Closed Bug 83747 Opened 23 years ago Closed 23 years ago

Trunk, M092 & N610 branch crash [@ libc.so.6 - ssl_DefSend()]

Categories

(Core :: Layout: Form Controls, defect)

x86
Linux
defect
Not set
critical

Tracking

()

VERIFIED DUPLICATE of bug 92123

People

(Reporter: jcarpenter0524, Assigned: ddrinan0264)

References

Details

(Keywords: crash, qawanted, topcrash)

Crash Data

This bug is a topcrasher for M091.
Added [@ libc.so.6 - ssl_DefSend()] for tracking.  

Here are some Comments from a recent crash:

 Incident ID: 31153115
 Build ID:  2001053108 LinuxIntel
 filling out a form, I tabbed, then started entering text, I think mozilla
 moved the focus to a text label instead of the textinput field 


Here is the stack trace:

Incident ID 31153115 
libc.so.6 + 0xcca42 (0x40540a42) 
libnspr4.so + 0x1d1a8 (0x401b71a8) 
libnspr4.so + 0xc57a (0x401a657a) 
ssl_DefSend() 
ssl3_SendRecord() 
SSL3_SendAlert() 
ssl_SecureClose() 
ssl_Close() 
nsSSLIOLayerClose() 
libnspr4.so + 0xb757 (0x401a5757) 
nsSocketTransport::CloseConnection() 
nsSocketTransport::Process() 
nsSocketTransportService::Run() 
nsThread::Main() 
libnspr4.so + 0x1f6ee (0x401b96ee) 
libpthread.so.0 + 0x6fc5 (0x401d5fc5)
added crash, topcrash keywords
Keywords: crash, topcrash
Severity: normal → critical
--> html forms
Assignee: morse → rods
Component: Form Manager → HTML Form Controls
QA Contact: tpreston → vladimire
I need an URL for this....
Status: NEW → ASSIGNED
there are several reports with this stack trace but this is the
only one with any commment.   we will have keep an eye on the
new incoming boxes on this one.  

maybe it should be reassigned to talkback-staff@netscape.com ?
I don't think aat this moment it looks like my bug, in fact, I think it ought 
start with someone who knos something about SSL.
adding qawanted to keywords
(nothing in today's talkback reports with a URL or any more comments for this.)
Keywords: qawanted
This crash has been around on the Trunk for a while also.  No recent crashes
show this, but I know it's still there.   I'll update the bug if I come across
any more info.  Just adding Trunk to summary for now.
Summary: M091 crash [@ libc.so.6 - ssl_DefSend()] → Trunk & M091 crash [@ libc.so.6 - ssl_DefSend()]
Some one needs to reassign this to someone else.
lets add some psm folks
Assignee: rods → ddrinan
Status: ASSIGNED → NEW
Keywords: nsenterprise
Here are the latest user comments/urls for this crash

     (31890784)	Comments: reading mail
     (31888185)	Comments: Changing to another IMAP folder.
     (31883537)	Comments: I pressed ctrl+m to create a new message. 
     (31877483)	Comments: Tried to print a page.  The printer was unknow. Crash.
     (31877245)	Comments: Tried to print a page.  stderr output:Width = 612 Height = 792dpi
72.000000 top 36 bottom 36 left 36 right 36sh: lp: command not foundFollowed by
crash.
     (31875869)	Comments: hmm
     (31862872)	Comments: trying to print an email message
     (31855862)	Comments: Starting up mozilla - mozilla still started
     (31852304)	Comments: Starting mozilla -
     (31849525)	URL: http://www.timwest.com
(31849525)
Comments: Entered the above URL. Page did not start loading; the browser just
hung. bulbul@fizzylogic.com
(31835688)
Comments: Saving a message to the Drafts folder of an IMAP mailbox.
     (31828664)	Comments: I simply launched Mozilla
     (31817973)	Comments: Opened news reader
     (31816309)	Comments: Shutdown (File->Quit)
     (31807230)	Comments: Changing IMAP folders.
     (31798373)	Comments: changing from composer to browser window
     (31792210)	Comments: Opening ftp server by dragging the URL to second mozilla window.
     (31789150)	Comments: I think when I closed the "main" Mozilla window the others were
unable to load pages anymore.
     (31789024)	Comments: crashed while attempting to read  mail that contined java files
includingan applet and its html
     (31781361)	URL: http://www.cart.com
(31781361)
Comments: I was opening this page for the firt time.  I see that it detects if
Macomedia flash is installed.  Jist after that it died. 
     (31777898)	URL: http://www.computerbuyers.com.au
(31777898)
Comments: Entered non-existant domain name
     (31774520)	Comments: news?
     (31766112)	Comments: Closing Mozilla
     (31756271)	Comments: In my IMAP mailbox
     (31753338)	URL: bugzilla.mozilla.com
     (31753338)	Comments: I had entered a bug id into the main page
     (31753338)	Comments:  windowmanager's close button had no effect.Also did not see any
relevant messages in the term session thatlaunched Mozilla.
     (31745359)	URL: News.bbc.co.uk
     (31745359)	Comments: It hung so I did a cntrl-c (I got this form on the
restart)probable bug 85025.  This and www.bbc.co.uk seem to cause this probelm
     (31719247)	URL: http://hotwired.lycos.com/webmonkey/
(31719247)
Comments: clicked "javascript code library".  browser stalled (throbber active
butno new page loaded)
     (31640102)	URL: https://www.newearth.org/~bac/postarchive/2001/06/14.php
(31612104)
Comments: As far as I can tell

