Closed
Bug 85542
Opened 23 years ago
Closed 22 years ago
M099 N622 Trunk crash loading plugins w/Plugger [@ plugger.so - ns4xPluginStreamListener::OnFileAvailable]
Categories
(Plugins Graveyard :: Plugger, defect)
Tracking
(Not tracked)
RESOLVED
WORKSFORME
People
(Reporter: greer, Unassigned)
References
()
Details
(Keywords: crash, qawanted, topcrash)
Crash Data
Attachments
(8 files)
Talkback data for the M091 build is showing a number of crashes related to a variety of plugins loaded with Plugger. (31467107) Comments: loading an AVI using Plugger 3.2 (31516416) URL: www.theshadowlands.net (31516416) Comments: embedded Midi was loading via plugger 3.3 (31542982) Comments: Trying to view a real-video feed (it appears to be a java feed) (31599060) URL: www.mp3.com/electronic (31519066) URL: www.userfriendly.org/bts (31519066) Comments: just load the page www.userfriendly.org/bts (31507903) URL: http://www.bmwfilms.com/site_layout/view_film.asp?FilmID=4 (31507903) Comments: Trying to view a film at BMWfilms.com (31505845) URL: http://members.optushome.com.au/glohrere (31505845) Comments: Crashed while attempting to load the page! From what I see using the opera web browser it's an MS specific site requiring more plugins than a casual read should ever require. Along with a .#quakenotice_ie{....crap....} or <layer id="quakenotice_ns" (31505845) Comments: width=300 height=55 bgcolor=lightyellow visibility=hide></layer>embedded nonsense in the *crash* url.Keep smiling .... Here's the stack: plugger.so + 0x3e3f (0x40ffce3f) plugger.so + 0x464f (0x40ffd64f) ns4xPluginStreamListener::OnFileAvailable() nsPluginStreamListenerPeer::OnFileAvailable() nsPluginStreamListenerPeer::OnStopRequest() PluginListener::OnStopRequest() nsDocumentOpenInfo::OnStopRequest() nsStreamListenerTee::OnStopRequest() nsHttpChannel::OnStopRequest() nsOnStopRequestEvent::HandleEvent() nsARequestObserverEvent::HandlePLEvent() PL_HandleEvent() PL_ProcessEventsBeforeID() processQueue() nsVoidArray::EnumerateForwards() nsAppShell::ProcessBeforeID() handle_gdk_event() libgdk-1.2.so.0 + 0x15b80 (0x402d7b80) libglib-1.2.so.0 + 0x10736 (0x40303736) libglib-1.2.so.0 + 0x10d43 (0x40303d43) libglib-1.2.so.0 + 0x10eec (0x40303eec) libgtk-1.2.so.0 + 0x785d9 (0x402585d9) nsAppShell::Run() nsAppShellService::Run() main1() main() libc.so.6 + 0x18cae (0x40430cae)
Adding crash, topcrash, qawanted keywords
Updated•23 years ago
|
Severity: normal → critical
Confirmed crash using plugger 3.2 & plugger 3.3 on: www.theshadowlands.net http://members.optushome.com.au/glohrere seems to cause the browser to exit completely with Xlib server errors Can't reproduce the crash at: www.mp3.com/electronic (can't get the video feed at all) www.userfriendly.org/bts (get errors about xanim not finding the file) http://www.bmwfilms.com/site_layout/view_film.asp?FilmID=4 real stream works fine. quicktime pops up dialog asking for plugin asking for (video/QuickTime) At least in the userfriendly & theshadowlands cases, I'm noticing that the file attempting to be streamed is empty. If I have disk cache enabled, the cache file is a zero-byte file. If I don't then (with userfriendly), the file attempting to be streamed is "".
Comment 8•23 years ago
|
||
plugger.so & ns4xPluginStreamListener::OnFileAvailable() not showing up on either the trunk or M091 topcrash reports for today.
Comment 9•23 years ago
|
||
I've built debugged plugger3.3 (thanks to the author hubbe@hubbe.net the sources is here http://fredrik.hubbe.net/plugger.html) and spent some time to debug it.It looks like gtk based mozilla does not like exit() system call in any plugger's child processes, maybe because it does flush out all open streams, just guessing... But calling _exit() instead of exit() seems does help mozilla stay alive. Plugger.c patch as well as linux libc-2.1.92 binary plugger3.3.so file is following. Also sent an email to author asking about patch proposal.
