On Thursday 29 May 2008 18:15:26 Ondrej Zary wrote:
quoted text
> On Thursday 29 May 2008 08:27:11 Ondrej Zary wrote:
> > On Thursday 29 May 2008, Andrew Morton wrote:
> > > On Thu, 29 May 2008 00:09:36 +0200
> > >
> > > "Rafael J. Wysocki" <rjw@sisk.pl> wrote:
> > > > On Wednesday, 28 of May 2008, Ondrej Zary wrote:
> > > > > Hello,
> > > > > I'm using hibernation on my desktop machine every day instead of
> > > > > power off. It mostly works but sometimes aborts with "no space left
> > > > > on device" error. Closing some programs and trying again usually
> > > > > fixes it - but recently, I got two oopses instead. I'm sending them
> > > > > because they're the same, only some details are different. Does
> > > > > anyone know what might be wrong?
> > > >
> > > > Thanks for the report, but I have no idea of what could go wrong.
> > > >
> > > > Rafael
> > > >
> > > > > ------------[ cut here ]------------
> > > > > Kernel BUG at c015610b [verbose debug info unavailable]
> > >
> > > Looks like this is
> > >
> > > 	BUG_ON(inode->i_state == I_CLEAR);
> > >
> > > Please do enable CONFIG_DEBUG_BUGVERBOSE.  Turning off this stuff
> > > doesn't gain much.
> > >
> > > > > invalid opcode: 0000 [#1]
> > > > > Modules linked in: snd_sb16 ppdev snd_opl3_synth snd_seq_midi_emul
> > > > > snd_opl3_lib snd_hwdep snd_sb16_dsp snd_sb_common snd_mpu401_uart
> > > > > snd_rawmidi 3c509 de2104x sr_mod cdrom [last unloaded: snd_sb16]
> > > > >
> > > > > Pid: 8634, comm: bash Not tainted (2.6.25.3-pentium #3)
> > > > > EIP: 0060:[<c015610b>] EFLAGS: 00210246 CPU: 0
> > > > > EIP is at iput+0x19/0x61
> > > > > EAX: c02db808 EBX: cf402c08 ECX: 0001ec9e EDX: 00000000
> > > > > ESI: cf402ba0 EDI: cf987c00 EBP: 00000000 ESP: c9a4deb4
> > > > >  DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068
> > > > > Process bash (pid: 8634, ti=c9a4c000 task=c3975000
> > > > > task.ti=c9a4c000) Stack: 00000000 c0163eb5 cf402bac ffffffe4
> > > > > c142b000 c0329881 cf80b6a0 c012e212 00000001 c38de7ac 00200286
> > > > > 00001000 00000000 00000001 00000000 00000000 c142b000 00000000
> > > > > 00000000 0001ec82 0001ec82 00200246 0001ec82 fffffa5d Call Trace:
> > > > >  [<c0163eb5>] __blkdev_put+0xc2/0xda
> > > > >  [<c012e212>] swsusp_write+0x307/0x311
> > > > >  [<c012c81b>] hibernate+0xb4/0x131
> > > > >  [<c012b9ad>] state_store+0x41/0xa3
> > > > >  [<c012b96c>] state_store+0x0/0xa3
> > > > >  [<c01babcb>] kobj_attr_store+0x18/0x1c
> > > > >  [<c0173dd0>] sysfs_write_file+0xab/0xd8
> > > > >  [<c0173d25>] sysfs_write_file+0x0/0xd8
> > > > >  [<c0148a58>] vfs_write+0x7f/0xec
> > > > >  [<c0148e9c>] sys_write+0x3c/0x63
> > > > >  [<c01039d2>] syscall_call+0x7/0xb
> > > > >  [<c02d0000>] i8042_probe+0x4c4/0x4db
> > > > >  =======================
> > > > > Code: 08 01 00 00 77 ff ff ff eb e5 e8 90 ad 17 00 31 c0 c3 53 85
> > > > > c0 89 c3 74 58 8b 80 8c 00 00 00 83 bb 08 01 00 00 40 8b 40 20 75
> > > > > 04 <0f> 0b eb fe 85 c0 74 0b 8b 50 10 85 d2 74 04 89 d8 ff d2 8d 43
> > > > > EIP: [<c015610b>] iput+0x19/0x61 SS:ESP 0068:c9a4deb4
> > >
> > > Beats me.  Somehow the swap device's blockdev inode got I_CLEAR set
> > > while swsusp_write() was playing with it.  Or during.
