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
444aa0fe
|
2019-06-08T13:53:17
|
|
Vulkan: Prefer immediate present mode for benchmarking.
This should give the highest possible scores. Note that immediate is
not always available.
Bug: angleproject:3163
Change-Id: Ifedabd39819625b57feda71ed83a7bee8fa0a42b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1638904
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
b407e1a0
|
2019-06-03T17:15:51
|
|
Vulkan: implement ES3 blit
Augment the resolve shaders to be able to stretch and blit too. The
UtilsVk resolve function is accordingly expanded to include blit.
Bug: angleproject:3200
Change-Id: I30b172a5e388089735ab494f55cbfdc2781a8bf9
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1635753
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
148ecd89
|
2019-05-28T14:14:42
|
|
Vulkan: Handle VK_ERROR_OUT_OF_DATE returned by vkAcquireNextImageKHR()
There is some new Android HWUI/SkiaGL code, that (in certain cases)
looks up the underlying Vulkan swapchain that ANGLE created, and adds
an image to it. This causes the next call to vkAcquireNextImageKHR() to
return VK_ERROR_OUT_OF_DATE and requires the swapchain to be recreated.
Bug: angleproject:3480
Test: Use ANGLE on Android and verify apps load correctly.
Change-Id: I21c86035664878e75d6a9dc769546747aa4c8256
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1632424
Commit-Queue: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
3f17f929
|
2019-06-05T13:13:12
|
|
Avoid recreating swapchain for preRotation
Android will return VK_SUBOPTIMAL_KHR if the orientation of the
screen is not optimial for the orientation of the device.
We were re-creating the swapchain when this happens but we weren't
changing the orientation and would get this return code every
time we call vkQueuePresent and re-create the swapchain each time.
Apparently recreating the swapchain causes visual artifacts on
some devices. Just ignoring VK_SUBOPTIMAL_KHR allows things to
work.
Bug: angleproject:3497
Need to implement proper preRotation support for longterm fix.
Bug: angleproject:3502
Change-Id: I0904eb60b742d24618d502c111510117758a8502
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1641206
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Courtney Goeltzenleuchter <courtneygo@google.com>
|
|
|
ab6a59b2
|
2019-05-21T21:26:26
|
|
Vulkan: Submit semaphores from glWaitSemaphoreEXT & glSignalSemaphoreEXT
Implement submission of client semaphores passed to glWaitSemaphoreEXT &
glSignalSemaphoreEXT.
This also relaxes the expectation that we will not flush() if there are
no commands. Signaling semaphores in particular requires queue submission
irrespective of whether there are any command buffers to submit. If there
are neither commands nor semaphores, we can still skip queue submission.
WebGL runs in Chrome with ANGLE & Vulkan interop as of this patch, albeit
with incorrect synchronization due to texture barriers not being
implemented yet. Quite a few flags are needed to try this:
GN args: angle_vulkan_conformant_configs_only=true
chrome \
--enable-features=UseSkiaRenderer,UiGpuRasterization \
--enable-gpu-rasterization \
--enable-oop-rasterization \
--enable-vulkan \
--use-gl=angle \
--use-angle=vulkan
Bug: angleproject:3289
Change-Id: I3d49c230a2fbf0cd2a2b943b05ded0e4604cc313
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1623815
Commit-Queue: Michael Spang <spang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
ee244c77
|
2019-05-06T10:30:18
|
|
Vulkan: Move command graph and garbage to ContextVk.
To support multithreading, contexts should manage their own command graphs
and garbage. This allows safe access to vulkan resources such as command pools
without thread synchronization.
BUG=angleproject:2464
Change-Id: I930149bc9f0793028761ee05ab50b8c0a4dec98a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1516515
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
ce9be8c7
|
2019-05-24T09:35:06
|
|
Vulkan: Have a cubemap as 2D-array view handy
Previously, only texture copies used a shader that performed
texelFetch(). To support cubemaps, a hack was used to temporarily
create a 2D array view.
With upcoming support for multisample resolve, more shaders will be
using texelFetch() all requiring this workaround. This change instead
makes sure that a separate view is created for cubemaps for the purpose
of being used with these shaders. As a result, we have three logical
views on textures and render targets:
- Draw: a view that can be used as a color/depth/stencil attachment
- Read: a view that can be used to sample from
- Fetch: a view that can be used to fetch from
The fetch view is generally the same as the read view, except for cube
maps.
