|
ecc378cc
|
2025-03-03T16:43:33
|
|
Reland "Dont use Subject/Observer for SwapchainImageChanged"
This is reland of the following CL
https://chromium-review.googlesource.com/c/angle/angle/+/6319893
Because we do deferred ANI (VkAcquireNextImage) call until image is
needed, we need a way to force Context to go through
FramebufferVk::syncState call (FramebufferVk::syncState calls
WindowSurfaceVk::getAttachmentRenderTarget, which end up calling ANI.
Right now we uses subject/observer mechanism, by sending
angle::SubjectMessage::SwapchainImageChanged to all observers of
WindowSurfaceVk. In this case it is egl::Surface. Then eglSurface
redirects this message to its observers, which are all gl::Framebuffer's
attachments: color, depth, stencil. Even though only color attachment
needs to be notified, but because we don't have a separate list of
observers, depth/stencil attachment also receive the notification and
they early out. Then gl::Framebuffer sets
DIRTY_BIT_COLOR_BUFFER_CONTENTS_0 dirty bit and send the
angle::SubjectMessage::DirtyBitsFlagged to Context, which dirty DrawFBO
and ReadFBO and dirty cached state. Note that this is specific for swap
image changed case, there is no surface property change (surface
property change will still trigger the subject/observer message with
SubjectMessage::SubjectChanged message, but this occurs rarely). This
gets worse for apps that uses multiple contexts, for the example
pokemon_masters_ex has three contexts, each context has its own default
frame buffer that attach to the same surface, and we never remove
non-current context from the observer list. This end up with
egl::Surface has 12 observers and for every frame, it loop over the list
of 12 observers and send message (virtual function call) to each of
them. Color attachment also ends up sending two messages to Context, one
for Read FBO and another for Draw FBO. There are total 21 virtual
function calls. Even for single context usage, you have 6 virtual
function calls, for every frame.
EGL spec says "an EGLSurface must be current on only one thread at a
time", any other context must call EGLMakeCurrent in order to use this
surface, which will add all necessary dirty bits at that time. So we
really only need to notify current context. In this CL,
SwapchainImageChanged no longer uses subject/observer mechanism, so this
message is removed.
This CL still uses subject/observer mechanism to send DirtyBitsFlagged
from Framebuffer back to context. We could call setDrawFramebufferDirty
and setReadFramebufferDirty directly, but that will require to remove
the "const" decoration out of gl::Context which generates too much code
diff, so onStateChange(angle::SubjectMessage::DirtyBitsFlagged) is still
used.
Bug: angleproject:400711938
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6319893
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Change-Id: I017b0e8934b5194a520828fa5c4af1d6e3ef9aac
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6404621
|
|
3cfc0ce2
|
2025-03-24T07:41:33
|
|
Revert "Vulkan:Dont use Subject/Observer for SwapchainImageChanged"
This reverts commit 48103cb2f2b292cb50cc5a29546b358b2e47fd29.
Reason for revert: assert fails
https://ci.chromium.org/ui/p/angle/builders/ci/android-arm64-exp-test/7085/overview
I 22:27:33.697 77.533s _RunTestsOnDevice(17221FDF6000A4) [ RUN ] EGLAndroidAutoRefreshTest.SwapCPUThrottling/ES3_Vulkan_NoFixture
INFO:root:ERR: SurfaceVk.cpp:3165 (getCurrentFramebuffer): ! Assert failed in getCurrentFramebuffer (../../src/libANGLE/renderer/vulkan/SurfaceVk.cpp:3165): mAcquireOperation.state == ImageAcquireState::Ready
Original change's description:
> Vulkan:Dont use Subject/Observer for SwapchainImageChanged
>
> Because we do deferred ANI (VkAcquireNextImage) call until image is
> needed, we need a way to force Context to go through
> FramebufferVk::syncState call (FramebufferVk::syncState calls
> WindowSurfaceVk::getAttachmentRenderTarget, which end up calling ANI.
