Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

[Pekka Enberg]

Hi Bob,

On Thu, Aug 20, 2009 at 4:56 AM, Bob Copeland<me@bobcopeland.com> wrote:
> Sorry I don't have a lot to go on here, but thought I'd report it anyway.
> I was just starting firefox and got the following two stack traces in
> succession.  The kernel is based on 894ef820b10d77e2d6d717342fc408bdd9825139.
>
> Some kind of memory/filesystem corruption?

Someone is stomping on SLAB internal data structures. Ext4 appears in
both stack traces so I guess we should CC linux-ext4. How easy is it
to reproduce this bug? One option is to try kmemcheck to see if it
catches the problem (see Documentation/kmemcheck.txt for details).

> [ 8904.113230] ------------[ cut here ]------------
> [ 8902.989751] kernel BUG at mm/slab.c:2869!
> [ 8902.989755] invalid opcode: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
> [ 8902.989762] last sysfs file: /sys/class/rfkill/rfkill0/state
> [ 8902.989765] Modules linked in: fuse af_packet ipt_REJECT xt_tcpudp nf_conntrack_ipv4 nf_defrag_ipv4 xt_state iptable_filter ip_tables x_tables acpi_cpufreq binfmt_misc dm_mirror dm_region_hash dm_log dm_multipath dm_mod kvm_intel kvm uinput arc4 ecb ath5k mac80211 i915 appletouch hid_apple usbhid snd_hda_codec_idt ath drm ohci1394 snd_hda_intel cfg80211 snd_hda_codec i2c_algo_bit ieee1394 snd_pcm video processor rfkill ehci_hcd sky2 snd_timer snd snd_page_alloc backlight uhci_hcd joydev sg ac battery thermal button output applesmc sr_mod cdrom input_polldev evdev unix [last unloaded: scsi_wait_scan]
> [ 8902.989844]
> [ 8902.989849] Pid: 253, comm: kswapd0 Not tainted (2.6.31-rc6 #179) MacBook1,1
> [ 8902.989852] EIP: 0060:[<c10995cc>] EFLAGS: 00010002 CPU: 0
> [ 8902.989860] EIP is at cache_free_debugcheck+0x1a9/0x28a
> [ 8902.989863] EAX: 009d4e2e EBX: 8ee08ec0 ECX: 31286ea0 EDX: 009d4e2e
> [ 8902.989867] ESI: d84156c5 EDI: f7034440 EBP: f590fcb0 ESP: f590fc80
> [ 8902.989870]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
> [ 8902.989874] Process kswapd0 (pid: 253, ti=f590e000 task=f72b0418 task.ti=f590e000)
> [ 8902.989877] Stack:
> [ 8902.989879]  f590fc88 c104aac4 00002184 c008f000 c10b8f69 f4205880 009d4e2e c008fd60
> [ 8902.989890] <0> c008fd68 f706af30 f7034440 e77e4ea0 f590fccc c1099911 c008fd68 00000296
> [ 8902.989901] <0> c008fd68 00000001 e77e4ea0 f590fcd8 c10b8f69 c008fd68 f590fcf4 c10b92f3
> [ 8902.989914] Call Trace:
> [ 8902.989920]  [<c104aac4>] ? trace_hardirqs_off+0xb/0xd
> [ 8902.989926]  [<c10b8f69>] ? free_buffer_head+0x21/0x62
