Messages in this thread | | | Date | Thu, 26 Feb 2004 08:40:43 +0100 | From | Jens Axboe <> | Subject | Re: [PATCH] iSeries virtual disk |
| |
On Thu, Feb 26 2004, Jeff Garzik wrote: > 4) why do you call blkdev_dequeue_request() in do_viodasd_request() > rather than viodasd_end_request() ? Or just use end_request() ?
It makes the queueing simpler (you could potentially leave the _last_ request on the queue, but it's probably not worth the hassle).
They should not call elv_next_request() only to bail if num_req_outstanding is too big, since it has side effects (moving request from io scheduler core to dispatch, which makes it ineligible for merging, sorting, etc).
do { if (too_many_queued) break;
rq = elv_next_request(q); if (!rq) break;
... }
> 12) don't you need to set blk_queue_max_phys_segments() too?
Yep
> P.S. I so wish that people had named the API function > dma_alloc_incoherent() rather than dma_alloc_noncoherent :)
;-)
-- Jens Axboe
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |