Skip to contentSkip to conversation

2.6.18-1.2689.fc6PAE: oops in ext3_clear_inode+0x52/0x8b

DISCUSSION Jeremy Fitzhardinge 5 participants 10 messages Sep 28, 2006 – Oct 3, 2006

lore
Share & tools
lore
#0 lore raw

I just filed this in the Redhat bugzilla, since its from the FC6 distro kernel. But since its fairly close to current kernel.org kernels, I thought it might be relevent.

The bug is https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=208488

Unfortunately this isn't a very useful report since it was a once-off, and there's a P-tainting module in there. But if anyone sees anything else like this, it's interesting.

The oops is:

BUG: unable to handle kernel paging request at virtual address 756e6547
printing eip:
f898bf73
*pde = 0014c7c0
Oops: 0002 [#1]
SMP last sysfs file: /devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
Modules linked in: ath_pci(U) usb_storage loop wlan_wep(U) snd_hda_codec nfsd exportfs lockd nfs_acl tun ppp_deflate zlib_deflate ppp_async crc_ccitt ppp_generic slhc airprime usbserial hci_usb cpufreq_powersave i915 drm cpufreq_conservative ipv6 autofs4 hidp rfcomm l2cap bluetooth sunrpc ipt_REJECT xt_state ip_conntrack nfnetlink xt_tcpudp iptable_filter ip_tables x_tables vfat fat dm_mirror dm_mod video sbs ibm_acpi i2c_ec dock button battery asus_acpi ac parport_pc lp parport snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq wlan_scan_sta(U) snd_seq_device snd_pcm_oss snd_mixer_oss snd_pcm ath_rate_sample(U) mmc_block snd_timer snd sg ohci1394 i2c_i801 wlan(U) ieee1394 sdhci soundcore pcspkr i2c_core snd_page_alloc serio_raw e1000 mmc_core ath_hal(U) ahci libata sd_mod scsi_mod ext3 jbd ehci_hcd ohci_hcd uhci_hcd
CPU:    0
EIP:    0060:[<f898bf73>]    Tainted: P      VLI
EFLAGS: 00013206   (2.6.18-1.2689.fc6PAE #1) EIP is at ext3_clear_inode+0x52/0x8b [ext3]
eax: c07dda80   ebx: c07dd994   ecx: 00000000   edx: 756e6547
esi: c07dda80   edi: 00000000   ebp: f3f0cda0   esp: f3f0cd94
ds: 007b   es: 007b   ss: 0068
Process Xorg (pid: 2862, ti=f3f0c000 task=f3cf4030 task.ti=f3f0c000)
Stack: c07dda80 c07ddc9c 0000001f f3f0cdb0 c04881e5 c07dda80 c07dda88 f3f0cdc8 
      c04884ba f3f0cdd8 c0bf3db0 00000000 00000080 f3f0cdec c04886c6 00000080 
      00000080 c07dddb8 f32d1428 0000620c f7ff62fc 000000a9 f3f0ce1c c045ba2d 
Call Trace:
[<c04881e5>] clear_inode+0xd8/0x129
[<c04884ba>] dispose_list+0x3c/0xc1
[<c04886c6>] shrink_icache_memory+0x187/0x1af
[<c045ba2d>] shrink_slab+0xd0/0x137
[<c045c285>] try_to_free_pages+0x159/0x218
[<c045860d>] __alloc_pages+0x199/0x287
[<c045f3b8>] __handle_mm_fault+0x1b2/0xb84
[<c0614a5a>] do_page_fault+0x2b1/0x5ba
[<c0404be9>] error_code+0x39/0x40
DWARF2 unwinder stuck at error_code+0x39/0x40
Leftover inexact backtrace:
[<c040537f>] show_stack_log_lvl+0x8a/0x95
[<c04054b7>] show_registers+0x12d/0x19a
[<c04056b4>] die+0x190/0x293
[<c0614c91>] do_page_fault+0x4e8/0x5ba
[<c0404be9>] error_code+0x39/0x40
[<c04881e5>] clear_inode+0xd8/0x129
[<c04884ba>] dispose_list+0x3c/0xc1
[<c04886c6>] shrink_icache_memory+0x187/0x1af
[<c045ba2d>] shrink_slab+0xd0/0x137
[<c045c285>] try_to_free_pages+0x159/0x218
[<c045860d>] __alloc_pages+0x199/0x287
[<c045f3b8>] __handle_mm_fault+0x1b2/0xb84
[<c0614a5a>] do_page_fault+0x2b1/0x5ba
[<c0404be9>] error_code+0x39/0x40
Code: 0f 94 c0 84 c0 74 07 89 d0 e8 c4 0e ae c7 c7 83 90 00 00 00 ff ff ff ff 8b 93 94 00 00 00 85 d2 74 24 83 fa ff 74 1f 85 d2 74 11 <f0> ff 0a 0f 94 c0 84 c0 74 07 89 d0 e8 96 0e ae c7 c7 83 94 00 EIP: [<f898bf73>] ext3_clear_inode+0x52/0x8b [ext3] SS:ESP 0068:f3f0cd94
#1 lore raw
Quoted 18 lines · Jeremy Fitzhardinge
BUG: unable to handle kernel paging request at virtual address 756e6547
printing eip:
f898bf73
*pde = 0014c7c0
Oops: 0002 [#1]
SMP last sysfs file: /devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
Modules linked in: ath_pci(U) usb_storage loop wlan_wep(U) snd_hda_codec nfsd

