Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 104439 - kernel panic with 2.6.12-r9 Pre-empt configured
Summary: kernel panic with 2.6.12-r9 Pre-empt configured
Status: RESOLVED NEEDINFO
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Core system (show other bugs)
Hardware: x86 Linux
: Highest critical
Assignee: Gentoo Kernel Bug Wranglers and Kernel Maintainers
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2005-08-31 18:11 UTC by M. Edward Borasky
Modified: 2005-09-17 10:19 UTC (History)
0 users

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description M. Edward Borasky 2005-08-31 18:11:46 UTC
I have been encountering a crash while attempting to emerge "rmagick" on a 512
MB Athlon TBird running the 2.6.12-r9 kernel. This first started happening in X
windows and showed up as a hard lockup, so I rebooted the system without xdm and
ran it from the console. I got an OOPS and a kernel panic labelled PRE-EMPT.

I had the kernel configured with pre-empt *and* pre-empt the big kernel lock.
This appears to be reproducible; something in the Ruby processing of rmagick
triggers this. I have rebuilt the kernel without those configuration options and
will attempt the emerge of rmagick again. I think it will work; it's worked on
my other two Gentoo machines.

I can try to track this down further, get all the stack trace data, etc., but I
wanted to check with the kernel maintainers to see if this was a known upstream
issue before I spend any more time on it. I have some free time over the weekend
to go back to this, now that I know what it is and how to trigger it reproducibly.
Comment 1 Daniel Drake (RETIRED) gentoo-dev 2005-09-01 06:20:02 UTC
Its hard to say with the little info you have provided. Oops messages say
PREEMPT just to indicate that it is enabled, not to say that it is a
PREEMPT-related problem (but it might be). We really need to see the oops
message. Also you should try to reproduce this on gentoo-sources-2.6.13
Comment 2 M. Edward Borasky 2005-09-01 06:52:54 UTC
I now know how to reproduce this ... I think. It turns out that what was
happening is that the rmagick (1.3.2) ebuild was trying to write a huge file to
disk and overflowed /var/tmp/portage (and hence /), causing the kernel to crash
every time. So this is definitely a case of the kernel not being able to defend
itself against a stupid user who tried to write more than a filesystem could
handle. Once I cleaned up /var/tmp/portage/ rmagick emerged just fine.

I don't know enough about the kernel to know whether the upstream folks know or
care about this condition, but if there's some way to capture the stack trace,
etc. for them, I'll go ahead and fill up /var/tmp/portage again, emerge rmagick
again and ship the data to you all. Otherwise, you can close this as "user error".
Comment 3 Daniel Drake (RETIRED) gentoo-dev 2005-09-01 12:02:50 UTC
You can either capture it over serial console, over netconsole, with a digital
camera, or with pen-and-paper. Make sure you are running gentoo-sources-2.6.13
too, and give us more info (which filesystem is in use? emerge --info? etc)
Comment 4 Daniel Drake (RETIRED) gentoo-dev 2005-09-17 10:19:59 UTC
Please reopen when you have time to provide us with more info (see comment #3)