Re: reiser4 bug

[Alexander Zarochentsev <zam@namesys.com>]

On Tuesday 11 April 2006 09:28, Matt Eaton wrote:
> I was running OpenOffice and tried saving a file.
>
> I'm using 2.6.16.1 + reiser4-for-2.6.16-1.patch.gz
>
> (Please help! I'm having to run openoffice on a different filesystem
> as this crash has occurred twice saving the same file - after a
> reboot)
>
> kernel BUG at fs/reiser4/plugin/file/tail_conversion.c:29!
> invalid opcode: 0000 [#1]

it is known problem. Fixed in 2.6.17-rc1-mm2 
(reiser4-have-get_exclusive_access-restart-transaction.patch).

I am afraid the patch from mm would not apply cleanly to 2.6.16.1 + 
reiser4-for-2.6.16-1.patch, 
you may try to apply only the following hunk:

@@ -20,13 +20,12 @@ void get_exclusive_access(unix_file_info
        assert("nikita-3047", LOCK_CNT_NIL(inode_sem_w));
        assert("nikita-3048", LOCK_CNT_NIL(inode_sem_r));
        /*
-        * "deadlock detection": sometimes we commit a transaction under
+        * "deadlock avoidance": sometimes we commit a transaction under
         * rw-semaphore on a file. Such commit can deadlock with another
         * thread that captured some block (hence preventing atom from 
being
         * committed) and waits on rw-semaphore.
         */
-       assert("nikita-3361", get_current_context()->trans->atom == 
NULL);
-       BUG_ON(get_current_context()->trans->atom != NULL);
+       txn_restart_current();
        LOCK_CNT_INC(inode_sem_w);
        down_write(&uf_info->latch);
        uf_info->exclusive_use = 1;
_


> PREEMPT SMP
> Modules linked in: eeprom lm85 hwmon_vid i2c_i801 lp vmnet parport_pc
> parport vmmon nfs lockd sunrpc snd_seq_midi snd_emu10k1_synth
> snd_emux_synth snd_seq_virmidi snd_seq_midi_emul snd_pcm_oss
> snd_mixer_oss snd_seq_oss snd_seq_midi_event snd_seq snd_hda_intel
> snd_hda_codec snd_emu10k1 snd_rawmidi snd_ac97_codec snd_ac97_bus
> snd_pcm snd_seq_device snd_timer snd_page_alloc snd_util_mem
> snd_hwdep snd soundcore binfmt_misc ntfs usblp ide_cd cdrom nvidia
> e100 mii uhci_hcd ehci_hcdCPU:    0
> EIP:    0060:[<c01c623c>]    Tainted: P      VLI
> EFLAGS: 00210282   (2.6.16.1 #1)
> EIP is at get_exclusive_access+0x1c/0x3d
> eax: df21b5dc   ebx: 00000001   ecx: df21b654   edx: c50e17d4
> esi: acb2e000   edi: c50e1780   ebp: 00000000   esp: e4481f30
> ds: 007b   es: 007b   ss: 0068
> Process soffice.bin (pid: 21440, threadinfo=e4480000 task=dc5bfa90)
> Stack: <0>c01c56f4 df21b5dc c013ff95 c253a7ac dda673c0 00000001
> 000037c2 00000000
>        df21b5dc df21b654 5058c680 e32c4680 acb2e000 e4481fa4 000037c2
> c014c900
>        e32c4680 acb2e000 000037c2 e4481fa4 e32c4680 fffffff7 bf823d43
> e4480000
> Call Trace:
>  [<c01c56f4>] write_unix_file+0x2a6/0x45c
>  [<c013ff95>] vma_link+0xbe/0xc5
>  [<c014c900>] vfs_write+0x87/0x11b
>  [<c014ca32>] sys_write+0x3b/0x63
>  [<c010274b>] sysenter_past_esp+0x54/0x75
> Code: 8d 43 14 e8 0b 38 13 00 e9 d1 fc ff ff 90 90 ba 00 e0 ff ff 21
> e2 8b 12 8b 92 c4 04 00 00 8b 44 24 04 8b 52 50 83 7a 10 00 74 08
> <0f> 0b 1d 00 60 b8 31 c0 ba 01 00 ff ff f0 0f c1 10 85 d2 0f 85
>
>
>
>
> !DSPAM:443b3e85155598791294130!

-- 
Alex.