From: Ming Lei <[email protected]>
To: Jens Axboe <[email protected]>
Cc: [email protected], [email protected],
Pavel Begunkov <[email protected]>,
Kevin Wolf <[email protected]>
Subject: Re: [PATCH 5/9] io_uring: support SQE group
Date: Wed, 24 Apr 2024 08:46:41 +0800 [thread overview]
Message-ID: <ZihWcV8+3rfyYxGI@fedora> (raw)
In-Reply-To: <[email protected]>
On Mon, Apr 22, 2024 at 12:27:28PM -0600, Jens Axboe wrote:
> On 4/7/24 7:03 PM, Ming Lei wrote:
> > SQE group is defined as one chain of SQEs starting with the first sqe that
> > has IOSQE_EXT_SQE_GROUP set, and ending with the first subsequent sqe that
> > doesn't have it set, and it is similar with chain of linked sqes.
> >
> > The 1st SQE is group leader, and the other SQEs are group member. The group
> > leader is always freed after all members are completed. Group members
> > aren't submitted until the group leader is completed, and there isn't any
> > dependency among group members, and IOSQE_IO_LINK can't be set for group
> > members, same with IOSQE_IO_DRAIN.
> >
> > Typically the group leader provides or makes resource, and the other members
> > consume the resource, such as scenario of multiple backup, the 1st SQE is to
> > read data from source file into fixed buffer, the other SQEs write data from
> > the same buffer into other destination files. SQE group provides very
> > efficient way to complete this task: 1) fs write SQEs and fs read SQE can be
> > submitted in single syscall, no need to submit fs read SQE first, and wait
> > until read SQE is completed, 2) no need to link all write SQEs together, then
> > write SQEs can be submitted to files concurrently. Meantime application is
> > simplified a lot in this way.
> >
> > Another use case is to for supporting generic device zero copy:
> >
> > - the lead SQE is for providing device buffer, which is owned by device or
> > kernel, can't be cross userspace, otherwise easy to cause leak for devil
> > application or panic
> >
> > - member SQEs reads or writes concurrently against the buffer provided by lead
> > SQE
>
> In concept, this looks very similar to "sqe bundles" that I played with
> in the past:
>
> https://git.kernel.dk/cgit/linux/log/?h=io_uring-bundle
Indeed, so looks it is something which io_uring needs.
>
> Didn't look too closely yet at the implementation, but in spirit it's
> about the same in that the first entry is processed first, and there's
> no ordering implied between the test of the members of the bundle /
> group.
Yeah.
>
> I do think that's a flexible thing to support, particularly if:
>
> 1) We can do it more efficiently than links, which are pretty horrible.
Agree, link is hard to use in async/.await of modern language per my
experience.
Also sqe group won't break link, and the group is thought as a whole
wrt. linking.
> 2) It enables new worthwhile use cases
> 3) It's done cleanly
> 4) It's easily understandable and easy to document, so that users will
> actually understand what this is and what use cases it enable. Part
> of that is actually naming, it should be readily apparent what a
> group is, what the lead is, and what the members are. Using your
> terminology here, definitely worth spending some time on that to get
> it just right and self evident.
All are nice suggestions, and I will follow above and make them in V2.
Thanks,
Ming
next prev parent reply other threads:[~2024-04-24 0:46 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-08 1:03 [RFC PATCH 0/9] io_uring: support sqe group and provide group kbuf Ming Lei
2024-04-08 1:03 ` [PATCH 1/9] io_uring: net: don't check sqe->__pad2[0] for send zc Ming Lei
2024-04-08 1:03 ` [PATCH 2/9] io_uring: support user sqe ext flags Ming Lei
2024-04-22 18:16 ` Jens Axboe
2024-04-23 13:57 ` Ming Lei
2024-04-29 15:24 ` Pavel Begunkov
2024-04-30 3:43 ` Ming Lei
2024-04-30 12:00 ` Pavel Begunkov
2024-04-30 12:56 ` Ming Lei
2024-04-30 14:10 ` Pavel Begunkov
2024-04-30 15:46 ` Ming Lei
2024-05-02 14:22 ` Pavel Begunkov
2024-05-04 1:19 ` Ming Lei
2024-04-08 1:03 ` [PATCH 3/9] io_uring: add helper for filling cqes in __io_submit_flush_completions() Ming Lei
2024-04-08 1:03 ` [PATCH 4/9] io_uring: add one output argument to io_submit_sqe Ming Lei
2024-04-08 1:03 ` [PATCH 5/9] io_uring: support SQE group Ming Lei
2024-04-22 18:27 ` Jens Axboe
2024-04-23 13:08 ` Kevin Wolf
2024-04-24 1:39 ` Ming Lei
2024-04-25 9:27 ` Kevin Wolf
2024-04-26 7:53 ` Ming Lei
2024-04-26 17:05 ` Kevin Wolf
2024-04-29 3:34 ` Ming Lei
2024-04-29 15:48 ` Pavel Begunkov
2024-04-30 3:07 ` Ming Lei
2024-04-29 15:32 ` Pavel Begunkov
2024-04-30 3:03 ` Ming Lei
2024-04-30 12:27 ` Pavel Begunkov
2024-04-30 15:00 ` Ming Lei
2024-05-02 14:09 ` Pavel Begunkov
2024-05-04 1:56 ` Ming Lei
2024-05-02 14:28 ` Pavel Begunkov
2024-04-24 0:46 ` Ming Lei [this message]
2024-04-08 1:03 ` [PATCH 6/9] io_uring: support providing sqe group buffer Ming Lei
2024-04-08 1:03 ` [PATCH 7/9] io_uring/uring_cmd: support provide group kernel buffer Ming Lei
2024-04-08 1:03 ` [PATCH 8/9] ublk: support provide io buffer Ming Lei
2024-04-08 1:03 ` [RFC PATCH 9/9] liburing: support sqe ext_flags & sqe group Ming Lei
2024-04-19 0:55 ` [RFC PATCH 0/9] io_uring: support sqe group and provide group kbuf Ming Lei
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=ZihWcV8+3rfyYxGI@fedora \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
/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