On Sun, Jul 08, 2007 at 12:09:11PM +0200, Markus Trippelsdorf wrote:
> On Sun, Jul 08, 2007 at 01:20:16AM -0700, Andrew Morton wrote:
> > On Sun, 8 Jul 2007 07:14:52 +0200 Markus Trippelsdorf <[EMAIL PROTECTED]>
> > wrote:
> >
> > ...
> > > touch src/.depend # to prevent unecessary warnings
> > > make: *** [dep] Killed
> >
> > ug. nd.dentry.d_inode (as set up by do_utimes()) is garbage. I don't know
> > what could have caused that.
> >
> >
> > > Running "touch
> > > /var/tmp/paludis/sys-libs/gpm-1.20.1-r6/work/gpm-1.20.1/src/.depend"
> > > triggers this oops.
> >
> > What type of filesystem is at /var/tmp?
> >
> > Is it repeatable after a reboot?
> >
>
> The filesystem at /var/tmp was xfs. And yes it was repeatable after a
> reboot...
>
> So I ran xfs_check on that partition and it reported errors. Then I ran
> xfs_repair and after that rebooted, but the problem was still there.
> Finally I switched the partition back to ext3, because it got too scary.
> But the oops is now repeatable even on this new ext3 partition.
I tested this further and it turned out that the Linus tree is also
affected. So I ran git-bisect, after I found out that version
2.6.21.6 was not affected by this bug.
This is the result:
gentoox2 linux # git bisect bad
1c710c896eb461895d3c399e15bb5f20b39c9073 is first bad commit
commit 1c710c896eb461895d3c399e15bb5f20b39c9073
Author: Ulrich Drepper <[EMAIL PROTECTED]>
Date: Tue May 8 00:33:25 2007 -0700
utimensat implementation
Implement utimensat(2) which is an extension to futimesat(2) in that it
a) supports nano-second resolution for the timestamps
b) allows to selectively ignore the atime/mtime value
c) allows to selectively use the current time for either atime or mtime
d) supports changing the atime/mtime of a symlink itself along the lines
of the BSD lutimes(3) functions
For this change the internally used do_utimes() functions was changed to
accept a timespec time value and an additional flags parameter.
Additionally the sys_utime function was changed to match compat_sys_utime
which already use do_utimes instead of duplicating the work.
Also, the completely missing futimensat() functionality is added. We have
such a function in glibc but we have to resort to using /proc/self/fd/*
which
not everybody likes (chroot etc).
Test application (the syscall number will need per-arch editing):
...
[EMAIL PROTECTED]: add missing i386 syscall table entry]
Signed-off-by: Ulrich Drepper <[EMAIL PROTECTED]>
Cc: Alexey Dobriyan <[EMAIL PROTECTED]>
Cc: Michael Kerrisk <[EMAIL PROTECTED]>
Cc: <[EMAIL PROTECTED]>
Signed-off-by: Andrew Morton <[EMAIL PROTECTED]>
Signed-off-by: Linus Torvalds <[EMAIL PROTECTED]>
HTH,
--
Markus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
