kernel panic related to ReiserFS (v3)
First, the disclaimer: I've been watching the arguments and accusations about reiser4, and am not trying to add fuel to either side ;)
This kernel panic happened a few minutes ago, and it is probably related to a firefox bug, but it shouldn't cause the kernel to panic. I'm running the 2.6.16.16 kernel with the latest squashfs (CVS) patch which supported this kernel series. The distro is Slamd64, and is on a 64-bit AMD X2 system. As you can probably guess, my main filesystem is ReiserFS. It is on an SATA drive. If more information is needed, let me know.
Before sending this in I did a search on the release announcements. It looks to me like the last reiserfs change was well before the .16 series. If I'm wrong, feel free to tell me.
Here is the relevant section from /var/log/syslog:
Jul 17 23:42:49 localhost kernel: Unable to handle kernel paging request
at ffff81023dfaf8b0 RIP:
Jul 17 23:42:49 localhost kernel: <ffffffff8016285b>{cache_alloc_refill
+250}
Jul 17 23:42:49 localhost kernel: PGD 8063 PUD 0
Jul 17 23:42:49 localhost kernel: Oops: 0000 [1] SMP
Jul 17 23:42:49 localhost kernel: CPU 1
Jul 17 23:42:49 localhost kernel: Modules linked in: usb_storage vmnetparport_pc parport vmmon snd_pcm_oss snd_mixer_oss md5 ipv6 ipt_recent ipt_REJECT xt_state iptable_filter nfs lockd nfs_acl sunrpc r8169 ohci1394 ieee1394 emu10k1_gp gameport 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 8250_pci 8250 serial_core tda9887 tuner cx8800 cx88xx video_buf ir_common tveeprom compat_ioctl32 v4l1_compat v4l2_common btcx_risc videodev nvidia forcedeth i2c_nforce2 pcmcia firmware_class yenta_socket rsrc_nonstatic pcmcia_core
Jul 17 23:42:49 localhost kernel: Pid: 7770, comm: firefox-bin Tainted:
P 2.6.16.16 #1
Jul 17 23:42:49 localhost kernel: RIP: 0010:[<ffffffff8016285b>]
<ffffffff8016285b>{cache_alloc_refill+250}
Jul 17 23:42:49 localhost kernel: RSP: 0018:ffff8100b9f7fa78 EFLAGS:
00010082
Jul 17 23:42:49 localhost kernel: RAX: 0000000080000e00 RBX:
ffff810004174940 RCX: 0000000000000016
Jul 17 23:42:49 localhost kernel: RDX: ffff81003e2450c8 RSI:
ffff81003dfac080 RDI: ffff81000417498c
Jul 17 23:42:49 localhost kernel: RBP: 0000000000000004 R08:
0000000000000001 R09: 000000000000000a
Jul 17 23:42:49 localhost kernel: R10: 0000000000000002 R11:
ffff8100b9f7fbc8 R12: ffff810004173800
Jul 17 23:42:49 localhost kernel: R13: ffff8100bfe4abc0 R14:
ffff8100b9f7fb50 R15: ffffffff801adb4c
Jul 17 23:42:49 localhost kernel: FS: 00002af9c9c25ec0(0000)
GS:ffff810004079440(0000) knlGS:00000000f44afbb0
Jul 17 23:42:49 localhost kernel: CS: 0010 DS: 0000 ES: 0000 CR0:
0000000080050033
Jul 17 23:42:49 localhost kernel: CR2: ffff81023dfaf8b0 CR3:
000000004b950000 CR4: 00000000000006e0
Jul 17 23:42:49 localhost kernel: Process firefox-bin (pid: 7770,
threadinfo ffff8100b9f7e000, task ffff81007d4282c0)
Jul 17 23:42:49 localhost kernel: Stack: 0000000000203246
0000000000203296 000000d0911857c0 00000000000000d0
Jul 17 23:42:49 localhost kernel: ffff8100bfe4abc0
ffff8100bf87d800 ffff810003e6cd38 ffff8100b9f7fb50
Jul 17 23:42:49 localhost kernel: ffffffff801adb4c
ffffffff80162756
Jul 17 23:42:49 localhost kernel: Call Trace:
