Skip to main content

Intermittent crashes in GFX

2 replies [Last post]
abysubin
Offline
Joined: 2011-07-31
Points: 0

Hi,
I'm observing two intermittent crashes while executing graphics test cases. The test case that is run is TC0713 of TDK (testVideoDeviceResolutionWithNoImpactGraphics). In this test case, we are setting the coherent configurations obtained from the graphics and video (for which ZERO_GRAPHICS_IMPACT property will be set as required) configurations which are different than the current graphics and video configurations. The gdb traces for the crashes are as follows

Crash#1:
Program received signal SIGSEGV, Segmentation fault.
[Switching to thread 1344.0x21cc]
0xfeeefeee in ?? ()
#0 0xfeeefeee in ?? ()
#1 0x18aaf284 in mpeos_gfxContextCreate (base=0x22656018, ctx=0x2080fd5c)
at ../common/gfxdfb/mpeos_context.c:482
#2 0x1bc66dd3 in Java_java_awt_MPEGraphics_pClone ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#3 0x1bc65db7 in CVMjniInvokeNative ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#4 0x1bbdc3be in CVMinvokeJNIHelper ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#5 0x1bc6593c in CVMgcUnsafeExecuteJavaMethod ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
(gdb)
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#7 0x1bc03cbc in CVMjniCallVoidMethod ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#8 0x1bc0d3be in start_func ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#9 0x18ad2c27 in threadStart (data=0x1b9c93f8)
at ../RI_Win32/mpeos_thread.c:167
#10 0x7c80b729 in KERNEL32!GetModuleFileNameA ()
from /cygdrive/c/WINDOWS/system32/kernel32.dll
#11 0x00000000 in ?? ()

Crash #2:
Program received signal SIGSEGV, Segmentation fault.
[Switching to thread 11412.0xa40]
0xfeeefeee in ?? ()
#0 0xfeeefeee in ?? ()
#1 0x18aaf398 in mpeos_gfxContextDelete (ctx=0x225a9e68)
at ../common/gfxdfb/mpeos_context.c:526
#2 0x1bc75db7 in CVMjniInvokeNative ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#3 0x1bbec3be in CVMinvokeJNIHelper ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#4 0x1bc7593c in CVMgcUnsafeExecuteJavaMethod ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#5 0x1bc139fc in CVMjniInvoke ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#6 0x1bc13cbc in CVMjniCallVoidMethod ()
from /cygdrive/d/OCAP_RI_Pjt/GIT_Checkout_NewUK/ri/RI_Stack/bin/CableLabs/sim
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#7 0x1bc1d3be in start_func ()
(gdb)
ulator/Win32/debug/env/cvm/bin/libcvm.dll
#8 0x18ad2c27 in threadStart (data=0x1b9d69d8)
at ../RI_Win32/mpeos_thread.c:167
#9 0x7c80b729 in KERNEL32!GetModuleFileNameA ()
from /cygdrive/c/WINDOWS/system32/kernel32.dll
#10 0x00000000 in ?? ()

Crash#1 has been seen to be happening during the clean-up operation of the test case. As part of clean up, we are setting the graphics and video configuration back to the original. Hence, we are primary surface to 640x480 and as part of this, the thread is releasing the Surface used. At this point, the crash occurs. This looks like AWT thread is trying to access that surface

Crash#2 happens while the test case is executing. This also looks like a race condition issue.

The logs for both these crashes are attached here with.

This issue can be reproduced easily on running the testcase after increasing the invocation count to 50.

Have anyone come across such crashes ? Please share your thoughts on this.

AttachmentSize
RILog_crash1.txt556.35 KB
RILog_crash2.txt861.14 KB

Reply viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
sarendt
Offline
Joined: 2009-07-21
Points: 0

We have been chasing a VERY similar issue off and on for some time. Please file a bug with all this info, and I will get on it.

ssathish
Offline
Joined: 2011-05-26
Points: 0

Filed an issue tracker# OCAP_RI-635.

Thanks