From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752941AbbIKSiG (ORCPT ); Fri, 11 Sep 2015 14:38:06 -0400 Received: from g2t4618.austin.hp.com ([15.73.212.83]:35552 "EHLO g2t4618.austin.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752037AbbIKSh7 (ORCPT ); Fri, 11 Sep 2015 14:37:59 -0400 X-Greylist: delayed 145905 seconds by postgrey-1.27 at vger.kernel.org; Fri, 11 Sep 2015 14:37:58 EDT From: Waiman Long To: Peter Zijlstra , Ingo Molnar , Thomas Gleixner , "H. Peter Anvin" Cc: x86@kernel.org, linux-kernel@vger.kernel.org, Scott J Norton , Douglas Hatch , Davidlohr Bueso , Waiman Long Subject: [PATCH v6 2/6] locking/pvqspinlock: Unconditional PV kick with _Q_SLOW_VAL Date: Fri, 11 Sep 2015 14:37:34 -0400 Message-Id: <1441996658-62854-3-git-send-email-Waiman.Long@hpe.com> X-Mailer: git-send-email 1.7.1 In-Reply-To: <1441996658-62854-1-git-send-email-Waiman.Long@hpe.com> References: <1441996658-62854-1-git-send-email-Waiman.Long@hpe.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org If _Q_SLOW_VAL has been set, the vCPU state must have been vcpu_hashed. The extra check at the end of __pv_queued_spin_unlock() is unnecessary and so is removed. Signed-off-by: Waiman Long Reviewed-by: Davidlohr Bueso --- kernel/locking/qspinlock_paravirt.h | 6 +----- 1 files changed, 1 insertions(+), 5 deletions(-) diff --git a/kernel/locking/qspinlock_paravirt.h b/kernel/locking/qspinlock_paravirt.h index c8e6e9a..f0450ff 100644 --- a/kernel/locking/qspinlock_paravirt.h +++ b/kernel/locking/qspinlock_paravirt.h @@ -267,7 +267,6 @@ static void pv_wait_head(struct qspinlock *lock, struct mcs_spinlock *node) } if (!lp) { /* ONCE */ - WRITE_ONCE(pn->state, vcpu_hashed); lp = pv_hash(lock, pn); /* @@ -275,11 +274,9 @@ static void pv_wait_head(struct qspinlock *lock, struct mcs_spinlock *node) * when we observe _Q_SLOW_VAL in __pv_queued_spin_unlock() * we'll be sure to be able to observe our hash entry. * - * [S] pn->state * [S] [Rmw] l->locked == _Q_SLOW_VAL * MB RMB * [RmW] l->locked = _Q_SLOW_VAL [L] - * [L] pn->state * * Matches the smp_rmb() in __pv_queued_spin_unlock(). */ @@ -364,8 +361,7 @@ __visible void __pv_queued_spin_unlock(struct qspinlock *lock) * vCPU is harmless other than the additional latency in completing * the unlock. */ - if (READ_ONCE(node->state) == vcpu_hashed) - pv_kick(node->cpu); + pv_kick(node->cpu); } /* * Include the architecture specific callee-save thunk of the -- 1.7.1