Closed
Bug 743938
Opened 12 years ago
Closed 12 years ago
crash in glClear @ WSEGL_GetDrawableParameters
Categories
(Firefox for Android Graveyard :: General, defect)
Tracking
(blocking-fennec1.0 +)
RESOLVED
DUPLICATE
of bug 752368
Tracking | Status | |
---|---|---|
blocking-fennec1.0 | --- | + |
People
(Reporter: xti, Assigned: bjacob)
References
Details
(Keywords: crash, reproducible, topcrash, Whiteboard: [native-crash][gfx][QA^] fixed by patch in bug 752368)
Crash Data
Attachments
(1 file, 1 obsolete file)
11.08 KB,
text/plain
|
Details |
This bug was filed from the Socorro interface and is report bp-02674a6e-e511-4ba9-9dd5-a14462120410 . ============================================================= Frame Module Signature Source 0 libpvrANDROID_WSEGL.so WSEGL_GetDrawableParameters android_ws.c:977 1 libIMGegl.so KEGLGetDrawableParameters generic_ws.c:735 2 libGLESv2_POWERVR_SGX540_120.so libGLESv2_POWERVR_SGX540_120.so@0x16b3a 3 libGLESv2_POWERVR_SGX540_120.so libGLESv2_POWERVR_SGX540_120.so@0x128de 4 libGLESv2_POWERVR_SGX540_120.so libGLESv2_POWERVR_SGX540_120.so@0x62e6 5 libGLESv2.so glClear 6 libxul.so mozilla::layers::LayerManagerOGL::Render GLContext.h:2024 7 libxul.so mozilla::layers::LayerManagerOGL::EndTransaction gfx/layers/opengl/LayerManagerOGL.cpp:454 8 libxul.so mozilla::layers::LayerManagerOGL::EndEmptyTransaction gfx/layers/opengl/LayerManagerOGL.cpp:427 9 libxul.so mozilla::layers::CompositorParent::Composite gfx/layers/ipc/CompositorParent.cpp:224 10 libxul.so RunnableMethod<mozilla::layers::CompositorParent, void , Tuple0>::Run ipc/chromium/src/base/tuple.h:383 11 libxul.so MessageLoop::RunTask ipc/chromium/src/base/message_loop.cc:318 12 libxul.so MessageLoop::DeferOrRunPendingTask ipc/chromium/src/base/message_loop.cc:326 13 libxul.so MessageLoop::DoWork ipc/chromium/src/base/message_loop.cc:426 14 libxul.so base::MessagePumpDefault::Run ipc/chromium/src/base/message_pump_default.cc:23 15 libxul.so MessageLoop::RunInternal ipc/chromium/src/base/message_loop.cc:208 16 libxul.so MessageLoop::Run ipc/chromium/src/base/message_loop.cc:201 17 libxul.so base::Thread::ThreadMain ipc/chromium/src/base/thread.cc:156 18 libxul.so ThreadFunc ipc/chromium/src/base/platform_thread_posix.cc:26 19 libc.so __thread_entry 20 libc.so pthread_create This crash occurred while Fennec was partially locked on kevs3d.co.uk/dev -- Firefox 14.0a1 (2012-04-09) Device: Samsung Galaxy S (Captivate) OS: Android 2.2
Updated•12 years ago
|
Summary: crash [@ WSEGL_GetDrawableParameters ] → crash in glClear @ WSEGL_GetDrawableParameters
Whiteboard: [native-crash][gfx]
Nicolae, can you place in specific STRs for this please?
Keywords: topcrash
Reporter | ||
Comment 2•12 years ago
|
||
(In reply to Naoki Hirata :nhirata from comment #1) > Nicolae, can you place in specific STRs for this please? I can only reproduce this bug by following these steps: 1. Open Fennec 2. Browse to http://kevs3d.co.uk/dev/ 3. Tap on Tabs Menu button 4. Tap on device Back button 5. Repeat steps 3-4 for a couple of times Expected result: No crash should occur after step 5 Actual result: Fennec crashes
Updated•12 years ago
|
Keywords: reproducible
Reporter | ||
Comment 3•12 years ago
|
||
I uploaded a video that might be useful: http://youtu.be/kOO6cEoiFIk
Updated•12 years ago
|
blocking-fennec1.0: --- → ?
