Re: 2.6.12-rc3: Bad page state at prep_new_page

[Andrew Morton]

Mikael Pettersson <mikpe@xxxxxxxxxxxxx> wrote:
>
> My trusty network server (440BX PentiumIII box serving NFS and News)
>  just threw the following at me:
> 
>  Bad page state at prep_new_page (in process 'python', page c10984e0)
>  flags:0x20021008 mapping:00000000 mapcount:0 count:0
>  Backtrace:
>   [<c0137555>] bad_page+0x75/0xb0
>   [<c01378aa>] prep_new_page+0x2a/0x70
>   [<c0137e60>] buffered_rmqueue+0xc0/0x1c0
>   [<c01380f0>] __alloc_pages+0xc0/0x400
>   [<c01425a9>] do_anonymous_page+0x79/0x150
>   [<c0142870>] do_no_page+0x1f0/0x360
>   [<c0142c2d>] handle_mm_fault+0x13d/0x170
>   [<c0110c92>] do_page_fault+0x2a2/0x6bf
>   [<c0111d12>] recalc_task_prio+0xc2/0x170
>   [<c026b49b>] schedule+0x31b/0x5c0
>   [<c01109f0>] do_page_fault+0x0/0x6bf
>   [<c0102dbf>] error_code+0x4f/0x54
>  Trying to fix it up, but a reboot is needed
> 
>  It's running 2.6.12-rc3 since last Friday, but had been running
>  2.6.12-rc2 for weeks before that w/o problems. 2.4 kernels and
>  most recent 2.6 kernels have always been rock solid on it.
> 

PG_swapcache is set.  It wasn't set when this page was returned to the page
allocator.

>  This is the first time this box has had a problem like this.
>  While I can't rule out bad memory (I'll memtest86 it when I
>  get a chance to), I'm inclined to suspect a 2.6 kernel bug.

Could be a bug, but it'd be a damn unusual one.  I wouldn't rule out a
bitflip.

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