Hi Andrew,
I seem to be running into truncate issues on 2.6.10-rc1-mm2.
Kernel compile hangs. Here is the sysrq-t output for the
process. Do you know ?
cc1 D 000053567a8130b2 0 16244 1 20486 12881
(NOTLB)
00000101a5a09d98 0000000000000006 0000000100000001 000001017b121870
0000000000076a00 00000101dfff7030 000001017b121b08
0000000000000001
000053567a81808b ffffffff80198db9
Call Trace:<ffffffff80198db9>{__d_lookup+297}
<ffffffff803fcb30>{__down_write+128}
<ffffffff8017f8a7>{do_truncate+71}
<ffffffff80198db9>{__d_lookup+297}
<ffffffff8018e2ff>{get_write_access+79}
<ffffffff8018fd7f>{may_open+511}
<ffffffff8019151f>{open_namei+815}
<ffffffff8017ebd7>{filp_open+39}
<ffffffff8017e994>{get_unused_fd+244}
<ffffffff8017ec4c>{sys_open+76}
<ffffffff801106ae>{system_call+126}
BTW, the compile is hung, so I tried killing it and did make
again. So you see 2 "cc"s. second one is waiting for the first
one to finish up.
Thanks,
Badari
Badari Pulavarty <[email protected]> wrote:
>
> I seem to be running into truncate issues on 2.6.10-rc1-mm2.
> Kernel compile hangs. Here is the sysrq-t output for the
> process. Do you know ?
Beats me. If we can find a workload which reproduces it in a sane amount
of time I can do a binary search.
On Wed, Nov 03, 2004 at 02:29:28PM -0800, Andrew Morton wrote:
> Badari Pulavarty <[email protected]> wrote:
> >
> > I seem to be running into truncate issues on 2.6.10-rc1-mm2.
> > Kernel compile hangs. Here is the sysrq-t output for the
> > process. Do you know ?
>
> Beats me. If we can find a workload which reproduces it in a sane amount
> of time I can do a binary search.
I have also had a report of this on a very recent SLES9 updates
kernel, from looking through the patches the SUSE folks have and
those in the -mm, the invalidate_inode_pages2 patch is in both
so it could be the cause, but thats just a guess so far (patch
name in -mm is invalidate_inode_pages-mmap-coherency-fix.patch).
I don't have a reproducible test case yet either.
cheers.
--
Nathan
I don't have a reproducible case either. Just randomly happens,
once in a while - requiring a force reboot of the machine.
Thanks,
Badari
On Wed, 2004-11-03 at 15:28, Nathan Scott wrote:
> On Wed, Nov 03, 2004 at 02:29:28PM -0800, Andrew Morton wrote:
> > Badari Pulavarty <[email protected]> wrote:
> > >
> > > I seem to be running into truncate issues on 2.6.10-rc1-mm2.
> > > Kernel compile hangs. Here is the sysrq-t output for the
> > > process. Do you know ?
> >
> > Beats me. If we can find a workload which reproduces it in a sane amount
> > of time I can do a binary search.
>
> I have also had a report of this on a very recent SLES9 updates
> kernel, from looking through the patches the SUSE folks have and
> those in the -mm, the invalidate_inode_pages2 patch is in both
> so it could be the cause, but thats just a guess so far (patch
> name in -mm is invalidate_inode_pages-mmap-coherency-fix.patch).
>
> I don't have a reproducible test case yet either.
>
> cheers.