running firefox crashes after first rdp connection to that machine
Categories
(Core :: Audio/Video: Playback, defect, P2)
Tracking
()
People
(Reporter: p.huiskens, Unassigned)
Details
Attachments
(3 files)
User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:103.0) Gecko/20100101 Firefox/103.0
Steps to reproduce:
boot pc and start firefox
go to another machine and connect to the first via rdp
Actual results:
as soon as the rdp connection is established and the desktop is shown, firefox is unresponsive
extra: if i restarted firefox and everything is responsive again, i can kill the rdp connection and connect again , firefox will stay responsive
Only the first time you connect to rdp it will die
Expected results:
it should be responsive
|
||
Comment 1•2 years ago
|
||
The Bugbug bot thinks this bug should belong to the 'Core::Networking' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.
|
||
Comment 2•2 years ago
|
||
Hi,
Could you try to get a process dump when Firefox hangs?
Thanks.
|
Reporter | |
Comment 3•2 years ago
|
||
(In reply to Kershaw Chang [:kershaw] from comment #2)
Hi,
Could you try to get a process dump when Firefox hangs?
Thanks.
i did make a dumpfile
i repeated what i described and firefox instantly crashed on rdp connection
Sadly the dump is kinda big ( 560 megs) so i decided to take a look at it, but it contains way to much private stuff to simply add this dump onto the internet
is there a specific part you are looking for?
that i might past here?
|
|
||
Comment 4•2 years ago
|
||
(In reply to p.huiskens from comment #3)
(In reply to Kershaw Chang [:kershaw] from comment #2)
Hi,
Could you try to get a process dump when Firefox hangs?
Thanks.i did make a dumpfile
i repeated what i described and firefox instantly crashed on rdp connection
Sadly the dump is kinda big ( 560 megs) so i decided to take a look at it, but it contains way to much private stuff to simply add this dump onto the internetis there a specific part you are looking for?
that i might past here?
Is it possible to provide us the stacktrace?
Or could you upload the entire dump to a cloud storage and send the link to necko@mozilla.com?
Thanks.
|
||
Comment 6•2 years ago
|
||
Thank you for the dump.
I used the Debug Diagnostics Tool to analyze it and generate the report.
Not sure if it's definitely correct.
Thread 0 - System ID 8196
Entry point firefox!IsSandboxedProcess+1a80
Create time 28/07/2022 19:07:57
Time spent in user mode 0 Days 00:00:03.656
Time spent in kernel mode 0 Days 00:00:01.578
This thread is not fully resolved and may or may not be a problem. Further analysis of these threads may be required.
ntdll!NtWaitForSingleObject+14
KERNELBASE!WaitForSingleObjectEx+8e
MMDevAPI!AXB::WaitForOperations+20e29
MMDevAPI!CMediaNotifications::UnregisterMediaCallback+f8
MMDevAPI!UnregisterMediaCallback+30
AudioSes!CAudioSessionControl::~CAudioSessionControl+23f
AudioSes!CAudioSessionControl::`vector deleting destructor'+14
AudioSes!Microsoft::WRL::Details::RuntimeClassImpl<Microsoft::WRL::RuntimeClassFlags<3>,1,1,0,IInspectable,IAudioSessionControl,IAudioSessionControl2,IAudioSessionControlInternal,ISimpleAudioVolume,IChannelAudioVolume,IAudioMeterInformation,IPropertyStore,ICrossVmMediaNotificationProducer,ICrossVmMediaNotificationConsumer,Microsoft::WRL::FtmBase>::Release+65
xul!soundtouch::SoundTouch::operator=+70d3fb
xul!GIFFT_TimingDistributionCancel+7ae764
xul!GIFFT_TimingDistributionCancel+4fc029
xul!GIFFT_TimingDistributionCancel+7d3524
xul!workerlz4_compress+6148bf
xul!XRE_GetBootstrap+1d35e
xul!XRE_GetBootstrap+17fa58
xul!XRE_GetBootstrap+17e9df
xul!JOG_RegisterMetric+65821
xul!JOG_RegisterMetric+d0192
xul!JOG_RegisterMetric+d0fbc
xul!GIFFT_TimingDistributionCancel+cf99d3
firefox!GetNtLoaderAPI+14bde
firefox!IsSandboxedProcess+1a08
kernel32!BaseThreadInitThunk+14
ntdll!RtlUserThreadStart+21
If this is the main thread, then presumably it hangs because of some audio API issue.
|
|
||
Comment 7•2 years ago
|
||
I was also unable to reproduce this bug locally.
Firefox still seems to work after connecting via RDP (from a mobile device).
|
|
||
Comment 8•2 years ago
|
||
The severity field is not set for this bug.
:valentin, could you have a look please?
For more information, please visit auto_nag documentation.
|
|
||
Comment 9•2 years ago
|
||
Hi,
Could you give us another hand with this bug?
It would be great if you could get the main thread call stack while Firefox is stuck.
