lkml.org 
[lkml]   [2004]   [Jun]   [8]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: msync() oops in 2.6.7-rc2-bk1
Date
Andrew Morton wrote:
> "Miquel van Smoorenburg" <miquels@cistron.nl> wrote:
>
>>I'm running a news server. The innd process uses mmap()s for several
>>files and uses msync() to force synchronization to disk every so
>>often. Suddenly, an msync() causes an oops (and innd SEGVs). This
>>is after the box has been up and running for 3 days:
>>
>># uname -a
>>Linux enterprise 2.6.7-rc2-bk1 #1 Mon May 31 15:03:52 CEST 2004 i686 GNU/Linux
>>
>> <1>Unable to handle kernel NULL pointer dereference at virtual address 00000000 printing eip:
>>c0149120
>>*pde = 00000000
>>Oops: 0002 [#5]
>>Modules linked in: e100 mii
>>CPU: 0
>>EIP: 0060:[<c0149120>] Not tainted
>>EFLAGS: 00010213 (2.6.7-rc2-bk1)
>>EIP is at __set_page_dirty_buffers+0x20/0xb0
>>eax: 00000000 ebx: f77a1e7c ecx: c15706e0 edx: eba5a83c
>>esi: 5ccfb000 edi: 00000000 ebp: 5d000000 esp: f44b5efc
>>ds: 007b es: 007b ss: 0068
>>Process innd (pid: 10936, threadinfo=f44b5000 task=d0eb2c70)
>>Stack: f77a1de4 00000004 00000000 c4af23ec c013143e c15706e0 c013da1c 5ccfb000
>> c4af23f0 c013db0f c4af23ec f7101900 5ccfb000 00000001 5cc00000 ce8b25d0
>> 00000000 5d000000 c013dbc3 ce8b25cc 5cc00000 5d000000 f7101900 00000001
>>Call Trace:
>> [<c013143e>] set_page_dirty+0x3e/0x50
>> [<c013da1c>] filemap_sync_pte+0x5c/0x80
>> [<c013db0f>] filemap_sync_pte_range+0xcf/0xf0
>> [<c013dbc3>] filemap_sync+0x93/0x100
>> [<c013dc96>] msync_interval+0x66/0xf0
>> [<c013de37>] sys_msync+0x117/0x123
>> [<c0103c7b>] syscall_call+0x7/0xb
>>
>>Code: 0f ba 28 01 8b 40 08 39 d0 75 f5 0f ba 29 04 19 c0 85 c0 75
>
>
> You have a page which has PG_private set, but page->private is NULL. And
> the machine is non-SMP, non-preempt, yes?
>
> I'd be wondering whether that machine has flipped a bit in page->flags,
> frankly. How old is it?
>
> Was the mmap of a regular file or of a block device?

I don't think so... I have a number of machines, also news servers,
which are producing various errors in filemap. In the RH2.4.21-15 kernel
it shows up in filemap_sync_pte_range, which is now in msync. It appears
to be a bad pte coming in, I see the error at pmd_bad sometimes.

I see it with several applications, all news, all mmap() heavily.

Sorry I can't tell you more, but it happens on multiple brand-new Xeon
systems. If he is seeing similar on an Athlon I would assume that it's
some "less traveled way" in the kernel, because I have a limited ability
to believe in coincidence.


--
-bill davidsen (davidsen@tmr.com)
"The secret to procrastination is to put things off until the
last possible moment - but no longer" -me
-
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/

\
 
 \ /
  Last update: 2005-03-22 14:03    [W:0.076 / U:0.424 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site