763896 - [adbe 3216155] Flash hang in F_1152915508 by quickly toggling Full Screen

Status: RESOLVED FIXED

(External Software Affecting Firefox Graveyard :: Flash (Adobe), defect)

Description

[Scoobidiver (away)]

It's #30 top crasher and #6 top plugin crasher in 13.0.
It's more common with Flash 11.3 with 64% of crashes.

Signature 	hang | BaseGetNamedObjectDirectory More Reports Search
UUID	46fadfcc-609f-453b-ab1e-51b282120612
Date Processed	2012-06-12 12:36:25
Process Type	plugin Version: Filename: NPSWF32_11_3_300_257.dll
Uptime	325
Install Age	1.9 days since version was first installed.
Install Time	2012-06-10 15:14:10
Product	Firefox
Version	13.0
Build ID	20120601045813
Release Channel	release
OS	Windows NT
OS Version	6.1.7601 Service Pack 1
Build Architecture	x86
Build Architecture Info	GenuineIntel family 6 model 28 stepping 10
Crash Reason	EXCEPTION_BREAKPOINT
Crash Address	0x77b67094
App Notes 	
AdapterVendorID: 0x8086, AdapterDeviceID: 0xa011, AdapterSubsysID: 3660103c, AdapterDriverVersion: 8.14.10.1929
D3D10 Layers? D3D10 Layers- D3D9 Layers? D3D9 Layers- 
EMCheckCompatibility	True

Frame 	Module 	Signature 	Source
0 	ntdll.dll 	KiFastSystemCallRet 	
1 	ntdll.dll 	NtWaitForMultipleObjects 	
2 	KERNELBASE.dll 	BaseGetNamedObjectDirectory 	
3 	kernel32.dll 	WaitForMultipleObjectsExImplementation 	
4 	user32.dll 	RealMsgWaitForMultipleObjectsEx 	
5 	user32.dll 	MsgWaitForMultipleObjects 	
6 	NPSWF32_11_3_300_257.dll 	F_1152915508___________________________________ 	F1814729968_______________________________________________________________________________________:115
7 	NPSWF32_11_3_300_257.dll 	F2166389_____________________________________________________________________ 	F_1046257564___________________________________________________________________:509
8 	NPSWF32_11_3_300_257.dll 	F_917831355____________________________________________ 	F_1046257564___________________________________________________________________:445
9 	NPSWF32_11_3_300_257.dll 	F1315696776________________________________ 	F_1046257564___________________________________________________________________:409
10 	NPSWF32_11_3_300_257.dll 	F81047063_______________________________________ 	F_863654669__________________________________________________________________________________:180
11 	NPSWF32_11_3_300_257.dll 	F845925699_____________________________________ 	F_1017111218___________________________________________________________________________________________:90
12 	NPSWF32_11_3_300_257.dll 	F1471391699_________________________________________ 	F520480456______________________________________________________________________________:327
13 	xul.dll 	mozilla::plugins::PluginScriptableObjectChild::AnswerHasProperty 	dom/plugins/ipc/PluginScriptableObjectChild.cpp:833
14 	xul.dll 	mozilla::plugins::PPluginScriptableObjectChild::OnCallReceived 	obj-firefox/ipc/ipdl/PPluginScriptableObjectChild.cpp:919
15 	xul.dll 	mozilla::plugins::PPluginModuleChild::OnCallReceived 	obj-firefox/ipc/ipdl/PPluginModuleChild.cpp:980
16 	xul.dll 	mozilla::ipc::RPCChannel::DispatchIncall 	ipc/glue/RPCChannel.cpp:517
17 	xul.dll 	mozilla::ipc::RPCChannel::Incall 	ipc/glue/RPCChannel.cpp:503
18 	xul.dll 	mozilla::ipc::RPCChannel::OnMaybeDequeueOne 	ipc/glue/RPCChannel.cpp:430
19 	xul.dll 	MessageLoop::RunTask 	ipc/chromium/src/base/message_loop.cc:318
20 	xul.dll 	MessageLoop::DeferOrRunPendingTask 	ipc/chromium/src/base/message_loop.cc:326
21 	xul.dll 	MessageLoop::DoWork 	ipc/chromium/src/base/message_loop.cc:426
22 	xul.dll 	base::MessagePumpForUI::DoRunLoop 	ipc/chromium/src/base/message_pump_win.cc:214
23 	xul.dll 	base::MessagePumpWin::RunWithDispatcher 	ipc/chromium/src/base/message_pump_win.cc:53
24 	xul.dll 	base::MessagePumpWin::Run 	ipc/chromium/src/base/message_pump_win.h:78
25 	xul.dll 	MessageLoop::RunInternal 	ipc/chromium/src/base/message_loop.cc:208
26 	xul.dll 	MessageLoop::RunHandler 	ipc/chromium/src/base/message_loop.cc:194
27 	xul.dll 	MessageLoop::Run 	ipc/chromium/src/base/message_loop.cc:175
28 	xul.dll 	XRE_InitChildProcess 	toolkit/xre/nsEmbedFunctions.cpp:513
29 	plugin-container.exe 	wmain 	toolkit/xre/nsWindowsWMain.cpp:107
30 	plugin-container.exe 	__tmainCRTStartup 	crtexe.c:552
31 	kernel32.dll 	BaseThreadInitThunk 	
32 	ntdll.dll 	__RtlUserThreadStart 	
33 	ntdll.dll 	_RtlUserThreadStart

