Re: [sporadic crash] blk: request botched

From: Jens Axboe
Date: Tue Mar 29 2011 - 07:38:42 EST


On 2011-03-29 13:17, Ingo Molnar wrote:
>
> FYI, i'm seeing a new block IO related boot failure. It starts by the kernel spewing:
>
> [ 84.434778] blk: request botched
> [ 84.437546] blk: request botched
> [ 84.441532] blk: request botched
>
> And after more dying noises, a colorful kernel crash in an apparently rarely
> excercised error handler:

I don't think it's the error handler being broken, it simply looks like
a request that is in a bad bad state thus causing the normal rq -> bio
-> bvec run through to bomb out.

So the 'request botched' is the real BUG here.

Is that part reproducible? If so, can you please try with this patch?

diff --git a/block/blk-core.c b/block/blk-core.c
index e0a0623..3045d0e 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -2163,7 +2163,8 @@ bool blk_update_request(struct request *req, int error, unsigned int nr_bytes)
* size, something has gone terribly wrong.
*/
if (blk_rq_bytes(req) < blk_rq_cur_bytes(req)) {
- printk(KERN_ERR "blk: request botched\n");
+ blk_dump_rq_flags(req, "request botched");
+ WARN_ON(1);
req->__data_len = blk_rq_cur_bytes(req);
}


--
Jens Axboe

--
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/