From: Changlong Xie <xiecl.fnst@cn.fujitsu.com>
To: Stefan Hajnoczi <stefanha@gmail.com>
Cc: Kevin Wolf <kwolf@redhat.com>, Fam Zheng <famz@redhat.com>,
zhanghailiang <zhang.zhanghailiang@huawei.com>,
qemu block <qemu-block@nongnu.org>,
Jiang Yunhong <yunhong.jiang@intel.com>,
Dong Eddie <eddie.dong@intel.com>,
qemu devel <qemu-devel@nongnu.org>,
"Michael R. Hines" <mrhines@linux.vnet.ibm.com>,
Max Reitz <mreitz@redhat.com>, Gonglei <arei.gonglei@huawei.com>,
Stefan Hajnoczi <stefanha@redhat.com>,
Paolo Bonzini <pbonzini@redhat.com>,
"Dr. David Alan Gilbert" <dgilbert@redhat.com>
Subject: Re: [Qemu-devel] [PATCH v15 8/9] Implement new driver for block replication
Date: Wed, 9 Mar 2016 11:29:21 +0800 [thread overview]
Message-ID: <56DF9891.4010207@cn.fujitsu.com> (raw)
In-Reply-To: <20160304173954.GF9130@stefanha-x1.localdomain>
On 03/05/2016 01:39 AM, Stefan Hajnoczi wrote:
> On Fri, Feb 05, 2016 at 12:18:07PM +0800, Changlong Xie wrote:
>> +static void replication_start(ReplicationState *rs, ReplicationMode mode,
>> + Error **errp)
>> +{
>> + BlockDriverState *bs = rs->opaque;
>> + BDRVReplicationState *s = bs->opaque;
>> + int64_t active_length, hidden_length, disk_length;
>> + AioContext *aio_context;
>> + Error *local_err = NULL;
>> +
>> + if (s->replication_state != BLOCK_REPLICATION_NONE) {
>
> Dereferencing bs and s is not thread-safe since the AioContext isn't
> acquired here. Unless you have other locking rules, the default is that
> everything in BlockDriverState and bs->opaque must be accessed with the
> AioContext acquired.
>
> Please apply this comment to the rest of the patch series, too.
> Functions that are not part of BlockDriver are probably called outside
> the AioContext and must acquire it before they are allowed to access
> anything else in bs.
>
>> + s->top_bs = get_top_bs(bs);
>> + if (!s->top_bs) {
>> + error_setg(errp, "Cannot get the top block driver state to do"
>> + " internal backup");
>> + return;
>> + }
>
> Traversing the BDS graph using the parent pointer is not safe - you
> cannot stash the parent pointer because it changes if a parent node is
> inserted/removed.
>
> I suggest passing the drive name as an argument so that bdrv_lookup_bs()
> can be used when installing the op blocker. Then the BdrvChild.parent
> pointer patch and get_top_bs() can be deleted.
>
> Jeff Cody is currently working on a new op blocker system. Hopefully
> this system will allow you to install blockers on bs instead of on the
> drive's top BDS. But let's not worry about that for now and just use
> the drive name as an argument.
>
>> + /*
>> + * Must protect backup target if backup job was stopped/cancelled
>> + * unexpectedly
>> + */
>> + bdrv_ref(s->hidden_disk->bs);
>
> Where is the matching bdrv_unref() call?
Two conditions
1. job is cancelled, so "local_err != 0"
449 if (local_err) {
450 error_propagate(errp, local_err);
451 backup_job_cleanup(s);
452 bdrv_unref(s->hidden_disk->bs);
453 return;
454 }
2. backup completed
backup_complete
bdrv_unref(s->target);
If i'm wrong, pls correct me.
Thanks
-Xie
>
next prev parent reply other threads:[~2016-03-09 3:28 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-02-05 4:17 [Qemu-devel] [PATCH v15 0/9] Block replication for continuous checkpoints Changlong Xie
2016-02-05 4:18 ` [Qemu-devel] [PATCH v15 1/9] unblock backup operations in backing file Changlong Xie
2016-02-05 4:18 ` [Qemu-devel] [PATCH v15 2/9] Store parent BDS in BdrvChild Changlong Xie
2016-02-05 4:18 ` [Qemu-devel] [PATCH v15 3/9] Backup: clear all bitmap when doing block checkpoint Changlong Xie
2016-02-05 4:18 ` [Qemu-devel] [PATCH v15 4/9] Link backup into block core Changlong Xie
2016-02-05 4:18 ` [Qemu-devel] [PATCH v15 5/9] docs: block replication's description Changlong Xie
2016-02-05 4:18 ` [Qemu-devel] [PATCH v15 6/9] auto complete active commit Changlong Xie
2016-02-05 4:18 ` [Qemu-devel] [PATCH v15 7/9] Introduce new APIs to do replication operation Changlong Xie
2016-02-15 0:57 ` Hailiang Zhang
2016-02-15 1:13 ` Wen Congyang
2016-02-19 8:41 ` Hailiang Zhang
2016-02-19 8:55 ` Wen Congyang
2016-02-15 2:25 ` Changlong Xie
2016-03-04 16:13 ` [Qemu-devel] [Qemu-block] " Stefan Hajnoczi
2016-03-08 3:08 ` Changlong Xie
2016-02-05 4:18 ` [Qemu-devel] [PATCH v15 8/9] Implement new driver for block replication Changlong Xie
2016-03-04 17:39 ` Stefan Hajnoczi
2016-03-08 3:30 ` Changlong Xie
2016-03-09 3:29 ` Changlong Xie [this message]
2016-03-04 17:53 ` Stefan Hajnoczi
2016-03-08 3:31 ` Changlong Xie
2016-03-09 10:35 ` Changlong Xie
2016-02-05 4:18 ` [Qemu-devel] [PATCH v15 9/9] support replication driver in blockdev-add Changlong Xie
2016-02-15 7:02 ` [Qemu-devel] [PATCH v15 0/9] Block replication for continuous checkpoints Changlong Xie
2016-02-17 9:32 ` Hailiang Zhang
2016-02-23 9:10 ` Changlong Xie
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=56DF9891.4010207@cn.fujitsu.com \
--to=xiecl.fnst@cn.fujitsu.com \
--cc=arei.gonglei@huawei.com \
--cc=dgilbert@redhat.com \
--cc=eddie.dong@intel.com \
--cc=famz@redhat.com \
--cc=kwolf@redhat.com \
--cc=mreitz@redhat.com \
--cc=mrhines@linux.vnet.ibm.com \
--cc=pbonzini@redhat.com \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@gmail.com \
--cc=stefanha@redhat.com \
--cc=yunhong.jiang@intel.com \
--cc=zhang.zhanghailiang@huawei.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).