NPSWF32_11_3_300_257.dll 	11.3.300.257 	4E78057B792547E7838DFC58366D62861

More reports at:
https://crash-stats.mozilla.com/report/list?signature=hang+|+BaseGetNamedObjectDirectory

Comment 1

[Scoobidiver (away)]

I add other related signatures:
https://crash-stats.mozilla.com/report/list?signature=hang+|+NtUserSetWindowsHookEx+|+RealMsgWaitForMultipleObjectsEx
https://crash-stats.mozilla.com/report/list?signature=hang+|+GetProcessHeap
https://crash-stats.mozilla.com/report/list?signature=hang+|+WaitForMultipleObjectsEx+|+RealMsgWaitForMultipleObjectsEx

There are about 7400 crashes in 13.0 making it #8 top crasher and #2 top plugin crasher.

Comment 2

[Alex Keybl [:akeybl]]

Sal/Hitomi - has the crash signature for this hang shown up in your about:crashes? I'm wondering if this is related to the repro steps you all provided for bug 763444 (which included a hang, I believe).

Anthony/Juan - please let us know if you run into this crash while trying the different repro steps for bug 763444.

Comment 3

[Alex Keybl [:akeybl]]

Also adding needURLs, to help repro in the future.

Comment 4

[Hitomi Kudo]

When I was trying to reproduce the top crash ( ) , four of the crash reports sitting in same session time frame were the followings:

https://crash-stats.mozilla.com/report/index/bp-7bbad6ed-6de9-4158-a8bf-acbf92120615
https://crash-stats.mozilla.com/report/index/bp-930aa5c7-0a26-422e-9c9c-32b982120615
https://crash-stats.mozilla.com/report/index/bp-662c35c8-b492-4a51-8cb9-61e4a2120615
https://crash-stats.mozilla.com/report/index/bp-7ab85a5d-3d9f-491a-bf36-cf8b52120615

The first two hang in mine do not look like same with this bug. But Hang must related to the crash we are chasing. Is there good way to share my about:crashes info with you?

Comment 5

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

Kudo-san, about:crashes is just text with dates and URLs... if you copy and paste as you did, that should work.  You do not necessarily need to copy the whole url.  bugzilla will automatically associate the link to our crash-stats server (Socorro).

The crashes that you are seeing is as follows respectively:
[@ hang | NtUserSetWindowsHookEx ] 
[@ hang | mozilla::plugins::PPluginInstanceParent::CallNPP_Destroy(short*) ] 
[@ F1398665248_____________________________ ] 
[@ F1398665248_____________________________ ] 

The last 2 crashes look like Bug 763237.

