From: Max Reitz <mreitz@redhat.com>
To: qemu-devel@nongnu.org
Cc: Kevin Wolf <kwolf@redhat.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Stefan Hajnoczi <stefanha@redhat.com>
Subject: Re: [Qemu-devel] [PATCH v2 07/22] block: Add bdrv_close_all() handlers
Date: Fri, 13 Feb 2015 18:02:39 -0500 [thread overview]
Message-ID: <54DE828F.7040604@redhat.com> (raw)
In-Reply-To: <1423507124-29809-8-git-send-email-mreitz@redhat.com>
On 2015-02-09 at 13:38, Max Reitz wrote:
> Every time a reference to a BlockBackend is taken, a notifier for
> bdrv_close_all() has to be deposited so the reference holder can
> relinquish its reference when bdrv_close_all() is called. That notifier
> should be revoked on a bdrv_unref() call.
>
> Add a Notifier * parameter to all the functions changing the reference
> count of a BlockBackend: blk_new(), blk_new_with_bs(), blk_new_open(),
> blk_ref() and blk_unref().
>
> By dropping the manual reference handling in hw/block/xen_disk.c, the
> blk_unref() in hw/ide/piix.c can be dropped as well.
>
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
> block/block-backend.c | 69 ++++++++++++++++++++++++++++++++++++------
> blockdev.c | 62 ++++++++++++++++++++++++++++++++++---
> device-hotplug.c | 2 +-
> hw/block/xen_disk.c | 16 +++++++---
> hw/ide/piix.c | 1 -
> include/sysemu/block-backend.h | 13 +++++---
> include/sysemu/blockdev.h | 3 ++
> nbd.c | 26 ++++++++++++++--
> qemu-img.c | 39 ++++++++++++------------
> qemu-io.c | 6 ++--
> qemu-nbd.c | 6 ++--
> 11 files changed, 189 insertions(+), 54 deletions(-)
[snip]
> diff --git a/qemu-nbd.c b/qemu-nbd.c
> index ac1e5d6..5e6e4b4 100644
> --- a/qemu-nbd.c
> +++ b/qemu-nbd.c
> @@ -694,7 +694,8 @@ int main(int argc, char **argv)
> }
>
> srcpath = argv[optind];
> - blk = blk_new_open("hda", srcpath, NULL, options, flags, &local_err);
> + /* the reference will be passed on nbd_export_new() */
> + blk = blk_new_open("hda", srcpath, NULL, options, flags, NULL, &local_err);
> if (!blk) {
> errx(EXIT_FAILURE, "Failed to blk_new_open '%s': %s", argv[optind],
> error_get_pretty(local_err));
> @@ -728,7 +729,9 @@ int main(int argc, char **argv)
> }
> }
>
> + /* nbd_export_new() takes the reference to blk */
> exp = nbd_export_new(blk, dev_offset, fd_size, nbdflags, nbd_export_closed);
> + blk_unref(blk, NULL);
>
> if (sockpath) {
> fd = unix_socket_incoming(sockpath);
> @@ -773,7 +776,6 @@ int main(int argc, char **argv)
> }
> } while (state != TERMINATED);
>
> - blk_unref(blk);
> if (sockpath) {
> unlink(sockpath);
> }
Because qemu-nbd does not install a handler for relinquishing its
reference, it has to make sure it drops that reference before a
bdrv_close_all() occurs. Therefore, err() and errx() calls after
blk_new_open() (except for the if (!blk) path) and before blk_unref(blk,
NULL) need to be preceded by a blk_unref(blk, NULL).
I'll fix it in v3.
Max
next prev parent reply other threads:[~2015-02-13 23:02 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-02-09 18:38 [Qemu-devel] [PATCH v2 00/22] block: Rework bdrv_close_all() Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 01/22] iotests: Move _filter_nbd into common.filter Max Reitz
2015-02-16 19:31 ` Eric Blake
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 02/22] iotests: Do not redirect qemu's stderr Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 03/22] iotests: Add test for eject under NBD server Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 04/22] quorum: Fix close path Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 05/22] block: Move BDS close notifiers into BB Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 06/22] block: Add bdrv_close_all() notifiers Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 07/22] block: Add bdrv_close_all() handlers Max Reitz
2015-02-13 23:02 ` Max Reitz [this message]
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 08/22] block: Use blk_remove_bs() in blk_delete() Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 09/22] blockdev: Use blk_remove_bs() in do_drive_del() Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 10/22] block: Make bdrv_close() static Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 11/22] block: Add blk_name_taken() Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 12/22] block: Add blk_next_inserted() Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 13/22] block: Add blk_commit_all() and blk_invalidate_cache_all() Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 14/22] block: Use BlockBackend more Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 15/22] blockdev: Add list of monitor-owned BlockBackends Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 16/22] blockdev: Remove blk_hide_on_behalf_of_do_drive_del() Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 17/22] block: Make bdrv_drain_one() public Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 18/22] block: Move some bdrv_*_all() functions to BB Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 19/22] block: Remove bdrv_states Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 20/22] blockdev: Keep track of monitor-owned BDS Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 21/22] block: Strip down bdrv_close_all() Max Reitz
2015-02-09 18:38 ` [Qemu-devel] [PATCH v2 22/22] iotests: Add test for multiple BB on BDS tree Max Reitz
2015-02-10 8:23 ` [Qemu-devel] [PATCH v2 00/22] block: Rework bdrv_close_all() Paolo Bonzini
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=54DE828F.7040604@redhat.com \
--to=mreitz@redhat.com \
--cc=kwolf@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@redhat.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).