> Right now we uses subject/observer mechanism, by sending
> angle::SubjectMessage::SwapchainImageChanged to all observers of
> WindowSurfaceVk. In this case it is egl::Surface. Then eglSurface
> redirects this message to its observers, which are all gl::Framebuffer's
> attachments: color, depth, stencil. Even though only color attachment
> needs to be notified, but because we don't have a separate list of
> observers, depth/stencil attachment also receive the notification and
> they early out. Then gl::Framebuffer sets
> DIRTY_BIT_COLOR_BUFFER_CONTENTS_0 dirty bit and send the
> angle::SubjectMessage::DirtyBitsFlagged to Context, which dirty DrawFBO
> and ReadFBO and dirty cached state. Note that this is specific for swap
> image changed case, there is no surface property change (surface
> property change will still trigger the subject/observer message with
> SubjectMessage::SubjectChanged message, but this occurs rarely). This
> gets worse for apps that uses multiple contexts, for the example
> pokemon_masters_ex has three contexts, each context has its own default
> frame buffer that attach to the same surface, and we never remove
> non-current context from the observer list. This end up with
> egl::Surface has 12 observers and for every frame, it loop over the list
> of 12 observers and send message (virtual function call) to each of
> them. Color attachment also ends up sending two messages to Context, one
> for Read FBO and another for Draw FBO. There are total 21 virtual
> function calls. Even for single context usage, you have 6 virtual
> function calls, for every frame.
>
> EGL spec says "an EGLSurface must be current on only one thread at a
> time", any other context must call EGLMakeCurrent in order to use this
> surface, which will add all necessary dirty bits at that time. So we
> really only need to notify current context. In this CL,
> SwapchainImageChanged no longer uses subject/observer mechanism, so this
> message is removed.
>
> This CL still uses subject/observer mechanism to send DirtyBitsFlagged
> from Framebuffer back to context. We could call setDrawFramebufferDirty
> and setReadFramebufferDirty directly, but that will require to remove
> the "const" decoration out of gl::Context which generates too much code
> diff, so onStateChange(angle::SubjectMessage::DirtyBitsFlagged) is still
> used.
>
> Bug: angleproject:400711938
> Change-Id: I61354516fd0aa307714b7abd30c6b6e45ff7b496
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6319893
> Commit-Queue: Charlie Lao <cclao@google.com>
> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Bug: angleproject:400711938
Change-Id: Ib7899d1ac63a1f86af0953a1d25922578c470fc9
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6387755
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
|
|
|
48103cb2
|
2025-03-03T16:43:33
|
|
Vulkan:Dont use Subject/Observer for SwapchainImageChanged
Because we do deferred ANI (VkAcquireNextImage) call until image is
needed, we need a way to force Context to go through
FramebufferVk::syncState call (FramebufferVk::syncState calls
WindowSurfaceVk::getAttachmentRenderTarget, which end up calling ANI.
Right now we uses subject/observer mechanism, by sending
angle::SubjectMessage::SwapchainImageChanged to all observers of
WindowSurfaceVk. In this case it is egl::Surface. Then eglSurface
redirects this message to its observers, which are all gl::Framebuffer's
attachments: color, depth, stencil. Even though only color attachment
needs to be notified, but because we don't have a separate list of
observers, depth/stencil attachment also receive the notification and
they early out. Then gl::Framebuffer sets
DIRTY_BIT_COLOR_BUFFER_CONTENTS_0 dirty bit and send the
angle::SubjectMessage::DirtyBitsFlagged to Context, which dirty DrawFBO
and ReadFBO and dirty cached state. Note that this is specific for swap
image changed case, there is no surface property change (surface
property change will still trigger the subject/observer message with
SubjectMessage::SubjectChanged message, but this occurs rarely). This
gets worse for apps that uses multiple contexts, for the example
pokemon_masters_ex has three contexts, each context has its own default
frame buffer that attach to the same surface, and we never remove
non-current context from the observer list. This end up with
egl::Surface has 12 observers and for every frame, it loop over the list
of 12 observers and send message (virtual function call) to each of
them. Color attachment also ends up sending two messages to Context, one
for Read FBO and another for Draw FBO. There are total 21 virtual
function calls. Even for single context usage, you have 6 virtual
function calls, for every frame.
EGL spec says "an EGLSurface must be current on only one thread at a
time", any other context must call EGLMakeCurrent in order to use this
surface, which will add all necessary dirty bits at that time. So we
really only need to notify current context. In this CL,
SwapchainImageChanged no longer uses subject/observer mechanism, so this
message is removed.
This CL still uses subject/observer mechanism to send DirtyBitsFlagged
from Framebuffer back to context. We could call setDrawFramebufferDirty
and setReadFramebufferDirty directly, but that will require to remove
the "const" decoration out of gl::Context which generates too much code
diff, so onStateChange(angle::SubjectMessage::DirtyBitsFlagged) is still
used.