Comment 6

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

The first 2 may be related...?  I am not sure.

For some reason I am not seeing the related URLs to the associated crasher.  Marcia, are you able to see any URLs?

Comment 7

[Paul Silaghi, QA [:pauly]]

(In reply to Alex Keybl [:akeybl] from comment #2)
> Sal/Hitomi - has the crash signature for this hang shown up in your
> about:crashes? I'm wondering if this is related to the repro steps you all
> provided for bug 763444 (which included a hang, I believe).
> 
> Anthony/Juan - please let us know if you run into this crash while trying
> the different repro steps for bug 763444.

Unable to reproduce this crash on Mozilla/5.0 (Windows NT 6.1; rv:13.0) Gecko/20100101 Firefox/13.0.1.
But the crash from bug 763444 is definitely seen on www.miniclip.com.

Comment 8

[Benjamin Smedberg]

Taking a look through these reports in crash-stats, it doesn't appear that there is any particular pattern about what the plugin or browser processes are doing at the time of the hang. In particular, the browser calls are all over the place:

PPluginModuleParent::CallPPluginInstanceConstructor
PPluginModuleParent::CallPPluginInstanceConstructor
PPluginInstanceParent::CallUpdateWindow
PPluginInstanceParent::CallNPP_HandleEvent
PPluginInstanceParent::CallNPP_GetValue_NPPVpluginScriptableNP

I currently suspect that this is just a generic signature/stack for when the our plugin process is waiting for its sandboxed Flash process to respond. As such, it probably replaces a bunch of different hangs which would have been reported as different signatures in prior versions of Flash. And it's probably not something that we can actively track or fix.

Comment 9

[Alex Keybl [:akeybl]]

(In reply to Benjamin Smedberg  [:bsmedberg] from comment #8)
> And it's
> probably not something that we can actively track or fix.

Do we need a bug on file to track splitting this signature out?

I'd still like to leave needURLs and qawanted on this bug, since if we can reproduce any hangs, we should file separate bugs. This can serve as a meta bug for 11.3 Flash hangs in that case.

Comment 10

[Benjamin Smedberg]

I currently don't see anything that we could split on, so another bug at this point wouldn't be helpful.

Comment 11

[Marcia Knous [:marcia]]

(In reply to Alex Keybl [:akeybl] from comment #9)
I don't see any URLs listed for any of the crash signatures in this bug.

> (In reply to Benjamin Smedberg  [:bsmedberg] from comment #8)
> > And it's
> > probably not something that we can actively track or fix.
> 
> Do we need a bug on file to track splitting this signature out?
> 
> I'd still like to leave needURLs and qawanted on this bug, since if we can
> reproduce any hangs, we should file separate bugs. This can serve as a meta
> bug for 11.3 Flash hangs in that case.

Comment 12

[Benjamin Smedberg]

The links in this bug will go to the plugin-side report which doesn't contain URLs. The matching browser-side report will contain a URL. And since we throttle release versions, the process of getting URLs for this might be:

* construct a query for the plugin-side signature such as "hang | BaseGetNamedObjectDirectory" on aurora or beta
* for each of those reports, click the link to retrieve the matching browser-side report and get the URL

Or we could ask the Socorro-folk to construct some sort of custom query to do this.

Comment 13

[chris hofmann]

using the process outlined in comment 12 and a couple of passes on the .csv files here are the top urls in the browser side reports of the plugin hang signature "hang | BaseGetNamedObjectDirectory"

