From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1761472AbcINKe0 (ORCPT <rfc822;w@1wt.eu>);
        Wed, 14 Sep 2016 06:34:26 -0400
Received: from merlin.infradead.org ([205.233.59.134]:48212 "EHLO
        merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1756012AbcINKeO (ORCPT
        <rfc822;linux-kernel@vger.kernel.org>);
        Wed, 14 Sep 2016 06:34:14 -0400
Date: Wed, 14 Sep 2016 12:33:48 +0200
From: Peter Zijlstra <peterz@infradead.org>
To: Waiman Long <waiman.long@hpe.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
        Jason Low <jason.low2@hpe.com>,
        Ding Tianhong <dingtianhong@huawei.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Will Deacon <Will.Deacon@arm.com>, Ingo Molnar <mingo@redhat.com>,
        Imre Deak <imre.deak@intel.com>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Davidlohr Bueso <dave@stgolabs.net>,
        Tim Chen <tim.c.chen@linux.intel.com>, Terry Rudd <terry.rudd@hpe.com>,
        "Paul E. McKenney" <paulmck@us.ibm.com>, Jason Low <jason.low2@hp.com>
Subject: Re: [PATCH -v3 07/10] mutex: Restructure wait loop
Message-ID: <20160914103348.GB5016@twins.programming.kicks-ass.net>
References: <20160905123636.450529224@infradead.org>
 <20160905124027.619747346@infradead.org>
 <57D31A3B.7000105@hpe.com>
 <57D31C67.8010809@hpe.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <57D31C67.8010809@hpe.com>
User-Agent: Mutt/1.5.23.1 (2014-03-12)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Fri, Sep 09, 2016 at 04:32:39PM -0400, Waiman Long wrote:

> >I think you need to remove the HANDOFF flag here in the error path too.
> >Maybe you should fix that in patch 6.
> >
> >It is also possible that __mutex_handoff() has been called at this point.
> >So this task may now be the owner of the lock. Maybe you should do a final
> >trylock here.
> 
> Alternatively, we can move the trylock in the main loop after the
> spin_lock_mutex(). In this case, we don't need an extra trylock in the error
> path.

I did something like so..

--- a/kernel/locking/mutex.c
+++ b/kernel/locking/mutex.c
@@ -629,12 +629,12 @@ __mutex_lock_common(struct mutex *lock,
 	if (__mutex_waiter_is_first(lock, &waiter))
 		__mutex_set_flag(lock, MUTEX_FLAG_WAITERS);
 
-	if (__mutex_trylock(lock, false))
-		goto remove_waiter;
-
 	lock_contended(&lock->dep_map, ip);
 
 	for (;;) {
+		if (__mutex_trylock(lock, first))
+			break;
+
 		/*
 		 * got a signal? (This code gets eliminated in the
 		 * TASK_UNINTERRUPTIBLE case.)
@@ -659,13 +659,9 @@ __mutex_lock_common(struct mutex *lock,
 			first = true;
 			__mutex_set_flag(lock, MUTEX_FLAG_HANDOFF);
 		}
-
-		if (__mutex_trylock(lock, first))
-			break;
 	}
 	__set_task_state(task, TASK_RUNNING);
 
-remove_waiter:
 	mutex_remove_waiter(lock, &waiter, task);
 	if (likely(list_empty(&lock->wait_list)))
 		__mutex_clear_flag(lock, MUTEX_FLAGS);