Bug: angleproject:400711938
Change-Id: I61354516fd0aa307714b7abd30c6b6e45ff7b496
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6319893
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
|
|
818915a9
|
2024-05-27T10:53:35
|
|
PPO: Propagate dirty uniforms via mPPOProgramExecutables
This eliminates the need for ProgramUniformUpdated notification
Update mDefaultUniformBlocksDirty rather than just add a check to
hasDirtyUniforms, as mDefaultUniformBlocksDirty can then be used
elsewhere (for example, calcUniformUpdateRequiredSpace)
Also adds ProgramExecutable.mIsPPO flag for an easy check for PPO
without inspecting mPPOProgramExecutables
Bug: angleproject:8666
Bug: b/335295728
Change-Id: I7725d02460104997df5c89a54d0e5ef3c3079946
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5569184
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
d05c9a5e
|
2024-01-25T13:01:49
|
|
Frontend support for QCOM foveated extensions
Add frontend state management to support foveated rendering extensions.
Bug: angleproject:8484
Test: Texture2D*Foveation*
Change-Id: I0e1be9f11b2d442207674562da760f5bfd7debc8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5208091
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: mohan maiya <m.maiya@samsung.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
ecc35205
|
2024-01-25T23:58:25
|
|
Move uniform block dirty bits to State
When glUniformBlockBinding changes the mapping from a program uniform
block to a buffer binding, all contexts in the share group need to
reprocess the affected block index. Prior to this change, the dirty
bits that indicated which blocks have their mapping redefined were
placed in the program executable, and were reset by the first context
that processed them. As a result, the other contexts in the share group
where not aware of such modifications.
Similarly, when a buffer changed in one context, the mapped program
blocks were marked dirty, with similar cross-context issues.
In this change, the dirty bits are moved to State, so every context
would react to these changes.
Bug: angleproject:8493
Change-Id: I5712002224cbc4a576bf2ac46e8e75f26ebc5b2a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5238991
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
a1143857
|
2023-12-18T15:24:42
|
|
Fix UBO dirty bits vs PPOs
This change fixes propagation of UBO dirty bits (such as through
glUniformBlockBinding and glBindBufferRange) to program pipeline
objects. Since PPOs concatenate the attached programs' UBOs in a list,
a map of program UBO indices to PPO UBO indices is introduced to
offset these dirty bits appropriately. Additionally, when the program's
executable's buffer bindings change (through glUniformBlockBinding), a
notification is send to the PPO to update its executable's buffer
binding accordingly (which is otherwise only updated during PPO link).
Bug: angleproject:8462
Change-Id: I4965ae23e6fc6cac0842e1643755e42e95d3d5cc
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5131418
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
|
|
|
7b0bb0f6
|
2023-09-01T13:52:28
|
|
Properly "install" program executables
According to GL:
- The program has an executable
- The executable is overwritten during link.
- After a failed link, queries of the executable may return
half-linked information
- On glUseProgram, the executable is installed in the context
- On glUseProgramStages, the executable is installed in the program
pipeline
- After a successful link, the executable is updated wherever the
previous executable of the program was installed.
This change implements exactly the above:
- The program's and the program pipeline's executables are now
shared_ptr. References to an executable in the context and PPO are
also through a shared_ptr. Installing an executable thus translates
to sharing the executable.
- The context and PPOs are made to not reference the program directly,
but work solely through the executable. As a result, the program is
free to create a new executable for link.
With this change, the link job will be free to modify the executable as
necessary because that will not be accessed until the link is done.
Note that previous changes made the backend executable accessed through
the frontend one, and moved all link results to the frontend and backend
executables as appropriate.
Bug: angleproject:6358
Bug: angleproject:8297
Change-Id: Ie636b23ff7420ad284d18b525ec4f5fb559dd9d1
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/4823089
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
|
|
45f91a8c
|
2023-05-08T13:11:44
|
|
Deinitialize PLS planes when their texture is deleted
The spec originally called for PLS planes to be converted to memoryless
when their texture was deleted, but this is not compatible with WebGL,
which does not support memoryless planes. Change the behavior to
deinitialize them instead.
This change requires the addition of a new observer message,
angle::SubjectMessage::TextureIDDeleted, which PLS uses to deinitialize
a plane when the app deletes its texture.