3645  
  81 apps.facebook.com onthefarm
  30 nav3.poker.zynga.com link
  29 zynga.com #
  22 apps.facebook.com texas_holdem
  19 apps.facebook.com playcastleville
  19 apps.facebook.com cityville
  17 apps.facebook.com bubblesafari
  16 www.hi5.com messages_center.html?uid=
  12 www.facebook.com dialog
  10 www.facebook.com 
   9 www.youporn.com watch
   8 www.odnoklassniki.ru game
   8 apps.facebook.com thesimssocial
   7 www.facebook.com ajax
   7 mail.google.com mail
   7 eknowledge.azc.uam.mx 
   7 apps.facebook.com playslingo
   7 apps.facebook.com playhappyfarm
   7 apps.facebook.com empiresandallies
   6 www.nicovideo.jp watch
   6 www.costabingo.com skin
   6 my.mail.ru apps
   5 oadsrv.com www
   5 me.zing.vn apps
   5 apps.facebook.com mafiawars-two
   5 apps.facebook.com hidden-chronicles
   5 apps.facebook.com cartown
   5 apps.facebook.com bubblewitch
   4 www.youtube.com watch?v=pwp1CH5R-w4&feature=fvst
   4 www.youjizz.com videos
   4 www.vplay.ro watch
   4 www.videosexo.com.br video
   4 www.terra.com.br portal
   4 www.bet365.com home
   4 apps.facebook.com avengersalliance
   3 www.putlocker.com file
   3 www.hulu.com watch
   3 www.facebook.com ?ref=tn_tnmn
   3 mymembervault.com video.php?mloc=videos&loc=toons&vidID=43552
   3 fv-zprod.farmville.com iframe_communication_proxy.php#height=2783&
   3 creatives.livejasmin.com pu
   3 apps.facebook.com pioneertrail
   3 apps.facebook.com gardensoftime
   3 apps.facebook.com cafeworld
   3 apps.facebook.com astrogarden

and here are the high frequency signatures on the browser side that are connected to plugin side "hang | BaseGetNamedObjectDirectory"

3603 hang | BaseGetNamedObjectDirectory
 223 hang | mozilla::plugins::PPluginInstanceParent::CallPBrowserStreamConstructor(mozilla::plugins::PBrowserStreamParent*, nsCString const&, unsigned int const&, unsigned int const&, mozilla::plugins::PStreamNotifyParent*, nsCString const&, nsCString const...
 216 hang | mozilla::plugins::PPluginScriptableObjectParent::CallHasProperty(mozilla::plugins::PPluginIdentifierParent*, bool*)
 155 hang | mozilla::plugins::PPluginInstanceParent::CallNPP_HandleEvent(mozilla::plugins::NPRemoteEvent const&, short*)
 138 hang | mozilla::plugins::PPluginInstanceParent::CallNPP_Destroy(short*)
 109 hang | WaitForSingleObjectEx | WaitForSingleObject | google_breakpad::ExceptionHandler::WriteMinidumpOnHandlerThread(_EXCEPTION_POINTERS*, MDRawAssertionInfo*)
  98 hang | mozilla::plugins::PPluginScriptableObjectParent::CallGetChildProperty(mozilla::plugins::PPluginIdentifierParent*, bool*, bool*, mozilla::plugins::Variant*, bool*)
  87 hang | mozilla::plugins::PPluginInstanceParent::CallNPP_SetWindow(mozilla::plugins::NPRemoteWindow const&)
  43 hang | mozilla::plugins::PPluginModuleParent::CallPPluginInstanceConstructor(mozilla::plugins::PPluginInstanceParent*, nsCString const&, unsigned short const&, InfallibleTArray<nsCString> const&, InfallibleTArray<nsCString> const&, short*)
  43 hang | mozilla::plugins::PPluginInstanceParent::CallUpdateWindow()
  39 hang | mozilla::plugins::PPluginScriptableObjectParent::CallHasMethod(mozilla::plugins::PPluginIdentifierParent*, bool*)
  32 hang | mozilla::plugins::PPluginInstanceParent::CallNPP_GetValue_NPPVpluginScriptableNPObject(mozilla::plugins::PPluginScriptableObjectParent**, short*)
  21 hang | mozilla::plugins::PPluginInstanceParent::Call__delete__(mozilla::plugins::PPluginInstanceParent*)
  20 hang | mozilla::plugins::PPluginScriptableObjectParent::CallInvoke(mozilla::plugins::PPluginIdentifierParent*, InfallibleTArray<mozilla::plugins::Variant> const&, mozilla::plugins::Variant*, bool*)
  11 hang
   9 hang | mozilla::plugins::PBrowserStreamParent::CallNPP_StreamAsFile(nsCString const&)
   1 hang | mozilla::plugins::PPluginInstanceParent::CallSetPluginFocus()
   1 hang | mozilla::plugins::PPluginInstanceParent::CallNPP_GetValue_NPPVpluginWantsAllNetworkStreams(bool*, short*)
   1 hang | WaitForSingleObjectEx | WaitForSingleObject | xul.dll@0x468f9a | xul.dll@0x6b18f7 | xul.dll@0x6b259b | xul.dll@0x76bded | _PeekMessage

