From: Damien Le Moal <email@example.com>
To: Bart Van Assche <firstname.lastname@example.org>, Jens Axboe <email@example.com>
Cc: firstname.lastname@example.org, Jaegeuk Kim <email@example.com>,
Avri Altman <firstname.lastname@example.org>,
Damien Le Moal <email@example.com>
Subject: Re: [PATCH 4/8] block/mq-deadline: Only use zone locking if necessary
Date: Tue, 10 Jan 2023 08:56:42 +0900 [thread overview]
Message-ID: <firstname.lastname@example.org> (raw)
On 1/10/23 08:51, Bart Van Assche wrote:
> On 1/9/23 15:46, Damien Le Moal wrote:
>> On 1/10/23 08:27, Bart Van Assche wrote:
>>> Measurements have shown that limiting the queue depth to one for zoned
>>> writes has a significant negative performance impact on zoned UFS devices.
>>> Hence this patch that disables zone locking from the mq-deadline scheduler
>>> for storage controllers that support pipelining zoned writes. This patch is
>>> based on the following assumptions:
>>> - Applications submit write requests to sequential write required zones
>>> in order.
>>> - It happens infrequently that zoned write requests are reordered by the
>>> block layer.
>>> - The storage controller does not reorder write requests that have been
>>> submitted to the same hardware queue. This is the case for UFS: the
>>> UFSHCI specification requires that UFS controllers process requests in
>>> order per hardware queue.
>>> - The I/O priority of all pipelined write requests is the same per zone.
>>> - Either no I/O scheduler is used or an I/O scheduler is used that
>>> submits write requests per zone in LBA order.
>>> If applications submit write requests to sequential write required zones
>>> in order, at least one of the pending requests will succeed. Hence, the
>>> number of retries that is required is at most (number of pending
>>> requests) - 1.
>> But if the mid-layer decides to requeue a write request, the workqueue
>> used in the mq block layer for requeuing is going to completely destroy
>> write ordering as that is outside of the submission path, working in
>> parallel with it... Does blk_queue_pipeline_zoned_writes() == true also
>> guarantee that a write request will *never* be requeued before hitting the
>> adapter/device ?
> We don't need the guarantee that reordering will never happen. What we
> need is that reordering happens infrequently (e.g. less than 1% of the
> cases). This is what the last paragraph before your reply refers to.
> Maybe I should expand that paragraph.
But my point is that if a request goes through the block layer requeue, it
will be out of order, and will be submitted out of order again, and will
fail again. Unless you stall dispatching, wait for all requeues to come
back in the scheduler, and then start trying again, I do not see how you
can guarantee that retrying the unaligned writes will ever succeed.
I am talking in the context of host-managed devices here.
Damien Le Moal
Western Digital Research
next prev parent reply other threads:[~2023-01-09 23:56 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-09 23:27 [PATCH 0/8] Enable zoned write pipelining for UFS devices Bart Van Assche
2023-01-09 23:27 ` [PATCH 1/8] block: Document blk_queue_zone_is_seq() and blk_rq_zone_is_seq() Bart Van Assche
2023-01-09 23:36 ` Damien Le Moal
2023-01-09 23:27 ` [PATCH 2/8] block: Introduce the blk_rq_is_seq_zone_write() function Bart Van Assche
2023-01-09 23:38 ` Damien Le Moal
2023-01-09 23:52 ` Bart Van Assche
2023-01-10 9:52 ` Niklas Cassel
2023-01-10 11:54 ` Damien Le Moal
2023-01-10 12:13 ` Niklas Cassel
2023-01-10 12:41 ` Damien Le Moal
2023-01-09 23:27 ` [PATCH 3/8] block: Introduce a request queue flag for pipelining zoned writes Bart Van Assche
2023-01-09 23:27 ` [PATCH 4/8] block/mq-deadline: Only use zone locking if necessary Bart Van Assche
2023-01-09 23:46 ` Damien Le Moal
2023-01-09 23:51 ` Bart Van Assche
2023-01-09 23:56 ` Damien Le Moal [this message]
2023-01-10 0:19 ` Bart Van Assche
2023-01-10 0:32 ` Damien Le Moal
2023-01-10 0:38 ` Jens Axboe
2023-01-10 0:41 ` Jens Axboe
2023-01-10 0:44 ` Bart Van Assche
2023-01-10 0:48 ` Jens Axboe
2023-01-10 0:56 ` Bart Van Assche
2023-01-10 1:03 ` Jens Axboe
2023-01-10 1:17 ` Bart Van Assche
2023-01-10 1:48 ` Jens Axboe
2023-01-10 2:24 ` Damien Le Moal
2023-01-10 3:00 ` Jens Axboe
2023-01-09 23:27 ` [PATCH 5/8] block/null_blk: Refactor null_queue_rq() Bart Van Assche
2023-01-09 23:27 ` [PATCH 6/8] block/null_blk: Add support for pipelining zoned writes Bart Van Assche
2023-01-09 23:27 ` [PATCH 7/8] scsi: Retry unaligned " Bart Van Assche
2023-01-09 23:51 ` Damien Le Moal
2023-01-09 23:55 ` Bart Van Assche
2023-01-09 23:27 ` [PATCH 8/8] scsi: ufs: Enable zoned write pipelining Bart Van Assche
2023-01-10 9:16 ` Avri Altman
2023-01-10 17:42 ` Bart Van Assche
2023-01-10 12:23 ` Bean Huo
2023-01-10 17:41 ` Bart Van Assche
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.