683967 - Firefox 9.0a1 Plugin Crash in mozilla::plugins::PluginModuleChild::ShouldContinueFromReplyTimeout [@ mozalloc_abort(char const* const) | msvcr80.dll@0xe456 ]

Status: VERIFIED FIXED

(Core :: General, defect)

Description

[Marcia Knous [:marcia]]

Seen while looking at trunk crash stats. Crashes started showing up using the 2011083000 build. https://crash-stats.mozilla.com/report/list?signature=mozalloc_abort%28char%20const*%20const%29%20|%20msvcr80.dll@0xe456

Possible pushlog regression range based on crash stats:

http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=33031c875984&tochange=e6591ea9b27b

Frame 	Module 	Signature [Expand] 	Source
0 	mozalloc.dll 	mozalloc_abort 	memory/mozalloc/mozalloc_abort.cpp:77
1 	msvcr80.dll 	msvcr80.dll@0xe456 	
2 	xul.dll 	mozilla::plugins::PluginModuleChild::ShouldContinueFromReplyTimeout 	dom/plugins/ipc/PluginModuleChild.cpp:531
3 	xul.dll 	mozilla::ipc::SyncChannel::ShouldContinueFromTimeout 	ipc/glue/SyncChannel.cpp:264
4 	xul.dll 	mozilla::ipc::SyncChannel::Send 	ipc/glue/SyncChannel.cpp:129
5 	xul.dll 	mozilla::ipc::RPCChannel::Send 	ipc/glue/RPCChannel.cpp:150
6 	xul.dll 	mozilla::plugins::PPluginInstanceChild::SendShow 	obj-firefox/ipc/ipdl/PPluginInstanceChild.cpp:863
7 	xul.dll 	mozilla::plugins::PluginInstanceChild::ShowPluginFrame 	dom/plugins/ipc/PluginInstanceChild.cpp:3266
8 	xul.dll 	mozilla::plugins::PluginInstanceChild::InvalidateRectDelayed 	dom/plugins/ipc/PluginInstanceChild.cpp:3334
9 	xul.dll 	MessageLoop::RunTask 	ipc/chromium/src/base/message_loop.cc:318
10 	xul.dll 	MessageLoop::DeferOrRunPendingTask 	ipc/chromium/src/base/message_loop.cc:326
11 	xul.dll 	MessageLoop::DoWork 	ipc/chromium/src/base/message_loop.cc:426
12 	xul.dll 	base::MessagePumpForUI::DoRunLoop 	ipc/chromium/src/base/message_pump_win.cc:214
13 	xul.dll 	base::MessagePumpWin::RunWithDispatcher 	ipc/chromium/src/base/message_pump_win.cc:53
14 	xul.dll 	base::MessagePumpWin::Run 	ipc/chromium/src/base/message_pump_win.h:78
15 	xul.dll 	MessageLoop::RunInternal 	ipc/chromium/src/base/message_loop.cc:208
16 	xul.dll 	MessageLoop::RunHandler 	
17 	xul.dll 	MessageLoop::Run 	ipc/chromium/src/base/message_loop.cc:175
18 	xul.dll 	XRE_InitChildProcess 	toolkit/xre/nsEmbedFunctions.cpp:512
19 	plugin-container.exe 	wmain 	toolkit/xre/nsWindowsWMain.cpp:107
20 	plugin-container.exe 	__tmainCRTStartup 	crtexe.c:594
21 	kernel32.dll 	BaseThreadInitThunk 	
22 	ntdll.dll 	__RtlUserThreadStart 	
23 	ntdll.dll 	_RtlUserThreadStart

Comment 1

[Marcia Knous [:marcia]]

Possible dupe of Bug 680130?

Comment 2

[Logan Rosen [:Logan]]

No, not a dupe. Bug 680130 is for the new type of signature that can appear in crashes, and this is an example of such a signature.

