assfail during unmount xfs 2.6.24.3 [from 2.6.24.y git]

assfail during unmount xfs 2.6.24.3 [from 2.6.24.y git]

[Chris Knadle]

[-- Attachment #1: Type: text/plain, Size: 4551 bytes --]

During the final unmount before reboot there was an assertion failure from XFS 
leading to the debug output below; I've included what was available on the 
screendump, which I handwrote + transcribed.  [I also have screenshots from a 
2.1 MP camera -- they're slightly ugly but usable.]  CONFIG_KEXEC was not 
compiled in, so I was uanble to get a crashdump via SysRq.  I'm sending this 
along in case someone working on the XFS fs can find + fix a bug.  I'm not 
sending the kernel config in this first post just as not to spam everybody 
with it if it's not needed -- if desired just ask and I'll send it.
Source used: 2.6.24.y from git -- 2.6.24.3.

This is on my Desktop box, x86 system -- single P4 CPU @ 2.6 GHz, IDE disk 
attached to an onboard HighPoint HPT370 controller, and running Debian Sid.

When replying please CC me, as I am not currently subscribed to the list.

Cheers.
-- Chris

-- 

Chris Knadle
Chris.Knadle@coredump.us

-------------------

Cleaning up ifdown....
Deactivating swap...done.
Unmounting local filesystems...done.
Assertion failed: atomic_read(&mp->m_active_trans) == 0, file: 
fs/xfs/xfs_vfsops.c, line: 708
------------[ cut here ]------------
kernel BUG at fs/xfs/support/debug.c:82!
infalid opcode: 0000 [#1]
Modules lined in: nvidia(P) xt_multiport iptable_filter ip_tables x_tables 
ppdev lp ac battery ipv6 ext2 mbcache joydev sidewinder kqemu loop 
snd_emu10k1_synth snd_emux_synth snd_seq_virmidi snd_seq_midi_emul 
snd_emu10k1 firmware_class snd_ac97_codec ac97_bus snd_util_mem snd_hwdep 
parport_pc parport snd_pcm_oss snd_pcm snd_page_alloc snd_mixer_oss psmouse 
rtc snd_deq_dummy pcspkr serio_raw evdev snd_seq_oss snd_seq_midi snd_rawmidi 
snd_seq_midi_event snd_seq snd_timer snd_seq_device emu10k1_gp gameport snd 
soundcore button i2c_i801 i2c_core intel_agp iTCO_wdt agpart shpchp 
pci_hotplug usbhid hid xfs ide_cd cdrom ata_piix ata_generic libata scsi_mod 
ide_disk floppy uhci_hcd usbcore hpt366 e1000 piix generic ide_core thermal 
processor fan