Bug: angleproject:3200
Change-Id: I21547f728c16f0aa8f0fcae152c400b5cc1565da
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1628585
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
6722009e
|
2019-05-20T11:12:53
|
|
Vulkan: Handle dirty RTs with state messages.
Prior to this CL we were handling dirty state change notifications by
flushing the RT Images just prior to use or just after they were
changed. This could lead to a few redundant checks in several places.
It also meant we needed an owner pointer from the RT to the parent
Image. This pointer would be null for Surfaces and Renderbuffers.
This cleans up the image flushing logic to be handled by dirty bit
notifications. When an app updates an attached Texture with TexSubImage
or related calls it will send a notification to the Framebuffer. The
Framebuffer then sets a dirty contents bit that is handled in the
implementation. In Vulkan this means flushing the dirty bits.
Requires adding a flag to the FramebufferImpl class to determine if we
need to syncState before we checkStatus. Adding the option allows us to
only call syncState for the GL back-end. Not calling syncState allows
the robust resource init operation to happen *before* we syncState.
Which in turn allows FramebuffeVk to initialize the VkImages in one go.
Added new regression tests for Texture updates. This might not cover
all cases. I found it was very hard to trigger some of the resource
update staging in TextureVk.
Bug: angleproject:3427
Change-Id: Idfa177436ba7fcb9d398f2b67922e085f778f82a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1601552
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
b16d69c3
|
2019-05-13T16:28:27
|
|
Vulkan: Add support for surface multisampling
A multisample image is created for the surface if multisampling is
enabled. Prior to present, this multisample image is resolved into the
swapchain image.
FramebufferVk::readPixelsImpl similarly has got the ability to resolve
the region of interest into a temporary image prior to readback.
Tests are added to render a point, line and a triangle on a 4x
multisampled surface.
Bug: angleproject:3204
Change-Id: I34aca502fa1918b5cbf000ff11521c350372e051
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1610188
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
f7033c55
|
2019-05-13T09:48:06
|
|
Vulkan: Request at least 3 images for MAILBOX.
The glmark2 benchmark score went down (on Android with Pixel devices)
when ANGLE changed VkSwapchainCreateInfoKHR::minImageCount from 3 down
to 2, when using VK_PRESENT_MODE_MAILBOX_KHR. In experimenting, this
was confirmed. This CL changes minImageCount back to 3.
Vulkan spec issue #1671 was filed in order to understand whether this
is a good/bad approach for other drivers/platforms.
Test: glmark2 on Pixel2.
Bug: angleproject: 3400
Change-Id: Ia4a72733eb648e4f53feeb8833b174d653fa5766
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1610242
Commit-Queue: Ian Elliott <ianelliott@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
776694cd
|
2019-05-08T10:28:55
|
|
Change all ANGLE workarounds to use struct definition with info.
Change each workaround from a simple bool to a struct with info
including name, workaround set, description, and bug IDs. This will help
with future workaround integration with Chrome.
Bug: angleproject:1621
Change-Id: Ia27c180abaf845e280060c803e5994cc3152a057
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1593917
Commit-Queue: Jonah Ryan-Davis <jonahr@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
c22ef61f
|
2019-04-09T09:55:14
|
|
Vulkan: Store fences instead of serials in SurfaceVK::SwapHistory
Previous swaps may not have been submitted by the same context as the current
swap and we can't safely wait on serials submitted by other contexts.
BUG=angleproject:2464
Change-Id: Ibcb23171feae038fd45abdefb80c4e8a373744e2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1562521
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>
|
|
|
be394bad
|
2019-05-02T13:41:19
|
|
Vulkan: Reset swap chain image dependency information when aquiring.
If an image was used by a different context previously, it's stored serial
is not relevent. We always know that any work involving the swap chain image
has been completed by the time we aquire a new one.
BUG=angleproject:2464
Change-Id: I0b12b9ff4b9a68593d661344d2d1e2c9d94e4755
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1592043
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>
|
|
|
0f861bd8
|
2019-04-26T16:00:23
|
|
Vulkan: Don't present on surface destruction.
The the user hasn't called SwapBuffers explicitly, we should not write to
the window surface. This is especially a problem when the native window has
already been destroyed.
BUG=angleproject:2464
Change-Id: Ib4289de471ba2ad10e5cc21a8c2af946642f138c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1586355
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>
|
|
|
28383fb2
|
2019-04-26T16:00:06
|
|
Vulkan: Wrap aquire semaphore in a vk::Scoped
If vkAcquireNextImageKHR fails, the aquireImageSemaphore would be leaked,
wrap it in a vk::Scoped so it is always cleaned up.