> [ 8902.989931]  [<c1099911>] ? kmem_cache_free+0x4c/0xdf
> [ 8902.989936]  [<c10b8f69>] ? free_buffer_head+0x21/0x62
> [ 8902.989941]  [<c10b92f3>] ? try_to_free_buffers+0x7f/0x8f
> [ 8902.989947]  [<c1122f6c>] ? jbd2_journal_try_to_free_buffers+0x137/0x13f
> [ 8902.989953]  [<c10fbce2>] ? ext4_releasepage+0x54/0x62
> [ 8902.989959]  [<c10784a2>] ? try_to_release_page+0x35/0x44
> [ 8902.989965]  [<c10827e1>] ? shrink_page_list+0x4b5/0x657
> [ 8902.989971]  [<c1049d9d>] ? put_lock_stats+0xd/0x21
> [ 8902.989976]  [<c104b8eb>] ? trace_hardirqs_on+0xb/0xd
> [ 8902.989982]  [<c1297839>] ? _spin_unlock_irq+0x32/0x47
> [ 8902.989987]  [<c1082bd0>] ? shrink_list+0x24d/0x514
> [ 8902.989993]  [<c1007cb3>] ? sched_clock+0x48/0x8d
> [ 8902.990000]  [<c1007cb3>] ? sched_clock+0x48/0x8d
> [ 8902.990004]  [<c1049d69>] ? get_lock_stats+0x11/0x38
> [ 8902.990009]  [<c1049d9d>] ? put_lock_stats+0xd/0x21
> [ 8902.990014]  [<c10830a8>] ? shrink_zone+0x211/0x2aa
> [ 8902.990020]  [<c1083682>] ? kswapd+0x3e2/0x587
> [ 8902.990021]  [<c1081686>] ? isolate_pages_global+0x0/0x18b
> [ 8902.990021]  [<c103ca60>] ? autoremove_wake_function+0x0/0x34
> [ 8902.990021]  [<c10832a0>] ? kswapd+0x0/0x587
> [ 8902.990021]  [<c103c78b>] ? kthread+0x70/0x75
> [ 8902.990021]  [<c103c71b>] ? kthread+0x0/0x75
> [ 8902.990021]  [<c10034f7>] ? kernel_thread_helper+0x7/0x10
> [ 8902.990021] Code: 0d 8b 47 2c 8b 55 e0 8b 4d ec 89 54 01 fc 8b 4d dc 8b 59 0c 8b 4d ec 29 d9 89 c8 f7 67 30 3b 57 38 89 45 e4 89 d0 89 55 e8 72 04 <0f> 0b eb fe 8b 4f 2c 0f af ca 8d 14 0b 39 55 ec 74 04 0f 0b eb
> [ 8902.990021] EIP: [<c10995cc>] cache_free_debugcheck+0x1a9/0x28a SS:ESP 0068:f590fc80
> [ 8902.990021] ---[ end trace 73f44c391fc78b0a ]---
> [ 8904.113230] ------------[ cut here ]------------
> [ 8904.113250] WARNING: at lib/list_debug.c:51 list_del+0x41/0x60()
> [ 8904.113256] Hardware name: MacBook1,1
> [ 8904.113262] list_del corruption. next->prev should be c03c0000, but was 8ed88ec8
> [ 8904.113267] Modules linked in: fuse af_packet ipt_REJECT xt_tcpudp nf_conntrack_ipv4 nf_defrag_ipv4 xt_state iptable_filter ip_tables x_tables acpi_cpufreq binfmt_misc dm_mirror dm_region_hash dm_log dm_multipath dm_mod kvm_intel kvm uinput arc4 ecb ath5k mac80211 i915 appletouch hid_apple usbhid snd_hda_codec_idt ath drm ohci1394 snd_hda_intel cfg80211 snd_hda_codec i2c_algo_bit ieee1394 snd_pcm video processor rfkill ehci_hcd sky2 snd_timer snd snd_page_alloc backlight uhci_hcd joydev sg ac battery thermal button output applesmc sr_mod cdrom input_polldev evdev unix [last unloaded: scsi_wait_scan]
