linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2] nvme: reset disk to the mpath node also when requeuing
@ 2021-06-02 13:00 Daniel Wagner
  2021-06-03  7:25 ` Christoph Hellwig
  0 siblings, 1 reply; 2+ messages in thread
From: Daniel Wagner @ 2021-06-02 13:00 UTC (permalink / raw)
  To: linux-nvme
  Cc: linux-kernel, Keith Busch, Jens Axboe, Christoph Hellwig,
	Sagi Grimberg, Daniel Wagner

Commit ce86dad222e9 ("nvme-multipath: reset bdev to ns head when
failover") moved the reset code where the bio is added to the
requeue_list for the failover path. But it left the original
bio_set_dev() in nvme_requeue_work().

Let's move the bio_set_dev() where we add the bio to the requeue_list
and avoid the double setting in case of the failover path. And this
makes the code more consistent.

Signed-off-by: Daniel Wagner <dwagner@suse.de>
---

v2: loop over all bios and reset them

 drivers/nvme/host/multipath.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/nvme/host/multipath.c b/drivers/nvme/host/multipath.c
index 127a17b4c13d..3334c6f23476 100644
--- a/drivers/nvme/host/multipath.c
+++ b/drivers/nvme/host/multipath.c
@@ -302,6 +302,7 @@ static blk_qc_t nvme_ns_head_submit_bio(struct bio *bio)
 	struct nvme_ns_head *head = bio->bi_bdev->bd_disk->private_data;
 	struct device *dev = disk_to_dev(head->disk);
 	struct nvme_ns *ns;
+	struct bio *b;
 	blk_qc_t ret = BLK_QC_T_NONE;
 	int srcu_idx;
 
@@ -324,6 +325,8 @@ static blk_qc_t nvme_ns_head_submit_bio(struct bio *bio)
 		dev_warn_ratelimited(dev, "no usable path - requeuing I/O\n");
 
 		spin_lock_irq(&head->requeue_lock);
+		for (b = bio; b; b = b->bi_next)
+			bio_set_dev(b, head->disk->part0);
 		bio_list_add(&head->requeue_list, bio);
 		spin_unlock_irq(&head->requeue_lock);
 	} else {
@@ -435,11 +438,6 @@ static void nvme_requeue_work(struct work_struct *work)
 		next = bio->bi_next;
 		bio->bi_next = NULL;
 
-		/*
-		 * Reset disk to the mpath node and resubmit to select a new
-		 * path.
-		 */
-		bio_set_dev(bio, head->disk->part0);
 		submit_bio_noacct(bio);
 	}
 }
-- 
2.29.2


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH v2] nvme: reset disk to the mpath node also when requeuing
  2021-06-02 13:00 [PATCH v2] nvme: reset disk to the mpath node also when requeuing Daniel Wagner
@ 2021-06-03  7:25 ` Christoph Hellwig
  0 siblings, 0 replies; 2+ messages in thread
From: Christoph Hellwig @ 2021-06-03  7:25 UTC (permalink / raw)
  To: Daniel Wagner
  Cc: linux-nvme, linux-kernel, Keith Busch, Jens Axboe,
	Christoph Hellwig, Sagi Grimberg

On Wed, Jun 02, 2021 at 03:00:39PM +0200, Daniel Wagner wrote:
> +	struct bio *b;
>  	blk_qc_t ret = BLK_QC_T_NONE;
>  	int srcu_idx;
>  
> @@ -324,6 +325,8 @@ static blk_qc_t nvme_ns_head_submit_bio(struct bio *bio)
>  		dev_warn_ratelimited(dev, "no usable path - requeuing I/O\n");
>  
>  		spin_lock_irq(&head->requeue_lock);
> +		for (b = bio; b; b = b->bi_next)
> +			bio_set_dev(b, head->disk->part0);
>  		bio_list_add(&head->requeue_list, bio);
>  		spin_unlock_irq(&head->requeue_lock);
>  	} else {

I can't see why we'd need this hunk.  bio->bi_bdev should have never been
changed to start with in this path.

> @@ -435,11 +438,6 @@ static void nvme_requeue_work(struct work_struct *work)
>  		next = bio->bi_next;
>  		bio->bi_next = NULL;
>  
> -		/*
> -		 * Reset disk to the mpath node and resubmit to select a new
> -		 * path.
> -		 */
> -		bio_set_dev(bio, head->disk->part0);
>  		submit_bio_noacct(bio);
>  	}

This hunk looks fine.

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2021-06-03  7:25 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-06-02 13:00 [PATCH v2] nvme: reset disk to the mpath node also when requeuing Daniel Wagner
2021-06-03  7:25 ` Christoph Hellwig

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).