Re: Kernel bug: mm/rmap.c:483

[Hugh Dickins]

On Mon, 24 Jan 2005, 0602@xxxxx wrote:
> 
> I've also experienced mentioned bug, kernel 2.6.10-ac8.
> The relevant syslog entries and .config are attached.
> I'm not subscribed, so please CC me.
> 
> Jan 23 20:01:14 kryton kernel: swap_free: Bad swap offset entry 00100000
> Jan 23 20:01:14 kryton kernel: swap_free: Bad swap offset entry 00110000
> Jan 23 20:01:14 kryton kernel: swap_free: Bad swap offset entry 00100000
> Jan 23 20:01:14 kryton kernel: kernel BUG at mm/rmap.c:483!
> Jan 23 20:02:24 kryton kernel: kernel BUG at fs/reiserfs/inode.c:357!
> Jan 23 20:11:29 kryton kernel: kernel BUG at fs/reiserfs/inode.c:357!
> Jan 23 20:25:53 kryton kernel: kernel BUG at mm/rmap.c:483!
> Jan 23 20:25:53 kryton kernel: kernel BUG at mm/rmap.c:483!
> Jan 23 20:29:15 kryton kernel: Bad page state at prep_new_page (in process 'httpd', page c1b29de0)
> Jan 23 20:29:15 kryton kernel: flags:0x40000110 mapping:00000000 mapcount:-1 count:0

I'm fairly confident that in your case, unlike Jay's,
memtest86 will soon show you're suffering from bad memory.

Those swap_free warnings suggest single-bit and double-bit memory errors;
which are likely to be responsible for all the rest of your symptoms.

It's natural that such errors will sometimes appear as swap_free warnings
and sometimes as rmap.c BUGs.  Increasingly it seems that the rmap.c BUG
is quite a good indicator of bad memory (though not necessarily so).  But
it concerns me that we don't see nearly so many swap_free warnings, which
should be nearly as common if bad memory were the sole cause.

Hugh
-
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/