Comment 10•23 years ago
|
||
Comment 11•23 years ago
|
||
Comment 12•23 years ago
|
||
*** Bug 88671 has been marked as a duplicate of this bug. ***
Comment 13•23 years ago
|
||
redhat plugger-3.3-3 work with Mozilla.
Comment 14•23 years ago
|
||
Comment 15•23 years ago
|
||
I've created a new patch to plugger (while working on bug 74080). The problem is that exit() calls all the atexit()/on_exit() handlers that Mozilla has registered; _exit() doesn't call them. My patch, rather than replacing exit() with _exit(), adds a new exit handler that calls _exit(). Since the handlers are called in the reverse order that they are added in, none of the older handlers get called after this one, so exit() can safely get called. Note that there are a few debugging statements added in addition to the core of the fix.
Comment 16•23 years ago
|
||
I agree, that should work too.
Comment 17•23 years ago
|
||
*** Bug 89881 has been marked as a duplicate of this bug. ***
Comment 18•23 years ago
|
||
*** Bug 90599 has been marked as a duplicate of this bug. ***
Comment 19•23 years ago
|
||
any chance to get the patch into the next milestone?
Comment 20•23 years ago
|
||
*** Bug 93739 has been marked as a duplicate of this bug. ***
Comment 21•23 years ago
|
||
Is my understanding correct that this is a plugger bug and we should resolve it as INVALID?
Comment 22•23 years ago
|
||
actually, this bug is about crash if plugin's child process call exit(), plugger is just good test case
Comment 23•23 years ago
|
||
OK, all patches here are for the plugger code. Can we get a fix in Mozilla then?
Comment 24•23 years ago
|
||
*** Bug 98525 has been marked as a duplicate of this bug. ***
Comment 25•23 years ago
|
||
*** Bug 92739 has been marked as a duplicate of this bug. ***
Reporter | ||
Comment 26•23 years ago
|
||
There are currently no Talkback incidents under the ns4xPluginStreamListener::OnFileAvailable signature on N620, M094 or Trunk. Was anything ever checked in for this bug? Is it WFM?
Comment 28•23 years ago
|
||
Updating summary with M094, since plugger.so has been a topcrasher for Mozilla 0.9.4.
Summary: M091 crash loading plugins w/Plugger [@ plugger.so] [@ ns4xPluginStreamListener::OnFileAvailable()] → M094 crash loading plugins w/Plugger [@ plugger.so] [@ ns4xPluginStreamListener::OnFileAvailable]
Comment 29•23 years ago
|
||
The error that I got in the shell window seems to indicate that it may be a problem because plugger doesn't have any type of image present to represent its plugin. The embed tag seems to assume that there is going to be a graphic, and this error almost indicates that a nonexistant image results in a null pixmap object, and when trying to call a pixmap function, a Gdk error is caused since it probably can't handle a null pixmap. Something to try would be to add an image to plugger or change mozilla so that it doesn't try to display an image if one is not given. Gdk-ERROR **: BadPixmap (invalid Pixmap parameter) serial 6652 error_code 4 request_code 56 minor_code 0 Gdk-ERROR **: BadDrawable (invalid Pixmap or Window parameter) serial 6653 error_code 9 request_code 62 minor_code 0
Reporter | ||
Comment 30•23 years ago
|
||
Updating the summary to M095 N620 Trunk, since all are showing incidents with the plugger.so signature. Attaching today's comments below.
Summary: M094 crash loading plugins w/Plugger [@ plugger.so] [@ ns4xPluginStreamListener::OnFileAvailable] → M095 N620 Trunk crash loading plugins w/Plugger [@ plugger.so] [@ ns4xPluginStreamListener::OnFileAvailable]
Reporter | ||
Comment 31•23 years ago
|
||
Comment 32•23 years ago
|
||
*** Bug 108299 has been marked as a duplicate of this bug. ***
Comment 33•23 years ago
|
||
-> mozilla 1.0.1 this is plugger's bug, but probably it's possible to implement something like on_exit() handler in mozilla to catch exit() call from plugins, for now this bug has two patches for plugger.c (attachment 57833 [details] & attachment 40402 [details] [diff] [review]) along with workable linux libc-2.1.92 binary plugger3.3.so (attachment 40403 [details]) I've posted some times ago.
