search for: request_fn

This patch set fixes the race when hot-unplug stressed disk. Asias He (3): virtio-blk: Call del_gendisk() before disable guest kick virtio-blk: Reset device after blk_cleanup_queue() virtio-blk: Use block layer provided spinlock drivers/block/virtio_blk.c | 25 ++++++------------------- 1 file changed, 6 insertions(+), 19 deletions(-) -- 1.7.10.2

This patch set fixes the race when hot-unplug stressed disk. Asias He (3): virtio-blk: Call del_gendisk() before disable guest kick virtio-blk: Reset device after blk_cleanup_queue() virtio-blk: Use block layer provided spinlock drivers/block/virtio_blk.c | 25 ++++++------------------- 1 file changed, 6 insertions(+), 19 deletions(-) -- 1.7.10.2

Block drivers like nbd and rbd unlock struct request_queue->queue_lock in their request_fn. I'd like to do the same in virtio_blk. After happily posting the patch, Michael Tsirkin pointed out an issue that I can't explain. This may affect existing block drivers that unlock the queue_lock too. What happens when the block device is removed (hot unplug or kernel module unloaded)...

Block drivers like nbd and rbd unlock struct request_queue->queue_lock in their request_fn. I'd like to do the same in virtio_blk. After happily posting the patch, Michael Tsirkin pointed out an issue that I can't explain. This may affect existing block drivers that unlock the queue_lock too. What happens when the block device is removed (hot unplug or kernel module unloaded)...

...f6892b3fc3726576cb42f17d1d6b5 upstream. blk_cleanup_queue() will call blk_drian_queue() to drain all the requests before queue DEAD marking. If we reset the device before blk_cleanup_queue() the drain would fail. 1) if the queue is stopped in do_virtblk_request() because device is full, the q->request_fn() will not be called. blk_drain_queue() { while(true) { ... if (!list_empty(&q->queue_head)) __blk_run_queue(q) { if (queue is not stoped) q->request_fn() } ... } } Do no reset the device before blk_cleanup_queue() gives the chance to start the qu...

...f6892b3fc3726576cb42f17d1d6b5 upstream. blk_cleanup_queue() will call blk_drian_queue() to drain all the requests before queue DEAD marking. If we reset the device before blk_cleanup_queue() the drain would fail. 1) if the queue is stopped in do_virtblk_request() because device is full, the q->request_fn() will not be called. blk_drain_queue() { while(true) { ... if (!list_empty(&q->queue_head)) __blk_run_queue(q) { if (queue is not stoped) q->request_fn() } ... } } Do no reset the device before blk_cleanup_queue() gives the chance to start the qu...

.../** - * blk_abort_queue -- Abort all request on given queue - * @queue: pointer to queue - * - */ -void blk_abort_queue(struct request_queue *q) -{ - unsigned long flags; - struct request *rq, *tmp; - LIST_HEAD(list); - - /* - * Not a request based block device, nothing to abort - */ - if (!q->request_fn) - return; - - spin_lock_irqsave(q->queue_lock, flags); - - elv_abort_queue(q); - - /* - * Splice entries to local list, to avoid deadlocking if entries - * get readded to the timeout list by error handling - */ - list_splice_init(&q->timeout_list, &list); - - list_for_each_entry_...

...imits - reset limits to default values * @lim: the queue_limits structure to reset * diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 2aa2466..e2d58bd 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -200,6 +200,7 @@ struct request_pm_state typedef void (request_fn_proc) (struct request_queue *q); typedef void (make_request_fn) (struct request_queue *q, struct bio *bio); +typedef void (abort_queue_fn) (struct request_queue *q); typedef int (prep_rq_fn) (struct request_queue *, struct request *); typedef void (unprep_rq_fn) (struct request_queue *, struct r...

...imits - reset limits to default values * @lim: the queue_limits structure to reset * diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 2aa2466..e2d58bd 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -200,6 +200,7 @@ struct request_pm_state typedef void (request_fn_proc) (struct request_queue *q); typedef void (make_request_fn) (struct request_queue *q, struct bio *bio); +typedef void (abort_queue_fn) (struct request_queue *q); typedef int (prep_rq_fn) (struct request_queue *, struct request *); typedef void (unprep_rq_fn) (struct request_queue *, struct r...

Holding the vblk->lock across kick causes poor scalability in SMP guests. If one CPU is doing virtqueue kick and another CPU touches the vblk->lock it will have to spin until virtqueue kick completes. This patch reduces system% CPU utilization in SMP guests that are running multithreaded I/O-bound workloads. The improvements are small but show as iops and SMP are increased. Khoa Huynh

Holding the vblk->lock across kick causes poor scalability in SMP guests. If one CPU is doing virtqueue kick and another CPU touches the vblk->lock it will have to spin until virtqueue kick completes. This patch reduces system% CPU utilization in SMP guests that are running multithreaded I/O-bound workloads. The improvements are small but show as iops and SMP are increased. Khoa Huynh

...ue to run * * Description: * See @blk_run_queue. This variant must be called with the queue lock * held and interrupts disabled. */ void __blk_run_queue(struct request_queue *q) { + lockdep_assert_held(q->queue_lock); + if (unlikely(blk_queue_stopped(q))) return; q->request_fn(q); } EXPORT_SYMBOL(__blk_run_queue); /** * blk_run_queue_async - run a single device queue in workqueue context @@ -372,30 +374,30 @@ void blk_drain_queue(struct request_queue *q, bool drain_all) /* * This function might be called on a queue which failed * driver init after queue...

...ue to run * * Description: * See @blk_run_queue. This variant must be called with the queue lock * held and interrupts disabled. */ void __blk_run_queue(struct request_queue *q) { + lockdep_assert_held(q->queue_lock); + if (unlikely(blk_queue_stopped(q))) return; q->request_fn(q); } EXPORT_SYMBOL(__blk_run_queue); /** * blk_run_queue_async - run a single device queue in workqueue context @@ -372,30 +374,30 @@ void blk_drain_queue(struct request_queue *q, bool drain_all) /* * This function might be called on a queue which failed * driver init after queue...

...ction: scsi_release_buffers() + * +@@ -1687,29 +1691,40 @@ u64 scsi_calculate_bounce_limit(struct S + } + EXPORT_SYMBOL(scsi_calculate_bounce_limit); + +-struct request_queue *scsi_alloc_queue(struct scsi_device *sdev) ++struct request_queue *__scsi_alloc_queue(struct Scsi_Host *shost, ++ request_fn_proc *request_fn) + { +- struct Scsi_Host *shost = sdev->host; + struct request_queue *q; + +- q = blk_init_queue(scsi_request_fn, NULL); ++ q = blk_init_queue(request_fn, NULL); + if (!q) + return NULL; + +- blk_queue_prep_rq(q, scsi_prep_fn); +- + blk_queue_max_hw_segments(q, shost->...