Branch: refs/heads/main
Home: https://github.com/WebKit/WebKit
Commit: 80f6e24ca1d10903db9e2c76dd6ea91249a45d9f
https://github.com/WebKit/WebKit/commit/80f6e24ca1d10903db9e2c76dd6ea91249a45d9f
Author: Matt Woodrow <mattwood...@apple.com>
Date: 2023-07-24 (Mon, 24 Jul 2023)
Changed paths:
M Source/WebKit/WebProcess/GPU/graphics/RemoteImageBufferProxy.cpp
Log Message:
-----------
Allocating a new IPC::Semaphore for each
RemoteImageBufferProxy::flushDrawingContextAsync adds significant overhead.
https://bugs.webkit.org/show_bug.cgi?id=259433
<rdar://112743547>
Reviewed by Dean Jackson.
Frequently we have an existing pending flush object, where the semaphore has
already been signaled.
This makes the m_signaled bool atomic, so that we can check for completed flush
objects without taking
the lock and blocking. This will still be false if the existing use of the
semaphore hasn't yet been signaled,
or if the waitFor call failed.
If we find a completed flush objects, then moves the semaphore across into the
new one to prevent a new allocation.
* Source/WebKit/WebProcess/GPU/graphics/RemoteImageBufferProxy.cpp:
(WebKit::RemoteImageBufferProxyFlushFence::tryTakeSemaphore):
(WebKit::RemoteImageBufferProxy::flushDrawingContextAsync):
(WebKit::RemoteImageBufferProxyFlushFence::WTF_GUARDED_BY_LOCK): Deleted.
Canonical link: https://commits.webkit.org/266265@main
_______________________________________________
webkit-changes mailing list
webkit-changes@lists.webkit.org
https://lists.webkit.org/mailman/listinfo/webkit-changes