BUG=angleproject:2464
Change-Id: Ic4a0e16c89ea7c35cf060e5601760422e673c080
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1585318
Reviewed-by: Tim Van Patten <timvp@google.com>
Commit-Queue: Geoff Lang <geofflang@chromium.org>
|
|
|
0631e19b
|
2019-04-18T16:09:12
|
|
Vulkan: Rename Vulkan "Texture" format to "Image"
Also adds some comments to vk_format_utils.h.
Bug: angleproject:3372
Change-Id: I529b9b189e4cdfd400c3c981a47139727d9954ab
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1565062
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
95c0fae6
|
2019-04-16T22:46:08
|
|
Vulkan: Clear surface images if emulated format
This was done for renderbuffers but was missing for surfaces.
Bug: angleproject:2722
Change-Id: I019805d6ca43eef86d2d46e7c72c1013803f2139
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1570149
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
a254fa28
|
2019-04-09T14:25:59
|
|
Rename getCurrentDisplay to getDisplay
There is always only one display that can be associated with a context,
so get*Current*Display is confusing.
Bug: None
Change-Id: Iff3a9fc5ad1154b046bb30d7f46a468802ba7fcc
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1558958
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
af0301a8
|
2019-04-09T10:43:04
|
|
Vulkan: fix CPU throttling frames to 2
Previously, the CPU was throttled to be at most N frames ahead, N being
the number of swapchain images. N is now fixed to 2, regardless of the
number of swapchain images. If N < 2, we would be stalling the CPU
unnecessarily, and if N > 2, the CPU could get too far ahead.
Effectively, here is how the throttling plays out with this commit:
Submit (Fence 1) + Present
Submit (Fence 2) + Present
Wait Fence 1
Submit (Fence 3) + Present
Wait Fence 2
Submit (Fence 4) + Present
Wait Fence 3
Submit (Fence 5) + Present
Wait Fence 4
...
Bug: angleproject:2942
Change-Id: I3b8c3bb88e52d62231306ec84aad50d2bf472d8c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1558681
Reviewed-by: Ian Elliott <ianelliott@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
65d10f3b
|
2019-03-21T16:30:31
|
|
Vulkan: Implement robust resource initialization
If a texture or renderbuffer needs to be cleared for robust access or
due to having emulated channels, it is immediately cleared. The former
relies on a front-end feature that optimizes robust access clears only
to levels and layers that are not fully initialized through data upload.
Bug: angleproject:2722
Change-Id: Icdab856eb4ffe963f78569b6d80d9ff5cb27ff9b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1535056
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
892d1805
|
2019-03-27T14:21:34
|
|
Vulkan: Have the WindowSurfaceVk own the submit semaphores.
Chaining of submit semaphores is only needed for window surfaces because they
are required for the first usage of the swap chain image and final present
of the image.
Move ownership of the submit semaphores from RendererVk to WindowSurfaceVk and
update all calls to finish and flush to be piped through a ContextVk which
tracks the currently bound window surface.
BUG=angleproject:2464
Change-Id: I4b3083124d7910a5dee297afc219e3a3f28057f2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1542257
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
02a579e9
|
2019-03-27T14:21:20
|
|
Vulkan: roll Vulkan repos to v1.1.102
Bug: angleproject:3320
Change-Id: Ia24f9bc2499f36c6aab810db00bb6408920222fd
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1541719
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
6854940d
|
2019-03-25T23:30:49
|
|
Vulkan: Correct RenderTargetVk extents to image mip
RenderTargetVk::getImageExtents() was always returning the base level
extents of the image it was viewing. This resulted in incorrect extents
being provided to various framebuffer- and renderpass-related
declarations. The function is renamed to `getExtents()` and returns the
appropriate mip extents.
Bug: angleproject:2361
Change-Id: I059a8d19fcb140c9095107d935aa3e5cb1852fc2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1537327
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
2660b503
|
2019-03-21T12:08:40
|
|
Vulkan: Restore CommandBuffer to namespace vk
Moved vk::CommandBuffer and vk::SecondaryCommandBuffer to vk::priv:: and
aliased vk::CommandBuffer to one or the other. This allows the rest of
the classes to continue seeing vk::CommandBuffer as they used to do.
Used a special alias for the primary command buffer that gets submitted
(vk::PrimaryCommandBuffer).