exportfs lockd nfs_acl tun ppp_deflate zlib_deflate ppp_async crc_ccitt ppp_generic slhc airprime usbserial hci_usb cpufreq_powersave i915 drm cpufreq_conservative ipv6 autofs4 hidp rfcomm l2cap bluetooth sunrpc ipt_REJECT xt_state ip_conntrack nfnetlink xt_tcpudp iptable_filter ip_tables x_tables vfat fat dm_mirror dm_mod video sbs ibm_acpi i2c_ec dock button battery asus_acpi ac parport_pc lp parport snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq wlan_scan_sta(U) snd_seq_device snd_pcm_oss snd_mixer_oss snd_pcm ath_rate_sample(U) mmc_block snd_timer snd sg ohci1394 i2c_i801 wlan(U) ieee1394 sdhci soundcore pcspkr i2c_core snd_page_alloc serio_raw e1000 mmc_core ath_hal(U) ahci libata sd_mod scsi_mod ext3 jbd ehci_hcd ohci_hcd uhci_hcd

Quoted 2 lines
CPU:    0
EIP:    0060:[<f898bf73>]    Tainted: P      VLI

Seems that one of the modules tainted the kernel. Any idea which?

Signature · 2 lines
James Morris
<[email protected]>
#2 lore raw
Quoted 2 lines · James Morris

Seems that one of the modules tainted the kernel. Any idea which?

ath_pci, the atheros madwifi driver. It was loaded, but not active (interface down). I think it's pretty unlikely to be a factor here.

J

#3 lore raw
In reply to Jeremy Fitzhardinge
Quoted 6 lines · On Thu, 28 Sep 2006 13:42:26 -0700 Jeremy Fitzhardinge <[email protected]> wrote:

I just filed this in the Redhat bugzilla, since its from the FC6 distro kernel. But since its fairly close to current kernel.org kernels, I thought it might be relevent.

The bug is https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=208488

Quoted 7 lines

Unfortunately this isn't a very useful report since it was a once-off, and there's a P-tainting module in there. But if anyone sees anything else like this, it's interesting.

The oops is:

BUG: unable to handle kernel paging request at virtual address 756e6547

756e6547 -> uneG. Matches "GenuineIntel".

That'll get written into a temporary page by the /proc/cpuinfo handler, so it might just be a use-uninitialised.