> [ 8904.113422] Pid: 6452, comm: firefox Tainted: G      D    2.6.31-rc6 #179
> [ 8904.113425] Call Trace:
> [ 8904.113433]  [<c1029a35>] warn_slowpath_common+0x6a/0x81
> [ 8904.113438]  [<c1156ff1>] ? list_del+0x41/0x60
> [ 8904.113443]  [<c1029a8a>] warn_slowpath_fmt+0x29/0x2c
> [ 8904.113447]  [<c1156ff1>] list_del+0x41/0x60
> [ 8904.113453]  [<c1099b3f>] free_block+0x7d/0x15a
> [ 8904.113457]  [<c10996f2>] ? cache_flusharray+0x45/0xf2
> [ 8904.113462]  [<c1099743>] cache_flusharray+0x96/0xf2
> [ 8904.113467]  [<c1099933>] kmem_cache_free+0x6e/0xdf
> [ 8904.113473]  [<c10b8f69>] free_buffer_head+0x21/0x62
> [ 8904.113477]  [<c10b92f3>] try_to_free_buffers+0x7f/0x8f
> [ 8904.113483]  [<c1122f6c>] jbd2_journal_try_to_free_buffers+0x137/0x13f
> [ 8904.113490]  [<c1106acf>] bdev_try_to_free_page+0x4c/0x5a
> [ 8904.113495]  [<c1106a83>] ? bdev_try_to_free_page+0x0/0x5a
> [ 8904.113500]  [<c10bebd7>] blkdev_releasepage+0x2b/0x36
> [ 8904.113506]  [<c10784a2>] try_to_release_page+0x35/0x44
> [ 8904.113512]  [<c10827e1>] shrink_page_list+0x4b5/0x657
> [ 8904.113518]  [<c104b8eb>] ? trace_hardirqs_on+0xb/0xd
> [ 8904.113524]  [<c1297839>] ? _spin_unlock_irq+0x32/0x47
> [ 8904.113529]  [<c1082bd0>] shrink_list+0x24d/0x514
> [ 8904.113535]  [<c103275f>] ? lock_timer_base+0x26/0x45
> [ 8904.113540]  [<c107f258>] ? get_dirty_limits+0x21/0x28a
> [ 8904.113562]  [<faaa16af>] ? mmu_shrink+0x1c/0xf4 [kvm]
> [ 8904.113566]  [<c104d6fa>] ? print_lock_contention_bug+0x11/0xb2
> [ 8904.113573]  [<c10830a8>] shrink_zone+0x211/0x2aa
> [ 8904.113578]  [<c1083cd3>] try_to_free_pages+0x1d1/0x2b2
> [ 8904.113583]  [<c1081686>] ? isolate_pages_global+0x0/0x18b
> [ 8904.113589]  [<c107e254>] __alloc_pages_nodemask+0x2d5/0x457
> [ 8904.113594]  [<c108001d>] __do_page_cache_readahead+0x10c/0x1c7
> [ 8904.113599]  [<c10800f4>] ra_submit+0x1c/0x21
> [ 8904.113604]  [<c1079897>] filemap_fault+0x1c5/0x39e
> [ 8904.113609]  [<c108a99e>] __do_fault+0x40/0x363
> [ 8904.113614]  [<c1299c10>] ? do_page_fault+0x1da/0x313
> [ 8904.113620]  [<c108c3a2>] handle_mm_fault+0x228/0x4ea
> [ 8904.113625]  [<c1299c10>] ? do_page_fault+0x1da/0x313
> [ 8904.113630]  [<c1299d3a>] do_page_fault+0x304/0x313
> [ 8904.113635]  [<c1299a36>] ? do_page_fault+0x0/0x313
> [ 8904.113639]  [<c1298133>] error_code+0x6b/0x70
> [ 8904.113644]  [<c1299a36>] ? do_page_fault+0x0/0x313
> [ 8904.113648] ---[ end trace 73f44c391fc78b0b ]---
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