Bug: angleproject:3136
Change-Id: I61236fd182230991db7395d05e3da3be5e3f45be
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1534456
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
a8ff8814
|
2019-03-05T07:06:32
|
|
Vulkan:Optimize SecondaryCommandBuffers
RELAND of this commit. Had to fix fuzzer build errors.
Optimize performance of SecondaryCommandBuffers and enable them as the
default build option.
To disable this set angle_enable_custom_vulkan_cmd_buffers=false in
your build args.
This CL enhances the PoolAllocator to have a "fast" mode that can
be enabled at class creation. This mode uses an alignment of 1 byte and
enables a fastAllocation() call that avoids some bookkeeping overhead.
The SecondaryCommandBuffer uses this fastAllocation() function.
Furthermore the fast path of fast allocate, using the current page,
is inlined for maximum speed.
Jamie Madill also updated the SecondaryCommandBuffers to pre-allocate
blocks so that the commands occur linearly in memory. This speeds up
processing with improved cache coherency and minimizes overhead when
recording commands.
Also the core Draw functions and their state updates are all inlined
as well as the common functions to initialize commands and to copy
command pointer data.
This change also includes some new, custom commands. One is
imageBarrier that is a specialized version of pipelineBarrier that only
performs a single image layout transition.
There are customized versions of various Draw commands to minimize
copying of parameters.
There are also specialized commands to bind[Graphics|Compute]Pipeline
that have the pipeline type built in to the command.
More custom commands and command data size optimizations will be made
in follow-on commits.
Bug: angleproject:3136
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1497418
Commit-Queue: Tobin Ehlis <tobine@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Change-Id: I621d8f8893308fca240b32390928e8ba0036cf06
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1535385
Reviewed-by: Tobin Ehlis <tobine@google.com>
|
|
|
896e7811
|
2019-03-22T14:56:33
|
|
Revert "Vulkan:Optimize SecondaryCommandBuffers"
This reverts commit 2219b18c984ed69251f3db3c7b5fd69a2fa68c77.
Reason for revert: Failing to compile on ASAN builders:
https://ci.chromium.org/p/chromium/builders/try/linux-libfuzzer-asan-rel/134782
Currently blocking roll.
Original change's description:
> Vulkan:Optimize SecondaryCommandBuffers
>
> Optimize performance of SecondaryCommandBuffers and enable them as the
> default build option.
> To disable this set angle_enable_custom_vulkan_cmd_buffers=false in
> your build args.
>
> This CL enhances the PoolAllocator to have a "fast" mode that can
> be enabled at class creation. This mode uses an alignment of 1 byte and
> enables a fastAllocation() call that avoids some bookkeeping overhead.
> The SecondaryCommandBuffer uses this fastAllocation() function.
> Furthermore the fast path of fast allocate, using the current page,
> is inlined for maximum speed.
> Jamie Madill also updated the SecondaryCommandBuffers to pre-allocate
> blocks so that the commands occur linearly in memory. This speeds up
> processing with improved cache coherency and minimizes overhead when
> recording commands.
> Also the core Draw functions and their state updates are all inlined
> as well as the common functions to initialize commands and to copy
> command pointer data.
>
> This change also includes some new, custom commands. One is
> imageBarrier that is a specialized version of pipelineBarrier that only
> performs a single image layout transition.
> There are customized versions of various Draw commands to minimize
> copying of parameters.
> There are also specialized commands to bind[Graphics|Compute]Pipeline
> that have the pipeline type built in to the command.
> More custom commands and command data size optimizations will be made
> in follow-on commits.
>
> Bug: angleproject:3136
> Change-Id: I35453cc2656bc8c51f0d84d1adef106900aca9a5
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1497418
> Commit-Queue: Tobin Ehlis <tobine@google.com>
> Reviewed-by: Jamie Madill <jmadill@chromium.org>
TBR=tobine@google.com,syoussefi@chromium.org,jmadill@chromium.org
Change-Id: I1c0bfe864ff343eb8ea6c88556523f8715c981d5
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: angleproject:3136
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1535998
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
2219b18c
|
2019-03-05T07:06:32
|
|
Vulkan:Optimize SecondaryCommandBuffers
Optimize performance of SecondaryCommandBuffers and enable them as the
default build option.
To disable this set angle_enable_custom_vulkan_cmd_buffers=false in
your build args.
This CL enhances the PoolAllocator to have a "fast" mode that can
be enabled at class creation. This mode uses an alignment of 1 byte and
enables a fastAllocation() call that avoids some bookkeeping overhead.