David, are you the correct owner for this bug?  If so, any progress?
This is a topcrasher for M092.  Here are the few entries that I found for this
crash that had useful comments:

libc.so.6 + 0xa1172 (0x40523172) 97cea164
         line 
        Build: 2001062823 CrashDate: 2001-07-01 UptimeMinutes: 928  Total: 1468 
        OS: Linux 2.2.16
         Detailed : http://climate/reports/incidenttemplate.cfm?bbid=32410167
         StackTrace:
http://climate/reports/stackcommentemail.cfm?dynamicBBID=32410167
     (32410167) URL: http://www.go.dlr.de/fresh/unix/src/misc/
     (32410167) Comments: I am not sure the URLis relevant. What is happening is
that mozulla doesn't crash but i seems to loose connectivity/transport.From that
point no more internet/ethernet transfer will happen either on the browser or
the mailer.stopping and restarting
     (32410167) Comments:  mozilla get rid of the problem (until next
time).RegardsJC


libc.so.6 + 0xa1172 (0x40523172) 97cea164
         line 
        Build: 2001062823 CrashDate: 2001-06-30 UptimeMinutes: 55  Total: 540 
        OS: Linux 2.2.16
         Detailed : http://climate/reports/incidenttemplate.cfm?bbid=32380925
         StackTrace:
http://climate/reports/stackcommentemail.cfm?dynamicBBID=32380925
     (32380925) Comments: browsing megapixel.netanother window was running the
mailer with an IMAP/SSL account and a noraml POP3 accountAt that time I don't
know what the problem is as nothing as crashed (yet) or no error is reported
(except for this window) 


This crash has been around for a long time...has anyone had a chance to work on it?
Summary: Trunk & M091 crash [@ libc.so.6 - ssl_DefSend()] → Trunk & M092 crash [@ libc.so.6 - ssl_DefSend()]
Here's another entry and it's stack trace:

libc.so.6 + 0xe5212 (0x40595212) 757bf776
         line 
        Build: 2001062823 CrashDate: 2001-06-30 UptimeMinutes: 6  Total: 6 
        OS: Linux 2.4.3-20mdk
         Detailed : http://climate/reports/incidenttemplate.cfm?bbid=32356835
         StackTrace:
http://climate/reports/stackcommentemail.cfm?dynamicBBID=32356835
     (32356835) URL: https://ecomm.royal.it:443/ecomm/f_preven.htm
     (32356835) Comments: Entering in a https page

Incident ID 32356835 
libc.so.6 + 0xe5212 (0x40595212) 
libnspr4.so + 0x1ce78 (0x401b6e78) 
libnspr4.so + 0xc42a (0x401a642a) 
ssl_DefSend() 
ssl3_SendRecord() 
SSL3_SendAlert() 
ssl_SecureClose() 
ssl_Close() 
nsSSLIOLayerClose() 
libnspr4.so + 0xb5f7 (0x401a55f7) 
nsSocketTransport::CloseConnection() 
nsSocketTransport::Process() 
nsSocketTransportService::Run() 
nsThread::Main() 
libnspr4.so + 0x1f3ee (0x401b93ee) 
libpthread.so.0 + 0x7817 (0x401d4817) 
I'm trying to reproduce this, but I'm not having much success. :(
Can we get QA to create a reproducible test case for this?  It crashes in the
depths of NSS/NSPR which leads me to believe we're probably trying to close down
a socket twice or something like that, but since I can't tell what type of
connection is causing this (ie, IMAP or https) it's almost impossible to figure
out what's causing this.
I saw a comment or two mention https connections.  If I find them again, I'll
post the complete entries.
PSM engineers have attempted to reproduce this crash without luck. We can't help
on this until we have a good way to reproduce.  Any help would be useful.
Blocks: 88268
According to the latest MozillaTrunk Talkback data, one particular user has been 
seeing this crash A LOT.  If you do this query:

http://climate/reports/searchemail.cfm?email=cb%40physicman.net

you'll find that most of this user's recent crashes are of this kind.  Here is 
the most recent entry from that user (the others have identical info and stack 
traces):

libc.so.6 + 0xcca92 (0x40559a92) 6b65fb8b
         line 
        Build: 2001070821 CrashDate: 2001-07-12 UptimeMinutes: 124  Total: 875 
        OS: Linux 2.4.7-pre2-xfs
         Detailed : http://climate/reports/incidenttemplate.cfm?bbid=32806370
         StackTrace: 
http://climate/reports/stackcommentemail.cfm?dynamicBBID=32806370

I don't know if this crash is the result of the Linux version he is using, but 
it might be worth looking into exactly what kind of Linux installation the 
"Linux 2.4.7-pre2-xfs" is.
Also a topcrash on N610 branch.

Here is the stacktrace from a crash on today's build:

Incident ID 33016663 
 Stack Signature
                libc.so.6 + 0xe7c32 (0x4057cc32) 63b29814 
 Trigger Time 
                2001-07-17 15:02:03 
 Build ID
                2001071704 
 Product ID
                Netscape6.10 
 Platform ID
                LinuxIntel 
 Stack Trace

libc.so.6 + 0xe7c32 (0x4057cc32) 
libnspr4.so + 0x1ce78 (0x401b6e78) 
libnspr4.so + 0xc42a (0x401a642a) 
ssl_DefSend() 
ssl3_SendRecord() 
SSL3_SendAlert() 
ssl_SecureClose() 
ssl_Close() 
nsSSLIOLayerClose() 
libnspr4.so + 0xb5f7 (0x401a55f7) 
nsSocketTransport::CloseConnection() 
nsSocketTransport::ClearSocketBS() 
nsSocketBS::SetTransport() 
nsSocketBS::~nsSocketBS() 
nsSocketBOS::~nsSocketBOS() 
nsSocketBOS::Release() 
nsCOMPtr_base::assign_with_AddRef() 
nsImapProtocol::OnStopRequest() 
nsOnStopRequestEvent::HandleEvent() 
nsARequestObserverEvent::HandlePLEvent() 
PL_HandleEvent() 
PL_ProcessPendingEvents() 
nsEventQueueImpl::ProcessPendingEvents() 
nsImapProtocol::Run() 
nsThread::Main() 
libnspr4.so + 0x1f3ee (0x401b93ee) 
libpthread.so.0 + 0x6d8c (0x401e2d8c) 


Comment from recent crash:
Incident 33025637     Comments: crash on shutdown
Summary: Trunk & M092 crash [@ libc.so.6 - ssl_DefSend()] → Trunk, M092 & N610 branch crash [@ libc.so.6 - ssl_DefSend()]
I think this bug should be marked a dup of 92123.  It's a newer bug, but more
people have looked at it and I think there's been a lot of progress there.  If
this is indeed a dup of that bug, please mark it so. Thanks.
My memory from looking at talkback reports was that this crash was a SIGPIPE
(not a SIGSEGV as most crashes are).  I'm not behind the firewall now so I can't
double-check.

It's useful to include the crash reason when filing talkback bugs.
*** Bug 92517 has been marked as a duplicate of this bug. ***
This is a dupe, since its SIGPIPE. Marking.

*** This bug has been marked as a duplicate of 92123 ***
Status: NEW → RESOLVED
Closed: 23 years ago
Resolution: --- → DUPLICATE
verifying
Status: RESOLVED → VERIFIED
Is there a web site that produces this crash? 
https://certificates.netscape.com/ does not anymore.  Without a reproducible
test case, this will *never* get fixed.

(I've tried the 3 sites mentioned in the talkback reports and none of them cause
my build from this morning to crash.)

The fact that the crash on https://certificates.netscape.com/ went away without
changing PSM code leads me to believe we're crashing in NSS because of a bug in
a different part of the code.  Perhaps trying to write to a closed socket that
only happens on slow connections.  

Anyone in QA have a modem they could test with?
javi: This bug is the dupe; your comment probably belongs in bug 92123. My
report (bug 92517) happened from inside B21 to our mailservers, so its not a
slow connection thing. Both times that occured (and the fact that this is a
SIGPIPE) suggests that we are writing to a closed socket. If you send the
talkback report, then, at least for my reports, the browser kept going normally.

Unfortunately, I couldn't reproduce this either, but it hasn't happened since I
stopped reading my mail via SSL.
Crash Signature: [@ libc.so.6 - ssl_DefSend()]
You need to log in before you can comment on or make changes to this bug.