[Bob Copeland]

On Thu, Aug 20, 2009 at 08:08:16AM +0300, Pekka Enberg wrote:
> Someone is stomping on SLAB internal data structures. Ext4 appears in
> both stack traces so I guess we should CC linux-ext4. How easy is it
> to reproduce this bug? One option is to try kmemcheck to see if it
> catches the problem (see Documentation/kmemcheck.txt for details).

I don't yet have a way to reproduce at will but I've anecdotally hit
this or some related bug 3 times in the last week or two without trying,
I only now captured a stack trace.  I can try some workload to exercise
the filesystem a bit more to see what turns up.

I'll try that and kmemcheck next.

-- 
Bob Copeland %% www.bobcopeland.com

Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

[Vegard Nossum]

2009/8/20 Bob Copeland <me@bobcopeland.com>:
> On Thu, Aug 20, 2009 at 08:08:16AM +0300, Pekka Enberg wrote:
>> Someone is stomping on SLAB internal data structures. Ext4 appears in
>> both stack traces so I guess we should CC linux-ext4. How easy is it
>> to reproduce this bug? One option is to try kmemcheck to see if it
>> catches the problem (see Documentation/kmemcheck.txt for details).
>
> I don't yet have a way to reproduce at will but I've anecdotally hit
> this or some related bug 3 times in the last week or two without trying,
> I only now captured a stack trace.  I can try some workload to exercise
> the filesystem a bit more to see what turns up.
>
> I'll try that and kmemcheck next.

Hm, I'm afraid kmemcheck gives some known false positives related to
bitfields in ext4 code, so in the case that something turned up, it
might be hard to distinguish it from those false positives.

Sorry :-/


Vegard
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

[Bob Copeland]

On Thu, Aug 20, 2009 at 02:02:49PM +0200, Vegard Nossum wrote:
> > I'll try that and kmemcheck next.
> 
> Hm, I'm afraid kmemcheck gives some known false positives related to
> bitfields in ext4 code, so in the case that something turned up, it
> might be hard to distinguish it from those false positives.