The SecondaryCommandBuffer uses this fastAllocation() function.
Furthermore the fast path of fast allocate, using the current page,
is inlined for maximum speed.
Jamie Madill also updated the SecondaryCommandBuffers to pre-allocate
blocks so that the commands occur linearly in memory. This speeds up
processing with improved cache coherency and minimizes overhead when
recording commands.
Also the core Draw functions and their state updates are all inlined
as well as the common functions to initialize commands and to copy
command pointer data.
This change also includes some new, custom commands. One is
imageBarrier that is a specialized version of pipelineBarrier that only
performs a single image layout transition.
There are customized versions of various Draw commands to minimize
copying of parameters.
There are also specialized commands to bind[Graphics|Compute]Pipeline
that have the pipeline type built in to the command.
More custom commands and command data size optimizations will be made
in follow-on commits.
Bug: angleproject:3136
Change-Id: I35453cc2656bc8c51f0d84d1adef106900aca9a5
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1497418
Commit-Queue: Tobin Ehlis <tobine@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
134425c7
|
2019-03-15T17:02:17
|
|
Vulkan:Integrate SecondaryCommandBuffers
Integrate the custom SecondaryCommandBuffer type into the CommandGraph
nodes by adding new ANGLE_USE_CUSTOM_VULKAN_CMD_BUFFERS define that can
be set in the BUILD gn args with angle_enable_custom_vulkan_cmd_buffers
set to "true."
Initially the custom cmd buffers are disabled by default.
This adds some support functions to SecondaryCommandBuffer to make the
integration easier by matching the wrapped cmd buffer interface:
initialize(), end(), valid().
Bug: angleproject:3136
Change-Id: Ib910554583192550757bb8ce89914e3ea8737988
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1526556
Commit-Queue: Tobin Ehlis <tobine@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
1581ff03
|
2019-02-28T20:59:26
|
|
Vulkan: disable VK_PRESENT_MODE_FIFO_KHR on Linux Intel
Causes random timeouts.
Bug: angleproject:2932, angleproject:3153
Change-Id: I69a200684afab9e00642850bfa8e8efd0da5ad1d
Reviewed-on: https://chromium-review.googlesource.com/c/1495229
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@google.com>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
|
|
|
a8b81590
|
2019-02-15T17:23:23
|
|
Vulkan: Free semaphores with other in flight resources
ANGLE calls WindowSurfaceVk::nextSwapchainImage() to prepare for
the next frame, which allocates an image (and other resources,
like semaphore). However, it's possible for
egl::Display::destroySurface() to be called before the next
submitFrame(), which prevents the image from being returned back
to the swapchain.
This fix refactors swapImpl() to separate out the present work, which
can be called as part of destroy() to return the allocated resources.
dEQP-EGL.functional.query_context.get_current_context.*
dEQP-EGL.functional.query_context.get_current_display.*
dEQP-EGL.functional.query_context.get_current_surface.r*
dEQP-EGL.functional.query_context.query_context.*
Bug: angleproject:3057
Change-Id: Ia943f2530e4f27d1a10ee907e810f0bbdb3bfa9a
Reviewed-on: https://chromium-review.googlesource.com/c/1474601
Reviewed-by: Ian Elliott <ianelliott@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
229fc83d
|
2019-01-21T18:09:15
|
|
Vulkan: Port renderer to Fuchsia (reland)
Add DisplayVk and WindowSurfaceVk subclasses for Fuchsia to the vulkan
renderer, as well as an implementation of OSWindow that renders
fullscreen for the test suite.
Disallow use of the vulkan loader from third_party as Fuchsia uses a fork
of the loader and has not sent those changes upstream yet.
Add a small wayland-inspired library libfuchsia-egl to provide a type
"struct fuchsia_egl_window" to use as EGLNativeWindowType. This type
combines a zx_handle_t to an image pipe channel and a surface size.
Image pipes can only be used once to create a VkSurfaceKHR. This means we
have to recreate the pipe in tests that call eglCreateWindowSurface more
than once with a single OSWindow, or the second call will fail. Add a
resetNativeWindow() method to accomplish this.
Reland disabling -Wextra-semi.
BUG=angleproject:2475
TEST=angle_end2end_tests on Fuchsia
Change-Id: Ie91715bcd760c6c04d4b8a02a91daa71e32ee30c
Reviewed-on: https://chromium-review.googlesource.com/c/1467603
Commit-Queue: Michael Spang <spang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
4a298703
|
2019-01-18T10:49:36
|
|
Vulkan: Support creating EGL images from non-zero mipmaps of textures.
