[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <k7g33ohjbc7tuo3vzhsdl4qzmbjafxncnvfduxggilmdb53mxc@3y37mw7jqw4z>
Date: Fri, 25 Jul 2025 17:56:21 +0200
From: Jan Kara <jack@...e.cz>
To: Tang Yizhou <yizhou.tang@...pee.com>
Cc: axboe@...nel.dk, hch@....de, jack@...e.cz, linux-block@...r.kernel.org,
linux-kernel@...r.kernel.org, tangyeechou@...il.com
Subject: Re: [PATCH 1/3] blk-wbt: Optimize wbt_done() for non-throttled writes
On Thu 24-07-25 16:29:59, Tang Yizhou wrote:
> From: Tang Yizhou <yizhou.tang@...pee.com>
>
> In the current implementation, the sync_cookie and last_cookie members of
> struct rq_wb are used only by read requests and not by non-throttled write
> requests. Based on this, we can optimize wbt_done() by removing one if
> condition check for non-throttled write requests.
>
> Signed-off-by: Tang Yizhou <yizhou.tang@...pee.com>
Nah, I'm undecided if this is actually worth it. But it looks correct and
doesn't make the code harder to follow so I guess feel free to add:
Reviewed-by: Jan Kara <jack@...e.cz>
Honza
> ---
> block/blk-wbt.c | 11 ++++++-----
> 1 file changed, 6 insertions(+), 5 deletions(-)
>
> diff --git a/block/blk-wbt.c b/block/blk-wbt.c
> index a50d4cd55f41..30886d44f6cd 100644
> --- a/block/blk-wbt.c
> +++ b/block/blk-wbt.c
> @@ -248,13 +248,14 @@ static void wbt_done(struct rq_qos *rqos, struct request *rq)
> struct rq_wb *rwb = RQWB(rqos);
>
> if (!wbt_is_tracked(rq)) {
> - if (rwb->sync_cookie == rq) {
> - rwb->sync_issue = 0;
> - rwb->sync_cookie = NULL;
> - }
> + if (wbt_is_read(rq)) {
> + if (rwb->sync_cookie == rq) {
> + rwb->sync_issue = 0;
> + rwb->sync_cookie = NULL;
> + }
>
> - if (wbt_is_read(rq))
> wb_timestamp(rwb, &rwb->last_comp);
> + }
> } else {
> WARN_ON_ONCE(rq == rwb->sync_cookie);
> __wbt_done(rqos, wbt_flags(rq));
> --
> 2.25.1
>
--
Jan Kara <jack@...e.com>
SUSE Labs, CR
Powered by blists - more mailing lists