On Tue, Sep 02 2008, Thomas Petazzoni wrote: > Hi Jens, > > Le Tue, 2 Sep 2008 08:46:49 +0200, > Jens Axboe <jens.axboe@xxxxxxxxxx> a écrit : > > > end_request() is deprecated, and only that one. You can use > > end_dequeued_request() or end_queued_request() to completely end a > > dequeued or queued request. Or you can use blk_end_request() to end > > nr_bytes of IO on that request. Those are the functions that you > > should use. And yes, you should use __blk_end_request() if you > > already have the queue lock grabbed. > > Thanks for this answer. However, it raises a few more questions: > > * end_queued_request() and end_dequeued_request() have exactly the > same code, so what's the point of having two separate functions ? They do act in the same way now, but that is because end_that_request_last() automatically detects whether a request needs to be dequeued or not. > * You say that end_*_request() are used to completely end a request, I > suppose it's because they call __end_request() with > rq->hard_nr_sectors << 9 (in the case of normal fs requests). But > end_request() also does this, and my understanding is that a driver > like drivers/block/z2ram.c only handles one segment per iteration > (req->current_nr_sectors sectors) of the loop in do_z2_request(). But > it uses end_request() at each iteration... which shouldn't completely > end the request, but simply move to the next segment (changing > rq->buffer, rq->sector, etc.). How does it work ? end_request() uses hard_cur_sectors, so it does indeed only act on the current segment of the request. end_*_request() act on the full request, hard_nr_sectors. -- Jens Axboe -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