Well I didn't get anything from ext4 so far.  I did hit one with
fsnotify:

WARNING: kmemcheck: Caught 32-bit read from freed memory (f34a443c)
eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee008a06f700011000
 a a a a a a a a a a a a a a a a a a a a a a a a f f f f f f f f
                                                         ^

Pid: 2745, comm: fsck.ext4 Not tainted (2.6.31-rc6 #2) MacBook1,1
EIP: 0060:[<c10f3656>] EFLAGS: 00010217 CPU: 0
EIP is at inotify_handle_event+0x76/0xc0
EAX: f34a443c EBX: f34a4438 ECX: 00000000 EDX: f6732000
ESI: f6559764 EDI: 00000000 EBP: f6733f0c ESP: c1527450
 DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
CR0: 8005003b CR2: f6c046d4 CR3: 367fb000 CR4: 000026d0
DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
DR6: ffff4ff0 DR7: 00000400
 [<c10f0d78>] fsnotify+0xa8/0x130
 [<c10c5e11>] __fput+0xb1/0x1e0
 [<c10c5f55>] fput+0x15/0x20
 [<c10c2ca7>] filp_close+0x47/0x80
 [<c10c2d54>] sys_close+0x74/0xc0
 [<c1002ec8>] sysenter_do_call+0x12/0x36
 [<ffffffff>] 0xffffffff

I think that is list_empty() here where %eax is list_head
and event_list->next is the read location... which definitely
doesn't look like a pointer, if I'm reading it correctly.

inotify_fsnotify.o:

        /* did event_priv get attached? */
        if (list_empty(&fsn_event_priv->event_list))
 143:   8d 43 04                lea    0x4(%ebx),%eax

        event_priv = kmem_cache_alloc(event_priv_cachep, GFP_KERNEL);
        if (unlikely(!event_priv))
                return -ENOMEM;

        fsn_event_priv = &event_priv->fsnotify_event_priv_data;
 146:   39 43 04                cmp    %eax,0x4(%ebx)     <=== read here
 149:   74 1d                   je     168 <inotify_handle_event+0x98>

-- 
Bob Copeland %% www.bobcopeland.com

Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

[Vegard Nossum]

2009/8/21 Bob Copeland <me@bobcopeland.com>:
> On Thu, Aug 20, 2009 at 02:02:49PM +0200, Vegard Nossum wrote:
>> > I'll try that and kmemcheck next.
>>
>> Hm, I'm afraid kmemcheck gives some known false positives related to
>> bitfields in ext4 code, so in the case that something turned up, it
>> might be hard to distinguish it from those false positives.
>
> Well I didn't get anything from ext4 so far.  I did hit one with
> fsnotify:
>
> WARNING: kmemcheck: Caught 32-bit read from freed memory (f34a443c)
> eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee008a06f700011000
>  a a a a a a a a a a a a a a a a a a a a a a a a f f f f f f f f
>                                                         ^
>
> Pid: 2745, comm: fsck.ext4 Not tainted (2.6.31-rc6 #2) MacBook1,1
> EIP: 0060:[<c10f3656>] EFLAGS: 00010217 CPU: 0
> EIP is at inotify_handle_event+0x76/0xc0
> EAX: f34a443c EBX: f34a4438 ECX: 00000000 EDX: f6732000
> ESI: f6559764 EDI: 00000000 EBP: f6733f0c ESP: c1527450
>  DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
> CR0: 8005003b CR2: f6c046d4 CR3: 367fb000 CR4: 000026d0
> DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
> DR6: ffff4ff0 DR7: 00000400
>  [<c10f0d78>] fsnotify+0xa8/0x130
>  [<c10c5e11>] __fput+0xb1/0x1e0
>  [<c10c5f55>] fput+0x15/0x20
>  [<c10c2ca7>] filp_close+0x47/0x80
>  [<c10c2d54>] sys_close+0x74/0xc0
>  [<c1002ec8>] sysenter_do_call+0x12/0x36
>  [<ffffffff>] 0xffffffff
>
> I think that is list_empty() here where %eax is list_head
> and event_list->next is the read location... which definitely
> doesn't look like a pointer, if I'm reading it correctly.

I think f34a443c is a valid pointer. On my machine, at least:

[    0.004000]     lowmem  : 0xc0000000 - 0xf73fe000   ( 883 MB)

>
> inotify_fsnotify.o:
>
>        /* did event_priv get attached? */
>        if (list_empty(&fsn_event_priv->event_list))
>  143:   8d 43 04                lea    0x4(%ebx),%eax
>
>        event_priv = kmem_cache_alloc(event_priv_cachep, GFP_KERNEL);
>        if (unlikely(!event_priv))
>                return -ENOMEM;
>
>        fsn_event_priv = &event_priv->fsnotify_event_priv_data;
>  146:   39 43 04                cmp    %eax,0x4(%ebx)     <=== read here
>  149:   74 1d                   je     168 <inotify_handle_event+0x98>

I can see somewhat of a race, I think:

1. userspace calls inotify_read(), where we wait for something to happen:

