From: Mike Snitzer <snitzer@redhat.com>
To: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: Christoph Hellwig <hch@lst.de>,
axboe@kernel.dk, agk@redhat.com, shli@kernel.org,
philipp.reisner@linbit.com, linux-block@vger.kernel.org,
linux-scsi@vger.kernel.org, drbd-dev@lists.linbit.com,
dm-devel@redhat.com, linux-raid@vger.kernel.org
Subject: Re: [PATCH 22/23] drbd: implement REQ_OP_WRITE_ZEROES
Date: Thu, 30 Mar 2017 19:15:50 -0400 [thread overview]
Message-ID: <20170330231550.GA3102@redhat.com> (raw)
In-Reply-To: <yq1pogy3i6u.fsf@oracle.com>
On Thu, Mar 30 2017 at 11:20am -0400,
Martin K. Petersen <martin.petersen@oracle.com> wrote:
> Mike Snitzer <snitzer@redhat.com> writes:
>
> > I can work on this now. Only question I have is: should DM thinp take
> > care to zero any misaligned head and tail? (I assume so but with all
> > the back and forth between Bart, Paolo and Martin I figured I'd ask
> > explicitly).
>
> Yep, let's make sure our semantics match the hardware ditto.
>
> - So write zeroes should behave deterministically and explicitly handle
> any blocks that can't be cleared via deprovisioning.
>
> - And discard can work at the discard granularity in a
> non-deterministic fashion.
I got pretty far along with implementing the DM thinp support for
WRITE_ZEROES in terms of thinp's DISCARD support (more of an
implementation detail.. or so I thought).
But while discussing this effort with Jeff Moyer he asked: shouldn't the
zeroed blocks be provisioned? This is a fairly embarassing question not
to be able to answer in the moment. So I clearly need to learn what the
overall intent of WRITE_ZEROES actually is.
If it is meant as a replacement for WRITE_SAME (as hch switched dm-io
over from WRITE_SAME with a payload of 0 to WRITE_ZEROES) and for the
backing mechanism for blkdev_issue_zeroout() then I guess I have my
answer. Unless DM thinp can guarantee that the discarded blocks will
always return zeroes (by zeroing before all partial block writes) my
discard based dm-thinp implementation of WRITE_ZEROES is a complete
throw-away (unless block zeroing is enabled.. which it never is because
performance sucks with it). So if an upper-level of the IO stack
(e.g. ext4) were to assume that a block will _definitely_ have zeroes
then DM thinp would fall short.
This is all to say: I don't see a quick way forward on implementing
performant WRITE_ZEROES support for DM thinp.
next prev parent reply other threads:[~2017-03-30 23:16 UTC|newest]
Thread overview: 70+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-03-23 14:33 RFC: always use REQ_OP_WRITE_ZEROES for zeroing offload Christoph Hellwig
2017-03-23 14:33 ` [PATCH 01/23] block: renumber REQ_OP_WRITE_ZEROES Christoph Hellwig
2017-03-28 16:12 ` Bart Van Assche
2017-03-30 8:53 ` hch
2017-03-23 14:33 ` [PATCH 02/23] block: implement splitting of REQ_OP_WRITE_ZEROES bios Christoph Hellwig
2017-03-23 14:33 ` [PATCH 03/23] sd: implement REQ_OP_WRITE_ZEROES Christoph Hellwig
2017-03-28 18:50 ` Bart Van Assche
2017-03-28 19:33 ` Mike Snitzer
2017-03-29 14:51 ` Paolo Bonzini
2017-03-29 16:28 ` Bart Van Assche
2017-03-29 16:53 ` Paolo Bonzini
2017-03-30 2:25 ` Martin K. Petersen
2017-03-23 14:33 ` [PATCH 04/23] md: support REQ_OP_WRITE_ZEROES Christoph Hellwig
2017-03-23 14:33 ` [PATCH 05/23] dm: " Christoph Hellwig
2017-03-23 14:33 ` [PATCH 06/23] dm-kcopyd: switch to use REQ_OP_WRITE_ZEROES Christoph Hellwig
2017-03-23 14:55 ` Mike Snitzer