Comment 14

[u279076]

(In reply to chris hofmann from comment #13)

Seems to indicate a very high correlation to Facebook and Zynga games.

Comment 15

[Benjamin Smedberg]

Well, that looks pretty much like normal usage to me, I don't think I'd draw any special correlations out of that data.

Comment 16

[chris hofmann]

but, it might help to reproduce by pounding away on that list of sites.

Comment 17

[Alex Keybl [:akeybl]]

No longer tracking for release - this is happening on all Firefox versions and we have no reason to believe there's anything we can do on our side.

QA should still try to reproduce the hangs, but this wont' block release.

Comment 19

[juan becerra [:juanb]]

There's a user who can reproduce a problem consistently and sometimes the signature is like the one here. See bug 723987

Comment 20

[smadayag]

unfortunately, i'm not able to reproduce the issue using the MirandaIM example.  

tested: Firefox 13.0.1 Flash Player 11.3.300.262 MirandIM v0.9.52

Comment 21

[kr0k0f4nt]

I just double checked my configuration, and I can also confirm I have the Firefox 13.0.1, Flash 11.3.300.262 and MirandaIM 0.9.52 and I can reproduce the issue (https://bugzilla.mozilla.org/show_bug.cgi?id=723987) each time.

Just make sure Miranda is already up and running in the tray. Open a clean new session of Firefox and go to youtube.com. Then just bring up the the contact list which has to be docked to the screen side and you will have your freeze.

I also want to point out here, I used a complete clean firefox profile with only the Flash Plugin enabled, nothing else! Here's the latest crash with that reproduction:

http://crash-stats.mozilla.com/report/index/bp-ae3f55a8-b0b1-4443-bc12-270ba2120628
http://crash-stats.mozilla.com/report/index/bp-a5124a0f-432c-480d-914c-cc6a82120628

Comment 23

[Robert Kaiser]

What we can say pretty much for sure is that hangs are almost twice as frequent as back when 11.2 was the most recent release, and it looks like most now come with this signature, the latter probably due to the architecture changes in Flash. The overall rise in hangs is something that would be good to get some track on.

Comment 24

[Marcia Knous [:marcia]]

I was able to get this crash while playing a video on Vimeo. In fact, it seems whenever I just let a video play on Vimeo on some machine I have issues with freezing.

STR:
1. Load vimeo.com. I was playing http://vimeo.com/45775119
2. Let the video play
3. Go full screen and then come back to the video
4. Crash 

https://crash-stats.mozilla.com/report/index/bp-984d1999-b1a5-4788-8031-e50872120717

Comment 25

[Scoobidiver (away)]

I can't reproduce on my machine with the STR in comment 24.

Comment 26

[Jeromie Clark]

Marcia, can you try disabling Hardware Acceleration in Flash, then in Firefox to see if either make a difference?  Also, if you could attach the output of dxdiag, that would be helpful.

Comment 27

[kr0k0f4nt]

Hey Jeromie,

I tried to disable the Hardware accleration and it still freezes like discribed in my bug report (https://bugzilla.mozilla.org/show_bug.cgi?id=723987). I also tried and updated my Nvidia Driver to the latest Version 301.42, but with no changes at all.