Quoted 15 lines

printing eip:
f898bf73
*pde = 0014c7c0

Oops: 0002 [#1]
SMP 
last sysfs file: /devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
Modules linked in: ath_pci(U) usb_storage loop wlan_wep(U) snd_hda_codec nfsd exportfs lockd nfs_acl tun ppp_deflate zlib_deflate ppp_async crc_ccitt ppp_generic slhc airprime usbserial hci_usb cpufreq_powersave i915 drm cpufreq_conservative ipv6 autofs4 hidp rfcomm l2cap bluetooth sunrpc ipt_REJECT xt_state ip_conntrack nfnetlink xt_tcpudp iptable_filter ip_tables x_tables vfat fat dm_mirror dm_mod video sbs ibm_acpi i2c_ec dock button battery asus_acpi ac parport_pc lp parport snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq wlan_scan_sta(U) snd_seq_device snd_pcm_oss snd_mixer_oss snd_pcm ath_rate_sample(U) mmc_block snd_timer snd sg ohci1394 i2c_i801 wlan(U) ieee1394 sdhci soundcore pcspkr i2c_core snd_page_alloc serio_raw e1000 mmc_core ath_hal(U) ahci libata sd_mod scsi_mod ext3 jbd ehci_hcd ohci_hcd uhci_hcd
CPU:    0
EIP:    0060:[<f898bf73>]    Tainted: P      VLI
EFLAGS: 00013206   (2.6.18-1.2689.fc6PAE #1) 

EIP is at ext3_clear_inode+0x52/0x8b [ext3]
eax: c07dda80 ebx: c07dd994 ecx: 00000000 edx: 756e6547
esi: c07dda80 edi: 00000000 ebp: f3f0cda0 esp: f3f0cd94
ds: 007b es: 007b ss: 0068
Process Xorg (pid: 2862, ti=f3f0c000 task=f3cf4030 task.ti=f3f0c000)

Quoted 13 lines
Stack: c07dda80 c07ddc9c 0000001f f3f0cdb0 c04881e5 c07dda80 c07dda88 f3f0cdc8 
       c04884ba f3f0cdd8 c0bf3db0 00000000 00000080 f3f0cdec c04886c6 00000080 
       00000080 c07dddb8 f32d1428 0000620c f7ff62fc 000000a9 f3f0ce1c c045ba2d 
Call Trace:
 [<c04881e5>] clear_inode+0xd8/0x129
 [<c04884ba>] dispose_list+0x3c/0xc1
 [<c04886c6>] shrink_icache_memory+0x187/0x1af
 [<c045ba2d>] shrink_slab+0xd0/0x137
 [<c045c285>] try_to_free_pages+0x159/0x218
 [<c045860d>] __alloc_pages+0x199/0x287
 [<c045f3b8>] __handle_mm_fault+0x1b2/0xb84
 [<c0614a5a>] do_page_fault+0x2b1/0x5ba
 [<c0404be9>] error_code+0x39/0x40

It's relatively common for that big inode LRU walk to wander off in the wrong direction and to start operating on random memory.

IOW: don't know. Something scribbled on memory somewhere.

#4 lore raw
Quoted 5 lines · Andrew Morton

756e6547 -> uneG. Matches "GenuineIntel".

That'll get written into a temporary page by the /proc/cpuinfo handler, so it might just be a use-uninitialised.

I was compiling a kernel at the time, so it could have come from the kernel source.

Quoted 5 lines

It's relatively common for that big inode LRU walk to wander off in the wrong direction and to start operating on random memory.

IOW: don't know. Something scribbled on memory somewhere.

Yep, that was as far as I got...

J

#5 lore raw
In reply to Andrew Morton
Quoted 9 lines · Andrew Morton
Quoted 6 lines · On Thu, 28 Sep 2006 13:42:26 -0700 Jeremy Fitzhardinge <[email protected]> wrote:

I just filed this in the Redhat bugzilla, since its from the FC6 distro kernel. But since its fairly close to current kernel.org kernels, I thought it might be relevent.