249         while (1) {
250                 prepare_to_wait(&group->notification_waitq, &wait,
TASK_INTERRUPTIBLE);
251
252                 mutex_lock(&group->notification_mutex);
253                 kevent = get_one_event(group, count);
254                 mutex_unlock(&group->notification_mutex);

2. an event occurs, and inotify_handle_event() calls
fsnotify_add_notify_event():

 64         ret = fsnotify_add_notify_event(group, event, fsn_event_priv);
 65         /* EEXIST is not an error */
 66         if (ret == -EEXIST)
 67                 ret = 0;

3. fsnotify_add_notify_event() adds the fsn_event_priv to the event,
and adds the event to the group, and finally wakes up anybody who is
waiting on &group->notification_waitq:

230         fsnotify_get_event(event);
231         list_add_tail(&holder->event_list, list);
232         if (priv)
233                 list_add_tail(&priv->event_list, &event->private_data_list);
234         spin_unlock(&event->lock);
235         mutex_unlock(&group->notification_mutex);
236
237         wake_up(&group->notification_waitq);

4. inotify_read() wakes up and frees the event:

253                 kevent = get_one_event(group, count);

5. inotify_handle_event() now dereferences the event_priv pointer,
which was already freed:

 69         /* did event_priv get attached? */
 70         if (list_empty(&fsn_event_priv->event_list))


I think that's it. Any thoughts? I put Eric Paris on Cc.


Vegard
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

[Vegard Nossum]

2009/8/21 Vegard Nossum <vegard.nossum@gmail.com>:
> 2009/8/21 Bob Copeland <me@bobcopeland.com>:
>> On Thu, Aug 20, 2009 at 02:02:49PM +0200, Vegard Nossum wrote:
>>> > I'll try that and kmemcheck next.
>>>
>>> Hm, I'm afraid kmemcheck gives some known false positives related to
>>> bitfields in ext4 code, so in the case that something turned up, it
>>> might be hard to distinguish it from those false positives.
>>
>> Well I didn't get anything from ext4 so far.  I did hit one with
>> fsnotify:
>>
>> WARNING: kmemcheck: Caught 32-bit read from freed memory (f34a443c)
>> eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee008a06f700011000
>>  a a a a a a a a a a a a a a a a a a a a a a a a f f f f f f f f
>>                                                         ^
>>
>> Pid: 2745, comm: fsck.ext4 Not tainted (2.6.31-rc6 #2) MacBook1,1
>> EIP: 0060:[<c10f3656>] EFLAGS: 00010217 CPU: 0
>> EIP is at inotify_handle_event+0x76/0xc0
>> EAX: f34a443c EBX: f34a4438 ECX: 00000000 EDX: f6732000
>> ESI: f6559764 EDI: 00000000 EBP: f6733f0c ESP: c1527450
>>  DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
>> CR0: 8005003b CR2: f6c046d4 CR3: 367fb000 CR4: 000026d0
>> DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
>> DR6: ffff4ff0 DR7: 00000400
>>  [<c10f0d78>] fsnotify+0xa8/0x130
>>  [<c10c5e11>] __fput+0xb1/0x1e0
>>  [<c10c5f55>] fput+0x15/0x20
>>  [<c10c2ca7>] filp_close+0x47/0x80
>>  [<c10c2d54>] sys_close+0x74/0xc0
>>  [<c1002ec8>] sysenter_do_call+0x12/0x36
>>  [<ffffffff>] 0xffffffff
>>
>> I think that is list_empty() here where %eax is list_head
>> and event_list->next is the read location... which definitely
>> doesn't look like a pointer, if I'm reading it correctly.
>
> I think f34a443c is a valid pointer. On my machine, at least:
>
> [    0.004000]     lowmem  : 0xc0000000 - 0xf73fe000   ( 883 MB)
>
>>
>> inotify_fsnotify.o:
>>
>>        /* did event_priv get attached? */
>>        if (list_empty(&fsn_event_priv->event_list))
>>  143:   8d 43 04                lea    0x4(%ebx),%eax
>>
>>        event_priv = kmem_cache_alloc(event_priv_cachep, GFP_KERNEL);
>>        if (unlikely(!event_priv))
>>                return -ENOMEM;
>>
>>        fsn_event_priv = &event_priv->fsnotify_event_priv_data;
>>  146:   39 43 04                cmp    %eax,0x4(%ebx)     <=== read here
>>  149:   74 1d                   je     168 <inotify_handle_event+0x98>
>
> I can see somewhat of a race, I think:
>
> 1. userspace calls inotify_read(), where we wait for something to happen:
>
> 249         while (1) {
> 250                 prepare_to_wait(&group->notification_waitq, &wait,
> TASK_INTERRUPTIBLE);
> 251
> 252                 mutex_lock(&group->notification_mutex);
> 253                 kevent = get_one_event(group, count);
> 254                 mutex_unlock(&group->notification_mutex);
>
> 2. an event occurs, and inotify_handle_event() calls
> fsnotify_add_notify_event():
>
>  64         ret = fsnotify_add_notify_event(group, event, fsn_event_priv);
>  65         /* EEXIST is not an error */
>  66         if (ret == -EEXIST)
>  67                 ret = 0;
>
> 3. fsnotify_add_notify_event() adds the fsn_event_priv to the event,
> and adds the event to the group, and finally wakes up anybody who is
> waiting on &group->notification_waitq:
>
> 230         fsnotify_get_event(event);
> 231         list_add_tail(&holder->event_list, list);
> 232         if (priv)
> 233                 list_add_tail(&priv->event_list, &event->private_data_list);
> 234         spin_unlock(&event->lock);
> 235         mutex_unlock(&group->notification_mutex);
> 236
> 237         wake_up(&group->notification_waitq);
>
> 4. inotify_read() wakes up and frees the event:
>
> 253                 kevent = get_one_event(group, count);
>
> 5. inotify_handle_event() now dereferences the event_priv pointer,
> which was already freed:
>
>  69         /* did event_priv get attached? */
>  70         if (list_empty(&fsn_event_priv->event_list))
>
>
> I think that's it. Any thoughts? I put Eric Paris on Cc.