Bug: chromium:1421437
Change-Id: I58fd91003747160f0a1abc1a8a7a87668890ba1f
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/4518565
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Chris Dalton <chris@rive.app>
|
|
a9fac249
|
2022-08-12T07:01:42
|
|
Use FixedVector for Framebuffer::mDirtyColorAttachmentBindings
Bug: chromium:1336126
Change-Id: Ie997af106de09468c8520187190a857db099dff9
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3829144
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Auto-Submit: Peng Huang <penghuang@chromium.org>
|
|
|
81554b66
|
2022-07-27T16:53:03
|
|
Vulkan: Reduce Framebuffer dirtyBits for swapchain only changes
For window system framebuffer, the change to it is very limited. Most
time it is only swapchain image changes. Right now we are also setting
depth/stencil buffer dirty and processing layer count update and color
mask and blending update when only color image changed. This CL avoids
setting depth/stencil dirty bit for swap chain image changes. It also
avoids color mask and blending update (they still gets updated for draw
buffer change or draw franmebuffer binding changes).
Bug: b/240475351
Change-Id: I0697a38d5939187244d67f01c0bc53fc28e11664
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3789471
Reviewed-by: Ian Elliott <ianelliott@google.com>
Commit-Queue: Charlie Lao <cclao@google.com>
Auto-Submit: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
b186991a
|
2022-05-26T11:50:56
|
|
Vulkan: Remove non-current vertex array from buffer's observer list
When app binds a buffer to many VAOs, any buffer modification have to
loop over all VAOs that uses that buffer, even though all these VAOs are
not current except one VAO. That causes significant CPU overhead.
black_desert_mobile has this type of usage. This CL checks number of
observers on each binding and if there are excessive observers, it
removes VAO from the buffer's observer list when it becomes non-current
and adds back to observer list when it becomes current.
This CL reduces CPU overhead of black_desert_mobile (as measured with
--offset --minimumgpuwork) from 21ms to 3.69ms, on par with native GLES
driver. Note that this CL only touches glBindVertexArrayObject call code
path, so no impact to apps not using VAOs. There is also no measurable
CPU time regression on a few other app traces that uses VAOs.
Bug: angleproject:6371
Change-Id: I6b1589f3e3c768d56c1c95cebdb577e3e256737f
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3669603
Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Charlie Lao <cclao@google.com>
|
|
|
ff011779
|
2022-05-13T12:36:24
|
|
Vulkan: Let texture buffer handle BufferVk's storage change
When buffer's storage changed due to glBufferData call, texture buffer
code should also respond to this and update the texture descriptor set.
This CL merges BufferVkStorageChanged message into
InternalMemoryAllocationChanged and removed BufferVkStorageChanged all
together.
Bug: angleproject:7283
Change-Id: I230ee7268634e747d06eab1954f5a76ecf84c9d6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3646955
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: mohan maiya <m.maiya@samsung.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
7c1346bd
|
2022-01-21T11:35:45
|
|
Vulkan: Consolidate uniforms code in ProgramExecutableVk.
De-duplicates several methods. Refactoring change only.
Bug: angleproject:3570
Change-Id: Ib68bc30b3e9b1087871bb268ea292677e0c9fe2e
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3408061
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
89e11878
|
2021-12-06T09:42:46
|
|
Vulkan: Fix the UAF issue with BufferData
* Fixed the use-after-free issue with stale buffer handles
after calling BeginTransformFeedback.
* Added an observer for TransformFeedbackVk to update the
buffer handles when buffer's storage is changed and the buffer
update type is StorageRedefined.
* Added a function to TransformFeedbackVk::onDestroy() to
release the counter buffers in order to avoid crash due to
TransformFeedbackVk::end() not being called, e.g., as a
result of no glEndTransformFeedback() calls.
Bug: chromium:1274316
Change-Id: I8ed477f36e6ff89dd4764bb59af564c69efe33e2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3321789
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com>
|
|
|
30878d2a
|
2021-09-24T09:23:56
|
|
Buffer: Notify contents changed in a separate observer list.
Buffer contents changed will only notify VertexArray. Also this
list is separated from the other observers because in the future
most of the time we won't need to notify the VertexArrays when
buffer contents changed, to avoid iterating over a large list of
VertexArray observers on simple SubData calls.