Status: NEW → ASSIGNED
Target Milestone: --- → mozilla1.0.1
Comment 34•23 years ago
|
||
Updating summary with M096, this has been a topcrasher with Mozilla 0.9.6 as well: plugger.so 30 85542 ASSI serge@netscape.com mozilla1.0.1 2001-11-29 BBID range: 17330 - 361720 Min/Max Seconds since last crash: 12 - 1076855 Min/Max Runtime: 6672 - 1115391 Crash data range: 2001-12-03 to 2001-12-11 Build ID range: 2001112012 to 2001112012 Keyword List : Stack Trace: plugger.so + 0x3e3f (0x426c0e3f) plugger.so + 0x464f (0x426c164f) ns4xPluginStreamListener::OnFileAvailable() nsPluginStreamListenerPeer::OnFileAvailable() nsPluginStreamListenerPeer::OnStopRequest() PluginListener::OnStopRequest() nsDocumentOpenInfo::OnStopRequest() nsHttpChannel::OnStopRequest() nsOnStopRequestEvent::HandleEvent() nsARequestObserverEvent::HandlePLEvent() PL_HandleEvent() PL_ProcessEventsBeforeID() processQueue() nsVoidArray::EnumerateForwards() nsAppShell::ProcessBeforeID() handle_gdk_event() (361720) URL: windows media mpg on of ufo at wtc bombing www2.justnet.ne.jp/~kiti/Ufo/wtc/wtc.htm (361720) Comments: boom! i think it is plugger (348463) Comments: i did view image on a png and it died (17330) Comments: testing plugger plugin
Summary: M095 N620 Trunk crash loading plugins w/Plugger [@ plugger.so] [@ ns4xPluginStreamListener::OnFileAvailable] → M096 N620 Trunk crash loading plugins w/Plugger [@ plugger.so] [@ ns4xPluginStreamListener::OnFileAvailable]
Comment 35•23 years ago
|
||
updating summary with M097, since this continues to be a topcrasher with Mozilla 0.9.7: Rank StackSignature Count 8 plugger.so 52 85542 ASSI serge@netscape.com mozilla1.0.1 Mon 19:23 ==================================================================================================== Count Offset Real Signature [ 2 plugger.so + 0x422f (0x412fb22f) ef8ab85e - ns4xPluginStreamListener::OnFileAvailable() ] [ 1 plugger.so + 0x422f (0x41a9522f) d7a6eab4 - ns4xPluginStreamListener::OnFileAvailable() ] [ 1 plugger.so + 0x422f (0x41a0022f) 59894ae2 - ns4xPluginStreamListener::OnFileAvailable() ] [ 1 plugger.so + 0x422f (0x4174422f) bb8fac6c - ns4xPluginStreamListener::OnFileAvailable() ] [ 1 plugger.so + 0x422f (0x416c722f) 2f1d9a3f - ns4xPluginStreamListener::OnFileAvailable() ] [ 1 plugger.so + 0x422f (0x4168b22f) fd1ae016 - ns4xPluginStreamListener::OnFileAvailable() ] Crash date range: 2002-01-10 to 2002-01-16 Min/Max Seconds since last crash: 25 - 59326 Min/Max Runtime: 14014 - 478703 Keyword List : Count Platform List 6 Linux 2.4.8-26mdk 1 Linux 2.4.3-20mdk Count Build Id List 7 2001122108 No of Unique Users 4 Stack trace(Frame) plugger.so + 0x422f (0x412fb22f) plugger.so + 0x4c5f (0x412fbc5f) ns4xPluginStreamListener::OnFileAvailable() nsPluginStreamListenerPeer::OnFileAvailable() nsPluginStreamListenerPeer::OnStopRequest() PluginListener::OnStopRequest() nsDocumentOpenInfo::OnStopRequest() nsHttpChannel::OnStopRequest() nsOnStopRequestEvent::HandleEvent() nsARequestObserverEvent::HandlePLEvent() PL_HandleEvent() PL_ProcessPendingEvents() nsEventQueueImpl::ProcessPendingEvents() event_processor_callback() our_gdk_io_invoke() ==================================================================================================== Count Offset Real Signature [ 1 plugger.so + 0x422f (0x4184122f) 6dc4c40f - ns4xPluginStreamListener::OnFileAvailable() ] [ 1 plugger.so + 0x422f (0x4170922f) be6c2989 - ns4xPluginStreamListener::OnFileAvailable() ] [ 1 plugger.so + 0x422f (0x416da22f) ef021aab - ns4xPluginStreamListener::OnFileAvailable() ] Crash date range: 2002-01-14 to 2002-01-16 Min/Max Seconds since last crash: 13232 - 88924 Min/Max Runtime: 263891 - 366219 Keyword List : Count Platform List 3 Linux 2.4.8-26mdk Count Build Id List 3 2001122108 No of Unique Users 1 Stack trace(Frame) plugger.so + 0x422f (0x4184122f) plugger.so + 0x4c5f (0x41841c5f) ns4xPluginStreamListener::OnFileAvailable() nsPluginStreamListenerPeer::OnFileAvailable() nsPluginStreamListenerPeer::OnStopRequest() PluginListener::OnStopRequest() nsDocumentOpenInfo::OnStopRequest() nsFTPChannel::OnStopRequest() DataRequestForwarder::OnStopRequest() nsOnStopRequestEvent::HandleEvent() nsARequestObserverEvent::HandlePLEvent() PL_HandleEvent() PL_ProcessPendingEvents() nsEventQueueImpl::ProcessPendingEvents() event_processor_callback() our_gdk_io_invoke() ==================================================================================================== Count Offset Real Signature [ 1 plugger.so + 0x422f (0x417fa22f) 68cc673a - ns4xPluginStreamListener::OnFileAvailable() ] Crash date range: 2002-01-14 to 2002-01-14 Min/Max Seconds since last crash: 73710 - 73710 Min/Max Runtime: 73710 - 73710 Keyword List : Count Platform List 1 Linux 2.4.8-26mdk Count Build Id List 1 2001122108 No of Unique Users 1 Stack trace(Frame) plugger.so + 0x422f (0x417fa22f) plugger.so + 0x4c5f (0x417fac5f) ns4xPluginStreamListener::OnFileAvailable() nsPluginStreamListenerPeer::OnFileAvailable() nsPluginStreamListenerPeer::OnStopRequest() nsHttpChannel::OnStopRequest() nsOnStopRequestEvent::HandleEvent() nsARequestObserverEvent::HandlePLEvent() PL_HandleEvent() PL_ProcessPendingEvents() nsEventQueueImpl::ProcessPendingEvents() event_processor_callback() our_gdk_io_invoke() If we have patches for this, why is the target milestone set to after Mozilla 1.0? Is it too risky? This has been a consitant topcrasher, so it would be nice to see this fixed sooner.
Summary: M096 N620 Trunk crash loading plugins w/Plugger [@ plugger.so] [@ ns4xPluginStreamListener::OnFileAvailable] → M097 N621 Trunk crash loading plugins w/Plugger [@ plugger.so - ns4xPluginStreamListener::OnFileAvailable]
Comment 36•23 years ago
|
||
the patches are not for mozilla code, its is for plugger.
Comment 37•23 years ago
|
||
*** Bug 121319 has been marked as a duplicate of this bug. ***
Comment 38•23 years ago
|
||
*** Bug 117725 has been marked as a duplicate of this bug. ***
Comment 39•23 years ago
|
||
*** Bug 123879 has been marked as a duplicate of this bug. ***
Reporter | ||
Comment 40•23 years ago
|
||
Adjusting summary, still in M098.
Summary: M097 N621 Trunk crash loading plugins w/Plugger [@ plugger.so - ns4xPluginStreamListener::OnFileAvailable] → M098 N621 Trunk crash loading plugins w/Plugger [@ plugger.so - ns4xPluginStreamListener::OnFileAvailable]
Comment 41•23 years ago
|
||
*** Bug 126105 has been marked as a duplicate of this bug. ***
Comment 42•23 years ago
|
||
*** Bug 123928 has been marked as a duplicate of this bug. ***
Comment 44•23 years ago
|
||
The crash in ns4xPluginStreamListener::OnFileAvailable is fixed by check in for bug 123974, at least, I'm unable to reproduce that crash using plugger-4.0 and I have not seen any TB reports with such signature since fix for 123974 was checked in. However plugger-3.x can still expose this crash.
