All of lore.kernel.org
 help / color / mirror / Atom feed
From: Stefan Hajnoczi <stefanha@redhat.com>
To: Prasad Pandit <ppandit@redhat.com>
Cc: kwolf@redhat.com, qemu-block@nongnu.org, qemu-devel@nongnu.org,
	mtosatti@redhat.com, Prasad Pandit <pjp@fedoraproject.org>
Subject: Re: [PATCH v3] linux-aio: add IO_CMD_FDSYNC command support
Date: Wed, 13 Mar 2024 11:18:39 -0400	[thread overview]
Message-ID: <20240313151839.GA545171@fedora> (raw)
In-Reply-To: <20240313084935.1412274-1-ppandit@redhat.com>

[-- Attachment #1: Type: text/plain, Size: 3767 bytes --]

On Wed, Mar 13, 2024 at 02:19:35PM +0530, Prasad Pandit wrote:
> From: Prasad Pandit <pjp@fedoraproject.org>
> 
> Libaio defines IO_CMD_FDSYNC command to sync all outstanding
> asynchronous I/O operations, by flushing out file data to the
> disk storage.
> 
> Enable linux-aio to submit such aio request. This helps to
> reduce latency induced via pthread_create calls by
> thread-pool (aio=threads).
> 
> Signed-off-by: Prasad Pandit <pjp@fedoraproject.org>
> ---
>  block/file-posix.c |  7 +++++++
>  block/linux-aio.c  | 22 +++++++++++++++++++++-
>  2 files changed, 28 insertions(+), 1 deletion(-)
> 
> v3: check if host kernel supports aio_fsync call
>   -> https://lists.nongnu.org/archive/html/qemu-devel/2024-03/msg03210.html
> 
> diff --git a/block/file-posix.c b/block/file-posix.c
> index 35684f7e21..a30494d1a8 100644
> --- a/block/file-posix.c
> +++ b/block/file-posix.c
> @@ -2453,6 +2453,8 @@ static inline bool raw_check_linux_io_uring(BDRVRawState *s)
>  #endif
>  
>  #ifdef CONFIG_LINUX_AIO
> +extern bool laio_has_fdsync(int);

Please declare this in include/block/raw-aio.h alongside the other laio APIs.

> +
>  static inline bool raw_check_linux_aio(BDRVRawState *s)
>  {
>      Error *local_err = NULL;
> @@ -2599,6 +2601,11 @@ static int coroutine_fn raw_co_flush_to_disk(BlockDriverState *bs)
>      if (raw_check_linux_io_uring(s)) {
>          return luring_co_submit(bs, s->fd, 0, NULL, QEMU_AIO_FLUSH);
>      }
> +#endif
> +#ifdef CONFIG_LINUX_AIO
> +    if (raw_check_linux_aio(s) && laio_has_fdsync(s->fd)) {

FDSYNC support should be probed at open() time and the result should be
stored in a new bool field like s->laio_supports_fdsync. That way the
cost of laio_has_fdsync() on every flush request is avoided.

> +        return laio_co_submit(s->fd, 0, NULL, QEMU_AIO_FLUSH, 0);
> +    }
>  #endif
>      return raw_thread_pool_submit(handle_aiocb_flush, &acb);
>  }
> diff --git a/block/linux-aio.c b/block/linux-aio.c
> index ec05d946f3..ed20a503e9 100644
> --- a/block/linux-aio.c
> +++ b/block/linux-aio.c
> @@ -67,6 +67,7 @@ struct LinuxAioState {
>      int event_max;
>  };
>  
> +bool laio_has_fdsync(int);
>  static void ioq_submit(LinuxAioState *s);
>  
>  static inline ssize_t io_event_ret(struct io_event *ev)
> @@ -384,6 +385,9 @@ static int laio_do_submit(int fd, struct qemu_laiocb *laiocb, off_t offset,
>      case QEMU_AIO_READ:
>          io_prep_preadv(iocbs, fd, qiov->iov, qiov->niov, offset);
>          break;
> +    case QEMU_AIO_FLUSH:
> +        io_prep_fdsync(iocbs, fd);
> +        break;
>      /* Currently Linux kernel does not support other operations */
>      default:
>          fprintf(stderr, "%s: invalid AIO request type 0x%x.\n",
> @@ -412,7 +416,7 @@ int coroutine_fn laio_co_submit(int fd, uint64_t offset, QEMUIOVector *qiov,
>      AioContext *ctx = qemu_get_current_aio_context();
>      struct qemu_laiocb laiocb = {
>          .co         = qemu_coroutine_self(),
> -        .nbytes     = qiov->size,
> +        .nbytes     = qiov ? qiov->size : 0,
>          .ctx        = aio_get_linux_aio(ctx),
>          .ret        = -EINPROGRESS,
>          .is_read    = (type == QEMU_AIO_READ),
> @@ -486,3 +490,19 @@ void laio_cleanup(LinuxAioState *s)
>      }
>      g_free(s);
>  }
> +
> +bool laio_has_fdsync(int fd)
> +{
> +    struct iocb cb;
> +    struct iocb *cbs[] = {&cb, NULL};
> +
> +    io_context_t ctx = 0;
> +    io_setup(1, &ctx);
> +
> +    /* check if host kernel supports IO_CMD_FDSYNC */
> +    io_prep_fdsync(&cb, fd);
> +    int ret = io_submit(ctx, 1, cbs);
> +
> +    io_destroy(ctx);
> +    return (ret == -EINVAL) ? false : true;
> +}
> -- 
> 2.44.0
> 

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

  reply	other threads:[~2024-03-13 15:19 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-03-13  8:49 [PATCH v3] linux-aio: add IO_CMD_FDSYNC command support Prasad Pandit
2024-03-13 15:18 ` Stefan Hajnoczi [this message]
2024-03-13 17:19   ` Prasad Pandit
2024-03-13 17:56     ` Stefan Hajnoczi

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=20240313151839.GA545171@fedora \
    --to=stefanha@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=mtosatti@redhat.com \
    --cc=pjp@fedoraproject.org \
    --cc=ppandit@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.