qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Stefan Hajnoczi <stefanha@gmail.com>
To: Dietmar Maurer <dietmar@proxmox.com>
Cc: kwolf@redhat.com, qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH 2/5] add basic backup support to block driver
Date: Thu, 22 Nov 2012 12:25:57 +0100	[thread overview]
Message-ID: <20121122112557.GI7598@stefanha-thinkpad.redhat.com> (raw)
In-Reply-To: <1353488464-82756-2-git-send-email-dietmar@proxmox.com>

On Wed, Nov 21, 2012 at 10:01:01AM +0100, Dietmar Maurer wrote:
Two comments from skimming the code, not a full review.

> +/* #define DEBUG_BACKUP */
> +
> +#ifdef DEBUG_BACKUP
> +#define DPRINTF(fmt, ...) \
> +    do { printf("backup: " fmt, ## __VA_ARGS__); } while (0)
> +#else
> +#define DPRINTF(fmt, ...) \
> +    do { } while (0)
> +#endif

Please don't #define debug macros like this.  It hides build breakage
because the debug code is #ifdef'd out.

Either use docs/tracing.txt or do the following so the code always gets
parsed by the compiler:

#define DEBUG_BACKUP 0
#define DPRINTF(fmt, ...) \
    do { \
        if (DEBUG_BACKUP) { \
            printf("backup: " fmt, ## __VA_ARGS__); \
	} \
    } while (0)

> +BackupInfo *
> +bdrv_backup_init(BlockDriverState *bs, BackupDumpFunc *backup_dump_cb,
> +                 BlockDriverCompletionFunc *backup_complete_cb,
> +                 void *opaque)
> +{
> +    assert(bs);
> +    assert(backup_dump_cb);
> +    assert(backup_complete_cb);
> +
> +    if (bs->backup_info) {
> +        DPRINTF("bdrv_backup_init already initialized %s\n",
> +                bdrv_get_device_name(bs));
> +        return NULL;
> +    }
> +
> +    BackupInfo *info = g_malloc0(sizeof(BackupInfo));
> +    int64_t bitmap_size;
> +    const char *devname = bdrv_get_device_name(bs);
> +
> +    if (!devname || !devname[0]) {
> +        return NULL;
> +    }
> +
> +    DPRINTF("bdrv_backup_init %s\n", bdrv_get_device_name(bs));
> +
> +    bitmap_size = bs->total_sectors +
> +        BACKUP_BLOCKS_PER_CLUSTER * BITS_PER_LONG - 1;
> +    bitmap_size /= BACKUP_BLOCKS_PER_CLUSTER * BITS_PER_LONG;
> +
> +    info->backup_dump_cb = backup_dump_cb;
> +    info->backup_complete_cb = backup_complete_cb;
> +    info->opaque = opaque;
> +    info->bitmap_size = bitmap_size;
> +    info->bitmap = g_new0(unsigned long, bitmap_size);
> +
> +    Error *errp;
> +    BackupBlockJob *job = block_job_create(&backup_job_type, bs, 0,
> +                                           backup_job_cleanup_cb, bs, &errp);

Is there a 1:1 relationship between BackupInfo and BackupBlockJob?  Then
it would be nicer to move BlockupInfo fields into BackupBlockJob (which
is empty right now).  Then you also don't need to add fields to
BlockDriverState because you know that if your blockjob is running you
can access it via bs->job, if necessary.  You can then drop BackupInfo
and any memory management code for it.

  reply	other threads:[~2012-11-22 11:26 UTC|newest]