Comment 45•23 years ago
|
||
Talkback folks: I've minused this bug because we can't find talk back data for it. If you still think this is a topcrash, please justify it for us!
Updated•23 years ago
|
Target Milestone: mozilla1.0.1 → mozilla1.1
Comment 46•23 years ago
|
||
Well, there appears to be plenty of talkback data for this crash. It was a topcrash with Mozilla 0.9.8 and I see quite a few crashes with Linux MozillaBranch build 20020207000 (what build that is i don't know). And it doesn't look like this was fixed with the checkin for bug 123974...i see a few crashes on the MozillaTrunk after the checkin date. Here is the most recent crash: Incident ID 3398750 Stack Signature plugger.so + 0x422f (0x41d0422f) f8eae59a Trigger Time 2002-02-26 14:38:10 Email Address sammy@jaton.nl URL visited Build ID 2002022508 Product ID MozillaTrunk Platform Operating System LinuxIntel Module Trigger Reason SIGSEGV: Segmentation Fault: (signal 11) User Comments Stack Trace plugger.so + 0x422f (0x41d0422f) plugger.so + 0x4c5f (0x41d04c5f) ns4xPluginStreamListener::OnFileAvailable() nsPluginStreamListenerPeer::OnFileAvailable() nsPluginStreamListenerPeer::OnStopRequest() PluginListener::OnStopRequest() nsDocumentOpenInfo::OnStopRequest() nsHttpChannel::OnStopRequest() nsOnStopRequestEvent::HandleEvent() nsARequestObserverEvent::HandlePLEvent() PL_HandleEvent() PL_ProcessPendingEvents() nsEventQueueImpl::ProcessPendingEvents() event_processor_callback() our_gdk_io_invoke() libglib-1.2.so.0 + 0x1001e (0x4039701e) libglib-1.2.so.0 + 0x117f3 (0x403987f3) libglib-1.2.so.0 + 0x11dd9 (0x40398dd9) libglib-1.2.so.0 + 0x11f8c (0x40398f8c) libgtk-1.2.so.0 + 0x94803 (0x402ad803) nsAppShell::Run() nsAppShellService::Run() main1() main() libc.so.6 + 0x1c507 (0x404e2507) And for internal folks...to see all the Talkback data...go to: http://climate/reports/searchstacksignature.cfm?stacksig=plugger.so to see all the crashes of this kind.
Reporter | ||
Comment 47•23 years ago
|
||
At least one of the crashing users states explicitly that he was using plugger 3.2 (3254868). -> see Serge's comment #44 Since we can't prevent the 3.x crashes, keeping the minus seems best for now so we can focus on higher priorities. Serge, how much more that can be done for this bug?
