linux-kernel - Re: BUG in strnlen

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <48121A37.5020504@cosmosbay.com>
Date:	Fri, 25 Apr 2008 19:51:51 +0200
From:	Eric Dumazet <dada1@...mosbay.com>
To:	Alexander van Heukelum <heukelum@...tmail.fm>
Cc:	Randy Dunlap <randy.dunlap@...cle.com>,
	lkml <linux-kernel@...r.kernel.org>
Subject: Re: BUG in strnlen

Alexander van Heukelum a écrit :
> On Fri, 25 Apr 2008 19:21:53 +0200, "Eric Dumazet" <dada1@...mosbay.com>
> said:
>   
>> Randy Dunlap a écrit :
>>     
>>> Hi,
>>>
>>> All of my daily testing (x86_64, 4 CPUs, 8 GB RAM)
>>> since (after) 2.6.25 is seeing this BUG:
>>> (i.e., 2.6.25 does not do this)
>>>
>>>
>>> BUG: unable to handle kernel paging request at ffffffffa00b7551
>>> IP: [<ffffffff80357aac>] strnlen+0x15/0x1f
>>> PGD 203067 PUD 207063 PMD 27e44f067 PTE 0
>>> Oops: 0000 [1] SMP 
>>> CPU 3 
>>> Modules linked in: hp_ilo parport_pc lp parport tg3 cciss ehci_hcd ohci_hcd uhci_hcd [last unloaded: reiserfs]
>>>       
>
> ------------------------------------------------------------------------------------------^^^^^^
>
>   
>>> Pid: 20926, comm: cat Not tainted 2.6.25-git5 #1
>>> RIP: 0010:[<ffffffff80357aac>]  [<ffffffff80357aac>] strnlen+0x15/0x1f
>>> RSP: 0018:ffff810274981cc8  EFLAGS: 00010297
>>> RAX: ffffffffa00b7551 RBX: ffff810274981d38 RCX: ffffffff80603719
>>> RDX: ffff810274981d68 RSI: fffffffffffffffe RDI: ffffffffa00b7551
>>> RBP: ffff810274981cc8 R08: 00000000ffffffff R09: 00000000000000c8
>>> R10: 0000000000000050 R11: 0000000000000246 R12: ffff8102364600cc
>>> R13: ffffffffa00b7551 R14: 0000000000000011 R15: 0000000000000010
>>> FS:  00007f956375d6f0(0000) GS:ffff81027f808980(0000) knlGS:00000000f7f7f6c0
>>> CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
>>> CR2: ffffffffa00b7551 CR3: 00000002734d5000 CR4: 00000000000006e0
>>> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>> DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
>>> Process cat (pid: 20926, threadinfo ffff810274980000, task ffff81026d18ce20)
>>> Stack:  ffff810274981d28 ffffffff80358d5a ffff810274981d28 0000000000000f34
>>>  ffff8102364600cc ffff810236461000 ffffffff80603719 ffff81024ac14f00
>>>  ffff81024ac14f00 0000000000000004 0000000000000000 0000000000000000
>>> Call Trace:
>>>  [<ffffffff80358d5a>] vsnprintf+0x31b/0x592
>>>  [<ffffffff802a78eb>] seq_printf+0x7e/0xa7
>>>  [<ffffffff8024c6fe>] ? debug_mutex_free_waiter+0x46/0x4a
>>>  [<ffffffff8053aaa2>] ? __down_read+0x17/0x92
>>>  [<ffffffff80539c25>] ? __mutex_lock_slowpath+0x1d8/0x1e5
>>>  [<ffffffff802886c2>] ? count_partial+0x45/0x4d
>>>  [<ffffffff80289a6d>] s_show+0x7e/0xcb
>>>  [<ffffffff802a7dd9>] seq_read+0x10b/0x298
>>>  [<ffffffff802c7dbb>] proc_reg_read+0x7b/0x95
>>>  [<ffffffff8028ec0b>] vfs_read+0xab/0x154
>>>  [<ffffffff8028f015>] sys_read+0x47/0x6f
>>>  [<ffffffff8020c182>] tracesys+0xd5/0xda
>>>
>>>
>>> Code: 48 8d 44 11 ff 40 38 30 74 0a 48 ff c8 48 39 d0 73 f3 31 c0 c9 c3 55 48 89 f8 48 89 e5 eb 03 48 ff c0 48 ff ce 48 83 fe ff 74 05 <80> 38 00 75 ef c9 48 29 f8 c3 55 31 c0 48 89 e5 eb 13 41 38 c8 
>>> RIP  [<ffffffff80357aac>] strnlen+0x15/0x1f
>>>  RSP <ffff810274981cc8>
>>> CR2: ffffffffa00b7551
>>>
>>>
>>> ---
>>>
>>>   
>>>       
>> My initial thoughts are :
>>
>> Fault address is  0xffffffffa00b7551 which is in module mapping space on 
>> x86_64
>>
>> strnlen() is OK
>>
>> Some module created a kmem_cache (with kmem_cache_create()).
>> slub or slab kept a pointer to the cache name in their internal
>> structures.
>> Module was unloaded but forgot to destroy kmem cache before unloading.
>>
>> Fault happens while doing "cat /proc/slabinfo", when trying to 
>> dereference cache name since module was unloaded and its memory unmapped.
>>
>> Next step is to find which module was unloaded ...
>>     
>
> The last one was reiserfs, apparently ;).
>   
Yes but reiserfs correctly destroys its cache at unload time.

Must be something else...




--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/