I guess it was fixed by this recently posted patch:

http://osdir.com/ml/linux-kernel/2009-08/msg05185.html

Was kmemcheck by any chance used to discover this race in the first place? ;-)


Vegard
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

[Eric Paris]

On Fri, 2009-08-21 at 10:28 +0200, Vegard Nossum wrote:
> 2009/8/21 Vegard Nossum <vegard.nossum@gmail.com>:
> > 2009/8/21 Bob Copeland <me@bobcopeland.com>:
> >> On Thu, Aug 20, 2009 at 02:02:49PM +0200, Vegard Nossum wrote:
> >>> > I'll try that and kmemcheck next.
> >>>
> >>> Hm, I'm afraid kmemcheck gives some known false positives related to
> >>> bitfields in ext4 code, so in the case that something turned up, it
> >>> might be hard to distinguish it from those false positives.
> >>
> >> Well I didn't get anything from ext4 so far.  I did hit one with
> >> fsnotify:
> >>
> >> WARNING: kmemcheck: Caught 32-bit read from freed memory (f34a443c)
> >> eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee008a06f700011000
> >>  a a a a a a a a a a a a a a a a a a a a a a a a f f f f f f f f
> >>                                                         ^
> >>
> >> Pid: 2745, comm: fsck.ext4 Not tainted (2.6.31-rc6 #2) MacBook1,1
> >> EIP: 0060:[<c10f3656>] EFLAGS: 00010217 CPU: 0
> >> EIP is at inotify_handle_event+0x76/0xc0
> >> EAX: f34a443c EBX: f34a4438 ECX: 00000000 EDX: f6732000
> >> ESI: f6559764 EDI: 00000000 EBP: f6733f0c ESP: c1527450
> >>  DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
> >> CR0: 8005003b CR2: f6c046d4 CR3: 367fb000 CR4: 000026d0
> >> DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
> >> DR6: ffff4ff0 DR7: 00000400
> >>  [<c10f0d78>] fsnotify+0xa8/0x130
> >>  [<c10c5e11>] __fput+0xb1/0x1e0
> >>  [<c10c5f55>] fput+0x15/0x20
> >>  [<c10c2ca7>] filp_close+0x47/0x80
> >>  [<c10c2d54>] sys_close+0x74/0xc0
> >>  [<c1002ec8>] sysenter_do_call+0x12/0x36
> >>  [<ffffffff>] 0xffffffff
> >>
> >> I think that is list_empty() here where %eax is list_head
> >> and event_list->next is the read location... which definitely
> >> doesn't look like a pointer, if I'm reading it correctly.
> >
> > I think f34a443c is a valid pointer. On my machine, at least:
> >
> > [    0.004000]     lowmem  : 0xc0000000 - 0xf73fe000   ( 883 MB)
> >
> >>
> >> inotify_fsnotify.o:
> >>
> >>        /* did event_priv get attached? */
> >>        if (list_empty(&fsn_event_priv->event_list))
> >>  143:   8d 43 04                lea    0x4(%ebx),%eax
> >>
> >>        event_priv = kmem_cache_alloc(event_priv_cachep, GFP_KERNEL);
> >>        if (unlikely(!event_priv))
> >>                return -ENOMEM;
> >>
> >>        fsn_event_priv = &event_priv->fsnotify_event_priv_data;
> >>  146:   39 43 04                cmp    %eax,0x4(%ebx)     <=== read here
> >>  149:   74 1d                   je     168 <inotify_handle_event+0x98>
> >
> > I can see somewhat of a race, I think:
> >
> > 1. userspace calls inotify_read(), where we wait for something to happen:
> >
> > 249         while (1) {
> > 250                 prepare_to_wait(&group->notification_waitq, &wait,
> > TASK_INTERRUPTIBLE);
> > 251
> > 252                 mutex_lock(&group->notification_mutex);
> > 253                 kevent = get_one_event(group, count);
> > 254                 mutex_unlock(&group->notification_mutex);
> >
> > 2. an event occurs, and inotify_handle_event() calls
> > fsnotify_add_notify_event():
> >
> >  64         ret = fsnotify_add_notify_event(group, event, fsn_event_priv);
> >  65         /* EEXIST is not an error */
> >  66         if (ret == -EEXIST)
> >  67                 ret = 0;
> >
> > 3. fsnotify_add_notify_event() adds the fsn_event_priv to the event,
> > and adds the event to the group, and finally wakes up anybody who is
> > waiting on &group->notification_waitq:
> >
> > 230         fsnotify_get_event(event);
> > 231         list_add_tail(&holder->event_list, list);
> > 232         if (priv)
> > 233                 list_add_tail(&priv->event_list, &event->private_data_list);
> > 234         spin_unlock(&event->lock);
> > 235         mutex_unlock(&group->notification_mutex);
> > 236
> > 237         wake_up(&group->notification_waitq);
> >
> > 4. inotify_read() wakes up and frees the event:
> >
> > 253                 kevent = get_one_event(group, count);
> >
> > 5. inotify_handle_event() now dereferences the event_priv pointer,
> > which was already freed:
> >
> >  69         /* did event_priv get attached? */
> >  70         if (list_empty(&fsn_event_priv->event_list))
> >
> >
> > I think that's it. Any thoughts? I put Eric Paris on Cc.
> 
> I guess it was fixed by this recently posted patch:
> 
> http://osdir.com/ml/linux-kernel/2009-08/msg05185.html
> 
> Was kmemcheck by any chance used to discover this race in the first place? ;-)

