Re: [2.6.11-mm3] umount: Scheduling while atomic

[Jesper Juhl]

On Fri, 18 Mar 2005, Andrew Morton wrote:

> Alexander Gran <alex@xxxxxxxxxxxxxxxxxxx> wrote:
> >
> > while umounting an ext2 partition on a usb hdd I'm getting:
> >  scheduling while atomic: umount/0x10000001/14941
> >   [<c0451392>] schedule+0x5f2/0x600
> >   [<c0451cc7>] cond_resched+0x27/0x40
> >   [<c0140af1>] invalidate_mapping_pages+0x81/0xe0
> >   [<c015b27d>] kill_bdev+0xd/0x20
> >   [<c015b315>] __set_blocksize+0x85/0xa0
> >   [<c015bba0>] __bd_release+0x70/0x80
> >   [<c015c458>] __close_bdev_excl+0x8/0x10
> >   [<c015a100>] deactivate_super+0x50/0x80
> >   [<c016f82b>] sys_umount+0x3b/0x90
> >   [<c0148c20>] do_munmap+0x120/0x150
> >   [<c016f895>] sys_oldumount+0x15/0x20
> >   [<c010300b>] sysenter_past_esp+0x54/0x75
> 
> hm, yes, that was a bug in blockdev-fixes-race-between-mount-umount.patch,
> but that patch got dropped because Linus fixed things differently.
> 
I just got a similar one when shutting down my box with 2.6.11-mm4, 
rebooted it a few times, and it seems to happen 1 in 4 times.  last few 
messages during shutdown are here :  

Turning off swap.
Unmounting local file systems.
Scheduling while atomic: umount/0x10000001/12240
[<c0103997>] dump_stack+0x17/0x20
[<c02dd927>] schedule+0x587/0x650
[<c02de3fa>] cond_resched+0x2a/0x50
[<c014283a>] invalidate_mapping_pages+0xda/0xe0
[<c015d7e0>] kill_bdev+0x10/0x30
[<c015d885>] __set_blocksize+0x85/0xb0
[<c015e1c2>] __bd_release+0x62/0x70
[<c015eaeb>] __close_bdev_excl+0xb/0x14
[<c015c487>] deactivate_super+0x67/0x90
[<c01730a7>] sys_umount+0x37/0x80
[<c0173109>] sys_oldumount+0x19/0x20
[<c0102b15>] syscall_call+0x7/0xb
Remounting root filesystem read-only.
Synchronizing SCSI cache for disk sda:
Power down.


Now I'm of to try and see if 2.6.12-rc1-mm1 still does this.


-- 
Jesper Juhl


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/