[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 7 Jul 2014 17:20:58 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Paolo Bonzini <pbonzini@...hat.com>
Cc: Waiman Long <waiman.long@...com>, tglx@...utronix.de,
mingo@...nel.org, linux-arch@...r.kernel.org,
linux-kernel@...r.kernel.org,
virtualization@...ts.linux-foundation.org,
xen-devel@...ts.xenproject.org, kvm@...r.kernel.org,
paolo.bonzini@...il.com, konrad.wilk@...cle.com,
boris.ostrovsky@...cle.com, paulmck@...ux.vnet.ibm.com,
riel@...hat.com, torvalds@...ux-foundation.org,
raghavendra.kt@...ux.vnet.ibm.com, david.vrabel@...rix.com,
oleg@...hat.com, gleb@...hat.com, scott.norton@...com,
chegu_vinod@...com
Subject: Re: [PATCH 10/11] qspinlock: Paravirt support
On Wed, Jun 18, 2014 at 02:03:12PM +0200, Paolo Bonzini wrote:
> Il 17/06/2014 00:08, Waiman Long ha scritto:
> >>+void __pv_queue_unlock(struct qspinlock *lock)
> >>+{
> >>+ int val = atomic_read(&lock->val);
> >>+
> >>+ native_queue_unlock(lock);
> >>+
> >>+ if (val & _Q_LOCKED_SLOW)
> >>+ ___pv_kick_head(lock);
> >>+}
> >>+
> >
> >Again a race can happen here between the reading and writing of the lock
> >value. I can't think of a good way to do that without using cmpxchg.
>
> Could you just use xchg on the locked byte?
I'll have to, indeed. This is racy.
Content of type "application/pgp-signature" skipped
Powered by blists - more mailing lists