public inbox for [email protected]
 help / color / mirror / Atom feed
From: Jens Axboe <[email protected]>
To: Pavel Begunkov <[email protected]>,
	[email protected], [email protected]
Subject: Re: [PATCH v2] io_uring: optimise sqe-to-req flags translation
Date: Sat, 18 Jan 2020 09:34:38 -0700	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <06bcf64774c4730b33d1ef65e4fcb67f381cae08.1579340590.git.asml.silence@gmail.com>

On 1/18/20 3:24 AM, Pavel Begunkov wrote:
> For each IOSQE_* flag there is a corresponding REQ_F_* flag. And there
> is a repetitive pattern of their translation:
> e.g. if (sqe->flags & SQE_FLAG*) req->flags |= REQ_F_FLAG*
> 
> Use the same numerical values/bits for them, and copy them instead of
> manual handling.
> 
> Signed-off-by: Pavel Begunkov <[email protected]>
> ---
> 
> v2: enum to generate bits (Jens Axboe)
>     Comments cross 80 chars, but IMHO more visually appealing
> 
> Crosses 
> 
>  fs/io_uring.c                 | 75 +++++++++++++++++++++--------------
>  include/uapi/linux/io_uring.h | 26 +++++++++---
>  2 files changed, 67 insertions(+), 34 deletions(-)
> 
> diff --git a/fs/io_uring.c b/fs/io_uring.c
> index ed1adeda370e..e3e2438a7480 100644
> --- a/fs/io_uring.c
> +++ b/fs/io_uring.c
> @@ -452,6 +452,49 @@ struct io_async_ctx {
>  	};
>  };
>  
> +enum {
> +	REQ_F_FIXED_FILE_BIT	= IOSQE_FIXED_FILE_BIT,
> +	REQ_F_IO_DRAIN_BIT	= IOSQE_IO_DRAIN_BIT,
> +	REQ_F_LINK_BIT		= IOSQE_IO_LINK_BIT,
> +	REQ_F_HARDLINK_BIT	= IOSQE_IO_HARDLINK_BIT,
> +	REQ_F_FORCE_ASYNC_BIT	= IOSQE_ASYNC_BIT,
> +
> +	REQ_F_LINK_NEXT_BIT,
> +	REQ_F_FAIL_LINK_BIT,
> +	REQ_F_INFLIGHT_BIT,
> +	REQ_F_CUR_POS_BIT,
> +	REQ_F_NOWAIT_BIT,
> +	REQ_F_IOPOLL_COMPLETED_BIT,
> +	REQ_F_LINK_TIMEOUT_BIT,
> +	REQ_F_TIMEOUT_BIT,
> +	REQ_F_ISREG_BIT,
> +	REQ_F_MUST_PUNT_BIT,
> +	REQ_F_TIMEOUT_NOSEQ_BIT,
> +	REQ_F_COMP_LOCKED_BIT,
> +};

Perfect

> +enum {
> +	/* correspond one-to-one to IOSQE_IO_* flags*/
> +	REQ_F_FIXED_FILE	= BIT(REQ_F_FIXED_FILE_BIT),	/* ctx owns file */

I'd put the comment on top of the line instead, these lines are way too
long.

> diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
> index 955fd477e530..cee59996b23a 100644
> --- a/include/uapi/linux/io_uring.h
> +++ b/include/uapi/linux/io_uring.h
> @@ -10,6 +10,7 @@
>  
>  #include <linux/fs.h>
>  #include <linux/types.h>
> +#include <linux/bits.h>
>  
>  /*
>   * IO submission data structure (Submission Queue Entry)
> @@ -45,14 +46,29 @@ struct io_uring_sqe {
>  	};
>  };
>  
> +enum {
> +	IOSQE_FIXED_FILE_BIT,
> +	IOSQE_IO_DRAIN_BIT,
> +	IOSQE_IO_LINK_BIT,
> +	IOSQE_IO_HARDLINK_BIT,
> +	IOSQE_ASYNC_BIT,
> +};
> +
>  /*
>   * sqe->flags
>   */
> -#define IOSQE_FIXED_FILE	(1U << 0)	/* use fixed fileset */
> -#define IOSQE_IO_DRAIN		(1U << 1)	/* issue after inflight IO */
> -#define IOSQE_IO_LINK		(1U << 2)	/* links next sqe */
> -#define IOSQE_IO_HARDLINK	(1U << 3)	/* like LINK, but stronger */
> -#define IOSQE_ASYNC		(1U << 4)	/* always go async */
> +enum {
> +	/* use fixed fileset */
> +	IOSQE_FIXED_FILE	= BIT(IOSQE_FIXED_FILE_BIT),

Let's please not use BIT() for the user visible part, though. And I'd
leave these as defines, sometimes apps do things ala:

#ifndef IOSQE_IO_LINK
#define IOSQE_IO_LINK ...
#endif

to make it easier to co-exist with old and new headers.

-- 
Jens Axboe


  reply	other threads:[~2020-01-18 16:34 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-17 22:22 [PATCH 1/2] io_uring: remove REQ_F_IO_DRAINED Pavel Begunkov
2020-01-17 22:22 ` [PATCH 2/2] io_uring: optimise sqe-to-req flags translation Pavel Begunkov
2020-01-17 22:41 ` [PATCH 0/2] optimise sqe-to-req flags Pavel Begunkov
2020-01-17 22:49   ` Jens Axboe
2020-01-17 23:14     ` Pavel Begunkov
2020-01-18  0:16       ` Jens Axboe
2020-01-18 10:24         ` [PATCH v2] io_uring: optimise sqe-to-req flags translation Pavel Begunkov
2020-01-18 16:34           ` Jens Axboe [this message]
2020-01-18 17:10             ` Pavel Begunkov
2020-01-18 17:22               ` [PATCH v3 1/1] " Pavel Begunkov
2020-01-18 20:46                 ` Jens Axboe
2020-01-19  7:46                   ` Pavel Begunkov
2020-01-19 17:36                     ` Jens Axboe
2020-01-18 20:45               ` [PATCH v2] " Jens Axboe
2020-01-17 22:50 ` [PATCH 1/2] io_uring: remove REQ_F_IO_DRAINED Jens Axboe

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