qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
To: "Daniel P. Berrangé" <berrange@redhat.com>
Cc: qemu-devel@nongnu.org, qemu-block@nongnu.org,
	Stefan Hajnoczi <stefanha@redhat.com>,
	Hailiang Zhang <zhang.zhanghailiang@huawei.com>,
	Juan Quintela <quintela@redhat.com>, Fam Zheng <fam@euphon.net>
Subject: Re: [PATCH 17/20] migration: remove the QEMUFileOps 'get_buffer' callback
Date: Thu, 9 Jun 2022 17:46:29 +0100	[thread overview]
Message-ID: <YqIj5YNn0XZZqyIw@work-vm> (raw)
In-Reply-To: <20220524110235.145079-18-berrange@redhat.com>

* Daniel P. Berrangé (berrange@redhat.com) wrote:
> This directly implements the get_buffer logic using QIOChannel APIs.
> 
> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
> ---
>  migration/qemu-file-channel.c | 29 -----------------------------
>  migration/qemu-file.c         | 18 ++++++++++++++++--
>  migration/qemu-file.h         |  9 ---------
>  3 files changed, 16 insertions(+), 40 deletions(-)
> 
> diff --git a/migration/qemu-file-channel.c b/migration/qemu-file-channel.c
> index 8ff58e81f9..7b32831752 100644
> --- a/migration/qemu-file-channel.c
> +++ b/migration/qemu-file-channel.c
> @@ -74,34 +74,6 @@ static ssize_t channel_writev_buffer(void *opaque,
>  }
>  
>  
> -static ssize_t channel_get_buffer(void *opaque,
> -                                  uint8_t *buf,
> -                                  int64_t pos,
> -                                  size_t size,
> -                                  Error **errp)
> -{
> -    QIOChannel *ioc = QIO_CHANNEL(opaque);
> -    ssize_t ret;
> -
> -    do {
> -        ret = qio_channel_read(ioc, (char *)buf, size, errp);
> -        if (ret < 0) {
> -            if (ret == QIO_CHANNEL_ERR_BLOCK) {
> -                if (qemu_in_coroutine()) {
> -                    qio_channel_yield(ioc, G_IO_IN);
> -                } else {
> -                    qio_channel_wait(ioc, G_IO_IN);
> -                }
> -            } else {
> -                return -EIO;
> -            }
> -        }
> -    } while (ret == QIO_CHANNEL_ERR_BLOCK);
> -
> -    return ret;
> -}
> -
> -
>  static QEMUFile *channel_get_input_return_path(void *opaque)
>  {
>      QIOChannel *ioc = QIO_CHANNEL(opaque);
> @@ -117,7 +89,6 @@ static QEMUFile *channel_get_output_return_path(void *opaque)
>  }
>  
>  static const QEMUFileOps channel_input_ops = {
> -    .get_buffer = channel_get_buffer,
>      .get_return_path = channel_get_input_return_path,
>  };
>  
> diff --git a/migration/qemu-file.c b/migration/qemu-file.c
> index a855ce33dc..e024b43851 100644
> --- a/migration/qemu-file.c
> +++ b/migration/qemu-file.c
> @@ -374,8 +374,22 @@ static ssize_t qemu_fill_buffer(QEMUFile *f)
>          return 0;
>      }
>  
> -    len = f->ops->get_buffer(f->ioc, f->buf + pending, f->total_transferred,
> -                             IO_BUF_SIZE - pending, &local_error);
> +    do {
> +        len = qio_channel_read(f->ioc,

Yes, I think that's OK - not that 'len' is an int where 'ret'
was a ssize_t; but I think our buffers are guranteed to be 'small'.


Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

> +                               (char *)f->buf + pending,
> +                               IO_BUF_SIZE - pending,
> +                               &local_error);
> +        if (len == QIO_CHANNEL_ERR_BLOCK) {
> +            if (qemu_in_coroutine()) {
> +                qio_channel_yield(f->ioc, G_IO_IN);
> +            } else {
> +                qio_channel_wait(f->ioc, G_IO_IN);
> +            }
> +        } else if (len < 0) {
> +            len = EIO;
> +        }
> +    } while (len == QIO_CHANNEL_ERR_BLOCK);
> +
>      if (len > 0) {
>          f->buf_size += len;
>          f->total_transferred += len;
> diff --git a/migration/qemu-file.h b/migration/qemu-file.h
> index 7ec105bf96..cd49184c00 100644
> --- a/migration/qemu-file.h
> +++ b/migration/qemu-file.h
> @@ -29,14 +29,6 @@
>  #include "exec/cpu-common.h"
>  #include "io/channel.h"
>  
> -/* Read a chunk of data from a file at the given position.  The pos argument
> - * can be ignored if the file is only be used for streaming.  The number of
> - * bytes actually read should be returned.
> - */
> -typedef ssize_t (QEMUFileGetBufferFunc)(void *opaque, uint8_t *buf,
> -                                        int64_t pos, size_t size,
> -                                        Error **errp);
> -
>  /*
>   * This function writes an iovec to file. The handler must write all
>   * of the data or return a negative errno value.
> @@ -77,7 +69,6 @@ typedef size_t (QEMURamSaveFunc)(QEMUFile *f,
>  typedef QEMUFile *(QEMURetPathFunc)(void *opaque);
>  
>  typedef struct QEMUFileOps {
> -    QEMUFileGetBufferFunc *get_buffer;
>      QEMUFileWritevBufferFunc *writev_buffer;
>      QEMURetPathFunc *get_return_path;
>  } QEMUFileOps;
> -- 
> 2.36.1
> 
-- 
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK



  reply	other threads:[~2022-06-09 18:54 UTC|newest]