<ffffffff801adb4c>{reiserfs_init_locked_inode+0}
Jul 17 23:42:49 localhost kernel:
<ffffffff80162756>{kmem_cache_alloc+75}
<ffffffff801ae02b>{reiserfs_find_actor+0}
Jul 17 23:42:49 localhost kernel:
<ffffffff801b7131>{reiserfs_alloc_inode+18}
<ffffffff8017b526>{alloc_inode+18}
Jul 17 23:42:49 localhost kernel: <ffffffff8017c3d3>{iget5_locked
+120} <ffffffff801ae07a>{reiserfs_iget+52}
Jul 17 23:42:49 localhost kernel:
<ffffffff801aae31>{reiserfs_lookup+204} <ffffffff80171cb0>{do_lookup
+194}
Jul 17 23:42:49 localhost kernel:
<ffffffff8017281f>{__link_path_walk+2386}
<ffffffff80172cf2>{link_path_walk+80}
Jul 17 23:42:49 localhost kernel:
<ffffffff80165357>{get_unused_fd+108} <ffffffff80172fc8>{do_path_lookup
+604}
Jul 17 23:42:49 localhost kernel:
<ffffffff8017311d>{__path_lookup_intent_open+78}
<ffffffff80173988>{open_namei+127}
Jul 17 23:42:49 localhost kernel: <ffffffff801651f6>{do_filp_open
+25} <ffffffff80165357>{get_unused_fd+108}
Jul 17 23:42:49 localhost kernel: <ffffffff801654e8>{do_sys_open
+68} <ffffffff8010a93e>{system_call+126}
Jul 17 23:42:49 localhost kernel:
Jul 17 23:42:49 localhost kernel: Code: 8b 44 86 30 89 46 24 49 89 54 cc
18 41 8b 45 5c 39 46 20 72
Jul 17 23:42:49 localhost kernel: RIP
<ffffffff8016285b>{cache_alloc_refill+250} RSP <ffff8100b9f7fa78>
Jul 17 23:42:49 localhost kernel: CR2: ffff81023dfaf8b0
Jul 17 23:42:49 localhost kernel: BUG: firefox-bin/7770, lock held at
task exit time!
Jul 17 23:42:49 localhost kernel: [ffff8100a97edd60] {inode_init_once}
Jul 17 23:42:49 localhost kernel: .. held by: firefox-bin: 7770
[ffff81007d4282c0, 116]
Jul 17 23:42:49 localhost kernel: ... acquired at:do_lookup+0x81/0x179
(sorry for the line-wrap, it looks bad even as I click "Send." :)
Thanks
--
Chris Largret <http://www.largret.com>
Chris> This kernel panic happened a few minutes ago, and it is Chris> probably related to a firefox bug, but it shouldn't cause the Chris> kernel to panic. I'm running the 2.6.16.16 kernel with the Chris> latest squashfs (CVS) patch which supported this kernel Chris> series. The distro is Slamd64, and is on a 64-bit AMD X2 Chris> system. As you can probably guess, my main filesystem is Chris> ReiserFS. It is on an SATA drive. If more information is Chris> needed, let me know.
Chris> Jul 17 23:42:49 localhost kernel: Modules linked in: usb_storage vmnet Chris> parport_pc parport vmmon snd_pcm_oss snd_mixer_oss md5 ipv6 ipt_recent Chris> ipt_REJECT xt_state iptable_filter nfs lockd nfs_acl sunrpc r8169 Chris> ohci1394 ieee1394 emu10k1_gp gameport snd_emu10k1 snd_rawmidi Chris> snd_ac97_codec snd_ac97_bus snd_pcm snd_seq_device snd_timer Chris> snd_page_alloc snd_util_mem snd_hwdep snd 8250_pci 8250 serial_core Chris> tda9887 tuner cx8800 cx88xx video_buf ir_common tveeprom compat_ioctl32 Chris> v4l1_compat v4l2_common btcx_risc videodev nvidia forcedeth i2c_nforce2 Chris> pcmcia firmware_class yenta_socket rsrc_nonstatic pcmcia_core Chris> Jul 17 23:42:49 localhost kernel: Pid: 7770, comm: firefox-bin Tainted:
Chris> P 2.6.16.16 #1
You've got a binary kernel module loaded here, please try to re-create this crash without the nvidia module loaded. We (hah! Not me actually... :-) can't debug this with such a module.
