Closed Bug 203735 Opened 21 years ago Closed 14 years ago

sudden crash kills xfs

Categories

(SeaMonkey :: General, defect)

Product:
SeaMonkey
Component:
General
Platform:
x86
Linux
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

(Not tracked)

Status:
RESOLVED WORKSFORME

People

(Reporter: pipian, Unassigned)

Details

(Keywords: crash, stackwanted, Whiteboard: bugday0420) 

User-Agent:       Mozilla/5.0 (compatible; Konqueror/3.1; Linux)
Build Identifier: http://ftp.mozilla.org/pub/mozilla/nightly/latest/mozilla-i686-pc-linux-gnu-sea.tar.gz (April 28th Nightly)

Upon accessing my bookmarks recently (it may not have been limited to the above release, 
but also to a previous one dating from soon after the 1.3 release), Mozilla locked up for several 
moments and then crashed.  The first time this happened (I am not quite certain if this also 
happened while looking at mail as well), all of my bookmarks were deleted (but only 
discovered upon upgrade to the listed nightly build, so I am uncertain whether the crash or the 
upgrade caused loss of the bookmarks).  The second time has just occured, and it does not 
appear that they were deleted again, though it may be so upon reboot.  In this case however, I 
was accessing my bookmarks when the same lockup and crash occurred.  In both cases, 
Mozilla fails to restart (for unknown reasons).  I am currently using Mandrake 9.1.  This may be 
split up into several bugs, but I am not exactly sure (as I am currently uncertain of the exact 
cause of the first crash.  The second crash was undoubtedly triggered by accessing 
bookmarks.) 

Reproducible: Sometimes

Steps to Reproduce:
did you send in a talkback report? (please provide a talkback ID once you do)

worksforme with linux trunk 2003042805
Keywords: crash
I'm not sure but it might be a general instability with my setup (that exhibits
itself primarily trhough a crash in Mozilla, that ultimately brings down X (If I
log out and log back in, I can't restart X, claiming that there are errors in
trying to find fonts).

In any case, I've also noted that talkback consistently fails to work for me
(which is why I have not been able to send any data).  But that would be tied
with <a href="http://bugzilla.mozilla.org/show_bug.cgi?id=107287">bug 107287</a>.
can you try the fixup in bug 107287 Comment 16? (or 11)
Moving this, as it appears to be a general bug (bookmarks aren't necessarily the
cause).  In addition, it appears that the crash (whatever it may be) kills the
xfs server (Or perhaps a sudden death of the xfs server kills Mozilla?),
possibly explaining why Mozilla fails to reopen and X fails to restart. 
Talkback reported the exceptions (I believe from when I tried to run the program
while xfs was dead) as TB19636730M and TB19636703G (if that helps anyone)
Component: Bookmarks → Browser-General
Summary: crash upon bookmarks access (possible cause of bookmark deletion as well) → sudden crash kills xfs
reassign
Assignee: chanial → asa
Keywords: stackwanted
QA Contact: cpetersen0953 → asa
Whiteboard: TB19636730M
I can confirm this - it's happened several times, but most recently a few
moments ago, using 1.4rc2. I'm on Mandrake 9.1. A few points:

1)I think actually, it may be XFS which is crashing, and killing mozilla!

2)The crash is instant - and without any notice of a segfault. All Moz. windows
just vanished.

3)The crash doesn't seem to be repeatable - returning moments later to the same
site is fine.

4)I have submitted a talkback - I'll give you the id if you can tell me where to
find it.

5)Moz won't restart because xfs is no longer running:
# service xfs status
xfs dead but pid file exists
However, this diagnostic is only available at a console; if the GUI is used to
start mozilla, nothing at all happens (which is hardly helpful!)
To add to my comment #6:

I've done some more checks:

1)There is nothing relevant at all in /var/log/messages or /var/log/syslog, nor
dmesg.

2)This failure cannot be simulated by either
   service xfs stop
or
   kill -9 [PID of xfs process]
while mozilla is running - new tabs open quite happily afterwards (with horrid
fonts).



Re #6)

Got the ID from talkback - TB21351997Y

Also another data point:

1)Start with everything working, mozilla open
2)kill -9 [PID of xfs]
3)Open link in new tab => works fine (with yukky fonts)
4)service xfs start
5)Open link in new tab => works fine (with yukky fonts)
6)Shutdown mozilla. 
7)Try to start mozilla => fails - console output says it can't find any fonts
8)Check xfs is running (it is).
9)Restart KDE; mozilla now works.
I have the same problem, but I do not have talkback installed so I wont be able
to provide debugging info.

Its mostly a hit and miss, however, it happens consistently (every time) I visit
this website:

http://www.upgradeware.com/english/product/xptmc/xptmc.htm

Mozilla logo stops animating, waits about 10 seconds, then mozilla window
closes. If I try to re-open mozilla it wont open.. If open it from an xterm
session i get the following console message:

Gtk-ERROR **: Unable to load default font.
aborting...

If I logout of KDE and try to login again, it cant initialize X.. When I
shutdown, all attempts to shutdown the XFS service fail.

I'm using the stock fonts of mozilla (adobe times/adobe helvetica), however I
*do* have truetype fonts installed but am not using them anymore in mozilla due
to the underline links not underlining bug.

