From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([208.118.235.92]:46544)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1TGtNo-0006cC-O7
	for qemu-devel@nongnu.org; Wed, 26 Sep 2012 11:18:44 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1TGtNi-0004Is-Vu
	for qemu-devel@nongnu.org; Wed, 26 Sep 2012 11:18:40 -0400
Received: from mx1.redhat.com ([209.132.183.28]:16439)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1TGtNi-0004IZ-Nq
	for qemu-devel@nongnu.org; Wed, 26 Sep 2012 11:18:34 -0400
Message-ID: <50631CC5.5090504@redhat.com>
Date: Wed, 26 Sep 2012 17:18:29 +0200
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <1348577046-12643-1-git-send-email-pbonzini@redhat.com>
	<CAFEAcA8=uonjFpLQx0ykx5AdBQw0ytbY2EwAy3p_XFX6PayzJQ@mail.gmail.com>
In-Reply-To: <CAFEAcA8=uonjFpLQx0ykx5AdBQw0ytbY2EwAy3p_XFX6PayzJQ@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH] coroutine: always use pooling
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Peter Maydell <peter.maydell@linaro.org>
Cc: kwolf@redhat.com, stefanha@gmail.com, qemu-devel@nongnu.org

Il 26/09/2012 16:34, Peter Maydell ha scritto:
>> It makes sense to use it for other implementations than ucontext, too.
>> >  Coroutine *qemu_coroutine_create(CoroutineEntry *entry)
>> >  {
>> > -    Coroutine *co = qemu_coroutine_new();
>> > +    Coroutine *co;
>> > +
>> > +    co = QSLIST_FIRST(&pool);
>> > +    if (co) {
>> > +        QSLIST_REMOVE_HEAD(&pool, pool_next);
>> > +        pool_size--;
>> > +    } else {
>> > +        co = qemu_coroutine_new();
>> > +    }
>> >      co->entry = entry;
>> >      return co;
>> >  }
> Since this is obviously going to blow up badly if it's called
> from multiple threads, is there some kind of assert we can add
> so that we fail obviously if somebody makes that coding error?
> [the difficulty is probably when your backend is the gthread
> one, at least I assume creating a coroutine inside a coroutine
> is allowed.]

Yes, it is; however, creating a coroutine outside the big QEMU lock is
not allowed.  Since one coroutine only runs at a time the code is safe,
just like the other global variables that are used in
qemu-coroutine-lock.c for example.

Paolo