We schedule additional composites as long as there is a CompositeUntil timestamp. These composites need a special tag and shouldn't carry the tags of whatever led us to check the timestamps. In addition, we should add the ASYNC_IMAGE tag when GetAndResetWillGenerateFrame returns true even if some other reason led us to check it.
Depends on D142233
Differential Revision: https://phabricator.services.mozilla.com/D142234
Also differentiate between skipped composites (too many pending frames) and discarded composites (paused or no display list).
Differential Revision: https://phabricator.services.mozilla.com/D142232
RecvClearCachedResources() does not handle parent commands. They are handled by a different transaction. It caused the problem. To avoid the problem, modifying RecvClearCachedResources() as to handle parent commands.
Differential Revision: https://phabricator.services.mozilla.com/D139568
For WebGPU, we produce the textures in the compositor process and the
content process doesn't need to be that involved except for hooking up
the texture to the display list. Currently this is done via an external
image ID.
Given that WebGPU needs to work with OffscreenCanvas, it would be best
if its display pipeline was consistent whether it was gotten from an
HTMLCanvasElement, OffscreenCanvas on the main thread, or on a worker
thread. As such, using an AsyncImagePipeline would be best.
However there is no real need to bounce the handles across process
boundaries. Hence this patch which adds CompositableInProcessManager.
This static class is responsible for collecting WebRenderImageHost
objects backed by TextureHost objects which do not leave the compositor
process. This will allow WebGPUParent to schedule compositions directly
in future patches.
Differential Revision: https://phabricator.services.mozilla.com/D138588
For WebGPU, we produce the textures in the compositor process and the
content process doesn't need to be that involved except for hooking up
the texture to the display list. Currently this is done via an external
image ID.
Given that WebGPU needs to work with OffscreenCanvas, it would be best
if its display pipeline was consistent whether it was gotten from an
HTMLCanvasElement, OffscreenCanvas on the main thread, or on a worker
thread. As such, using an AsyncImagePipeline would be best.
However there is no real need to bounce the handles across process
boundaries. Hence this patch which adds CompositableInProcessManager.
This static class is responsible for collecting WebRenderImageHost
objects backed by TextureHost objects which do not leave the compositor
process. This will allow WebGPUParent to schedule compositions directly
in future patches.
Differential Revision: https://phabricator.services.mozilla.com/D138588
When we encounter an error updating resources, or processing parent
commands, we actually bail out midway. This is problematic because some
resource updates and commands depend on follow-ups to release resources
and such. Also, when the next display list update comes in, we may now
have an incomplete resource set and see display errors / crashes that
may not be easily traceable to the previous failure.
This patch makes us always finish processing a transaction, even if we
encountered errors. This will put us in a more consistent state at
least. It adds new asserts to try to catch these issues sooner and in
automation. It also ensures we do a print to the critical log in all of
these cases so we know this issue was encountered.
Differential Revision: https://phabricator.services.mozilla.com/D135613
Now that most calls to `profiler_thread_is_being_profiled` have been updated, the `ThreadProfilingFeatures` can be made compulsory, to force callers to think about what they really want to know about the current profiling state.
Differential Revision: https://phabricator.services.mozilla.com/D130010
`profiler_thread_is_being_profiled` is used a lot for markers, so it makes sense to have a specialized version, which is a bit shorter, and lives in ProfilerMarkers.h.
Differential Revision: https://phabricator.services.mozilla.com/D130009
Now that most calls to `profiler_thread_is_being_profiled` have been updated, the `ThreadProfilingFeatures` can be made compulsory, to force callers to think about what they really want to know about the current profiling state.
Differential Revision: https://phabricator.services.mozilla.com/D130010
`profiler_thread_is_being_profiled` is used a lot for markers, so it makes sense to have a specialized version, which is a bit shorter, and lives in ProfilerMarkers.h.
Differential Revision: https://phabricator.services.mozilla.com/D130009
The default size on most platforms is 256*256 which corresponds to the default blob tile size. I didn't check on android so I set it to 512*512 above which we never batch the upload so that the behavior is unchanged, but I suspect that a smaller threshold like 256*256 would also work better there.
On Windows with heavy blob image workloads, not batching gives a 20%-30% improvement to the time spent in update_texture_cache.
Differential Revision: https://phabricator.services.mozilla.com/D129516
This patch adds plumbing to keep track of why we request frames to be rendered.
This information is then displayed in gecko profile markers on the renderer thread as well as in profiler HUD counters (See "Render reasons" in profiler.rs).
