908520 - Nexus 4 debug build endlessly crashes on startup in GonkDisplayJB::Post

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: Hardware, defect)

Description

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Something is going wrong here. This is today's master branch build, --enable-debug, --disable-optimize.

#0  0xb10d14b8 in reset (displays=0xae2ed9f4, ctx=0xb6a3b400, 
    numDisplays=<optimized out>)
    at hardware/qcom/display/libhwcomposer/hwc.cpp:98
#1  hwc_prepare (dev=0xb6a3b400, numDisplays=2, displays=0xae2ed9f4)
    at hardware/qcom/display/libhwcomposer/hwc.cpp:214
#2  0x000206de in mozilla::GonkDisplayJB::Post (this=0xb6a1a380, 
    buf=0xac0879c0, fence=97)
    at ../../../../gecko/widget/gonk/libdisplay/GonkDisplayJB.cpp:217
#3  0x00020536 in mozilla::GonkDisplayJB::SwapBuffers (this=0xb6a1a380, 
    dpy=0x1, sur=0xadcbdee0)
    at ../../../../gecko/widget/gonk/libdisplay/GonkDisplayJB.cpp:179
#4  0xb417691a in mozilla::gl::GLContextEGL::SwapBuffers (this=0xa9d20000)
    at ../../../gecko/gfx/gl/GLContextProviderEGL.cpp:599
#5  0xb411717c in mozilla::layers::CompositorOGL::EndFrame (this=0xadc13dc0)
    at ../../../gecko/gfx/layers/opengl/CompositorOGL.cpp:1301
#6  0xb413dae4 in mozilla::layers::LayerManagerComposite::Render (
    this=0xadcab110)
    at ../../../gecko/gfx/layers/composite/LayerManagerComposite.cpp:365
#7  0xb413d29c in mozilla::layers::LayerManagerComposite::EndTransaction (
    this=0xadcab110, aCallback=0, aCallbackData=0x0, 
    aFlags=mozilla::layers::LayerManager::END_DEFAULT)
    at ../../../gecko/gfx/layers/composite/LayerManagerComposite.cpp:210
#8  0xb413d198 in mozilla::layers::LayerManagerComposite::EndEmptyTransaction (
    this=0xadcab110, aFlags=mozilla::layers::LayerManager::END_DEFAULT)
    at ../../../gecko/gfx/layers/composite/LayerManagerComposite.cpp:175
#9  0xb411c192 in mozilla::layers::CompositorParent::Composite (
    this=0xaa806c00) at ../../../gecko/gfx/layers/ipc/CompositorParent.cpp:532
#10 0xb2f906ac in DispatchToMethod<WebCore::ReverbConvolver, void (WebCore::ReverbConvolver::*)()> (obj=0xaa806c00, method=&virtual table offset 200, arg=...)
    at ../../../../../gecko/ipc/chromium/src/base/tuple.h:383
#11 0xb2f90620 in RunnableMethod<WebCore::ReverbConvolver, void (WebCore::ReverbConvolver::*)(), Tuple0>::Run (this=0xac8484c0)
    at ../../../../../gecko/ipc/chromium/src/base/task.h:307
#12 0xb40578b4 in MessageLoop::RunTask (this=0xae2edddc, task=0xac8484c0)
    at ../../../gecko/ipc/chromium/src/base/message_loop.cc:338
#13 0xb4057910 in MessageLoop::DeferOrRunPendingTask (this=0xae2edddc, 
    pending_task=...)
    at ../../../gecko/ipc/chromium/src/base/message_loop.cc:346
#14 0xb4057c68 in MessageLoop::DoWork (this=0xae2edddc)
    at ../../../gecko/ipc/chromium/src/base/message_loop.cc:446
#15 0xb405b5ce in base::MessagePumpDefault::Run (this=0xaefa7ce0, 
    delegate=0xae2edddc)
    at ../../../gecko/ipc/chromium/src/base/message_pump_default.cc:23
#16 0xb40574ec in MessageLoop::RunInternal (this=0xae2edddc)
    at ../../../gecko/ipc/chromium/src/base/message_loop.cc:220
#17 0xb4057486 in MessageLoop::RunHandler (this=0xae2edddc)
    at ../../../gecko/ipc/chromium/src/base/message_loop.cc:213
#18 0xb4057466 in MessageLoop::Run (this=0xae2edddc)
    at ../../../gecko/ipc/chromium/src/base/message_loop.cc:187
#19 0xb4060848 in base::Thread::ThreadMain (this=0xae78b2b0)
    at ../../../gecko/ipc/chromium/src/base/thread.cc:160
#20 0xb4065e4a in ThreadFunc (closure=0xae78b2b0)
    at ../../../gecko/ipc/chromium/src/base/platform_thread_posix.cc:39
#21 0xb6e13a5c in __thread_entry (func=0xb4065e31 <ThreadFunc(void*)>, 
    arg=0xae78b2b0, tls=0xae2edf00) at bionic/libc/bionic/pthread_create.cpp:92

In frame 2 mHwc is null, but we clearly called hwc_prepare with a non-null value. Somehow mHwc is getting overwritten, and maybe additional memory corruption is then responsible for the actual crash.

Comment 1

[Gnaneswar]

Tested with Nexus-4 device , It seems that issue is not reproducable on gecko commit ID: 2041a6624be1706b676d489e86fa9789d238a39e

Please share the detailed steps to reproduce it.

Comment 2

[BMO Automation]

Firefox OS is not being worked on