Comment 5•12 years ago
|
||
No reason why Benoit has to be the only person looking into GL issues - Jeff Gilbert can too!
Comment 6•12 years ago
|
||
This has dropped to #22 recently; please renom if it spikes. In the mean time, Jeff or Benoit can look at this!
blocking-fennec1.0: ? → -
Comment 7•12 years ago
|
||
It's #14 in 14.0a1 data over the last week and #3 in early 15.0a1 data, so still a topcrash.
Keywords: topcrash
Updated•12 years ago
|
blocking-fennec1.0: - → ?
Updated•12 years ago
|
blocking-fennec1.0: ? → +
Assignee | ||
Comment 9•12 years ago
|
||
I'm looking into this now, sorry for the delay, but is there a known testcase? How can I reproduce?
Assignee | ||
Comment 10•12 years ago
|
||
Oh OK, saw comment 2. Looking.
Assignee | ||
Comment 11•12 years ago
|
||
(In reply to Cristian Nicolae (:xti) from comment #2) > (In reply to Naoki Hirata :nhirata from comment #1) > > Nicolae, can you place in specific STRs for this please? > > I can only reproduce this bug by following these steps: > > 1. Open Fennec > 2. Browse to http://kevs3d.co.uk/dev/ > 3. Tap on Tabs Menu button In Step 3, what do you call the 'Tabs Menu button'? Is it the big '+' button that opens a new tab? I can't see another button in the Fennec UI.
Assignee | ||
Comment 12•12 years ago
|
||
Attached is a stack trace for a crash I get on Nexus S, when I follow these STR, interpreting 'Tabs Menu' as the big '+' button that opens a new tab. On a Galaxy S, I don't get a crash, but I get this: E/libEGL ( 3469): call to OpenGL ES API with no current context (logged once per thread) After which compositing stops working. Will try to get a stack trace for this once I've pulled the symbols off that Galaxy S.
Assignee | ||
Comment 13•12 years ago
|
||
Running into serious trouble with GDB on the Galaxy S, filed android-gdb bug: https://github.com/darchons/android-gdb/issues/5 Continuing debugging on Nexus S for now.
Assignee | ||
Comment 14•12 years ago
|
||
The Nexus S crash above was bug 754056 which has a patch now. With that patch, I can now reproduce the present WSEGL_GetDrawableParameters crash on the Nexus S. Debugging.
Depends on: 754056
Assignee | ||
Comment 15•12 years ago
|
||
Here are the stacks of the compositor thread and of the main thread during the crash. The main thread is doing a box blur while the compositor thread is doing a glClear().
Attachment #622942 -
Attachment is obsolete: true
Assignee | ||
Comment 16•12 years ago
|
||
The patch in bug 754426 removes the spurious non-threadsafe-refcounting assert failures before the crash, but doesn't change the crash itself.
Comment 17•12 years ago
|
||
It's #1 unfixed top crasher in the first days of 14.0b1.
Reporter | ||
Updated•12 years ago
|
Whiteboard: [native-crash][gfx] → [native-crash][gfx][QA^]
Assignee | ||
Comment 18•12 years ago
|
||
I know how to predict when this crash will imminently happen: that's when SwapBuffers() fails in LayerManagerOGL::Render(). Currently it doesn't check the return value of SwapBuffers(). Only question is what should we do when it fails. Recreate the LayerManager?
Assignee | ||
Comment 19•12 years ago
|
||
This crash is fixed by the patch in bug 752368.
Assignee | ||
Updated•12 years ago
|
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Whiteboard: [native-crash][gfx][QA^] → [native-crash][gfx][QA^] fixed by patch in bug 752368
Assignee | ||
Comment 20•12 years ago
|
||
Oops, didn't really mean to set as RESOLVED until more people had a chance to verify it. But let's call it a dupe.
Resolution: FIXED → DUPLICATE
Assignee | ||
Comment 21•12 years ago
|
||
Here's Vlad's tryserver build with the patch, to try: http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/vladimir@pobox.com-0e539f1d4ce4/try-android/
Updated•3 years ago
|
Product: Firefox for Android → Firefox for Android Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•