Store a mip offset in TextureVK to apply to all operations on the ImageHelper.
There is no need to store the mip offset in RenderbufferVk because it creates
the resource with the mip offset on the call to setStorageEGLImageTarget.
Store a mipmap level in the RenderTargetVk object so that clear operations
will target the correct mipmap of the image.
BUG=angleproject:2668
Change-Id: Ie976e3dd3a8de8135a7fbb8c84bd51eec0dddce8
Reviewed-on: https://chromium-review.googlesource.com/c/1422059
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
3f0ff409
|
2019-02-12T03:08:51
|
|
Revert "Vulkan: Port renderer to Fuchsia"
This reverts commit 991d1cfb5e5a2f20a46961da3ad85b95da30b88a.
Reason for revert: Failing compile on the fuchsia bots on the ANGLE
auto-roller.
https://ci.chromium.org/p/chromium/builders/luci.chromium.try/fuchsia_x64/209198
In file included from ../../third_party/angle/util/fuchsia/ScenicWindow.cpp:10:
In file included from ../../third_party/angle/util/fuchsia/ScenicWindow.h:13:
In file included from gen/third_party/fuchsia-sdk/sdk/fidl/fuchsia/ui/policy/cpp/fidl.h:5:
In file included from ../../third_party/fuchsia-sdk/sdk/pkg/fidl_cpp/include/lib/fidl/cpp/internal/header.h:12:
../../third_party/fuchsia-sdk/sdk/pkg/fit/include/lib/fit/function.h:135:6: error: extra ';' after member function definition [-Werror,-Wextra-semi]
};
^
1 error generated.
Original change's description:
> Vulkan: Port renderer to Fuchsia
>
> Add DisplayVk and WindowSurfaceVk subclasses for Fuchsia to the vulkan
> renderer, as well as an implementation of OSWindow that renders
> fullscreen for the test suite.
>
> Disallow use of the vulkan loader from third_party as Fuchsia uses a fork
> of the loader and has not sent those changes upstream yet.
>
> Add a small wayland-inspired library libfuchsia-egl to provide a type
> "struct fuchsia_egl_window" to use as EGLNativeWindowType. This type
> combines a zx_handle_t to an image pipe channel and a surface size.
>
> Image pipes can only be used once to create a VkSurfaceKHR. This means we
> have to recreate the pipe in tests that call eglCreateWindowSurface more
> than once with a single OSWindow, or the second call will fail. Add a
> resetNativeWindow() method to accomplish this.
>
> BUG=angleproject:2475
> TEST=angle_end2end_tests on Fuchsia
>
> Change-Id: I71a613a362dd1c8aada49a3c02ae461e064457bf
> Reviewed-on: https://chromium-review.googlesource.com/c/1446496
> Commit-Queue: Michael Spang <spang@chromium.org>
> Reviewed-by: Jamie Madill <jmadill@chromium.org>
TBR=spang@chromium.org,geofflang@chromium.org,jmadill@chromium.org,syoussefi@chromium.org
Change-Id: I2d9abefa9db5363ba63a17c1773d0e147040d055
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: angleproject:2475
Reviewed-on: https://chromium-review.googlesource.com/c/1465761
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
991d1cfb
|
2019-01-21T18:09:15
|
|
Vulkan: Port renderer to Fuchsia
Add DisplayVk and WindowSurfaceVk subclasses for Fuchsia to the vulkan
renderer, as well as an implementation of OSWindow that renders
fullscreen for the test suite.
Disallow use of the vulkan loader from third_party as Fuchsia uses a fork
of the loader and has not sent those changes upstream yet.
Add a small wayland-inspired library libfuchsia-egl to provide a type
"struct fuchsia_egl_window" to use as EGLNativeWindowType. This type
combines a zx_handle_t to an image pipe channel and a surface size.
Image pipes can only be used once to create a VkSurfaceKHR. This means we
have to recreate the pipe in tests that call eglCreateWindowSurface more
than once with a single OSWindow, or the second call will fail. Add a
resetNativeWindow() method to accomplish this.