How ever, I can't reproduce Marcia's STR.

Comment 28

[kr0k0f4nt]

Here's the related crash with the latest update, as discribed in my last post:

http://crash-stats.mozilla.com/report/index/bp-3c489ca9-3c61-4e5a-b61b-cc5ee2120720
http://crash-stats.mozilla.com/report/index/bp-4cc1ef65-df6f-445e-9ea3-e036b2120717

Comment 31

[Scoobidiver (away)]

bug 776562 has STR.

Comment 32

[_w_]

I can produce this bug with firefox 15.0b2 and flash 11,3,300,268 by watching any video on vimeo and then switching to fullscreen and back a few times (3x usually does it for me) in short succession.
This also happens with other sites with embedded video and switching back and forth to fullscreen, but not all of them (e.g. youtube does not seem to be affected)

Comment 33

[smadayag]

thanks _w_, jeromie adde your str to our internal bug, #3216155.

Comment 34

[chris nuuja]

I think the Vimeo fullscreen bug is different.   Are you going fullscreen by right clicking (or cntrl-clicking) to bring up the context menu, then selecting "go fullscreen"?   If so, that crash (which could happen on any site that presents a context menu option to go fullscreen)  is new to FP 11.3, it could not happen in FP 11.2.  This bug was reported as reproducing with FP 11.2.

Comment 35

[Marcia Knous [:marcia]]

In my testing I have seen the problem going full screen using the context menu - what is the Adobe bug for the crash that happens on any site that presents a context menu option?

(In reply to chris nuuja from comment #34)
> I think the Vimeo fullscreen bug is different.   Are you going fullscreen by
> right clicking (or cntrl-clicking) to bring up the context menu, then
> selecting "go fullscreen"?   If so, that crash (which could happen on any
> site that presents a context menu option to go fullscreen)  is new to FP
> 11.3, it could not happen in FP 11.2.  This bug was reported as reproducing
> with FP 11.2.

Comment 36

[Jeromie Clark]

The issue was directly related to the interaction of context menu and Full-Screen mode.  The resolution was to directly invoke Full Screen mode via context menu as of Flash Player 11.3.300.268.  

Content authors can work around this limitation and achieve the same end-result by listening for a click on the context menu item, then invoking FullScreen mode via ActionScript.

Comment 37

[Jeromie Clark]

The bugs that tracked this change are internal: 
Flash Player 11.3 - 3292736
Flash Player 11.4 - 3286863

They surfaced as STRs in one of the top few crash signatures, but did not account for the high volume of total signatures encountered.

Comment 38

[Jeromie Clark]

*** (correction, copy/paste mangling) The resolution was to disable the ability directly invoke Full Screen mode via context menu as of Flash Player 11.3.300.268.

Comment 39

[_w_]

Sorry for the late reply, I am not checking this daily.
No I do not use the context menu to switch to full screen. Just the normal 'switch to full screen' button also found on various other videosites. There are a variety of videosites which exhibit this behaviour for me. (One example is: http://www.feuerwehrleben.de/category/feuerwehr-videos/ the site is in german, but the  embedded videos behave like described in my previous post.)
Btw I run FF on Laptop with Intel HD + Nvidia graphics card (Optimus - switchable graphics) by default it runs with the Intel video card. If I switch to the Nvidia one for FF these crashes do not seem to appear.

Comment 40

[_w_]

Correction: if run FF with the Nvidia card a lot of strange things happen (blank white screen, some crashes and I never even see a video at most just sound) so i can not really tell if it exhibits the same behaviour. Apologies for not checking this thoroughly enough beforehand.

Comment 41

[Jeromie Clark]

Thanks, that's helpful information.  

Please attach the output of dxdiag, so we can get the exact details on your card and driver version. 

Go to Start > Run > dxdiag (or Start > and type dxdiag in the search box)
Click Save All Information, and attach the output to the bug

Thanks!

Comment 42

[Jeromie Clark]

After generating the logs, please try updating to the latest driver versions for your graphics cards and let us know if the issue persists.