No, it was found by Linus' stellar eye.  I haven't tried kmemcheck since
my last report that I couldn't get a vmware server machine to boot with
kmemcheck=1

http://osdir.com/ml/linux-kernel/2009-08/msg03797.html

I'll give it another shot today.

-Eric

Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

[Eric Paris]

On Fri, 2009-08-21 at 08:46 +0200, Vegard Nossum wrote:
> 2009/8/21 Bob Copeland <me@bobcopeland.com>:
> > On Thu, Aug 20, 2009 at 02:02:49PM +0200, Vegard Nossum wrote:
> >> > I'll try that and kmemcheck next.
> >>
> >> Hm, I'm afraid kmemcheck gives some known false positives related to
> >> bitfields in ext4 code, so in the case that something turned up, it
> >> might be hard to distinguish it from those false positives.
> >
> > Well I didn't get anything from ext4 so far.  I did hit one with
> > fsnotify:
> >
> > WARNING: kmemcheck: Caught 32-bit read from freed memory (f34a443c)
> > eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee008a06f700011000
> >  a a a a a a a a a a a a a a a a a a a a a a a a f f f f f f f f
> >                                                         ^
> >
> > Pid: 2745, comm: fsck.ext4 Not tainted (2.6.31-rc6 #2) MacBook1,1
> > EIP: 0060:[<c10f3656>] EFLAGS: 00010217 CPU: 0
> > EIP is at inotify_handle_event+0x76/0xc0
> > EAX: f34a443c EBX: f34a4438 ECX: 00000000 EDX: f6732000
> > ESI: f6559764 EDI: 00000000 EBP: f6733f0c ESP: c1527450
> >  DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
> > CR0: 8005003b CR2: f6c046d4 CR3: 367fb000 CR4: 000026d0
> > DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
> > DR6: ffff4ff0 DR7: 00000400
> >  [<c10f0d78>] fsnotify+0xa8/0x130
> >  [<c10c5e11>] __fput+0xb1/0x1e0
> >  [<c10c5f55>] fput+0x15/0x20
> >  [<c10c2ca7>] filp_close+0x47/0x80
> >  [<c10c2d54>] sys_close+0x74/0xc0
> >  [<c1002ec8>] sysenter_do_call+0x12/0x36
> >  [<ffffffff>] 0xffffffff
> >
> > I think that is list_empty() here where %eax is list_head
> > and event_list->next is the read location... which definitely
> > doesn't look like a pointer, if I'm reading it correctly.
> 
> I think f34a443c is a valid pointer. On my machine, at least:
> 
> [    0.004000]     lowmem  : 0xc0000000 - 0xf73fe000   ( 883 MB)
> 
> >
> > inotify_fsnotify.o:
> >
> >        /* did event_priv get attached? */
> >        if (list_empty(&fsn_event_priv->event_list))
> >  143:   8d 43 04                lea    0x4(%ebx),%eax
> >
> >        event_priv = kmem_cache_alloc(event_priv_cachep, GFP_KERNEL);
> >        if (unlikely(!event_priv))
> >                return -ENOMEM;
> >
> >        fsn_event_priv = &event_priv->fsnotify_event_priv_data;
> >  146:   39 43 04                cmp    %eax,0x4(%ebx)     <=== read here
> >  149:   74 1d                   je     168 <inotify_handle_event+0x98>
> 
> I can see somewhat of a race, I think:
> 
> 1. userspace calls inotify_read(), where we wait for something to happen:
> 
> 249         while (1) {
> 250                 prepare_to_wait(&group->notification_waitq, &wait,
> TASK_INTERRUPTIBLE);
> 251
> 252                 mutex_lock(&group->notification_mutex);
> 253                 kevent = get_one_event(group, count);
> 254                 mutex_unlock(&group->notification_mutex);
> 
> 2. an event occurs, and inotify_handle_event() calls
> fsnotify_add_notify_event():
> 
>  64         ret = fsnotify_add_notify_event(group, event, fsn_event_priv);
>  65         /* EEXIST is not an error */
>  66         if (ret == -EEXIST)
>  67                 ret = 0;
> 
> 3. fsnotify_add_notify_event() adds the fsn_event_priv to the event,
> and adds the event to the group, and finally wakes up anybody who is
> waiting on &group->notification_waitq:
> 
> 230         fsnotify_get_event(event);
> 231         list_add_tail(&holder->event_list, list);
> 232         if (priv)
> 233                 list_add_tail(&priv->event_list, &event->private_data_list);
> 234         spin_unlock(&event->lock);
> 235         mutex_unlock(&group->notification_mutex);
> 236
> 237         wake_up(&group->notification_waitq);
> 
> 4. inotify_read() wakes up and frees the event:
> 
> 253                 kevent = get_one_event(group, count);
> 
> 5. inotify_handle_event() now dereferences the event_priv pointer,
> which was already freed:
> 
>  69         /* did event_priv get attached? */
>  70         if (list_empty(&fsn_event_priv->event_list))
> 
> 
> I think that's it. Any thoughts? I put Eric Paris on Cc.

Yes, that's a race.  It's already fixed in Linus' git tree (Linus
pointed out what an idiot I am on that one)

-Eric

Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

[Bob Copeland]

On Fri, Aug 21, 2009 at 08:46:37AM +0200, Vegard Nossum wrote:
> 2009/8/21 Bob Copeland <me@bobcopeland.com>:
> > WARNING: kmemcheck: Caught 32-bit read from freed memory (f34a443c)
> > eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee008a06f700011000
> >  a a a a a a a a a a a a a a a a a a a a a a a a f f f f f f f f
> >
> > I think that is list_empty() here where %eax is list_head
> > and event_list->next is the read location... which definitely
> > doesn't look like a pointer, if I'm reading it correctly.
> 
> I think f34a443c is a valid pointer. On my machine, at least:

I meant list->next, which was 00100100 here.  (By the way, my first
time using kmemcheck, very nice tool!)

[race condition explanation snipped]

> I think that's it. Any thoughts? I put Eric Paris on Cc.

Thanks!  With a fresh pull containing the fsnotify fix the kmemcheck
warning is gone.  So unless I see anything else I'll assume that this
one is already fixed, sorry for the false alarm.

-- 
Bob Copeland %% www.bobcopeland.com

Re: 2.6.32-rc6 BUG at mm/slab.c:2869!

[Pekka Enberg]

On Fri, Aug 21, 2009 at 9:46 AM, Vegard Nossum<vegard.nossum@gmail.com> wrote:
> I think that's it. Any thoughts? I put Eric Paris on Cc.

Nice catch by kmemcheck! I don't think that explains the original slab
corruption, though. Bob, you might want to try again with the fsnotify
fix applied.

                        Pekka