kernel crash on mounting / parition (xfs File System) under all kernel since 2.6.15-1.1826.2.5_FC5 (this one is not affected). I've just tested kernel-2.6.15-1.1871_FC5 and crash already occurs. How can I give you some trace of the kernel crash? I've got a message saying automatic reboot will occur in 120 secondes or something like that.
there should be a bunch of text before the countdown, which is the actual panic data. XFS problems are better reported upstream however, as we won't spend any time actively fixing them in Fedora, but rather inherit them from upstream fixes. The XFS maintainers can be found at xfs-masters.com (For their info, kernel-2.6.15-1.1871_FC5 is based on upstream 2.6.16-rc1-git4, with no additional XFS patches).
Created attachment 123866 [details] crash xfs
OK. I'll post on sgi bugzilla. Nevertheless, this is the crash in JPEG. Thx
Thanks, we've got this bug over at http://oss.sgi.com/bugzilla/show_bug.cgi?id=472 now.
Just FYI for the RH guys: I put 2.6.15-1.1881_FC5 on my x86_64 box running RHEL4, and ran our latest xfstests test suite against it, with the latest xfsprogs. Things generally work, most tests are passing, but certainly nothing oopsed in the many many filesystem mounts that happen during the tests. We're going to need more info from the oops before we'd be able to make any progress on this. p.s. could we get the oss.sgi.com bugzilla entry in the external bugzilla reference list...?
The patch at http://oss.sgi.com/cgi-bin/cvsweb.cgi/xfs-linux/linux-2.6/xfs_buf.c.diff?r1=1.218;r2=1.219 should address this.
I confirm, this patch fix this oops. Fedora guys, could you add this little patch ? Thx
I'd rather you get it upstream. FC5test is still rebasing to Linus' tree on a daily basis, and this sounds like definite 2.6.16rc3 material.