2017-03-23 14:56 ` Christoph Hellwig
2017-03-23 15:10 ` Mike Snitzer
2017-03-27 9:12 ` Christoph Hellwig
2017-03-23 14:33 ` [PATCH 07/23] block: stop using blkdev_issue_write_same for zeroing Christoph Hellwig
2017-03-23 14:33 ` [PATCH 08/23] block: add a flags argument to (__)blkdev_issue_zeroout Christoph Hellwig
2017-03-23 14:33 ` [PATCH 09/23] block: add a REQ_UNMAP flag for REQ_OP_WRITE_ZEROES Christoph Hellwig
2017-03-23 14:33 ` [PATCH 10/23] block: add a new BLKDEV_ZERO_NOFALLBACK flag Christoph Hellwig
2017-03-23 14:33 ` [PATCH 11/23] block_dev: use blkdev_issue_zerout for hole punches Christoph Hellwig
2017-03-28 16:50 ` Bart Van Assche
2017-03-30 8:59 ` hch
2017-03-23 14:33 ` [PATCH 12/23] sd: handle REQ_UNMAP Christoph Hellwig
2017-03-28 16:48 ` Bart Van Assche
2017-03-29 14:57 ` Paolo Bonzini
2017-03-30 9:02 ` hch
2017-03-30 15:28 ` Martin K. Petersen
2017-03-30 17:30 ` hch
2017-03-31 2:19 ` Martin K. Petersen
2017-03-31 7:18 ` hch
2017-03-23 14:33 ` [PATCH 13/23] nvme: implement REQ_OP_WRITE_ZEROES Christoph Hellwig
2017-03-23 14:33 ` [PATCH 14/23] zram: " Christoph Hellwig
2017-03-23 14:33 ` [PATCH 15/23] loop: " Christoph Hellwig
2017-03-23 14:33 ` [PATCH 16/23] brd: remove discard support Christoph Hellwig
2017-03-23 14:33 ` [PATCH 17/23] rbd: remove the discard_zeroes_data flag Christoph Hellwig
2017-03-23 14:33 ` [PATCH 18/23] rsxx: " Christoph Hellwig
2017-03-23 14:33 ` [PATCH 19/23] mmc: " Christoph Hellwig
2017-03-23 14:33 ` [PATCH 20/23] block: stop using discards for zeroing Christoph Hellwig
2017-03-23 14:33 ` [PATCH 21/23] drbd: make intelligent use of blkdev_issue_zeroout Christoph Hellwig
2017-03-23 14:33 ` [PATCH 22/23] drbd: implement REQ_OP_WRITE_ZEROES Christoph Hellwig
2017-03-30 10:06 ` Lars Ellenberg
2017-03-30 11:44 ` Christoph Hellwig
2017-03-30 12:50 ` [Drbd-dev] " Lars Ellenberg
2017-03-30 13:49 ` Mike Snitzer
2017-03-30 15:20 ` Martin K. Petersen
2017-03-30 23:15 ` Mike Snitzer [this message]
2017-03-31 2:34 ` Martin K. Petersen
2017-03-31 7:17 ` Christoph Hellwig
2017-03-23 14:33 ` [PATCH 23/23] block: remove the discard_zeroes_data flag Christoph Hellwig
2017-03-28 17:00 ` Bart Van Assche
2017-03-29 14:52 ` Paolo Bonzini
2017-03-30 9:06 ` hch
2017-03-30 15:29 ` Martin K. Petersen
2017-03-30 17:29 ` hch
2017-03-23 15:54 ` RFC: always use REQ_OP_WRITE_ZEROES for zeroing offload Lars Ellenberg
2017-03-23 17:02 ` Mike Snitzer
2017-03-23 22:53 ` Lars Ellenberg
2017-03-29 14:57 ` Paolo Bonzini
2017-03-27 9:10 ` Christoph Hellwig
2017-03-27 14:03 ` Mike Snitzer
2017-03-27 14:57 ` Christoph Hellwig
2017-03-27 15:08 ` [Drbd-dev] " Bart Van Assche
2017-03-30 9:04 ` Christoph Hellwig
2017-03-30 15:12 ` Mike Snitzer
2017-03-30 15:22 ` Martin K. Petersen
2017-03-30 15:38 ` Mike Snitzer
Reply instructions:
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:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20170330231550.GA3102@redhat.com \
--to=snitzer@redhat.com \
--cc=agk@redhat.com \
--cc=axboe@kernel.dk \
--cc=dm-devel@redhat.com \
--cc=drbd-dev@lists.linbit.com \
--cc=hch@lst.de \
--cc=linux-block@vger.kernel.org \
--cc=linux-raid@vger.kernel.org \
--cc=linux-scsi@vger.kernel.org \
--cc=martin.petersen@oracle.com \
--cc=philipp.reisner@linbit.com \
--cc=shli@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* 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 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).