Title: [122737] trunk/Source
- Revision
- 122737
- Author
- [email protected]
- Date
- 2012-07-16 11:01:34 -0700 (Mon, 16 Jul 2012)
Log Message
[chromium] Incorrect assertion: Replicas will cause a RenderPass to be removed twice
https://bugs.webkit.org/show_bug.cgi?id=91328
Reviewed by Adrienne Walker.
Source/WebCore:
We asserted that we would never attempt to remove a render pass that had
already been removed. This was incorrect as a surface with a replica has
two quads and both may cause us to attempt its removal. We must handle
this case gracefully.
Test: CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit
* platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
(WebCore::CCLayerTreeHostImpl::CullRenderPassesWithNoQuads::shouldRemoveRenderPass):
Source/WebKit/chromium:
Add replicas to the surfaces in the test.
* tests/CCLayerTreeHostTest.cpp:
(WTF::CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit::CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit):
(WTF::CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit::beginTest):
(WTF::CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit::afterTest):
(CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit):
Modified Paths
Diff
Modified: trunk/Source/WebCore/ChangeLog (122736 => 122737)
--- trunk/Source/WebCore/ChangeLog 2012-07-16 17:51:06 UTC (rev 122736)
+++ trunk/Source/WebCore/ChangeLog 2012-07-16 18:01:34 UTC (rev 122737)
@@ -1,3 +1,20 @@
+2012-07-16 Dana Jansens <[email protected]>
+
+ [chromium] Incorrect assertion: Replicas will cause a RenderPass to be removed twice
+ https://bugs.webkit.org/show_bug.cgi?id=91328
+
+ Reviewed by Adrienne Walker.
+
+ We asserted that we would never attempt to remove a render pass that had
+ already been removed. This was incorrect as a surface with a replica has
+ two quads and both may cause us to attempt its removal. We must handle
+ this case gracefully.
+
+ Test: CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit
+
+ * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
+ (WebCore::CCLayerTreeHostImpl::CullRenderPassesWithNoQuads::shouldRemoveRenderPass):
+
2012-07-16 Ilya Tikhonovsky <[email protected]>
Web Inspector: native memory: fix instrumentation for string members
Modified: trunk/Source/WebCore/platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp (122736 => 122737)
--- trunk/Source/WebCore/platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp 2012-07-16 17:51:06 UTC (rev 122736)
+++ trunk/Source/WebCore/platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp 2012-07-16 18:01:34 UTC (rev 122737)
@@ -447,8 +447,11 @@
{
const CCRenderPass* renderPass = findRenderPassById(quad.renderPassId(), frame);
size_t passIndex = frame.renderPasses.find(renderPass);
- ASSERT(passIndex != notFound);
+ bool renderPassAlreadyRemoved = passIndex == notFound;
+ if (renderPassAlreadyRemoved)
+ return false;
+
// If any quad or RenderPass draws into this RenderPass, then keep it.
const CCQuadList& quadList = frame.renderPasses[passIndex]->quadList();
for (CCQuadList::constBackToFrontIterator quadListIterator = quadList.backToFrontBegin(); quadListIterator != quadList.backToFrontEnd(); ++quadListIterator) {
Modified: trunk/Source/WebKit/chromium/ChangeLog (122736 => 122737)
--- trunk/Source/WebKit/chromium/ChangeLog 2012-07-16 17:51:06 UTC (rev 122736)
+++ trunk/Source/WebKit/chromium/ChangeLog 2012-07-16 18:01:34 UTC (rev 122737)
@@ -1,3 +1,18 @@
+2012-07-16 Dana Jansens <[email protected]>
+
+ [chromium] Incorrect assertion: Replicas will cause a RenderPass to be removed twice
+ https://bugs.webkit.org/show_bug.cgi?id=91328
+
+ Reviewed by Adrienne Walker.
+
+ Add replicas to the surfaces in the test.
+
+ * tests/CCLayerTreeHostTest.cpp:
+ (WTF::CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit::CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit):
+ (WTF::CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit::beginTest):
+ (WTF::CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit::afterTest):
+ (CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit):
+
2012-07-13 Eric Penner <[email protected]>
[chromium] Add 'self-managed' option to CCPrioritizedTexture to enable render-surface and canvas use cases.
Modified: trunk/Source/WebKit/chromium/tests/CCLayerTreeHostTest.cpp (122736 => 122737)
--- trunk/Source/WebKit/chromium/tests/CCLayerTreeHostTest.cpp 2012-07-16 17:51:06 UTC (rev 122736)
+++ trunk/Source/WebKit/chromium/tests/CCLayerTreeHostTest.cpp 2012-07-16 18:01:34 UTC (rev 122737)
@@ -2267,7 +2267,9 @@
CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit()
: m_rootLayer(ContentLayerChromiumWithUpdateTracking::create(&m_mockDelegate))
, m_surfaceLayer1(ContentLayerChromiumWithUpdateTracking::create(&m_mockDelegate))
+ , m_replicaLayer1(ContentLayerChromiumWithUpdateTracking::create(&m_mockDelegate))
, m_surfaceLayer2(ContentLayerChromiumWithUpdateTracking::create(&m_mockDelegate))
+ , m_replicaLayer2(ContentLayerChromiumWithUpdateTracking::create(&m_mockDelegate))
{
}
@@ -2283,6 +2285,9 @@
m_surfaceLayer2->setForceRenderSurface(true);
m_surfaceLayer2->setOpacity(0.5);
+ m_surfaceLayer1->setReplicaLayer(m_replicaLayer1.get());
+ m_surfaceLayer2->setReplicaLayer(m_replicaLayer2.get());
+
m_rootLayer->addChild(m_surfaceLayer1);
m_surfaceLayer1->addChild(m_surfaceLayer2);
m_layerTreeHost->setRootLayer(m_rootLayer);
@@ -2321,14 +2326,18 @@
// Clear layer references so CCLayerTreeHost dies.
m_rootLayer.clear();
m_surfaceLayer1.clear();
+ m_replicaLayer1.clear();
m_surfaceLayer2.clear();
+ m_replicaLayer2.clear();
}
private:
MockContentLayerDelegate m_mockDelegate;
RefPtr<ContentLayerChromiumWithUpdateTracking> m_rootLayer;
RefPtr<ContentLayerChromiumWithUpdateTracking> m_surfaceLayer1;
+ RefPtr<ContentLayerChromiumWithUpdateTracking> m_replicaLayer1;
RefPtr<ContentLayerChromiumWithUpdateTracking> m_surfaceLayer2;
+ RefPtr<ContentLayerChromiumWithUpdateTracking> m_replicaLayer2;
};
SINGLE_AND_MULTI_THREAD_TEST_F(CCLayerTreeHostTestSurfaceNotAllocatedForLayersOutsideMemoryLimit)
_______________________________________________
webkit-changes mailing list
[email protected]
http://lists.webkit.org/mailman/listinfo/webkit-changes