Comment 48•23 years ago
|
||
I've tried several urls with plugger-4.0 all work fine bug 121319 http://bart.w-wa.pl/mozilla-crash.html bug 117725 http://www.apple.com/itunes/theater/radio.html bug 123928 http://www.ece.cmu.edu/research/movies/spin.shtml However this url works with plugger-4.0 but crashes with plugger-3.3 http://members.optushome.com.au/glohrere with stack trace similar to TB reports 3398750, 3398737, 3377299 Program received signal SIGPIPE, Broken pipe. #0 0x406b8584 in __libc_write () from /lib/i686/libc.so.6 #1 0x403215f8 in __DTOR_END__ () at eval.c:41 #2 0x42ad185e in NPP_Write (instance=0x898e424, stream=0x8986a60, offset=6535, len=3336, buf=0x8a033f0) at plugger.c:1124 #3 0x42ad24c5 in Private_Write () from /u/serge/prj/plugins/plugger-3.3/plugger.so #4 0x41ee72d0 in ns4xPluginStreamListener::OnDataAvailable (this=0x8986a48, pluginInfo=0x898d3f0, input=0x89cc6f0, length=3336) at ns4xPluginInstance.cpp:390 #5 0x41eee878 in nsPluginStreamListenerPeer::OnDataAvailable (this=0x898d398, request=0x898d480, aContext=0x0, aIStream=0x89cc6f0, sourceOffset=6535, aLength=3336) at nsPluginHostImpl.cpp:2198 #6 0x4099367b in nsStreamListenerTee::OnDataAvailable (this=0x89cc6d0, request=0x898d480, context=0x0, input=0x898da50, offset=6535, count=3336) at nsStreamListenerTee.cpp:56 #7 0x409d5496 in nsHttpChannel::OnDataAvailable (this=0x898d480, request=0x89899dc, ctxt=0x0, input=0x898da50, offset=6535, count=3336) at nsHttpChannel.cpp:2543 #8 0x409923a7 in nsOnDataAvailableEvent::HandleEvent (this=0x42b017b8) at nsStreamListenerProxy.cpp:193
Comment 49•23 years ago
|
||
*** Bug 130567 has been marked as a duplicate of this bug. ***
Comment 50•23 years ago
|
||
Is this a topcrash on Windows or Mac? I don't understand "plugger 3.3 vs plugger 4.0", but can users upgrade this the plugger 4.0 easily?
Comment 51•23 years ago
|
||
this is unix only bug. >can users upgrade this the plugger 4.0 easily? yes, from the author's page http://fredrik.hubbe.net/plugger.html
Comment 52•23 years ago
|
||
We should remove qawanted since this is reproducible with plugger 3.3.
Keywords: qawanted
Comment 53•23 years ago
|
||
Updating summary with M099 and N622...there are still crashes being reported for Mozilla 0.9.9 and Netscape 6.22. And it looks like we are still crashing with Plugger 4: Incident ID 4549849 Stack Signature plugger.so + 0x422f (0x4117b22f) c0e71c09 Trigger Time 2002-03-27 21:13:18 Email Address URL visited http://www.apple.com/trailers Build ID 2002031423 Product ID Netscape6.22 Platform Operating System LinuxIntel Module Trigger Reason SIGSEGV: Segmentation Fault: (signal 11) User Comments trying to get plugger 4.0 to work for a quicktime page Stack Trace plugger.so + 0x422f (0x4117b22f) plugger.so + 0x4c5f (0x4117bc5f) ns4xPluginStreamListener::OnFileAvailable() nsPluginStreamListenerPeer::OnFileAvailable() nsPluginStreamListenerPeer::OnStopRequest() nsHttpChannel::OnStopRequest() nsOnStopRequestEvent::HandleEvent() nsARequestObserverEvent::HandlePLEvent() PL_HandleEvent() PL_ProcessPendingEvents() nsEventQueueImpl::ProcessPendingEvents() event_processor_callback() our_gdk_io_invoke() Is this something we can check for and prevent in our code or is it a problem with pugger itself? Adding qawanted keyword again...although we know of people crashing...we still don't have a solid reproducible testcase.
Keywords: qawanted
Summary: M098 N621 Trunk crash loading plugins w/Plugger [@ plugger.so - ns4xPluginStreamListener::OnFileAvailable] → M099 N622 Trunk crash loading plugins w/Plugger [@ plugger.so - ns4xPluginStreamListener::OnFileAvailable]
Comment 54•22 years ago
|
||
*** Bug 142573 has been marked as a duplicate of this bug. ***
Comment 55•22 years ago
|
||
This stack signature is not showing up in talkback at all now. The last report submitted had a submission date of 2002-04-27. Marking WFM.
Status: ASSIGNED → RESOLVED
Closed: 22 years ago
Resolution: --- → WORKSFORME
Comment 56•22 years ago
|
||
could bug 155924 be a valid testcase ?
Component: Plug-ins → Other
Product: Core → Plugins
QA Contact: shrir → other
Target Milestone: mozilla1.1alpha → ---
Version: Trunk → unspecified
Assignee: srgchrpv → nobody
Component: Other → Plugger
QA Contact: other → plugger
Assignee | ||
Updated•13 years ago
|
Crash Signature: [@ plugger.so - ns4xPluginStreamListener::OnFileAvailable]
Assignee | ||
Updated•8 years ago
|
Product: Plugins → Plugins Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•