From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [140.186.70.92] (port=51946 helo=eggs.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1OMkBp-0004br-EQ for qemu-devel@nongnu.org; Thu, 10 Jun 2010 12:01:10 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.69) (envelope-from ) id 1OMkBo-000104-8V for qemu-devel@nongnu.org; Thu, 10 Jun 2010 12:01:09 -0400 Received: from mx1.redhat.com ([209.132.183.28]:14537) by eggs.gnu.org with esmtp (Exim 4.69) (envelope-from ) id 1OMkBo-0000zg-0A for qemu-devel@nongnu.org; Thu, 10 Jun 2010 12:01:08 -0400 Received: from int-mx01.intmail.prod.int.phx2.redhat.com (int-mx01.intmail.prod.int.phx2.redhat.com [10.5.11.11]) by mx1.redhat.com (8.13.8/8.13.8) with ESMTP id o5AG14D5028675 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Thu, 10 Jun 2010 12:01:05 -0400 From: Markus Armbruster References: <1275497729-13120-1-git-send-email-armbru@redhat.com> <1275497729-13120-8-git-send-email-armbru@redhat.com> <4C10F478.6040905@redhat.com> Date: Thu, 10 Jun 2010 18:00:59 +0200 In-Reply-To: <4C10F478.6040905@redhat.com> (Kevin Wolf's message of "Thu, 10 Jun 2010 16:19:36 +0200") Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Subject: [Qemu-devel] Re: [PATCH 07/13] blockdev: Means to destroy blockdev only if made with drive_init() List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Kevin Wolf Cc: qemu-devel@nongnu.org, kraxel@redhat.com Kevin Wolf writes: > Am 02.06.2010 18:55, schrieb Markus Armbruster: >> All drives are still made that way. They get destroyed along with >> their device. That's inappropriate for the alternative way to make >> blockdevs that will appear later in this series. These won't have a >> DriveInfo. >> >> blockdev_detach() destroys the blockdev only if it has a DriveInfo. >> >> blockdev_attach() does nothing for now. It'll be fleshed out later. >> >> Signed-off-by: Markus Armbruster >> --- >> blockdev.c | 35 +++++++++++++++++++++++++++++++++++ >> blockdev.h | 7 +++++++ >> 2 files changed, 42 insertions(+), 0 deletions(-) >> >> diff --git a/blockdev.c b/blockdev.c >> index ace74e4..f90d4fc 100644 >> --- a/blockdev.c >> +++ b/blockdev.c >> @@ -1,8 +1,12 @@ >> /* >> * QEMU host block devices >> * >> + * Copyright (C) 2010 Red Hat Inc. >> * Copyright (c) 2003-2008 Fabrice Bellard >> * >> + * Authors: >> + * Markus Armbruster , >> + * >> * This work is licensed under the terms of the GNU GPL, version 2 or >> * later. See the COPYING file in the top-level directory. >> */ >> @@ -17,6 +21,37 @@ >> >> static QTAILQ_HEAD(drivelist, DriveInfo) drives = QTAILQ_HEAD_INITIALIZER(drives); >> >> +static int blockdev_del_dinfo(BlockDriverState *bs) >> +{ >> + DriveInfo *dinfo, *next_dinfo; >> + int res = 0; >> + >> + QTAILQ_FOREACH_SAFE(dinfo, &drives, next, next_dinfo) { >> + if (dinfo->bdrv == bs) { >> + qemu_opts_del(dinfo->opts); >> + QTAILQ_REMOVE(&drives, dinfo, next); >> + qemu_free(dinfo); >> + res = 1; >> + } >> + } >> + >> + return res; > > Can it happen that a BlockDriverState belongs to multiple DriveInfos? If > no, why not returning in the loop? Wouldn't need a FOREACH_SAFE then, too. No, that shouldn't happen. Defensive coding, I don't want to leave dinfos with dangling dinfo->bdrv around. Maybe I should put an assert(!res) before the qemu_opts_del(). Or just forget about it, and simplify like you suggest. > It's not worth respinning because of this one, but there were more > comments and I think you'll send a v2 for the actual -blockdev option > anyway once we have decided how to do it. > > I have applied patches 1 to 6 now, and I think I could safely go on > until patch 9 if the minor improvements that were mentioned in comments > are made. I'd ignore patch 10 to 13 for now. > > Is this what you would have expected or should I handle anything in a > different way? No, that suits me fine. I definitely need to respin from part 8 on (commit message too terse).