public inbox for [email protected]
 help / color / mirror / Atom feed
From: Pavel Begunkov <[email protected]>
To: Noah Goldstein <[email protected]>
Cc: Jens Axboe <[email protected]>,
	"open list:IO_URING" <[email protected]>
Subject: Re: [PATCH 12/13] io_uring: cache task struct refs
Date: Thu, 27 May 2021 23:13:00 +0100	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <CAFUsyfJ-1iKiSA31C5cgZ3fDqskVEAdQmzSf8nbcVV1ParPvDA@mail.gmail.com>

On 5/27/21 10:51 PM, Noah Goldstein wrote:
> On Mon, May 24, 2021 at 7:51 PM Pavel Begunkov <[email protected]> wrote:
[...]
>>
>> -       percpu_counter_add(&current->io_uring->inflight, nr);
>> -       refcount_add(nr, &current->usage);
>> +       tctx = current->io_uring;
>> +       tctx->cached_refs -= nr;
>> +       if (unlikely(tctx->cached_refs < 0)) {
>> +               unsigned int refill = -tctx->cached_refs + IO_TCTX_REFS_CACHE_NR;
> 
> Might be cleared to use:
> 
> unsigned int refill =  IO_TCTX_REFS_CACHE_NR - tctx->cached_refs;

=======================================================================
Did think about it, but left as is to emphasize that it's negative and
we convert it to positive, so all operations are with positive (and
unsigned due to C rules). Code generation will be same.


>> +
>> +               percpu_counter_add(&tctx->inflight, refill);
>> +               refcount_add(refill, &current->usage);
>> +               tctx->cached_refs += refill;
>> +       }
>>         io_submit_state_start(&ctx->submit_state, nr);
>>

-- 
Pavel Begunkov

  reply	other threads:[~2021-05-27 22:13 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-05-24 23:50 [PATCH for-next 00/13] 5.14 batch 2 Pavel Begunkov
2021-05-24 23:51 ` [PATCH 01/13] io-wq: embed wqe ptr array into struct io_wq Pavel Begunkov
2021-05-24 23:51 ` [PATCH 02/13] io-wq: remove unused io-wq refcounting Pavel Begunkov
2021-05-24 23:51 ` [PATCH 03/13] io_uring: refactor io_iopoll_req_issued Pavel Begunkov
2021-05-24 23:51 ` [PATCH 04/13] io_uring: rename function *task_file Pavel Begunkov
2021-05-24 23:51 ` [PATCH 05/13] io-wq: replace goto while Pavel Begunkov
2021-05-27 21:48   ` Noah Goldstein
2021-05-27 22:18     ` Pavel Begunkov
2021-05-24 23:51 ` [PATCH 06/13] io-wq: don't repeat IO_WQ_BIT_EXIT check by worker Pavel Begunkov
2021-05-24 23:51 ` [PATCH 07/13] io-wq: simplify worker exiting Pavel Begunkov
2021-05-24 23:51 ` [PATCH 08/13] io_uring: hide rsrc tag copy into generic helpers Pavel Begunkov
2021-05-24 23:51 ` [PATCH 09/13] io_uring: remove rsrc put work irq save/restore Pavel Begunkov
2021-05-24 23:51 ` [PATCH 10/13] io_uring: add helpers for 2 level table alloc Pavel Begunkov
2021-05-27 21:43   ` Noah Goldstein
2021-05-27 22:14     ` Pavel Begunkov
2021-05-24 23:51 ` [PATCH 11/13] io_uring: don't vmalloc rsrc tags Pavel Begunkov
2021-05-24 23:51 ` [PATCH 12/13] io_uring: cache task struct refs Pavel Begunkov
2021-05-27 21:51   ` Noah Goldstein
2021-05-27 22:13     ` Pavel Begunkov [this message]
2021-05-24 23:51 ` [PATCH 13/13] io_uring: unify SQPOLL and user task cancellations Pavel Begunkov

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 \
    [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