In case the driver set RQF_DONTPREP flag, requeue the request as head as
it is likely that the backing storage already had a request to an
adjacent region, so getting the requeued request out as soon as possible
may give us some performance benefit.
Signed-off-by: Christian Loehle <[email protected]>
---
block/blk-mq.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 33292c01875d..d863c826fb23 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1429,7 +1429,7 @@ static void blk_mq_requeue_work(struct work_struct *work)
* merge.
*/
if (rq->rq_flags & RQF_DONTPREP)
- blk_mq_request_bypass_insert(rq, false, false);
+ blk_mq_request_bypass_insert(rq, true, false);
else
blk_mq_sched_insert_request(rq, true, false, false);
}
--
2.37.3
Hyperstone GmbH | Reichenaustr. 39a | 78467 Konstanz
Managing Director: Dr. Jan Peter Berns.
Commercial register of local courts: Freiburg HRB381782
On Wed, 26 Oct 2022 at 09:30, Christian Löhle <[email protected]> wrote:
>
> In case the driver set RQF_DONTPREP flag, requeue the request as head as
> it is likely that the backing storage already had a request to an
> adjacent region, so getting the requeued request out as soon as possible
> may give us some performance benefit.
This is a bit handwavy to me. Would you mind extending this with a
small concrete example, to better understand the benefit?
>
> Signed-off-by: Christian Loehle <[email protected]>
Other than the above nitpick, I can't think of any obvious unwanted
side-effects this could have. So, feel free to add:
Reviewed-by: Ulf Hansson <[email protected]>
Kind regards
Uffe
> ---
> block/blk-mq.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index 33292c01875d..d863c826fb23 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -1429,7 +1429,7 @@ static void blk_mq_requeue_work(struct work_struct *work)
> * merge.
> */
> if (rq->rq_flags & RQF_DONTPREP)
> - blk_mq_request_bypass_insert(rq, false, false);
> + blk_mq_request_bypass_insert(rq, true, false);
> else
> blk_mq_sched_insert_request(rq, true, false, false);
> }
> --
> 2.37.3
> Hyperstone GmbH | Reichenaustr. 39a | 78467 Konstanz
> Managing Director: Dr. Jan Peter Berns.
> Commercial register of local courts: Freiburg HRB381782
>
Jens could you consider this patch?
As far as I can see Barts series has not been merged in any form that would fix my problem and even if it does, requeuing RQF_DONTPREP as head seems on par at worst and a performance improvement at best.
-----Original Message-----
From: Christian L?hle
Sent: Mittwoch, 26. Oktober 2022 09:29
To: [email protected]; [email protected]; [email protected]; [email protected]; '[email protected]' <[email protected]>
Cc: 'Avri Altman' <[email protected]>; [email protected]; [email protected]; 'Christian L?hle' <[email protected]>
Subject: [PATCH 1/3] block: Requeue req as head if driver touched it
In case the driver set RQF_DONTPREP flag, requeue the request as head as it is likely that the backing storage already had a request to an adjacent region, so getting the requeued request out as soon as possible may give us some performance benefit.
Signed-off-by: Christian Loehle <[email protected]>
---
block/blk-mq.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/block/blk-mq.c b/block/blk-mq.c index 33292c01875d..d863c826fb23 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1429,7 +1429,7 @@ static void blk_mq_requeue_work(struct work_struct *work)
* merge.
*/
if (rq->rq_flags & RQF_DONTPREP)
- blk_mq_request_bypass_insert(rq, false, false);
+ blk_mq_request_bypass_insert(rq, true, false);
else
blk_mq_sched_insert_request(rq, true, false, false);
}
--
2.37.3
Hyperstone GmbH | Reichenaustr. 39a | 78467 Konstanz
Managing Director: Dr. Jan Peter Berns.
Commercial register of local courts: Freiburg HRB381782