qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Denis V. Lunev" <den@openvz.org>
To: Paolo Bonzini <pbonzini@redhat.com>,
	Stefan Hajnoczi <stefanha@gmail.com>
Cc: qemu-devel@nongnu.org, Stefan Hajnoczi <stefanha@redhat.com>,
	qemu-stable@nongnu.org
Subject: Re: [Qemu-devel] [PATCH 1/4] fifolock: create rfifolock_is_locked helper
Date: Mon, 2 Nov 2015 17:02:18 +0300	[thread overview]
Message-ID: <56376CEA.5060301@openvz.org> (raw)
In-Reply-To: <56376B57.3090806@redhat.com>

On 11/02/2015 04:55 PM, Paolo Bonzini wrote:
>
> On 02/11/2015 14:39, Denis V. Lunev wrote:
>>> This is thread-safe:
>>>
>>> bool owner;
>>>
>>> qemu_mutex_lock(&r->lock);
>>> owner = r->nesting > 0 && qemu_thread_is_self(&r->owner_thread);
>>> qemu_mutex_unlock(&r->lock);
>>>
>>> return owner;
>> yep, I know.
>>
>> But I do not want to take the lock for check.
> You can use a trylock.  If it fails, it is definitely safe to return false.
>
>> IMHO it would be better to
>>
>> @@ -68,11 +68,16 @@ void rfifolock_lock(RFifoLock *r)
>>   void rfifolock_unlock(RFifoLock *r)
>>   {
>>       qemu_mutex_lock(&r->lock);
>> -    assert(r->nesting > 0);
>> -    assert(qemu_thread_is_self(&r->owner_thread));
>> +    assert(rfifolock_is_owner(r));
>>       if (--r->nesting == 0) {
>> +        qemu_thread_clear(&r->owner_thread);
>>           r->head++;
>>           qemu_cond_broadcast(&r->cond);
>>       }
>>       qemu_mutex_unlock(&r->lock);
>>   }
>> +
>> +bool rfifolock_is_owner(RFifoLock *r)
>> +{
>> +    return r->nesting > 0 && qemu_thread_is_self(&r->owner_thread);
>> +}
>>
>> which does not require lock and thread safe.
> I think it requires memory barriers though.  But it can be simplified:
> if you clear owner_thread, you do not need to check r->nesting in
> rfifolock_is_owner.
>
> Clearing owner_thread can be done with a simple memset.
this does not require memory barrier as call to qemu_call_broadcast
will do the job just fine.

The check for ownership is actually enough as:
- current thread could be set in the current thread only and
   cleared in the same current thread only. This is not racy at all :)
- count check is moved here for convenience only by request of
   Stefan, it is not required at all to make a decision with after
   clearing the thread.

OK, I can use memset for sure if it will be accepted :)
This was my first opinion.

Den

  reply	other threads:[~2015-11-02 14:02 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-10-28 15:01 [Qemu-devel] [PATCH v3 0/4] dataplane snapshot fixes Denis V. Lunev
2015-10-28 15:01 ` [Qemu-devel] [PATCH 1/4] fifolock: create rfifolock_is_locked helper Denis V. Lunev
2015-10-30 15:41   ` Stefan Hajnoczi
2015-10-30 20:30     ` Denis V. Lunev
2015-11-02 13:10       ` Stefan Hajnoczi
2015-11-01 13:55     ` Denis V. Lunev
2015-11-02 13:12       ` Stefan Hajnoczi
2015-11-02 13:39         ` Denis V. Lunev
2015-11-02 13:55           ` Paolo Bonzini
2015-11-02 14:02             ` Denis V. Lunev [this message]
2015-10-28 15:01 ` [Qemu-devel] [PATCH 2/4] aio_context: create aio_context_is_locked helper Denis V. Lunev
2015-10-30 15:42   ` Stefan Hajnoczi
2015-10-28 15:01 ` [Qemu-devel] [PATCH 3/4] io: add locking constraints check into bdrv_drain to ensure locking Denis V. Lunev
2015-10-30 15:43   ` Stefan Hajnoczi
2015-10-28 15:01 ` [Qemu-devel] [PATCH 4/4] migration: add missed aio_context_acquire into HMP snapshot code Denis V. Lunev
2015-10-28 15:33   ` Juan Quintela
2015-10-28 15:57     ` Denis V. Lunev
2015-10-30 15:52   ` Stefan Hajnoczi
2015-11-03 14:48     ` Juan Quintela
2015-11-03 15:30       ` Stefan Hajnoczi
2015-11-03 15:36         ` Denis V. Lunev

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=56376CEA.5060301@openvz.org \
    --to=den@openvz.org \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-stable@nongnu.org \
    --cc=stefanha@gmail.com \
    --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).