Comment 43

[chris nuuja]

Ah, there are known bugs in the Optimus drivers that might be responsible here.   Can you reproduce the problem in a browser other than Firefox?   Also, can you try updating your Optimus drivers?  I believe the Optimus Control panel should have an option to update drivers .   Optimus can use both gpus at the same time, its not always clear what gpu is being used for what.

Finally can you report what version of Optimus drivers you currently have installed?    it is possible that you actually have the latest and the latest has a bug.

Comment 44

[blinky]

The crash occurs in the link https://dl.dropbox.com/u/95157096/e08ac2f5.html

Crash reports:
http://crash-stats.mozilla.com/report/index/bp-b977626c-b9f8-4314-8b4a-092962120731

http://crash-stats.mozilla.com/report/index/bp-96631173-636e-4c83-999c-4d0fb2120731

Comment 45

[_w_]

Attachment: DxDiag output

Comment 46

[_w_]

Ok, I attached the output of DxDiag, the Nvidia card is not mentioned in there...
The Intel drivers are the newest available from the Intel website.

The Nvidia drivers are really outdated (285.62) but should not interfere (There is a tray icon telling you wether the nvidia card is running any software, and when i normally run firefox it says 'No activity'). They only come into play when I explicitly start FF on the card via context menu. Sure then FF exhibits strange behaviour, but I think these are different bugs alltogether. Sorry for the false lead.

I looked through all of the components in System Manager but there is no separate Optimus driver. The Nvidia controlcenter also only shows the driver for the NVS 4200M installed in my system und no separate 'optimus' driver.

Finally I can not reproduce this behaviour in chrome.

Comment 47

[Johnny Stenback (:jst)]

I'm seeing what looks exactly like this when loading http://www.cnn.com/2012/07/28/living/rock-climber-honnold/index.html?hpt=hp_c3 here, and John Schoenick also reproduced the same problem. The hang reports I got were bp-34d69ed2-044f-4911-81fb-525532120801 and https://crash-stats.mozilla.com/report/index/bp-8a6f9665-698d-4080-a52b-22f832120801.

What's happening in this case is that the flash plugin asks us to evaluate some JS code, that JS code then ends up setting innerHTML on an element, which ends up destroying a DOM plugin element, which tears down the plugin, and the plugin deadlocks in that call. That is all clear from reading the stack reports in the crashes, and I've seen the same exact thing using the mozilla profiler too.

Adobe folks, what more information is needed here for you guys to be able to fix this deadlock in your code?

Comment 48

[Johnny Stenback (:jst)]

Sorry, STR for me is not simply loading that page, I load it, and then click in the video to play the video, and after that the plugin process hangs and eventually gets killed by Firefox.

Comment 49

[Jeromie Clark]

We believe that this hang Johnny is seeing is fixed in current Flash Player 11.4 builds.  I've emailed a copy to Johnny for feedback.

Comment 50

[Marcia Knous [:marcia]]

If we can get the build that it is fixed in, we can test on our end to confirm.

(In reply to Jeromie Clark from comment #49)
> We believe that this hang Johnny is seeing is fixed in current Flash Player
> 11.4 builds.  I've emailed a copy to Johnny for feedback.

Comment 51

[Ioana (away)]

This issue reproduced on my machine with Flash 11.3.300.265, but it doesn't reproduce anymore with Flash 11.4.400.252.

Comment 52

[Jeromie Clark]

In case anyone is waiting on the build where this is fixed, it's publicly available here:
http://labs.adobe.com/technologies/flashplatformruntimes/flashplayer11-4/

Comment 53

[[github robot]]

Commit pushed to master at https://github.com/mozilla/socorro

https://github.com/mozilla/socorro/commit/9ad8c3024324004268424aae7f9b953ef0725466
Add BaseGetNamedObjectDirectory to the append list for better signatures from bug 763896

Comment 54

[nick5990]

Using STR of comment 24:

I can't reproduce on Nightly v17.0a1 (11/08/12), Windows XP Mode, Flash Player v11.3.300.268

I also can't reproduce using Firefox v14.0.1, Windows 7 Ulimate SP1, Flash Player v11.3.300.270, MSI GTX 560Ti v131.42.