> > >
> > > I guess we could add I_CLEAR checks on resume_bdev into
> > > kernel/power/swap.c in various places.
> > >
> > > Had there been any swapoffs before or during this suspend?
> >
> > Yes, you're right. I almost forgot that. I have two swaps - one 256MB
> > swap partition (the machine has 256MB RAM) and one 128MB swap file. The
> > silly thing is that the partition fills up first while the swapfile
> > remains empty (ok, it's possible to change the priority - but haven't
> > tried that yet). So when the hibernation failed, I tried to free the swap
> > partition using swapoff and swapon. Sometimes, it crashes during the
> > swapoff - like yesterday (will post later).
>
> OK, here's the other crash. I was wrong, it was during swapon, not swapoff.
> I'll try to reproduce both of them.
>
> PM: Not enough free swap
> Restarting tasks ... done.
> ata2.01: configured for PIO2
> BUG: unable to handle kernel NULL pointer dereference at 00000030
> IP: [<c0163812>] bd_claim+0x14/0x3e
> *pde = 00000000
> Oops: 0000 [#1]
> Modules linked in: usbhid snd_opl3_lib snd_hwdep snd_sb16_dsp snd_sb_common
> snd_mpu401_uart snd_rawmidi 3c509 ppdev sr_mod cdrom de2104x [last
> unloaded: snd_sb16]
>
> Pid: 10517, comm: swapon Not tainted (2.6.25.3-pentium #3)
> EIP: 0060:[<c0153812>] EFLAGS: 00210207 CPU: 0
> EIP is at bd_claim+0x14/0x3e
> EAX: 00000000 EBX: cf402ca0 ECX: cf402ba0 EDX: c0142178
> ESI: 00000000 EDI: 00000000 EBP: c03daaac ESP: c7095f48
>  DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068
> Process swapon (pid: 10517, ti=c7094000 task=c201a550 task.ti=c7094000)
> Stack: c0142308 c013b28c 00000000 000000cf 00000000 c70e7000 cf402ba0
> cd825340 cf435b40 0000095c c71cab7c 00000001 cf402c08 b7e57180 c201a550
> 00000004 c010c2d4 00000000 cf435b74 c7095fb8 cf435b40 00000000 1c9b5d63
> 0860d648 Call Trace:
>  [<c0142308>] sys_swapon+0x190/0x7be
>  [<c013b28c>] handle_mm_fault+0x202/0x438
>  [<c010c2d4>] do_page_fault+0x205/0x559
>  [<c01039d2>] syscall_call+0x7/0xb
>  =======================
> Code: 03 48 28 8d 42 a8 8b 50 58 8d 74 26 00 3d 34 67 37 c0 75 e3 89 c8 c3
> 89 c1 8b 40 30 39 d0 74 19 85 c0 75 2b 8b 41 40 39 c8 74 0e <8b> 40 30 3d
> fe 37 16 c0 74 04 85 c0 75 16 8b 41 40 ff 40 34 c7
> EIP: [<c0163812>] bd_claim+0x15/0x3e SS:ESP 0068:c7095f48
> ---[ end trace 24e66613592015ef ]---

It crashed again yesterday - during swapoff after failed hibernation. Have no 
dump as it completely crashed after entering "dmesg".
Although there were almost no applications running (KDE, Konsole and Pidgin), 
there was not enough swap space available, while I can sometimes hibernate 
fine with KMail and Konqueror running too. I suspect that something is 
leaking memory here. Don't know if the (attached) output from /proc/slabinfo 
and /proc/vmstat is enough to tell.


-- 
Ondrej Zary