The key is the 'Tainted' flag.
John
Quoted 13 lines · John Stoffel
Chris> Jul 17 23:42:49 localhost kernel: Modules linked in:
usb_storagevmnet Chris>parport_pcparport vmmonsnd_pcm_osssnd_mixer_ossmd5 ipv6ipt_recentChris>ipt_REJECTxt_stateiptable_filternfs lockdnfs_aclsunrpc r8169 Chris> ohci1394 ieee1394emu10k1_gpgameportsnd_emu10k1snd_rawmidiChris>snd_ac97_codecsnd_ac97_bussnd_pcmsnd_seq_devicesnd_timerChris>snd_page_allocsnd_util_memsnd_hwdepsnd 8250_pci 8250serial_coreChris> tda9887 tuner cx8800 cx88xxvideo_bufir_commontveepromcompat_ioctl32Chris>v4l1_compatv4l2_commonbtcx_riscvideodev nvidia forcedethi2c_nforce2Chris> pcmciafirmware_classyenta_socketrsrc_nonstaticpcmcia_coreChris> Jul 17 23:42:49 localhost kernel: Pid: 7770, comm: firefox-bin Tainted:
Chris> P 2.6.16.16 #1
Quoted 5 lines
You've got a binary kernel module loaded here, please try to re-create this crash without the nvidia module loaded. We (hah! Not me actually... :-) can't debug this with such a module.
The key is the 'Tainted' flag.
Even though the back-trace doesn't touch the video drivers? I wish I could reproduce this easily. I have been using this kernel and firefox for several weeks, and this is the first time it has panic'd on me.
Thanks, I'll see what I can do.
--
Chris Largret <http://www.largret.com>
Quoted 19 lines · Chris Largret
Quoted 17 lines · John Stoffel
Chris> Jul 17 23:42:49 localhost kernel: Modules linked in:
usb_storagevmnet Chris>parport_pcparport vmmonsnd_pcm_osssnd_mixer_ossmd5 ipv6ipt_recentChris>ipt_REJECTxt_stateiptable_filternfs lockdnfs_aclsunrpc r8169 Chris> ohci1394 ieee1394emu10k1_gpgameportsnd_emu10k1snd_rawmidiChris>snd_ac97_codecsnd_ac97_bussnd_pcmsnd_seq_devicesnd_timerChris>snd_page_allocsnd_util_memsnd_hwdepsnd 8250_pci 8250serial_coreChris> tda9887 tuner cx8800 cx88xxvideo_bufir_commontveepromcompat_ioctl32Chris>v4l1_compatv4l2_commonbtcx_riscvideodev nvidia forcedethi2c_nforce2Chris> pcmciafirmware_classyenta_socketrsrc_nonstaticpcmcia_coreChris> Jul 17 23:42:49 localhost kernel: Pid: 7770, comm: firefox-bin Tainted:
Chris> P 2.6.16.16 #1You've got a binary kernel module loaded here, please try to re-create this crash without the nvidia module loaded. We (hah! Not me actually... :-) can't debug this with such a module.
Quoted 5 lines
Quoted 1 line
The key is the 'Tainted' flag.
Even though the back-trace doesn't touch the video drivers? I wish I could reproduce this easily. I have been using this kernel and firefox for several weeks, and this is the first time it has panic'd on me.
Yes, all kernel modules have full reign on the system so they can overwrite memory not related to them at all. The nVidia could have overwritten some piece of reiserfs's memory and continued on it's own merry way leaving reiserfs to panic the next time it looked there. Personally, I use the nVidia driver all the time and have never had a problem so it's probably not the culprit, but the only way to be sure is to repeat the problem without it ever having been loaded. Booting up and rmmod'ing the module isn't good enough, it has to be a 100% clean boot.
Jim.