I notice the title of the website uses a multibyte font.. perhaps this is a bug
with mozilla/xfs and multibyte? I did not have this problem with Mozilla 1.2
that came out-of-the-box from RedHat 9.

Build information:
1.4 final
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030624
running on RedHat 9.0
Installed from the installer bin available on the mozilla webpage
Hmm. I just opened up the website that i had mentioned in a previous post in
KDE's integrated browser (it loads O.K.), but the page is gone --  looks like
its a 404 page thats all multibyte.

I got the website as a link from slashdot about some AMD motherboard upgrade and
this site does produce the crash consistently for me regardless of its
now-lack-of-content and

Since theres really no other content on the page since it appears to have been
removed, the only thing thats left on the page is the 404 message in Multibyte??
or maybe the way it handles 404 pages? Just takin a stab in the dark.


This has happened to me *again* (I think it's now happened 4 times - once in
1.3, and thrice in 1.4!)
I can't be sure which dies first - xfs or mozilla. Either way, it's instant. No
Segfault; just vanished. All other programs remain intact.

There's nothing helpful in dmesg, var/log/messages or .xsession_errors.

It's happened to me on various websites, including slashdot and google. The
crash always occurs as a page is loading.

(I've reported this bug before with more details, but I can't find it anywhere -
sorry)
This is quite odd. I just reinstalled 1.2.1 and it doesnt work either. I
compiled 1.4 on my machine and ran it, same problem. But im sure ive went to
websites with (chineese?) font sets in 1.2.1 that comes installed out of the box
with redhat 9..
anyways.

when I ran the 1.4 final that I compiled on my machine and went to the website,
here's the errors that are produced before it crashes:

Document http://www.upgradeware.com/english/product/xptmc/xptmc.htm loaded 
successfully
WARNING: NS_ENSURE_TRUE(focusedShell && doc) failed, file 
nsTypeAheadFind.cpp, line 562
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file 
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432
Document http://www.upgradeware.com/tw/index.htm loaded successfully
Gdk-ERROR **: BadValue (integer parameter out of range for operation)
  serial 31825 error_code 2 request_code 75 minor_code 0

Hopefully someone will look at this.
And again! This is now quite a common cause of crashes for me - and Moz is
normally extremely stable. I'll willingly try to debug this if someone can tell
me how. Incidentally, this time, I didn't get a talkback, and restarting the
dead xfs (service xfs stop; service xfs start) wasn't sufficient - I had to
restart KDE.

Richard
I've finally happened upon a site that will "reliably" create this crash.
http://www.contec.com/products/products.cgi?HTML=DTL&TYPE=0&KATA=PI-32L(PCI)H&BCD1=1&BCD2=4&BCD3=1

Go to the site. Wait about 20 seconds. Moz will stall, then show most of the
page, then quit. No talkback this time. Doing "service xfs status" gives "xfs
dead but pid file exists". Restarting xfs allows Moz to re-open. Re-visitng that
page will crash it again.

I'm using Moz 1.4 on Mandrake 9.1. The page is fine in Konqueror.


I can confirm that this still happens in Moz 1.5. Restarting XFS always seems to
do the trick for me, however.
URL: N/A
Whiteboard: TB19636730M
Re: my comment #14, trying that webpage in Moz 1.6 (mandrake 9.2) does not now
cause a crash.
Product: Browser → Seamonkey
*** This bug has been confirmed by popular vote. ***
Status: UNCONFIRMED → NEW
Ever confirmed: true
Assignee: asa → general
QA Contact: asa → general
MASS-CHANGE:
This bug report is registered in the SeaMonkey product, but has been without a comment since the inception of the SeaMonkey project. This means that it was logged against the old Mozilla suite and we cannot determine that it's still valid for the current SeaMonkey suite. Because of this, we are setting it to an UNCONFIRMED state.

If you can confirm that this report still applies to current SeaMonkey 2.x nightly builds, please set it back to the NEW state along with a comment on how you reproduced it on what Build ID, or if it's an enhancement request, why it's still worth implementing and in what way.
If you can confirm that the report doesn't apply to current SeaMonkey 2.x nightly builds, please set it to the appropriate RESOLVED state (WORKSFORME, INVALID, WONTFIX, or similar).
If no action happens within the next few months, we move this bug report to an EXPIRED state.

Query tag for this change: mass-UNCONFIRM-20090614
Status: NEW → UNCONFIRMED
I cannot reproduce.
Build identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.9) Gecko/20100319 Lightning/1.0b1 SeaMonkey/2.0.4
I suspect this bug can be closed as expired now. XFS is long-gone, and the bug is over 6 years old; I've certainly not seen it for a very very long time.
Whiteboard: bugday0420 [CLOSEME WFM 2010-05-15 see comment #19]
> I suspect this bug can be closed as expired now. XFS is long-gone
Closing as WFM.
Status: UNCONFIRMED → RESOLVED
Closed: 14 years ago
Resolution: --- → WORKSFORME
Whiteboard: bugday0420 [CLOSEME WFM 2010-05-15 see comment #19] → bugday0420
You need to log in before you can comment on or make changes to this bug.