Bug 136005

Summary: udev segfault on boot
Product: [Fedora] Fedora Reporter: Neal Becker <ndbecker2>
Component: udevAssignee: Harald Hoyer <harald>
Status: CLOSED CURRENTRELEASE QA Contact:
Severity: high Docs Contact:
Priority: medium    
Version: 3CC: barryn, bojan, davej, harald, luc.lalonde, michal, oliva
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: udev-039-1 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2006-02-15 07:25:13 EST Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Bug Depends On:    
Bug Blocks: 123268    

Description Neal Becker 2004-10-16 13:34:03 EDT
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (compatible; Konqueror/3.3; Linux) (KHTML, like Gecko)

Description of problem:
After updating to kernel-2.6.8-1.624, system won't boot.  I got:

udev segfault @ 30, rip 424d6c rsp 7fbfffc8e0 error 4


Version-Release number of selected component (if applicable):
kernel-2.6.8-1.624


How reproducible:
Always

Steps to Reproduce:
1.boot
2.
3.
    

Additional info:
Comment 1 Bill Nottingham 2004-10-17 23:15:03 EDT
*** Bug 136083 has been marked as a duplicate of this bug. ***
Comment 2 Michal Jaegermann 2004-10-17 23:53:12 EDT
A segfault reported in bug #136083 was with kernel-2.6.8-1.610
Comment 3 Bojan Smojver 2004-10-18 06:22:22 EDT
*** Bug 136111 has been marked as a duplicate of this bug. ***
Comment 4 Bojan Smojver 2004-10-18 06:24:13 EDT
Just to mention that here, in my case (bug 136111, marked duplicate
here) does not cause systems not to boot. They just behave very
strange (i.e. a lot of stuff doesn't work).
Comment 5 Harald Hoyer 2004-10-18 07:46:53 EDT
*** Bug 136058 has been marked as a duplicate of this bug. ***
Comment 6 Harald Hoyer 2004-10-18 08:07:34 EDT
you have to remake your initrd with mkinitrd!
Comment 7 David Nielsen 2004-10-18 12:56:03 EDT
I updated to 18-10-04, which supposedly fixes this (?), did 
mkinitrd -f /boot/initrd-2.6.8-1.624.img 2.6.8-1.624 
and I still get the udevstart error and it fails to find the lvm2
volumes - rendering my system unbootable, it works perfectly fine on
.541 though.
Comment 8 Harald Hoyer 2004-10-19 04:40:16 EDT
> I updated to 18-10-04
I hope you mean udev-039-1 or udev-039-3
Comment 9 Barry K. Nathan 2004-10-19 05:11:26 EDT
Perhaps he meant a 20041018 rawhide snapshot; that still has udev-038-2.
Comment 10 Harald Hoyer 2004-10-19 16:12:18 EDT
*** Bug 136391 has been marked as a duplicate of this bug. ***
Comment 11 Dave Jones 2004-10-19 23:37:08 EDT
my x86-64 box has around 40 or so segfaults during boot.
udev[10]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff708 error 14
udev[11]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff708 error 14
udev[12]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff708 error 14
udev[14]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff688 error 14
udev[15]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff688 error 14
udev[16]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff688 error 14
udev[17]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff688 error 14
udev[18]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff688 error 14
udev[19]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff688 error 14
udev[20]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff688 error 14
udev[21]: segfault at ffffffffff600400 rip ffffffffff600400 rsp
0000007fbffff688 error 14
....

Using udev-039-3 with what will be the new rawhide kernel as of
tomorrow.  The kernel was installed /after/ the udev upgrade, so there
can't be any 'need to rebuild initrd' type problems.

Comment 12 Bojan Smojver 2004-10-20 00:44:19 EDT
Just an FYI, this particular problem (udevstart failing at boot time)
does go away with udev-0.39-3 after I removed kernel-2.6.8-1.624 and
then put it back on (i.e. the initrd was build by rpm itself).

However, the weird problems that I mentioed (e.g. Evolution can't get
Exchange/IMAP mail) are still there, which would point to this
particular kernel. The 2.6.8-1.610 kernel does all that with no
problems at all.
Comment 13 Harald Hoyer 2004-10-20 10:59:18 EDT
jeremy, please remove the hotplug symlink in the initrd and
echo -n "/sbin/udev" > /proc/sys/kernel/hotplug
instead of
echo -n "/sbin/hotplug" > /proc/sys/kernel/hotplug
Comment 14 Harald Hoyer 2004-10-20 11:05:49 EDT
best solution would be, if nash(hotplug) could spool the hotplug
events somehow and replay them (call /sbin/udev with the correct
environment variables), after /dev is mounted :)
Comment 15 Harald Hoyer 2004-10-20 11:07:12 EDT
even better would be, if nash could save the hotplug events in /dev
and the replay is done after ROOT is mounted..
Comment 16 Jeremy Katz 2004-10-20 11:22:48 EDT
gregkh was explicitly against setting the hotplug handler to be udev
when I talked to him at OLS...
Comment 17 Harald Hoyer 2004-10-20 13:23:27 EDT
davej, please retry with udev-093-4
Comment 18 Harald Hoyer 2004-10-20 13:27:15 EDT
udev segfaulted because of openlog(3) called in initramfs with an
empty /dev, so I disabled logging in udev-093-4
Comment 19 Michal Jaegermann 2004-10-21 16:32:44 EDT
Just updated to kernel-2.6.9-1.640 and udev-039-3 (that what is
available as of today) and on the top of 28 segfaults :-) I got
also
.....
Probing IDE interface ide0...
request_module: runaway loop modprobe net-pf-1
request_module: runaway loop modprobe net-pf-1
request_module: runaway loop modprobe net-pf-1
request_module: runaway loop modprobe net-pf-1
request_module: runaway loop modprobe net-pf-1
Probing IDE interface ide1...
.....

which seem to be coming from udev too.  Or this is initrd?

After all this excitment the machine started.
Comment 20 Harald Hoyer 2004-10-22 04:17:06 EDT
*** Bug 136745 has been marked as a duplicate of this bug. ***
Comment 21 Harald Hoyer 2004-10-22 04:19:23 EDT
Michal, please read comment #17 .. udev >= 039-4 should fix the issue
with the segfaults.
For the runaway loops, please file a seperate bugzilla for mkinitrd.
Comment 22 Neal Becker 2004-10-22 11:00:15 EDT
udev-0.39-6 seems to fix it for me 
Comment 23 Michal Jaegermann 2004-10-22 11:47:44 EDT
> Michal, please read comment #17 .. udev >= 039-4 should fix the issue
As you can read in comment #19 this was mentioned only as an aside.
Still twenty eight segfaults is an impressive number. :-)

> For the runaway loops, please file a seperate bugzilla for mkinitrd.
Will do. I was not sure where is the source of that problem.
Comment 24 Nadav Kavalerchik 2005-05-10 08:32:15 EDT
i'm getting the same "/sbin/udevstart exited abnormally" error on boot. 
using : "2.6.11-1.1287_FC4" kernel and "udev-057-4" udev rpm. 
Comment 25 Harald Hoyer 2005-05-10 12:08:06 EDT
Nadav, use "udev-057-5" rebuild your initrd with mkinitrd