Comment 3

[Henrik Gemal]

After a while on this site:
http://davidwalsh.name/gmail-php-imap
I crashes with the above signature

Comment 4

[Marcia Knous [:marcia]]

mozalloc_abort(char const* const) | _RTC_Terminate plugin signature started appearing in the same time frame so it is likely related.

Comment 5

[u88484]

I get this crash playing games on google+.  I leave the game open when putting the computer to sleep, after waking up the computer the plugin crashes.

Comment 6

[Benjamin Smedberg]

Comment 5 is a situation that jimm and I discussed and we need to fix. Comment 3 is probably different.

Comment 7

[Jim Mathies [:jimm]]

(In reply to Benjamin Smedberg  [:bsmedberg] from comment #6)
> Comment 5 is a situation that jimm and I discussed and we need to fix.

bug 679240 was filed on that. Not surprising the frequency of this problem would increase now that we are looking for timeouts on both sides of the connection.

Comment 8

[kolubinowicki]

Attachment: See comment 8 for more details

With last nightly i have problem with playing video when opened at least 2-3 tabs with Youtube and pause each video for buffering. After few seconds, all the video starts playing in the background and also the plugin container have memory leak when growth to about 900-1500MB it crashes the Nvidia video driver that sometimes don't properly recover. Crash signature point to these bug https://crash-stats.mozilla.com/report/index/bp-5ab369ac-867f-46c7-a983-d3ab52110920

Comment 9

[Andrew Truong [:feer56] [:andrewtruong] [:atruong]]

Hello,
I had bug 566062 and was told to try nightly. Nightly did the same thing and had adobe flash player crash. https://crash-stats.mozilla.com/report/index/bp-cde24d56-e1a0-4326-a8f9-01d022111001

Thank You

Comment 10

[u88484]

I still get this crash all the time but 99% of my 30 crash reports fail to submit.

Comment 11

[Scoobidiver (away)]

It's #15 top crasher in 9.0b2.

Comment 12

[Wayne Mery (:wsmwk)]

encountered this with http://syndicate.sellpoint.net/Syndicate/AssetRedirect?tt_aq=0-359487-51160-150&ext_url=true&cdntype=10&go=http%3A%2F%2Fassetsw.sellpoint.net%2FIA%2Fpanasonic%2Ftelephones%2Fkxtg7620%2FFlash_Tour%2Findex.html

Comment 13

[Scoobidiver (away)]

Dupe of bug 681385?

Comment 14

[Marcia Knous [:marcia]]

Not tracking for Firefox 9.

Comment 15

[Wayne Mery (:wsmwk)]

#14 crash for both 9.0b2 and 9.0b3

(In reply to Scoobidiver from comment #13)
> Dupe of bug 681385?

seems likely, given that this signature picks up where mozalloc_abort(char const* const) | NS_DebugBreak_P | mozilla::plugins::PluginModuleChild::ShouldContinueFromReplyTimeout()  dies out https://crash-stats.mozilla.com/report/list?product=Firefox&query_search=signature&query_type=exact&query=mozalloc_abort%28char%20const*%20const%29%20|%20NS_DebugBreak_P%20|%20mozilla%3A%3Aplugins%3A%3APluginModuleChild%3A%3AShouldContinueFromReplyTimeout%28%29&reason_type=contains&date=10%2F02%2F2011%2004%3A01%3A16&range_value=16&range_unit=weeks&hang_type=any&process_type=any&do_query=1&admin=1&signature=mozalloc_abort%28char%20const*%20const%29%20|%20NS_DebugBreak_P%20|%20mozilla%3A%3Aplugins%3A%3APluginModuleChild%3A%3AShouldContinueFromReplyTimeout%28%29

is Bug 680130 actually going to be a meta bug?

Comment 16

[Keir Hardie]

Getting this consistently when loading Words With Friends on Facebook in 10b1 on XP, the last 9 beta never did this. On my third attempt however I noticed that despite the crash it still works, and if anything I think the performance might be better than before.

Comment 17

[Sheila Mooney]

This is sitting at #18 on 9.0.1 but at #2 on 10.0b2 with pretty high volume - over 10K in a week. I think we need to get someone on it.