BUG=angleproject:2475
TEST=angle_end2end_tests on Fuchsia
Change-Id: I71a613a362dd1c8aada49a3c02ae461e064457bf
Reviewed-on: https://chromium-review.googlesource.com/c/1446496
Commit-Queue: Michael Spang <spang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
0029dfe2
|
2019-02-08T18:45:07
|
|
Vulkan: Respect EGL_HEIGHT & EGL_WIDTH in SurfaceVk
If currentExtent is unspecified and EGL_WIDTH & EGL_HEIGHT are provided,
we'll use a surface size of 0xFFFFFFFF. That's not going to work and
appears to be a typo. Use the size specified in EGL_WIDTH & EGL_HEIGHT
instead.
BUG=angleproject:2475
TEST=angle_end2end_tests on Fuchsia with an earlier version of
CL 1446496 PS12. It is N/A with later patches.
Change-Id: I7e67b2c796318bed512dd6d5d22ef2183916416d
Reviewed-on: https://chromium-review.googlesource.com/c/1461637
Commit-Queue: Michael Spang <spang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
38833111
|
2019-02-06T16:19:49
|
|
Vulkan: implement eglSwapInterval
Uses FIFO with interval 1 and mailbox/immediate with interval 0 (with a
fallback to FIFO). An end2end test is added to make sure the present
mode is correctly set.
Bug: angleproject:2932
Change-Id: I45d2b1e551b5c63c42ca3c8964bd5e62abd2d459
Reviewed-on: https://chromium-review.googlesource.com/c/1456622
Reviewed-by: Ian Elliott <ianelliott@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
f256339a
|
2019-01-15T09:50:44
|
|
Vulkan: Implement eglBindTexImage
Now that vk::ImageHelper is stored as a pointer, it can be shared between an
offscreen surface and texture.
BUG=angleproject:3073
Change-Id: I91e520259106eef497950b8b2e622fbf910f7444
Reviewed-on: https://chromium-review.googlesource.com/c/1412234
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
ab5acbd5
|
2019-01-23T13:58:09
|
|
Vulkan: implement swapchain resizing
Vulkan allows a swapchain to be created based off of an older swapchain
that's still presenting, to support seamless window resizing. The old
swapchain will remain alive (though no image can be acquired from it)
and automatically cleaned once no image is being presented from it.
The retired swapchain can be destroyed once all operations on its images
are completed. We store the old swapchain next to the serial of
submission that's used for CPU throttling and defer the destroy call
until after `finishToSerial` for that serial is called.
Bug: angleproject:2942
Change-Id: Ic62a5a57b712ffa2b087f5fecde0dc8942194075
Reviewed-on: https://chromium-review.googlesource.com/c/1435634
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
|
|
|
eb47e2c4
|
2019-02-05T16:26:41
|
|
Vulkan: improve calculation of swapchain image count
Based on present mode, different values for the image count make sense.
In mailbox mode, the driver will increase the number as necessary. In
immediate mode, two buffers suffice. In fifo mode, we would ideally
want triple-buffering.
Bug: angleproject:2932
Change-Id: I45eec83d0292eea472b931fe81e49cde6d105c27
Reviewed-on: https://chromium-review.googlesource.com/c/1454290
Reviewed-by: Ian Elliott <ianelliott@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
7dafe3eb
|
2019-01-28T11:39:15
|
|
Vulkan: optimize image memory barriers
Each image was tracking its current layout, but not the pipeline stage
it was used. Additionally, the barrier access masks were inferred from
the layout. This incurred two inefficiencies:
- The src pipeline stage mask often included all stages, causing
unnecessarily heavy barriers.
- The access masks included all possible accesses by a layout, which in
some cases was overkill, like VK_ACCESS_MEMORY_WRITE_BIT for
VK_IMAGE_LAYOUT_GENERAL (which will eventually used for compute shader
output).
This change instead creates an enum where each element represents the
layout, the stage and access masks when transitioning into the layout
and the stage and access masks when transitioning out of that layout.
The image will instead track a value of this enum (instead of
VkImageLayout), which allows it to create the layout transition barriers
as tight as possible, since it includes all the necessary information.
Bug: angleproject:2999
Change-Id: I91535ce06d10530a6fc217ad3b94b7e288521e25
Reviewed-on: https://chromium-review.googlesource.com/c/1440074
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
f3e823db
|
2019-01-14T12:40:34
|
|
Vulkan: Store ImageHelper as a pointer in TextureVk and RenderbufferVk
Storing ImageHelper as a pointer allows the storage to be swapped or shared
with other objects.
