On 4/8/23 08:58, Bart Van Assche wrote:
> Make sure that zoned writes are submitted in LBA order.
>
> Cc: Damien Le Moal <damien.lemoal@xxxxxxxxxxxxxxxxxx>
> Cc: Christoph Hellwig <hch@xxxxxx>
> Cc: Ming Lei <ming.lei@xxxxxxxxxx>
> Cc: Mike Snitzer <snitzer@xxxxxxxxxx>
> Signed-off-by: Bart Van Assche <bvanassche@xxxxxxx>
> ---
> block/mq-deadline.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/block/mq-deadline.c b/block/mq-deadline.c
> index 50a9d3b0a291..891ee0da73ac 100644
> --- a/block/mq-deadline.c
> +++ b/block/mq-deadline.c
> @@ -798,7 +798,7 @@ static void dd_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
>
> trace_block_rq_insert(rq);
>
> - if (at_head) {
> + if (at_head && !blk_rq_is_seq_zoned_write(rq)) {
> list_add(&rq->queuelist, &per_prio->dispatch);
> rq->fifo_time = jiffies;
> } else {
Looks OK, but I would prefer us addressing the caller site using at_head = true,
as that is almost always completely wrong for sequential zone writes.
That would reduce the number of places we check for blk_rq_is_seq_zoned_write().
