Open Bug 731812 Opened 10 years ago Updated 1 year ago

crash in MMDevAPI.dll when inserting or removing a usb headset


(Core :: Widget: Win32, defect, P5)

Windows 7




(Reporter: blizzard, Unassigned)



(Keywords: crash, regression, Whiteboard: [tbird crash][tpi:+])

Crash Data

Firefox crashes reliably whenever I plug in or remove a USB headset.

Example crash report:
Assignee: nobody → netzen
It might be related to bug 691355.
Severity: normal → critical
Crash Signature: [@ CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*)]
Keywords: crash
Hardware: x86_64 → x86
Blocks: 691355
These are different crashes by the way, I could 100% reproduce the one from bug 691355 before its fix, but not this one.  Also the crash signature is different.  Looking into it soon.
Crash Signature: [@ CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*)] → [@ CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*)] [@ @0x0 | CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*) ] [@ _chkstk | CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*) ]
Jim are you able to reproduce this?
I don't have the right headphones. Chris, care to share info on the hardware you're using? Maybe we can pick up a a pair for testing.
Assignee: netzen → nobody
Crash Signature: [@ CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*)] [@ @0x0 | CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*) ] [@ _chkstk | CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*) ] → [@ CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*)] [@ @0x0 | CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*) ] [@ _chkstk | CLocalEndpointEnumerator::OnMediaNotification(MEDIA_NOTIFICATION_BLOCK*) ] [@…
Summary: crash when inserting or removing a usb headset → crash in MMDevAPI.dll when inserting or removing a usb headset
Whiteboard: [tbird crash]
(In reply to Jim Mathies [:jimm] from comment #4)
> I don't have the right headphones. Chris, care to share info on the hardware
> you're using? Maybe we can pick up a a pair for testing.
I just experienced this myself:

I'm using a Plantronics Audio 655 DSP headset.
Can you get it consistently? If so please indicate the steps to reproduce. If so I'll also pickup one of these headsets to fix.
(In reply to Brian R. Bondy [:bbondy] from comment #7)
> Can you get it consistently? If so please indicate the steps to reproduce.
> If so I'll also pickup one of these headsets to fix.
Unfortunately, no, I haven't seen it consistently. :-(
Hi, it seems not to be only related to Firefox, because I have it with IE11 too.
Flash Player or Microsoft (Win7) bug ?

Philips SHB4000/10 bluetooth headset
LogiLink usb bluetooth v4.0 dongle

STR: sometimes on loading

Internet Explorer

Nicht mehr funktionsfähig

‎20.‎05.‎2014 20:15

Der Bericht wurde gesendet.

Pfad der fehlerhaften Anwendung:	C:\Program Files (x86)\Internet Explorer\iexplore.exe

Problemereignisame:	APPCRASH
Anwendungsname:	IEXPLORE.EXE
Anwendungsversion:	11.0.9600.17041
Anwendungszeitstempel:	531807e4
Fehlermodulname:	MMDevApi.dll
Fehlermodulversion:	6.1.7601.17514
Fehlermodulzeitstempel:	4ce7b892
Ausnahmecode:	c0000005
Ausnahmeoffset:	00018d99
Betriebsystemversion:	6.1.7601.
Gebietsschema-ID:	1031
Zusatzinformation 1:	a7aa
Zusatzinformation 2:	a7aa91f17ea749d42a4de3b390fa5b3d
Zusatzinformation 3:	a7aa
Zusatzinformation 4:	a7aa91f17ea749d42a4de3b390fa5b3d

Weitere Informationen über das Problem
Bucket-ID:	177971817
I was able to reproduce this once on Windows 8.1 with a Logitech USB headset with Firefox 28 and Flash Player, but that was once in about 25 tries.  I haven't hit it since.

I've opened Adobe 3770891 and we'll try to triage on our side.  I'm not totally convinced that this is a Flash issue, but we're definitely happy to look into it further.
No luck with four open tabs, Firefox 29.0.1 and Flash Player with Win7 x64 after 15 minutes.
I pulled a bunch of representative stacks from the crash reporter and resolved the symbols, but I don't see a smoking gun.  The Flash segments of the stack are all over the place, but the top three MMDevAPI.dll calls are all consistent.  

I've sent this off to our audio team for further analysis, but I'm leaning towards a CoreAudio bug based on what I'm seeing.  

If you guys end up with employee minidumps that you can share, that would be helpful.  I haven't seen this bug since.  I'm wondering if it happens when Windows recognizes the headset as a new device, as I believe that's when I saw it, I had used a different USB port and it was going through the install sequence for the headset in the background.  

I'll go back and delete the headset devices and see if I can capture another minidump tomorrow.
Crash Signature: ] [@ _chkstk | ATL::CAtlList<MEDIA_NOTIFICATION_BLOCK*, ATL::CElementTraits<MEDIA_NOTIFICATION_BLOCK*> >::GetFreeNode() ] → ] [@ _chkstk | ATL::CAtlList<MEDIA_NOTIFICATION_BLOCK*, ATL::CElementTraits<MEDIA_NOTIFICATION_BLOCK*> >::GetFreeNode() ] [@ CLocalEndpointEnumerator::OnMediaNotification] [@ @0x0 | CLocalEndpointEnumerator::OnMediaNotification ] [@ _chkstk | CLocalE…
Priority: -- → P5
Whiteboard: [tbird crash] → [tbird crash][tpi:+]
Duplicate of this bug: 1312312
Is there nothing we can do?   If not we should close invalid.

Unfortunately for thunderbird, the combined signatures makes this a top 40 crash. For most users this is a one-off crash, never to be seen again. But there are rare exceptions - for example (from the same user)
28c33dcf-347c-4a00-a04e-688b50170531	2017-05-31 00:12:56 	PL_HashTableDestroy | SECOID_Shutdown 
e3e627ef-adf1-4961-9612-ba3e70170523	2017-05-23 15:30:52 	CLocalEndpointEnumerator::OnMediaNotification 
> 0807fff0-8f21-4998-970a-1c9e10170523	2017-05-23 15:29:40 	mozilla::widget::AudioSession::Start  
cff24507-a0b9-4812-9776-f3a5b0170523	2017-05-23 15:29:24 	CAudioSessionControl::QueueStreamSwitch 

And CTypeMarshal::Marshal seems to be another signature - from bp-7c441a62-0d26-4184-b969-726880170801 - which added to the other signatures would drive the ranking into top 30 crash
Flags: needinfo?(jmathies)
See Also: → 691355
Looks similar to bug 1339259 which I'm trying to find some time to work on.
Depends on: 1339259
Flags: needinfo?(jmathies)

Bugbug thinks this bug is a regression, but please revert this change in case of error.

Keywords: regression
You need to log in before you can comment on or make changes to this bug.