Comment 56

[smadayag]

we'll need to monitor crash-stats here also as FP11.4 adoption ramps up...

Comment 57

[Bob Clary [:bc] (inactive)]

see bug 566062 comment 40

Comment 58

[blinky]

Steps to reproduce:
1. Open the page https://dl.dropbox.com/u/95157096/85f61cf7/f3d6ffe9.html
2. Wait for 3 minutes.


Crash reports:

bp-f51910d4-4e3d-4a8c-93b4-9ddb52120902

bp-56ddae4c-284a-4b48-aaea-83e572120902

Comment 59

[Scoobidiver (away)]

(In reply to blinky from comment #58)
Nightly also crashes: bp-a1bce8ba-8517-46c1-be3b-df22f2120902.

Comment 60

[Jim Jeffery not reading bug-mail 1/2/11]

I got a slow script warning, but no crash, even after 15 mins on the page noted in comment #58

Latest m-c hourly build: 
https://hg.mozilla.org/mozilla-central/rev/3d9424eb6eb4

Comment 61

[blinky]

(In reply to Jim Jeffery not reading bug-mail 1/2/11 from comment #60)
> I got a slow script warning, but no crash, even after 15 mins on the page
> noted in comment #58
> 
> Latest m-c hourly build: 
> https://hg.mozilla.org/mozilla-central/rev/3d9424eb6eb4

Try this page https://dl.dropbox.com/u/95157096/85f61cf7/e34e9a87.html

Comment 62

[Jim Jeffery not reading bug-mail 1/2/11]

(In reply to blinky from comment #61)
> (In reply to Jim Jeffery not reading bug-mail 1/2/11 from comment #60)
> > I got a slow script warning, but no crash, even after 15 mins on the page
> > noted in comment #58
> > 
> > Latest m-c hourly build: 
> > https://hg.mozilla.org/mozilla-central/rev/3d9424eb6eb4
> 
> Try this page https://dl.dropbox.com/u/95157096/85f61cf7/e34e9a87.html

Yes, same as above... got a slow script warning twice, then eventually flash crashed.

Comment 63

[Hitomi Kudo]

By following the reproduction step in comment #58, I got this crash with FP 11.4.402.265:

https://crash-stats.mozilla.com/report/index/bp-ae5a5116-957d-4881-a43a-9a4b02120904

https://crash-stats.mozilla.com/report/index/bp-c97a2c20-dd5d-4375-9df0-719f82120904

Comment 64

[smadayag]

thanks for the STR in comment #58.  can we break that out into it's own bug so we can track it separately?  we can reproduce, but the signatures from Firefox 15/18 are similar to Hitomi's (comment #63).  thanks...

Comment 65

[Scoobidiver (away)]

(In reply to smadayag from comment #64)
> thanks for the STR in comment #58.  can we break that out into it's own bug
> so we can track it separately?
I filed bug 788423.

I close this one as fixed based on comment 52.

Comment 66

[banakon]

its sure that it has been fixed, I just obtined one, after switching from full screen.
https://crash-stats.mozilla.com/report/index/ade1e84b-1c30-44b7-b888-a93632131115

Comment 67

[candaules]

https://crash-stats.mozilla.com/report/index/e92cc55e-0802-4e87-b31f-50f0f2131117

full screen. google maps requiring (?!!) flash! hang when switching to other tabs.

Comment 68

[candaules]

https://crash-stats.mozilla.com/report/index/33595682-9e81-427b-855d-87d132131221

Comment 69

[banakon]

https://crash-stats.mozilla.com/report/index/ad6c6b5e-facf-4eb0-bc0f-4156e2140121

not fiexd yet...

Comment 70

[BMO Automation]

Version and milestone values are being reset to defaults as part of product refactoring.