archive mirror
 help / color / mirror / Atom feed
From: Ming Lei <>
To: Jan Kara <>
Cc: Jens Axboe <>,, Khazhy Kumykov <>,
	Paolo Valente <>
Subject: Re: [PATCH 1/2] block: Do not merge recursively in elv_attempt_insert_merge()
Date: Fri, 21 May 2021 21:12:14 +0800	[thread overview]
Message-ID: <YKexrshH5i7mvF6U@T590> (raw)
In-Reply-To: <>

On Fri, May 21, 2021 at 01:53:54PM +0200, Jan Kara wrote:
> On Fri 21-05-21 08:42:16, Ming Lei wrote:
> > On Fri, May 21, 2021 at 12:33:52AM +0200, Jan Kara wrote:
> > > Most of the merging happens at bio level. There should not be much
> > > merging happening at request level anymore. Furthermore if we backmerged
> > > a request to the previous one, the chances to be able to merge the
> > > result to even previous request are slim - that could succeed only if
> > > requests were inserted in 2 1 3 order. Merging more requests in
> > 
> > Right, but some workload has this kind of pattern.
> > 
> > For example of qemu IO emulation, it often can be thought as single job,
> > native aio, direct io with high queue depth. IOs is originated from one VM, but
> > may be from multiple jobs in the VM, so bio merge may not hit much because of IO
> > emulation timing(virtio-scsi/blk's MQ, or IO can be interleaved from multiple
> > jobs via the SQ transport), but request merge can really make a difference, see
> > recent patch in the following link:
> > 
> >
> Oh, request merging definitely does make a difference. But the elevator
> hash & merge logic I'm modifying here is used only by BFQ and MQ-DEADLINE
> AFAICT. And these IO schedulers will already call blk_mq_sched_try_merge()
> from their \.bio_merge handler which gets called from blk_mq_submit_bio().
> So all the merging that can happen in the code I remove should have already
> happened. Or am I missing something?

There might be at least two reasons:

1) when .bio_merge() is called, some requests are kept in plug list, so
the bio may not be merged to requests in scheduler queue; when flushing plug
list and inserts these requests to scheduler queue, we have to try to
merge them further

2) only blk_mq_sched_try_insert_merge() is capable of doing aggressive
request merge, such as, when req A is merged to req B, the function will
continue to try to merge req B with other in-queue requests, until no
any further merge can't be done; neither blk_mq_sched_try_merge() nor
blk_attempt_plug_merge can do such aggressive request merge.


  reply	other threads:[~2021-05-21 13:12 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-05-20 22:33 [PATCH 0/2] block: Fix deadlock when merging requests with BFQ Jan Kara
2021-05-20 22:33 ` [PATCH 1/2] block: Do not merge recursively in elv_attempt_insert_merge() Jan Kara
2021-05-21  0:42   ` Ming Lei
2021-05-21 11:53     ` Jan Kara
2021-05-21 13:12       ` Ming Lei [this message]
2021-05-21 13:44         ` Jan Kara
2021-05-20 22:33 ` [PATCH 2/2] blk: Fix lock inversion between ioc lock and bfqd lock Jan Kara
2021-05-21  0:57   ` Ming Lei
2021-05-21  3:29     ` Khazhy Kumykov
2021-05-21  6:54       ` Ming Lei
2021-05-21 12:05         ` Jan Kara
2021-05-21 13:36           ` Ming Lei
2021-05-21 13:47             ` Jan Kara

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:

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=YKexrshH5i7mvF6U@T590 \ \ \ \ \ \ \

* 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).