Comment 18

[Benjamin Smedberg]

This appears to be the intentional crash when a plugin detects that its parent (Firefox) process is not responding. That code was disabled on previous release branches, and until we can get matching browser-side stacks for the hangs (bug 679238) we should probably continue disabling this for release branches.

Comment 19

[Jim Mathies [:jimm]]

(In reply to Benjamin Smedberg  [:bsmedberg] from comment #18)
> This appears to be the intentional crash when a plugin detects that its
> parent (Firefox) process is not responding. That code was disabled on
> previous release branches, and until we can get matching browser-side stacks
> for the hangs (bug 679238) we should probably continue disabling this for
> release branches.

Lets just disable it on trunk until the issues with the crash reporter code can be worked out. These crashes aren't telling us anything without parent side stacks. 

I'll work up a patch and post it here.

Comment 20

[Jim Mathies [:jimm]]

Attachment: patch

Comment 21

[Jim Mathies [:jimm]]

Comment on attachment 585823 [details] [diff] [review]
patch

Disabling some crash generation code that has in the past been enabled on trunk but not branch.

Comment 22

[Marcia Knous [:marcia]]

I realize that work is going on to disable this, but I did hit this crash yesterday in the lab while testing Facebook: https://crash-stats.mozilla.com/report/index/bp-b06c36c7-774d-4d97-abfc-b84492120104.

I had a Facebook game loaded, left the computer for several hours, and when I came back I got this crash. I mentioned this to smooney during the meeting since at that time she was looking for STR for this crash.

Comment 23

[Jim Mathies [:jimm]]

(In reply to Marcia Knous [:marcia] from comment #22)
> I realize that work is going on to disable this, but I did hit this crash
> yesterday in the lab while testing Facebook:
> https://crash-stats.mozilla.com/report/index/bp-b06c36c7-774d-4d97-abfc-
> b84492120104.
> 
> I had a Facebook game loaded, left the computer for several hours, and when
> I came back I got this crash. I mentioned this to smooney during the meeting
> since at that time she was looking for STR for this crash.

Was the system suspended (in sleep or hibernate mode) when you came back?

Comment 24

[Alex Keybl [:akeybl]]

Comment on attachment 585823 [details] [diff] [review]
patch

[Triage Comment]
Approved for Aurora.

From https://bugzilla.mozilla.org/show_bug.cgi?id=683967#c17 though, it sounds like FF10 is also affected. Should the patch land there as well?

Comment 25

[Robert Kaiser]

It surely affects 10 as well, I have seen it high up throughout its aurora and beta lifetime. I think the patch needs to go into beta as well.

Comment 26

[Jim Mathies [:jimm]]

Comment on attachment 585823 [details] [diff] [review]
patch

Yep, we need this in 10 as well.

Comment 27

[Jim Mathies [:jimm]]

https://hg.mozilla.org/mozilla-central/rev/5413ba3ed406
https://hg.mozilla.org/releases/mozilla-aurora/rev/2f332e07843a

still need beta approval.

Note, I'm closing this bug out, but this crash will likely remain for other reasons. So we'll likely need to create a new bug for this once crash stats settles down.

Comment 28

[Alex Keybl [:akeybl]]

Comment on attachment 585823 [details] [diff] [review]
patch

[Triage Comment]
Approved for beta as well, based upon low risk vs high reward of mitigating this regression.

Comment 29

[Jim Mathies [:jimm]]

https://hg.mozilla.org/releases/mozilla-beta/rev/34030f77fd48

Comment 30

[Virgil Dicu [:virgil] [QA]]

Mozilla/5.0 (Windows NT 6.1; WOW64; rv:11.0) Gecko/20100101 Firefox/11.0

No reports for any 11 beta in the last 4 weeks with this crash signature: http://bit.ly/w3QKKN

Also tried playing google plus games and visiting sites which were reported to have crashed Firefox with this signature(comment 3, 5, 12). No crashes.