I found some instructions here
Thanks!
|
|
Reporter | |
Comment 10•2 years ago
|
||
i am not sure if this bug still exists.
today is the first day since i reported this issue, that it did not occur
Firefox did not crash , currently on version 103.0.2
I'll see tomorrow what happens, and if it does crash I'll gt the main thread call stack
|
|
Reporter | |
Comment 11•2 years ago
|
||
sadly today it just crashed as hard as before
Stack:
ntoskrnl.exe!KeSynchronizeExecution+0x5b66
ntoskrnl.exe!KeWaitForMutexObject+0x1460
ntoskrnl.exe!KeWaitForMutexObject+0x98f
ntoskrnl.exe!KeWaitForMutexObject+0x233
ntoskrnl.exe!ExWaitForRundownProtectionRelease+0x7dd
ntoskrnl.exe!KeWaitForMutexObject+0x3a29
ntoskrnl.exe!KeWaitForMutexObject+0x1787
ntoskrnl.exe!KeWaitForMutexObject+0x98f
ntoskrnl.exe!KeWaitForMutexObject+0x233
ntoskrnl.exe!ObWaitForSingleObject+0x91
ntoskrnl.exe!NtWaitForSingleObject+0x6a
ntoskrnl.exe!setjmpex+0x7c48
ntdll.dll!ZwWaitForSingleObject+0x14
KERNELBASE.dll!WaitForSingleObjectEx+0x8e
MMDevApi.dll!Ordinal14+0xa85d
MMDevApi.dll!Ordinal11+0x158
MMDevApi.dll!Ordinal11+0x30
AUDIOSES.DLL!DllCanUnloadNow+0x1af7
AUDIOSES.DLL!DllCanUnloadNow+0x1854
AUDIOSES.DLL!DllGetClassObject+0x1095
xul.dll!soundtouch::SoundTouch::operator=+0x70431b
xul.dll!GIFFT_TimingDistributionCancel+0x7acb57
xul.dll!GIFFT_TimingDistributionCancel+0x7ab7d3
xul.dll!GIFFT_TimingDistributionCancel+0x4fc55b
xul.dll!GIFFT_TimingDistributionCancel+0x7d23fd
xul.dll!workerlz4_compress+0x617d2f
xul.dll!XRE_GetBootstrap+0x1d1ee
xul.dll!XRE_GetBootstrap+0x1826d8
xul.dll!XRE_GetBootstrap+0x18165f
xul.dll!JOG_RegisterMetric+0x658c1
xul.dll!JOG_RegisterMetric+0xd0832
xul.dll!JOG_RegisterMetric+0xd165c
xul.dll!GIFFT_TimingDistributionCancel+0xcf7543
firefox.exe!GetNtLoaderAPI+0x14b9e
firefox.exe!IsSandboxedProcess+0x1a08
KERNEL32.DLL!BaseThreadInitThunk+0x14
ntdll.dll!RtlUserThreadStart+0x21
|
|
Reporter | |
Comment 12•2 years ago
|
||
i created another dumpfile ( specifically from the tread that was hanging according to taskmanager) and created an analasys of that , i'll ad that to the files
|
|
Reporter | |
Comment 13•2 years ago
|
||
new analasys from a new dump
|
|
||
Comment 14•2 years ago
|
||
Thank you, the stack seems similar to the one from my analysis in comment 6.
I am not really sure how we get from GIFFT_TimingDistributionCancel to SoundTouch, but the hang does seem to be somewhere in the audio library.
If I were to speculate, I suspect that RDP client also uses some of the same audio APIs and it hangs Firefox when it tries to use some of those same resources. 🤷
Paul, this is a shot in the dark - you added the SoundTouch library; do you have any idea what could be causing this hang? Is this a windows bug?
|
|
Reporter | |
Comment 15•2 years ago
|
||
today i started my rdp session with remote audio forwarding disabled,
Firefox did not crash
|
|
||
Comment 16•2 years ago
|
||
Thank you for the info. Seems we have a good lead.
Moving to Audio/Video for further investigation.
|
||
Comment 17•2 years ago
|
||
(In reply to Valentin Gosu [:valentin] (he/him) from comment #14)
Paul, this is a shot in the dark - you added the SoundTouch library; do you have any idea what could be causing this hang? Is this a windows bug?
Yeah, this happens often, it's often in stack traces because of identical code folding, not sure why it's picked.
That said, there's definitely something fishy here, theres IAudioSessionControl stuff on the stack, which is one of the only audio-related things we do on the main thread. David, would you mind having a look?
Reporter, is this a recent regression on your side, i.e., did it work well before, in a previous Firefox release?
|
|
||
Updated•2 years ago
|
|
||
Comment 18•2 years ago
|
||
Very familiar. This sounds like bug 1777518. A fix landed (in bug 1755700) in Fx 105 and was uplifted to 104 beta. @p.huiskens, would you be able to try to reproduce with the beta, or with the 104 release when it's released?
|
|
Reporter | |
Comment 19•2 years ago
|
||
(In reply to Paul Adenot (:padenot) from comment #17)
Reporter, is this a recent regression on your side, i.e., did it work well before, in a previous Firefox release?
it happened recently, i think about a month to 1.5 month ago, but i'm not fully sure when it exactly happened the first time
(In reply to David Parks
[:handyman] from comment #18)
Very familiar. This sounds like bug 1777518. A fix landed (in bug 1755700) in Fx 105 and was uplifted to 104 beta. @p.huiskens, would you be able to try to reproduce with the beta, or with the 104 release when it's released?
I'll install the beta right away, will report back tomorrow
|
|
Reporter | |
Comment 20•2 years ago
|
||
currently with audio enabled and beta installed ( it says after update that the Firefox version is 104.0) it did not crash
|
|
||
Comment 21•2 years ago
|
||
(In reply to p.huiskens from comment #20)
currently with audio enabled and beta installed ( it says after update that the Firefox version is 104.0) it did not crash
Well, that's awesome. Thanks for helping us with the bug report! I'm going to dupe this bug but if the same problem comes back, please let us know.
Description
•