From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:49474)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <stefanha@redhat.com>) id 1UzgNj-0002Kq-Bj
	for qemu-devel@nongnu.org; Thu, 18 Jul 2013 01:04:01 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <stefanha@redhat.com>) id 1UzgNd-0002XC-Ce
	for qemu-devel@nongnu.org; Thu, 18 Jul 2013 01:03:59 -0400
Received: from mx1.redhat.com ([209.132.183.28]:34022)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <stefanha@redhat.com>) id 1UzgNd-0002X2-67
	for qemu-devel@nongnu.org; Thu, 18 Jul 2013 01:03:53 -0400
Date: Thu, 18 Jul 2013 13:03:30 +0800
From: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20130718050330.GA26971@stefanha-thinkpad.redhat.com>
References: <1373885705-13722-1-git-send-email-pl@kamp.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1373885705-13722-1-git-send-email-pl@kamp.de>
Subject: Re: [Qemu-devel] [PATCHv3] block-migration: efficiently encode zero
	blocks
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Peter Lieven <pl@kamp.de>
Cc: kwolf@redhat.com, pbonzini@redhat.com, owasserm@redhat.com, qemu-devel@nongnu.org, quintela@redhat.com

On Mon, Jul 15, 2013 at 12:55:05PM +0200, Peter Lieven wrote:
> @@ -114,16 +115,29 @@ static void blk_mig_unlock(void)
>  static void blk_send(QEMUFile *f, BlkMigBlock * blk)
>  {
>      int len;
> +    uint64_t flags = BLK_MIG_FLAG_DEVICE_BLOCK;
> +
> +    if (migrate_zero_blocks() && buffer_is_zero(blk->buf, BLOCK_SIZE)) {
[...]
> +bool migrate_zero_blocks(void)
> +{
> +    MigrationState *s;
> +
> +    s = migrate_get_current();
> +
> +    return s->enabled_capabilities[MIGRATION_CAPABILITY_ZERO_BLOCKS];
> +}

blk_send() is called without locks held.  It would be safer and cleaner
to stash bool migrate_zero_blocks in BlkMigBlock in init_blk_migration()
instead of accessing migrate_get_current() without locks held.

This eliminates the assumption that accessing migrate_get_current() is
safe without locks.

Besides this locking issue I'm happy with the code.

Stefan