From: Fam Zheng <famz@redhat.com>
To: Stefan Hajnoczi <stefanha@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>, Jeff Cody <jcody@redhat.com>,
qemu-devel@nongnu.org, Max Reitz <mreitz@redhat.com>,
vsementsov@parallels.com, John Snow <jsnow@redhat.com>
Subject: Re: [Qemu-devel] [PATCH v3 09/15] blockjob: Move BlockJobDeferToMainLoopData into BlockJob
Date: Tue, 14 Jul 2015 18:36:26 +0800 [thread overview]
Message-ID: <20150714103626.GG27873@ad.nay.redhat.com> (raw)
In-Reply-To: <20150714100311.GE17927@stefanha-thinkpad.redhat.com>
On Tue, 07/14 11:03, Stefan Hajnoczi wrote:
> On Fri, Jul 10, 2015 at 11:46:46AM +0800, Fam Zheng wrote:
> > diff --git a/blockjob.c b/blockjob.c
> > index 11b48f5..e057dd5 100644
> > --- a/blockjob.c
> > +++ b/blockjob.c
> > @@ -92,6 +92,10 @@ void block_job_completed(BlockJob *job, int ret)
> > assert(!job->completed);
> > job->completed = true;
> > job->ret = ret;
> > + if (job->defer_to_main_loop_data.bh) {
> > + qemu_bh_delete(job->defer_to_main_loop_data.bh);
> > + job->defer_to_main_loop_data.bh = NULL;
> > + }
> > job->cb(job->opaque, ret);
> > block_job_release(bs);
> > }
>
> This doesn't make sense to me:
>
> 1. This function is called from a defer_to_main_loop BH so
> job->defer_to_main_loop_data.bh should already be running here.
>
> In fact, we've already called qemu_bh_delete() in
> block_job_defer_to_main_loop_bh(). This call is pointless but you
> could change it to an assertion and assign bh = NULL in
> block_job_defer_to_main_loop_bh().
>
> 2. If there was a BH scheduled, why is it safe to silently delete it?
> Wouldn't the caller rely on the BH code executing to clean up, for
> example?
>
> If you drop this if statement, is it necessary to move
> BlockJobDeferToMainLoopData into BlockJob at all? Maybe you can just
> drop this patch.
You're right, I agree this patch is superfluous and can be dropped.
Fam
>
> > @@ -344,44 +348,36 @@ BlockErrorAction block_job_error_action(BlockJob *job, BlockDriverState *bs,
> > return action;
> > }
> >
> > -typedef struct {
> > - BlockJob *job;
> > - QEMUBH *bh;
> > - AioContext *aio_context;
> > - BlockJobDeferToMainLoopFn *fn;
> > - void *opaque;
> > -} BlockJobDeferToMainLoopData;
> > -
> > static void block_job_defer_to_main_loop_bh(void *opaque)
> > {
> > - BlockJobDeferToMainLoopData *data = opaque;
> > + BlockJob *job = opaque;
> > + /* Copy the struct in case job get released in data.fn() */
> > + BlockJobDeferToMainLoopData data = job->defer_to_main_loop_data;
>
> A comment is warranted since this access is only safe due to the
> following:
>
> The job may still be executing in another thread (the AioContext hasn't
> been acquired by us yet), but job->defer_to_main_loop_data isn't
> modified by the other thread after the qemu_bh_schedule() call.
>
> Additionally, the atomic_xchg memory barriers in qemu_bh_schedule() and
> aio_bh_poll() to ensure that this thread sees the latest
> job->defer_to_main_loop_data data.
>
> Access to other job fields would probably not be safe here!
next prev parent reply other threads:[~2015-07-14 10:36 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-07-10 3:46 [Qemu-devel] [PATCH v3 00/15] block: incremental backup transactions using BlockJobTxn Fam Zheng
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 01/15] qapi: Add transaction support to block-dirty-bitmap operations Fam Zheng
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 02/15] iotests: add transactional incremental backup test Fam Zheng
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 03/15] block: rename BlkTransactionState and BdrvActionOps Fam Zheng
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 04/15] block: keep bitmap if incremental backup job is cancelled Fam Zheng
2015-07-13 19:48 ` John Snow
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 05/15] backup: Extract dirty bitmap handling as a separate function Fam Zheng
2015-07-13 23:06 ` John Snow
2015-07-14 2:46 ` Fam Zheng
2015-07-14 8:26 ` Stefan Hajnoczi
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 06/15] blockjob: Add .txn_commit and .txn_abort transaction actions Fam Zheng
2015-07-13 23:06 ` John Snow
2015-07-14 8:35 ` Stefan Hajnoczi
2015-07-14 9:26 ` Fam Zheng
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 07/15] blockjob: Add "completed" and "ret" in BlockJob Fam Zheng
2015-07-13 23:08 ` John Snow
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 08/15] blockjob: Simplify block_job_finish_sync Fam Zheng
2015-07-13 23:08 ` John Snow
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 09/15] blockjob: Move BlockJobDeferToMainLoopData into BlockJob Fam Zheng
2015-07-14 10:03 ` Stefan Hajnoczi
2015-07-14 10:36 ` Fam Zheng [this message]
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 10/15] block: add block job transactions Fam Zheng
2015-07-13 23:12 ` John Snow
2015-07-14 3:04 ` Fam Zheng
2015-07-14 15:05 ` John Snow
2015-07-14 10:27 ` Stefan Hajnoczi
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 11/15] blockdev: make BlockJobTxn available to qmp 'transaction' Fam Zheng
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 12/15] block/backup: support block job transactions Fam Zheng
2015-07-13 23:14 ` John Snow
2015-07-14 3:12 ` Fam Zheng
2015-07-14 10:32 ` Stefan Hajnoczi
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 13/15] iotests: 124 - transactional failure test Fam Zheng
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 14/15] qmp-commands.hx: Update the supported 'transaction' operations Fam Zheng
2015-07-10 3:46 ` [Qemu-devel] [PATCH v3 15/15] tests: add BlockJobTxn unit test Fam Zheng
2015-07-13 23:14 ` John Snow
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=20150714103626.GG27873@ad.nay.redhat.com \
--to=famz@redhat.com \
--cc=jcody@redhat.com \
--cc=jsnow@redhat.com \
--cc=kwolf@redhat.com \
--cc=mreitz@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@redhat.com \
--cc=vsementsov@parallels.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.