The bug is https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=208488

Unfortunately this isn't a very useful report since it was a once-off, and there's a P-tainting module in there. But if anyone sees anything else like this, it's interesting.

Quoted 8 lines
Quoted 3 lines

The oops is:

BUG: unable to handle kernel paging request at virtual address 756e6547

756e6547 -> uneG. Matches "GenuineIntel".

That'll get written into a temporary page by the /proc/cpuinfo handler, so it might just be a use-uninitialised.

But strangely enough, it's the second report we've seen with this exact backtrace, and the same "Genu" ascii string where the i_default_acl should be.

Both boxes had been through a suspend-to-ram recently, just in case that might matter.

Seems like something more than random chance...

-Eric

#6 lore raw
Quoted 23 lines · Eric Sandeen
Quoted 13 lines · Andrew Morton
Quoted 10 lines · On Thu, 28 Sep 2006 13:42:26 -0700 Jeremy Fitzhardinge <[email protected]> wrote:

I just filed this in the Redhat bugzilla, since its from the FC6 distro kernel. But since its fairly close to current kernel.org kernels, I thought it might be relevent.

The bug is https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=208488

Unfortunately this isn't a very useful report since it was a once-off, and there's a P-tainting module in there. But if anyone sees anything else like this, it's interesting.

Quoted 8 lines
Quoted 3 lines

The oops is:

BUG: unable to handle kernel paging request at virtual address 756e6547

756e6547 -> uneG. Matches "GenuineIntel".

That'll get written into a temporary page by the /proc/cpuinfo handler, so it might just be a use-uninitialised.

Quoted 8 lines

But strangely enough, it's the second report we've seen with this exact backtrace, and the same "Genu" ascii string where the i_default_acl should be.

Both boxes had been through a suspend-to-ram recently, just in case that might matter.

Seems like something more than random chance...

Can you give a pointer to the other report?

Quoted 1 line

-Eric

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#7 lore raw
Quoted 8 lines · Adrian Bunk
Quoted 7 lines · Eric Sandeen
Quoted 5 lines · Andrew Morton
Quoted 1 line
BUG: unable to handle kernel paging request at virtual address 756e6547

756e6547 -> uneG. Matches "GenuineIntel".

That'll get written into a temporary page by the /proc/cpuinfo handler, so it might just be a use-uninitialised.

Quoted 7 lines
Quoted 5 lines

But strangely enough, it's the second report we've seen with this exact backtrace, and the same "Genu" ascii string where the i_default_acl should be.

Both boxes had been through a suspend-to-ram recently, just in case that might matter.

Seems like something more than random chance...

Can you give a pointer to the other report?

Sure, https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=207658

-Eric

#8 lore raw
Quoted 14 lines · Eric Sandeen
Quoted 13 lines · Adrian Bunk
Quoted 12 lines · Eric Sandeen
Quoted 6 lines · Andrew Morton
Quoted 1 line
BUG: unable to handle kernel paging request at virtual address 756e6547

756e6547 -> uneG. Matches "GenuineIntel".

That'll get written into a temporary page by the /proc/cpuinfo handler, so
it might just be a use-uninitialised.

But strangely enough, it's the second report we've seen with this exact backtrace, and the same "Genu" ascii string where the i_default_acl should be.

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

Both boxes had been through a suspend-to-ram recently, just in case that might matter.

Seems like something more than random chance...

Can you give a pointer to the other report?

Sure, https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=207658

Another tainted kernel...

Quoted 1 line

-Eric

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#9 lore raw
Quoted 1 line · Adrian Bunk

Another tainted kernel...

Yes. One was tainted with a force, and another tainted with proprietary modules.

I was quick to dismiss the first one due to the taint, but the likelihood of both taints being the root cause of the exact same corruption seems low. Still, nigh impossible to work with.

-Eric