Bug: angleproject:6371
Change-Id: I6142a5757868caf33dc6594a1975f817f80553e8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3182701
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
|
|
68c0da83
|
2021-09-15T12:00:08
|
|
Vulkan: Inform frontend when new buffer is allocated
When a buffer is mapped with GL_MAP_INVALIDATE_BUFFER_BIT while it's
currently in use, the Vulkan backend will allocate a new buffer, map it,
and return the pointer to the new buffer. This was missing a call to
inform the frontend that a new buffer was allocated, causing the old
buffer data to be accessed in subsequent draw calls.
The fix is to add a onStateChange(angle::SubjectMessage::SubjectGhosted)
call when the new buffer is allocated, to inform the frontend.
Bug: angleproject:5971
Bug: angleproject:6396
Test: TextureBufferTestES31.MapTextureBufferInvalidateThenWrite
Change-Id: I9984d1049ab4d6a2066f4440fc710c9b93ff6ab8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3163244
Commit-Queue: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
9158436e
|
2021-07-31T18:26:16
|
|
Vulkan: glMapBuffer(): Create new buffer (Buffer Ghosting)
When glMapBuffer() is called, if the buffer is in use but not being
written to by the GPU:
1.) Create a new buffer.
2.) Copy the contents of the old buffer into the new buffer.
3.) Map the new buffer and return the pointer.
Creating a new buffer prevents ending the renderpass and flushing the
commands to allow the in-use buffer to be mapped. This change increases
Idle Heroes performance from 40FPS to 125FPS.
Bug: angleproject:5971
Test: VulkanPerformanceCounterTest.MappingGpuReadOnlyBufferGhostsBuffer
Test: BufferDataTest.MapWriteArrayBufferDataDrawQuad
Test: BufferDataTest.MapWriteArrayBufferDataDrawArrays
Change-Id: I1d433d179f9f5110a948f191c5aedda5397acac8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3065799
Commit-Queue: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
|
|
|
a8f1a916
|
2021-08-26T17:53:56
|
|
GetTexImage: Remove syncState calls & add early error exit.
Instead of calling syncState to force the Vulkan staged textures to
flush, we ensure the Texture is marked initialized by sending a
message up to the Texture when there are no more pending updates.
Also adds handling for textures with extra mips. Instead of crashing,
we now return an error with a warning. Also adds a test with a
texture that has extra (outside the mip chain) defined levels.
Also adds a test and handling for empty texture levels.
This should lead to more consistent behaviour in the trace test
serialization logic, where often the prior logic would behave
differently depending on if GetImage was called from MEC or from
the serialization code.
Also updates the generateMipmap logic to syncState after calling
initializeContents so the updates happen in the correct order on
Vulkan.
Bug: angleproject:5133
Bug: angleproject:6336
Change-Id: Iab1c2a22b9f9435e7f3c0f870c86bbed3460482e
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3123168
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
Reviewed-by: Tim Van Patten <timvp@google.com>
|
|
16a61bbb
|
2021-08-05T23:06:22
|
|
D3D: Fix not notifying RenderTarget release in TextureD3D
This could lead to use-after-free for the RenderTarget object.
Bug: chromium:1234829
Change-Id: I73d4547b8f09f2f2cf3f7f8394f7f573fe5a4ef5
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3063858
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
0261b1d3
|
2021-03-31T10:48:08
|
|
PPO: Fix updating sampler uniforms between draws
Updating sampler uniforms when using PPOs is currently broken, since the
Context/State use the currently bound ProgramExecutable which belongs to
the PPO, but the updates only happen to the Program's executable that
the uniform belongs to.
This change updates Program::updateSamplerUniform() to update any PPO
ProgramExecutables with the updated texture information when a Program's
sampler uniforms are updated, so the Context/State use the correct data.
Bug: b/182409935
Test: ProgramPipelineTest31.SampleTextureAThenTextureB
Test: SamplersTest31.SampleTextureAThenTextureB
Change-Id: I3c4e156c6e0c781e706f321f0bd12baf484ff42a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2797951
Commit-Queue: Tim Van Patten <timvp@google.com>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
323c5f24
|
2021-03-29T17:47:53
|
|
Validate PPO sampler uniforms
"Command & Conquer: Rivals" uses PPOs and was hitting the following
assert:
angle::Result ContextVk::updateActiveTextures(const gl::Context *context)
{
...
for (size_t textureUnit : activeTextures)
{
gl::Texture *texture = textures[textureUnit];
gl::TextureType textureType = textureTypes[textureUnit];
ASSERT(textureType != gl::TextureType::InvalidEnum);
This is the same assert that is generated by the test
ProgramPipelineTest31.DifferentTextureTypes which is currently being
skipped since it's known to fail.