Thread overview: 73+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-11-21  9:01 [Qemu-devel] [PATCH 1/5] RFC: Efficient VM backup for qemu (v1) Dietmar Maurer
2012-11-21  9:01 ` [Qemu-devel] [PATCH 2/5] add basic backup support to block driver Dietmar Maurer
2012-11-22 11:25   ` Stefan Hajnoczi [this message]
2012-11-22 11:29     ` Dietmar Maurer
2012-11-23  8:56     ` Dietmar Maurer
2012-11-21  9:01 ` [Qemu-devel] [PATCH 3/5] introduce new vma archive format Dietmar Maurer
2012-11-21 16:06   ` Eric Blake
2012-11-21 17:56     ` Dietmar Maurer
2012-11-21  9:01 ` [Qemu-devel] [PATCH 4/5] add backup related monitor commands Dietmar Maurer
2012-11-21 16:16   ` Eric Blake
2012-11-21 17:59     ` Dietmar Maurer
2012-11-21 18:49     ` Dietmar Maurer
2012-11-21  9:01 ` [Qemu-devel] [PATCH 5/5] add regression tests for backup Dietmar Maurer
2012-11-21 10:48 ` [Qemu-devel] [PATCH 1/5] RFC: Efficient VM backup for qemu (v1) Kevin Wolf
2012-11-21 11:10   ` Dietmar Maurer
2012-11-21 12:37     ` Kevin Wolf
2012-11-21 13:23       ` Paolo Bonzini
2012-11-23  7:42         ` Dietmar Maurer
2012-11-23  9:18           ` Paolo Bonzini
2012-11-23  9:28             ` Dietmar Maurer
2012-11-23  8:12         ` Dietmar Maurer
2012-11-23  9:01           ` Dietmar Maurer
2012-11-23  9:05             ` Dietmar Maurer
2012-11-23  9:15               ` Paolo Bonzini
2012-11-23  9:17                 ` Dietmar Maurer
2012-11-23  9:55               ` Kevin Wolf
2012-11-23 10:55                 ` Markus Armbruster
2012-11-21 13:25       ` Dietmar Maurer
2012-11-21 13:58         ` Kevin Wolf
2012-11-21 15:47           ` Dietmar Maurer
2012-11-23  7:38       ` Dietmar Maurer
2012-11-23  9:08         ` Kevin Wolf
2012-11-23  9:21           ` Dietmar Maurer
2012-11-23  9:31             ` Dietmar Maurer
2012-11-23 10:29               ` Kevin Wolf
2012-11-26  5:51                 ` Dietmar Maurer
2012-11-26 12:07                   ` Paolo Bonzini
2012-11-27  6:20                     ` Dietmar Maurer
2012-11-27  7:15                       ` Dietmar Maurer
2012-11-27  8:48                         ` Kevin Wolf
2012-11-27 10:24                           ` Dietmar Maurer
2012-11-21 11:23   ` Dietmar Maurer
2012-11-22 11:12 ` Stefan Hajnoczi
2012-11-22 11:26   ` Dietmar Maurer
2012-11-22 12:44     ` Stefan Hajnoczi
2012-11-22 12:55       ` Dietmar Maurer
2012-11-22 15:30         ` Stefan Hajnoczi
2012-11-22 15:58           ` Dietmar Maurer
2012-11-22 17:02             ` Stefan Hajnoczi
2012-11-22 17:34               ` Dietmar Maurer
2012-11-22 11:40   ` Dietmar Maurer
2012-11-22 15:42     ` Stefan Hajnoczi
2012-11-22 12:00   ` Dietmar Maurer
2012-11-22 15:45     ` Stefan Hajnoczi
2012-11-22 15:56       ` Dietmar Maurer
2012-11-22 16:37         ` Stefan Hajnoczi
2012-11-22 12:03   ` Dietmar Maurer
2012-11-22 17:16   ` Stefan Hajnoczi
2012-11-22 17:46     ` Dietmar Maurer
2012-11-23  5:23       ` Stefan Hajnoczi
2012-11-23  5:25         ` Stefan Hajnoczi
2012-11-23  6:18           ` Dietmar Maurer
2012-11-23  6:13         ` Dietmar Maurer
2012-11-22 17:50     ` Dietmar Maurer
2012-11-23  5:21       ` Stefan Hajnoczi
2012-11-22 18:05     ` Dietmar Maurer
2012-11-23  5:19       ` Stefan Hajnoczi
2012-11-23  6:05         ` Dietmar Maurer
2012-11-22 18:15     ` Dietmar Maurer
2012-11-27 10:09 ` Wenchao Xia
2012-11-27 10:37   ` Dietmar Maurer
2012-11-28  9:39     ` Wenchao Xia
2012-11-28 11:08       ` Dietmar Maurer

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=20121122112557.GI7598@stefanha-thinkpad.redhat.com \
    --to=stefanha@gmail.com \
    --cc=dietmar@proxmox.com \
    --cc=kwolf@redhat.com \
    --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 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).