From: Eric Blake <eblake@redhat.com>
To: Hanna Reitz <hreitz@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>,
Vladimir Sementsov-Ogievskiy <v.sementsov-og@mail.ru>,
qemu-devel@nongnu.org, qemu-block@nongnu.org
Subject: Re: [PATCH v2] block/stream: Drain subtree around graph change
Date: Fri, 25 Mar 2022 09:45:16 -0500 [thread overview]
Message-ID: <20220325144516.aueecyzlstclsxra@redhat.com> (raw)
In-Reply-To: <20220324140907.17192-1-hreitz@redhat.com>
On Thu, Mar 24, 2022 at 03:09:07PM +0100, Hanna Reitz wrote:
> When the stream block job cuts out the nodes between top and base in
> stream_prepare(), it does not drain the subtree manually; it fetches the
> base node, and tries to insert it as the top node's backing node with
> bdrv_set_backing_hd(). bdrv_set_backing_hd() however will drain, and so
> the actual base node might change (because the base node is actually not
> part of the stream job) before the old base node passed to
> bdrv_set_backing_hd() is installed.
>
...
>
> Therefore, we need to keep the whole subtree drained in
> stream_prepare(), so that the graph modification it performs is
> effectively atomic, i.e. that the base node it fetches is still the base
> node when bdrv_set_backing_hd() sets it as the top node's backing node.
>
> Verify this by asserting in said 030's test case that the base node is
> always the top node's immediate backing child when both jobs are done.
>
> Signed-off-by: Hanna Reitz <hreitz@redhat.com>
> ---
> v2:
> - Oops, the base can be NULL. Would have noticed if I had ran all test
> cases from 030, and not just test_overlapping_5()...
> That means that keeping a strong reference to the base node must be
> conditional, based on whether there even is a base node or not.
> (I mean, technically we do not even need to keep a strong reference to
> that node, given that we are in a drained section, but I believe it is
> better style to do it anyway.)
I agree with the conclusion that we don't need a strong reference once
you fix the bigger problem of the lock-by-drain, but that it is better
style to include it anyway.
> ---
> block/stream.c | 15 ++++++++++++++-
> tests/qemu-iotests/030 | 5 +++++
> 2 files changed, 19 insertions(+), 1 deletion(-)
Reviewed-by: Eric Blake <eblake@redhat.com>
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
next prev parent reply other threads:[~2022-03-25 14:46 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-24 14:09 [PATCH v2] block/stream: Drain subtree around graph change Hanna Reitz
2022-03-24 18:49 ` John Snow
2022-03-25 8:50 ` Hanna Reitz
2022-03-25 14:45 ` Eric Blake [this message]
2022-03-25 16:37 ` Vladimir Sementsov-Ogievskiy
2022-03-28 7:44 ` Hanna Reitz
2022-03-28 8:09 ` Hanna Reitz
2022-03-28 10:24 ` Vladimir Sementsov-Ogievskiy
2022-03-29 8:54 ` Hanna Reitz
2022-03-29 9:55 ` Vladimir Sementsov-Ogievskiy
2022-03-29 12:15 ` Hanna Reitz
2022-03-30 9:40 ` Emanuele Giuseppe Esposito
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=20220325144516.aueecyzlstclsxra@redhat.com \
--to=eblake@redhat.com \
--cc=hreitz@redhat.com \
--cc=kwolf@redhat.com \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=v.sementsov-og@mail.ru \
/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).