Re: Kernel BUG at mm/slab.c:2564 - 2.6.16-rc5-g7b14e3b5

[Linus Torvalds]

On Thu, 2 Mar 2006, Paolo Ornati wrote:
> 
> Something is happened again here!

I think you have bad ram.

> Slab corruption: start=ffff81000d0ffb30, len=104
> Redzone: 0x5a2cf071/0x5a2cf071.
> Last user: [<ffffffff8015caac>](end_bio_bh_io_sync+0x35/0x39)
> 000: 6b 6b 6b 2b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b

> Slab corruption: start=ffff81000d0ffb30, len=104
> Redzone: 0x5a2cf071/0x5a2cf071.
> Last user: [<ffffffff8015caac>](end_bio_bh_io_sync+0x35/0x39)
> 000: 6b 6b 6b 2b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b

It's the same corruption both times, and the exact same slab entry.

And it's a single-bit error: the "2b" should be a "6b".

Now, if could have been a software error, clearing that one bit, but the 
thing is, that is the first word in a "struct bio", which should be a 
"sector_t bi_sector". The entries around it are also "struct bio"s, and we 
don't do any bit-operations on anything in that area (on "bi_flags", yes).

The fact that it was the very same bit both times (not just the same 
offset: the same physical address) makes me suspect bad RAM.

		Linus
-
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/