This CL refactors sampler validation into the ProgramExecutable to allow
PPOs to take advantage of the shared code and behave correctly, since
the necessary data is already copied into the PPO's ProgramExecutable
via ProgramExecutable::updateActiveSamplers(). This also takes advantage
of the subject/observer pattern that's already established between
programs and PPOs to ensure only the PPOs that the program is a part of
are updated when a program's sampler uniforms are updated.
Bug: angleproject:3570
Bug: b/182409935
Test: ProgramPipelineTest31.DifferentTextureTypes
Change-Id: I3d34efd66dc85e7ff23a8422cb14d5f90a5f7085
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2792862
Commit-Queue: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
|
|
|
e3b9dbef
|
2021-03-17T18:01:02
|
|
Update PPO's executable when attached program is re-linked
When a program that's being used by a PPO is successfully re-linked, the
PPO's ProgramExecutable needs to be updated. This takes advantage of the
subject/observer pattern that's already established between programs and
PPOs to ensure only the PPOs that the program is a part of are updated.
Bug: b/182409935
Test: ProgramPipelineTest31.ModifyAndRelinkShader
Change-Id: Idcc11e7e819b4a9bef02bdd71afc8ea58111acc6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2770684
Commit-Queue: Tim Van Patten <timvp@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
6eb7756d
|
2020-08-07T17:41:39
|
|
Vulkan: tell ContextVk when swapchain is re-created
For an app that only draws to the swapchain, if the swapchain is
recreated with a different rotation (as done by the ANGLE perf tests
when switching from Angry Birds 2 to Candy Crush), ContextVk is not
informed, and so the new rotation is ignored. Use the
subject-observer pattern to set the appropriate dirty bits.
Test: run_angle_perftests --gtest_filter=TracePerfTest.Run/vulkan_angry*:*vulkan_candy* --verbose --local-output
Bug: angleproject:4910
Bug: b/163126746
Change-Id: Ib5303e9c4095db1b3e736911f483589e40a73d0c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2341768
Commit-Queue: Ian Elliott <ianelliott@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
|
|
|
1c88cf27
|
2020-03-18T18:00:01
|
|
Add optimized rendering feedback loop tracking.
This can be used both for WebGL and the Vulkan back-end workaround for
Manhattan. Uses the recently added tracking for Textures being bound as
samplers. Then caches this information in the Framebuffer using the
Subject/Observer pattern.
Bug: angleproject:4490
Change-Id: I08bef0a1b95c4333da19c2dae1f02a993e5835e1
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2109335
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
|
|
|
124f78c2
|
2019-06-18T11:48:24
|
|
Remove gl::Context parameter from Observer functions.
It was only used in exactly one instance in VertexArray. Instead we can
cache a bool and avoid needing to pass it around.
Will make signaling dirty easier in the Vulkan back-end.
Bug: angleproject:3539
Change-Id: Ia570aec051a24a5280df49edc4345c54022b46ec
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1663838
Reviewed-by: Jonah Ryan-Davis <jonahr@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
e4faae21
|
2019-05-10T08:27:00
|
|
Rename state change notification messages.
This makes the style use CamelCase instead of ALL_CAPS. It also cleans
up some of the naming. It also changes some uses of the messages in
some of the objects to hopefully be more consistent. See the comments
added to the enum SubjectMessage in Observer.h for more details.
Bug: angleproject:3427
Change-Id: I6dff4f6d335ecf1a27e48df65743b1490bd3025a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1600411
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
cf9383ed
|
2019-01-31T19:52:49
|
|
Optimize VertexArray::bindVertexBufferImpl.
Reduces the buffer nullptr checks to a single location. Also optimizes
how the transform feedback binding counter is changed.
Improves the score of the draw call vertex array change test.
Bug: angleproject:3014
Change-Id: I08ff341e08194a407c24143965a3d568e92b97b7
Reviewed-on: https://chromium-review.googlesource.com/c/1406891
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
b980c563
|
2018-11-27T11:34:27
|
|
Reformat all cpp and h files.
This applies git cl format --full to all ANGLE sources.