Pid: 25313, comm: mount Tainted: P   (2.6.24.3-686-initrd-crk1 #1)
EIP: 0060:[>f8a67a3f>] EFLAGS: 00010292 CPU: 0
EIP is at assfail+0x1b/0x1f [xfs]
EAX: 00000061 EBX: f7faa400 ECX: ffffffff EDX: c034bda0
ESI: dfc97000 EDI: f7faa400 EBP: db175e14 ESP: db175dcc
 DS: 007b ES: 007b FS: 0000 GS: 033 SS: 0068
Process mount (pid: 25313, ti=db174000 task=f7c32560 task.ti=db174000)
Stack: f8a6cc60 f8a6c6b0 f8a6a135 000002c4 f8a57e38 f7faa400 f8a57ee7 00000000
       dfc97000 f7faa400 db175e14 f8a66ebc 00000001 f8a7fd20 f7fbd400 00000000
       00000001 c0162bd2 00000001 f7fbd43c ffffffff f7fbd400 c0174c1b 00000000
Call Trace:
 [<f8a57e38>] xfs_attr_quiesce+0x46/0x61 [xfs]
 [<f8a57ee7>] xfs_mntupdate+0x84/0xbb [xfs]
 [<f8a66ebc>] xfs_fs_remount+0x3d/0x57 [xfs]
 [<c0162bd2>] do_remount_sb+0xb5/0x101 [xfs]
 [<c0174c1b>] do_mount+0x5a9/0x65b
 [<c0148ec9>] __alloc_pages+0x5f/0x360
 [<c014430c>] find_lock_page+0x15/0x67
 [<c0150dd8>] handle_mm_fault+0x25a/0x52d
 [<c0148ec9>] __alloc_pages+0x5f/0x360
 [<c0173707>] copy_mount_options+0x28/0x113
 [<c0168b45>] getname+0x87/0xaf
 [<c0174d3f>] sys_mount+0x72/0xa4
 [<c0103cf2>] sysenter_past_esp+0x5f/0x85
 =======================
Code: d7 6b c7 83 c3 08 81 fb c8 02 a8 f8 75 ee 5b c3 83 ec 10 89 4c 24 0c 89
 54 24 08 89 44 24 04 c7 04 24 60 cc a6 f8 e8 0d ed 6a c7 <0f> 0b eb fe 56 53
 83 ec 0c 89 c6 83 e6 07 9c 5b fa 89 0c 24 89
EIP: [<f8a67a3f>] assfail+0x1b/0x1f [xfs] SS:ESP 0068:db175dcc
---[ end trace 65bd78ca1bf60304 ]---
WARNING: at kernel/exit.c:917 do_exit()
Pid: 25313, comm: mount Tainted: P         D 2.6.24.3-686-initrd-crk1 #1
 [<c0118ceb>] do_exit+0x669/0x7a4
 [<c0116767>] printk+0x1b/0x1f
 [<c010477a>] do_trap+0x0/0xbd
 [<c0104a56>] do_invalid_op+0x0/0x8a
 [<f8a67a3f>] assfail+0x1b/0x1f [xfs]
 [<f8a57e38>] xfs_attr_quiesce+0x46/0x61 [xfs]
 [<f8a57ee7>] xfs_mntupdate+0x84/0xbb [xfs]
 [<f8a66ebc>] xfs_fs_remount+0x3d/0x57 [xfs]
 [<c0162bd2>] do_remount_sb+0xb5/0x101
 [<c0174c1b>] do_mount+0x5a9/0x65b
 [<c0148ec9>] __alloc_pages+0x5f/0x360
 [<c014430c>] find_lock_page+0x15/0x67
 [<c0150dd8>] handle_mm_fault+0x25a/0x52d
 [<c0148ec9>] __alloc_pages+0x5f/0x360
 [<c0173707>] copy_mount_options+0x28/0x113
 [<c0168b45>] getname+0x87/0xaf
 [<c0174d3f>] sys_mount+0x72/0xa4
 [<c0103cf2>] sysenter_past_esp+0x5f/0x85
========================
/etc/rc6.d/S60umountroot: line17: 25313 Segmentation fault mount 
$MOUNT_FORCE_OPT -n -o remount,ro -t dummytype dummydev / 2>/dev/null

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: assfail during unmount xfs 2.6.24.3 [from 2.6.24.y git]

[David Chinner]

On Thu, Mar 06, 2008 at 11:29:07PM -0500, Chris Knadle wrote:
> During the final unmount before reboot there was an assertion failure from XFS 
> leading to the debug output below; I've included what was available on the 
> screendump, which I handwrote + transcribed.  [I also have screenshots from a 
> 2.1 MP camera -- they're slightly ugly but usable.]  CONFIG_KEXEC was not 
> compiled in, so I was uanble to get a crashdump via SysRq.  I'm sending this 
> along in case someone working on the XFS fs can find + fix a bug.  I'm not 
> sending the kernel config in this first post just as not to spam everybody 
> with it if it's not needed -- if desired just ask and I'll send it.
> Source used: 2.6.24.y from git -- 2.6.24.3.
> 
> This is on my Desktop box, x86 system -- single P4 CPU @ 2.6 GHz, IDE disk 
> attached to an onboard HighPoint HPT370 controller, and running Debian Sid.
> 
> When replying please CC me, as I am not currently subscribed to the list.
......
> Assertion failed: atomic_read(&mp->m_active_trans) == 0, file: 
> fs/xfs/xfs_vfsops.c, line: 708