Thread overview: 51+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-05-24 11:02 [PATCH 00/20] migration: remove QEMUFileOps concept and assume use of QIOChannel Daniel P. Berrangé
2022-05-24 11:02 ` [PATCH 01/20] io: add a QIOChannelNull equivalent to /dev/null Daniel P. Berrangé
2022-05-24 21:14   ` Eric Blake
2022-06-16 16:26     ` Daniel P. Berrangé
2022-05-24 11:02 ` [PATCH 02/20] migration: switch to use QIOChannelNull for dummy channel Daniel P. Berrangé
2022-05-24 21:15   ` Eric Blake
2022-05-24 11:02 ` [PATCH 03/20] migration: remove unreachble RDMA code in save_hook impl Daniel P. Berrangé
2022-05-25 12:29   ` Eric Blake
2022-06-08 17:59   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 04/20] migration: rename rate limiting fields in QEMUFile Daniel P. Berrangé
2022-06-09  9:29   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 05/20] migration: rename 'pos' field in QEMUFile to 'bytes_processed' Daniel P. Berrangé
2022-06-09  9:51   ` Dr. David Alan Gilbert
2022-06-09  9:57     ` Daniel P. Berrangé
2022-06-09  9:59       ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 06/20] migration: rename qemu_ftell to qemu_file_total_transferred Daniel P. Berrangé
2022-06-09 10:23   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 07/20] migration: rename qemu_update_position to qemu_file_credit_transfer Daniel P. Berrangé
2022-06-09 10:29   ` Dr. David Alan Gilbert
2022-06-09 12:56   ` Peter Maydell
2022-06-09 13:02     ` Daniel P. Berrangé
2022-06-09 13:15       ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 08/20] migration: introduce a QIOChannel impl for BlockDriverState VMState Daniel P. Berrangé
2022-05-24 11:02 ` [PATCH 09/20] migration: convert savevm to use QIOChannelBlock for VMState Daniel P. Berrangé
2022-06-09 14:57   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 10/20] migration: stop passing 'opaque' parameter to QEMUFile hooks Daniel P. Berrangé
2022-06-09 15:00   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 11/20] migration: hardcode assumption that QEMUFile is backed with QIOChannel Daniel P. Berrangé
2022-06-09 15:01   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 12/20] migration: introduce new constructors for QEMUFile Daniel P. Berrangé
2022-06-09 15:33   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 13/20] migration: remove unused QEMUFileGetFD typedef Daniel P. Berrangé
2022-06-09 16:03   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 14/20] migration: remove the QEMUFileOps 'shut_down' callback Daniel P. Berrangé
2022-06-09 16:12   ` Dr. David Alan Gilbert
2022-06-09 16:14     ` Daniel P. Berrangé
2022-06-09 16:17       ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 15/20] migration: remove the QEMUFileOps 'set_blocking' callback Daniel P. Berrangé
2022-06-09 16:21   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 16/20] migration: remove the QEMUFileOps 'close' callback Daniel P. Berrangé
2022-06-09 16:40   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 17/20] migration: remove the QEMUFileOps 'get_buffer' callback Daniel P. Berrangé
2022-06-09 16:46   ` Dr. David Alan Gilbert [this message]
2022-06-09 17:09     ` Daniel P. Berrangé
2022-05-24 11:02 ` [PATCH 18/20] migration: remove the QEMUFileOps 'writev_buffer' callback Daniel P. Berrangé
2022-06-09 16:51   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 19/20] migration: remove the QEMUFileOps 'get_return_path' callback Daniel P. Berrangé
2022-06-09 16:54   ` Dr. David Alan Gilbert
2022-05-24 11:02 ` [PATCH 20/20] migration: remove the QEMUFileOps abstraction Daniel P. Berrangé
2022-06-09 16:59   ` Dr. David Alan Gilbert
2022-06-09 17:10     ` Daniel P. Berrangé

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=YqIj5YNn0XZZqyIw@work-vm \
    --to=dgilbert@redhat.com \
    --cc=berrange@redhat.com \
    --cc=fam@euphon.net \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=quintela@redhat.com \
    --cc=stefanha@redhat.com \
    --cc=zhang.zhanghailiang@huawei.com \
    /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;
as well as URLs for NNTP newsgroup(s).