Bug: angleproject:2986
Change-Id: Ib504e618c1589332a37e97696cdc3515d739308f
Reviewed-on: https://chromium-review.googlesource.com/c/1351367
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
cd0a0a3c
|
2018-10-18T18:41:57
|
|
Introduce SubjectBindingPointer.
We can share the same pointer for the subject binding and the binding
pointer. This further allows us to optimize buffer re-binding. The
shared memory increases cache coherency and reduces the number of
instructions needed.
Bug: angleproject:2891
Change-Id: Id3162fa79de203f75989e7289ea02cb2ea1bec73
Reviewed-on: https://chromium-review.googlesource.com/c/1270217
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
|
|
|
956ab4d9
|
2018-10-10T16:13:03
|
|
Optimize several functions for the Program perf test.
This gives the same or slightly better performance in the ProgramDraw
perf test. Also only set the Program object as dirty when there are
dirty bits set in the Program itself.
Bug: angleproject:2877
Change-Id: I07b428b40d3e3c24e0a42c970524756b6dc3a30e
Reviewed-on: https://chromium-review.googlesource.com/c/1271475
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
d511948b
|
2018-09-18T11:32:42
|
|
Use FastVector in angle::Subject.
This improves object binding performance. We no longer need to check
two containers when manipulating the bound subjects/observers.
Bug: angleproject:2763
Change-Id: I0ccb2e76be47bd8a28fba5a0c3eff857aa166bb7
Reviewed-on: https://chromium-review.googlesource.com/1227795
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
d69a5f12
|
2018-08-01T11:34:48
|
|
Cache VertexArray::hasMappedBuffer.
This can be updated in several places. Also adds a test which covers
some of the paths.
Bug: angleproject:2746
Change-Id: Id119e527fd0064998d7ad5011a9d8376e7b9dab0
Reviewed-on: https://chromium-review.googlesource.com/1153569
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
7267aa65
|
2018-04-17T15:28:21
|
|
Optimize ValidateDrawAttribs: Part 3.
This is a small optimization for the WebGL compatibility mode.
Instead of scanning the list of attributes for a Transform feedback
conflict, it can quickly check a cached mask. This should save a lot
of cycles on the fast path.
Bug: angleproject:1391
Change-Id: Icb8d095493a629dbff0e93872357e5bf7c7458ae
Reviewed-on: https://chromium-review.googlesource.com/1011236
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
|
|
|
3418fe80
|
2018-04-09T14:08:19
|
|
Optimize Subject's binding method.
Using FixedVector, we can remove the need for dynamic allocation in
most of the use cases. Only when we exceed a certain amount of fixed
storage do we need to spill into the heap.
Bug: angleproject:2389
Bug: chromium:829906
Change-Id: Ib5f7073b58876bfd2400bd87b743bddcee5e5dc8
Reviewed-on: https://chromium-review.googlesource.com/1002884
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Luc Ferron <lucferron@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
|
d7518622
|
2018-03-27T09:44:31
|
|
Buffer11: Refactor Subject/Observer pattern.
Instead of having a direct/static observer distinction, add two
messages for 'Contents Changed' and 'Storage Changed'. This makes
Buffer11 itself the subject with two different message handling
cases in the onSubjectStateChange methods.
Bug: angleproject:2389
Change-Id: I645cd4b7cc7ce51cb7f48a01c7fc72939cbe89fe
Reviewed-on: https://chromium-review.googlesource.com/957940
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
|
|
879f90d2
|
2018-03-01T11:45:20
|
|
Vulkan: Avoid recopying the data every time we draw line loops
Use the Observer pattern to get notified when the BufferVk we've copied our data
from is changing. We will only recopy the data if anything has changed, but otherwise
we'll keep drawing with the same index buffer we've created previously.
Bug: angleproject:2335
Change-Id: Ib65677b4d5ec90c46a5e3b975fffd1fddeed59e7
Reviewed-on: https://chromium-review.googlesource.com/948622
Commit-Queue: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
6f683089
|
2018-02-28T00:35:16
|
|
Rename signal utils to Observer.
This completes the basic refactor to the Observer pattern.
Bug: angleproject:2372
Change-Id: I810deff7c7e39baa64b57ce2a79cd732b1af7c34
Reviewed-on: https://chromium-review.googlesource.com/940862
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Luc Ferron <lucferron@chromium.org>
|