BUG=angleproject:2668
Change-Id: I2e51f24737be59ffe9f472e9b0b592774a792cd1
Reviewed-on: https://chromium-review.googlesource.com/c/1409404
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
08f8fa66
|
2019-01-10T14:25:47
|
|
Vulkan: Merge the PixelBuffer logic into ImageHelper.
To support EGL images efficiently, the pixel buffer needs to be tightly coupled
with the image it will upload data to allowing textures to upload data to EGL
images sourced from renderbuffers. The data will be uploaded to the image in
the correct order if it comes from multiple sources and the check for pending
uploads does not need to traverse EGL image siblings.
BUG=angleproject:2668
Change-Id: Iebf59ed1e358ddc76843b8bcfac39b0363f66a3f
Reviewed-on: https://chromium-review.googlesource.com/c/1405710
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
c2ee2ccb
|
2019-01-11T10:39:15
|
|
Vulkan: Don't support EGL_SWAP_BEHAVIOR_PRESERVED_BIT
dEQP-EGL.functional.preserve_swap.preserve.no_read_before_swap#no_draw
dEQP-EGL.functional.preserve_swap.preserve.no_read_before_swap#post_clear
dEQP-EGL.functional.preserve_swap.preserve.no_read_before_swap#post_render
dEQP-EGL.functional.preserve_swap.preserve.no_read_before_swap#pre_clear
dEQP-EGL.functional.preserve_swap.preserve.no_read_before_swap#pre_clear_post_clear
dEQP-EGL.functional.preserve_swap.preserve.no_read_before_swap#pre_clear_post_render
dEQP-EGL.functional.preserve_swap.preserve.no_read_before_swap#pre_render
dEQP-EGL.functional.preserve_swap.preserve.no_read_before_swap#pre_render_post_clear
dEQP-EGL.functional.preserve_swap.preserve.no_read_before_swap#pre_render_post_render
Bug: angleproject:3051 angleproject:2716
Change-Id: I3607f45cd2f447fb0a95f91ffd9aa93000dfa18c
Reviewed-on: https://chromium-review.googlesource.com/c/1407248
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Tim Van Patten <timvp@google.com>
|
|
|
b5ba549a
|
2019-01-02T15:19:22
|
|
Vulkan: Shader path for texture copy when image is not initialized
This change implements staging image/texture copies when the destination
image is not yet fully initialized. With this change, CPU readback for
glCopyTex[Sub]Image2D and glCopy[Sub]TextureCHROMIUM should happen only
if the texture formats don't allow a fragment-shader based copy.
Bug: angleproject:2958
Change-Id: I04087e14ea8fb6fbc731598c5493e44651c22c01
Reviewed-on: https://chromium-review.googlesource.com/c/1393909
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
29b49417
|
2019-01-07T14:03:06
|
|
Make copy texture test more extensive
By doing the copy multiple times, we exercise both paths where the
destination is already initialized and when it's not.
This adds tests for all combinations of formats and flags.
Bug: angleproject:2958
Change-Id: I56afb44496acd1b4d5a8527f4dbee29afbac9c81
Reviewed-on: https://chromium-review.googlesource.com/c/1398643
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
f83a28a6
|
2018-12-09T03:48:34
|
|
Vulkan: Shader path for framebuffer-to-texture copy
Part 1 in a series of changes to perform image copies on the GPU.
Bug: angleproject:2958
Change-Id: I6264a880865c4738c0866f2dc71af63425fc4118
Reviewed-on: https://chromium-review.googlesource.com/c/1370724
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
|
bcb78908
|
2018-12-19T11:46:29
|
|
Implement eglSwapBuffersWithDamage on top of VK_KHR_incremental_present
Bug: angleproject:2510
Change-Id: I12b0877f787dbcb48e2890f54ba4bc8ebe8294b4
Reviewed-on: https://chromium-review.googlesource.com/c/1383373
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Ian Elliott <ianelliott@google.com>
|
|
|
7c985f5c
|
2018-11-29T18:16:17
|
|
Make angle::Result an enum.
This moves away from a class type to a value type. This should improve
performance when using angle::Result as a return value. Previously the
generated code would return a pointer instead of a value.
Improves performance in the most targeted microbenchmark by 10%. In
more realistic scanarios it will have a smaller improvement. Also
simplifies the class implementation and usage.
Includes some unrelated code generation changes.
Bug: angleproject:2491
Change-Id: Ifcf86870bf1c00a2f73c39ea6e4f05ca705050aa
Reviewed-on: https://chromium-review.googlesource.com/c/1356139
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|