From: Stefan Priebe - Profihost AG <s.priebe@profihost.ag>
To: Stefan Hajnoczi <stefanha@gmail.com>
Cc: Josh Durgin <josh.durgin@inktank.com>,
qemu-devel@nongnu.org, Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [Qemu-devel] [PATCH] rbd block driver fix race between aio completition and aio cancel
Date: Thu, 22 Nov 2012 11:01:20 +0100 [thread overview]
Message-ID: <50ADF7F0.8030002@profihost.ag> (raw)
In-Reply-To: <20121121090751.GC17466@stefanha-thinkpad.redhat.com>
Hello,
i've send a new patch which hopefully cares about all your comments.
[PATCH] rbd block driver fix race between aio completition and aio cancel
Greets
Stefan
Am 21.11.2012 10:07, schrieb Stefan Hajnoczi:
> On Mon, Nov 19, 2012 at 09:39:45PM +0100, Stefan Priebe wrote:
>> @@ -376,9 +376,7 @@ static void qemu_rbd_complete_aio(RADOSCB *rcb)
>> RBDAIOCB *acb = rcb->acb;
>> int64_t r;
>>
>> - if (acb->cancelled) {
>> - qemu_vfree(acb->bounce);
>> - qemu_aio_release(acb);
>> + if (acb->bh) {
>> goto done;
>> }
>
> When is acb->bh != NULL here?
>
>>
>> @@ -406,9 +404,12 @@ static void qemu_rbd_complete_aio(RADOSCB *rcb)
>> acb->ret = r;
>> }
>> }
>> + acb->status = acb->ret;
>
> How about initializing acb->ret with -EINPROGRESS and using it instead
> of adding a new status field?
>
>> @@ -573,7 +574,10 @@ static void qemu_rbd_close(BlockDriverState *bs)
>> static void qemu_rbd_aio_cancel(BlockDriverAIOCB *blockacb)
>> {
>> RBDAIOCB *acb = (RBDAIOCB *) blockacb;
>> - acb->cancelled = 1;
>> +
>> + while (acb->status == -EINPROGRESS) {
>> + qemu_aio_wait();
>> + }
>> }
>
> Need to take care that acb is still valid (not yet released!) when the
> while loop iterates.
>
> One way of doing this is to mark the acb as cancelled so the completion
> handler won't release it. Then the cancellation function can release
> the acb - it's the last piece of code that needs a reference to acb. In
> this case the acb->cancelled field is useful.
>
>> @@ -642,10 +646,11 @@ static void rbd_aio_bh_cb(void *opaque)
>> qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
>> }
>> qemu_vfree(acb->bounce);
>> - acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
>> qemu_bh_delete(acb->bh);
>> acb->bh = NULL;
>>
>> + acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
>> +
>> qemu_aio_release(acb);
>> }
>>
>
> What does this hunk do?
>
next prev parent reply other threads:[~2012-11-22 10:01 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-11-19 20:39 [Qemu-devel] (no subject) Stefan Priebe
2012-11-19 20:39 ` [Qemu-devel] [PATCH] rbd block driver fix race between aio completition and aio cancel Stefan Priebe
2012-11-21 8:56 ` Josh Durgin
2012-11-21 9:07 ` Stefan Hajnoczi
2012-11-21 9:19 ` Stefan Priebe - Profihost AG
2012-11-21 15:34 ` Paolo Bonzini
2012-11-22 10:01 ` Stefan Priebe - Profihost AG [this message]
-- strict thread matches above, loose matches on Subject: below --
2012-11-22 10:00 Stefan Priebe
2012-11-24 19:54 ` Blue Swirl
2012-11-24 20:21 ` Stefan Priebe
2012-11-27 22:42 ` Josh Durgin
2012-11-29 15:24 ` Paolo Bonzini
2012-11-29 13:58 ` Stefan Hajnoczi
2012-11-29 14:32 ` Stefan Priebe - Profihost AG
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=50ADF7F0.8030002@profihost.ag \
--to=s.priebe@profihost.ag \
--cc=josh.durgin@inktank.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@gmail.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).