Re: rq_flag_bits and bio bi_rw flags

[Jens Axboe]

On Thu, Aug 17 2006, Matt Domsch wrote:
> Jens, I'm hoping you can clear something up for me.
> 
> ll_rw_blk.c:
> void blk_rq_bio_prep(request_queue_t *q, struct request *rq, struct
> bio *bio)
> {
>         /* first three bits are identical in rq->flags and bio->bi_rw
> 	*/
>         rq->flags |= (bio->bi_rw & 7);
> 
> That was recently changed to be the first 2 bits not 3.  But it made
> me look.
> 
> /*
>  * bio bi_rw flags
>  *
>  * bit 0 -- read (not set) or write (set)
>  * bit 1 -- rw-ahead when set
>  * bit 2 -- barrier
>  * bit 3 -- fail fast, don't want low level driver retries
>  * bit 4 -- synchronous I/O hint: the block layer will unplug immediately
>  */
> #define BIO_RW          0
> #define BIO_RW_AHEAD    1
> #define BIO_RW_BARRIER  2
> #define BIO_RW_FAILFAST 3
> #define BIO_RW_SYNC     4
> 
> 
> /*
>  * first three bits match BIO_RW* bits, important
>  */
> enum rq_flag_bits {
>         __REQ_RW,               /* not set, read. set, write */
>         __REQ_FAILFAST,         /* no low level driver retries */
>         __REQ_SORTED,           /* elevator knows about this request
> 	*/
> 
> 
> The first bit matches.  The second doesn't (BIO_RW_AHEAD vs
> __REQ_FAILFAST).  And obviously you just fixed the third bit.
> 
> Should BIO_RW_FAILFAST == __REQ_FAILFAST instead?

It does match, a rw-ahead request implies fail fast since we don't want
to be retrying read-head type IO. Could do with a better comment though,
but it's fully on purpose.

-- 
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/