Differential Revision: https://phabricator.services.mozilla.com/D127274
Right now adding a dynamic pref for webrender involves plumbing in gfxPlatform, gfxVars, CompositBridgeParent, WebRenderBridgeParent, WebRenderAPI, the bindings, and then messaging the pref change to the right thread in WebRender.
That's quite a bit time consuming for something we tend to do for multiple things.
We already have a pretty good system for boolean debug prefs where the pref only needs to be listed in gfxPlatform and in the DebugPref enum of webrender_api/lib.rs, which has led us to use it for non-debug purposes.
This patch adds a system similar to the DebugFlags with values passed in a bitfield to be able to use gfxVars. In WebRender the parameters are key-value pairs which flow through the pipeline so that any component can react to them. A followup will add integer parameters.
The patch also moves a few prefs to this system and adds a pref to use PBO uploads.
Differential Revision: https://phabricator.services.mozilla.com/D126100
The prefs name is gfx.webrender.blob-tile-size and can be changed at runtime.
It changes the behavior of a wrench reftest that was ensuring that we don't crash with unreasonable tile sizes. The new behavior (sanitizing the tile size) means we can render the image while we would previously skip it.
Differential Revision: https://phabricator.services.mozilla.com/D124789
This will allow experimenting with different representations of
the spatial tree (such as interning and/or providing stable
indices during display list building). It may also simplify
future changes to the public API to expose the spatial tree
directly.
As part of these changes, refactor how the debug representation
for the capture format is (de)serialized, to make it simpler to
add different payload vector types in future.
Differential Revision: https://phabricator.services.mozilla.com/D122183
This will allow experimenting with different representations of
the spatial tree (such as interning and/or providing stable
indices during display list building). It may also simplify
future changes to the public API to expose the spatial tree
directly.
As part of these changes, refactor how the debug representation
for the capture format is (de)serialized, to make it simpler to
add different payload vector types in future.
Differential Revision: https://phabricator.services.mozilla.com/D122183
This will allow experimenting with different representations of
the spatial tree (such as interning and/or providing stable
indices during display list building). It may also simplify
future changes to the public API to expose the spatial tree
directly.
As part of these changes, refactor how the debug representation
for the capture format is (de)serialized, to make it simpler to
add different payload vector types in future.
Differential Revision: https://phabricator.services.mozilla.com/D122183
This will allow experimenting with different representations of
the spatial tree (such as interning and/or providing stable
indices during display list building). It may also simplify
future changes to the public API to expose the spatial tree
directly.
As part of these changes, refactor how the debug representation
for the capture format is (de)serialized, to make it simpler to
add different payload vector types in future.
Differential Revision: https://phabricator.services.mozilla.com/D122183
We need to track window compositor state in various cases, when compositor is paused after window creation,
when compositor is paused for hidden window (EGLSurface is missing) and paused compositor when layout is not updated yet
and drawing will lead to incorrect / obsoleted window content.
In this patch we track compositor state by WindowCompositorState atomic enum (because it can be accessed from Compositor thread)
and two private methods (PauseCompositorHiddenWindow and ResumeCompositorHiddenWindow) and two public ones (PauseCompositor() and ResumeCompositor).
Private interface is used by nsWindow when EGLSurface is missing so we can't render to window - it's initial compositor pause and
pause when window is hidden.
Public interface is used by nsWindow and WebRender compositor to disable/enable rendering when window content is updated.
When compositor is disabled by public interface it will be enabled automatically after a timeout (1s) or when content layout is updated.
Changes in the patch:
- Implement GtkCompositorWidget::RemoteLayoutSizeUpdated() to notify GtkCompositorWidget widget about layout size update in content process.
- Track nsWindow compositor state by WindowCompositorState
- Implement nsWindow::PauseCompositorHiddenWindow() and ResumeCompositorHiddenWindow() to handle internal nsWindow states.
- Implement nsWindow::PauseCompositor() and nsWindow::ResumeCompositor() to allow compositor pause/resume during content layout updates.
- Use timeout to resume compositor paused by public interface as a fallback when content is not updated or it takes too long.
Differential Revision: https://phabricator.services.mozilla.com/D121650
Move the extra_data to be a specific cache_data separate vec in
the display list payload.
This shouldn't change any functionality, but serves as a proof
of concept for future changes which will introduce several other
separated payload vectors.
Differential Revision: https://phabricator.services.mozilla.com/D121937