Known problem. Race in the VFS w.r.t. read-only remounts:

http://marc.info/?l=linux-kernel&m=120106649923499&w=2

The fix for the problem lies outside XFS:

http://marc.info/?l=linux-kernel&m=120109304227035&w=2

Cheers,

Dave.
-- 
Dave Chinner
Principal Engineer
SGI Australian Software Group

Re: assfail during unmount xfs 2.6.24.3 [from 2.6.24.y git]

[Chris Knadle]

[-- Attachment #1: Type: text/plain, Size: 1446 bytes --]

On Friday 07 March 2008, David Chinner wrote:
> On Thu, Mar 06, 2008 at 11:29:07PM -0500, Chris Knadle wrote:
> > During the final unmount before reboot there was an assertion failure
> > from XFS 
...
> > When replying please CC me, as I am not currently subscribed to the list.
...
> > Assertion failed: atomic_read(&mp->m_active_trans) == 0, file:
> > fs/xfs/xfs_vfsops.c, line: 708

   Thank you for taking the time to look and reply to this.

> Known problem. Race in the VFS w.r.t. read-only remounts:
>
> http://marc.info/?l=linux-kernel&m=120106649923499&w=2

   At the bottom of the above message there's a patch which is not currently 
merged in either the linux-2.6.24.y.git or linux-2.6.git trees.  That's 
perhaps as it should be, but I would like to know if it's worth trying 
merging the patch locally as an attempted workaround until such time as 
2.6.25 is released and stable.

> The fix for the problem lies outside XFS:
>
> http://marc.info/?l=linux-kernel&m=120109304227035&w=2

   If I understand the above correctly, it sounds like the per-vfsmount is 
going to be a new feature committed to 2.6.25, but probably not backported to 
prior kernel trees.  If this feature was was added at 2.6.23.9 then if I 
understand corectly it's only an issue between that version and 2.6.25; is 
that correct?

   Thanks again.
   Cheers.

   -- Chris

-- 

Chris Knadle
Chris.Knadle@coredump.us

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: assfail during unmount xfs 2.6.24.3 [from 2.6.24.y git]

[David Chinner]

On Sat, Mar 08, 2008 at 03:55:08PM -0500, Chris Knadle wrote:
> On Friday 07 March 2008, David Chinner wrote:
> > On Thu, Mar 06, 2008 at 11:29:07PM -0500, Chris Knadle wrote:
> > > During the final unmount before reboot there was an assertion failure
> > > from XFS 
> ...
> > > When replying please CC me, as I am not currently subscribed to the list.
> ...
> > > Assertion failed: atomic_read(&mp->m_active_trans) == 0, file:
> > > fs/xfs/xfs_vfsops.c, line: 708
> 
>    Thank you for taking the time to look and reply to this.
> 
> > Known problem. Race in the VFS w.r.t. read-only remounts:
> >
> > http://marc.info/?l=linux-kernel&m=120106649923499&w=2
> 
>    At the bottom of the above message there's a patch which is not currently 
> merged in either the linux-2.6.24.y.git or linux-2.6.git trees.  That's 
> perhaps as it should be, but I would like to know if it's worth trying 
> merging the patch locally as an attempted workaround until such time as 
> 2.6.25 is released and stable.

Sure, the patch is harmless - I've had it running in my local tree for
quite some time now.

> 
> > The fix for the problem lies outside XFS:
> >
> > http://marc.info/?l=linux-kernel&m=120109304227035&w=2
> 
>    If I understand the above correctly, it sounds like the per-vfsmount is 
> going to be a new feature committed to 2.6.25, but probably not backported to 
> prior kernel trees.

Right.

> If this feature was was added at 2.6.23.9 then if I 
> understand corectly it's only an issue between that version and 2.6.25; is 
> that correct?

Only if the per-vfsmount writer counts got merged in 2.6.25-rcX. I'm not sure
that they did. Christoph?

Cheers,

Dave.
-- 
Dave Chinner
Principal Engineer
SGI Australian Software Group

Re: assfail during unmount xfs 2.6.24.3 [from 2.6.24.y git]

[Christoph Hellwig]

On Mon, Mar 10, 2008 at 09:18:58AM +1100, David Chinner wrote:
> Only if the per-vfsmount writer counts got merged in 2.6.25-rcX. I'm not sure
> that they did. Christoph?

No, it didn't go in.  The new target is 2.6.26.