Date: Wed, 6 Sep 2017 11:06:31 +0200
From: Peter Zijlstra <peterz@infradead.org>
To: Thomas Gleixner <tglx@linutronix.de>
Cc: "chengjian (D)" <cj.chengjian@huawei.com>, huawei.libin@huawei.com,
        mingo@redhat.com, dvhart@infradead.org, linux-kernel@vger.kernel.org
Subject: Re: a competition when some threads acquire futex
Message-ID: <20170906090631.tcer7olhogwxfrlh@hirez.programming.kicks-ass.net>
References: <3e497508-876f-6474-3f3d-acab06a63b55@huawei.com>
 <alpine.DEB.2.20.1709052131130.2393@nanos>
 <555186f2-8240-f7d0-e1b0-9ad1a67ff34c@huawei.com>
 <alpine.DEB.2.20.1709061033280.1843@nanos>
 <20170906085608.ogz4jhv2pieybzob@hirez.programming.kicks-ass.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20170906085608.ogz4jhv2pieybzob@hirez.programming.kicks-ass.net>
User-Agent: NeoMutt/20170609 (1.8.3)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 552
Lines: 16

On Wed, Sep 06, 2017 at 10:56:08AM +0200, Peter Zijlstra wrote:

> Right, but even if it was a coherent patch, I'm not sure it makes sense.
> 
> futex_wait() / futex_wake() don't make ordering guarantees and in
> general you don't get to have wakeup preemption if you don't run a
> PREEMPT kernel.
> 
> So what makes this wakeup so special? Any changelog would need to have a
> convincing argument.

Also, even on !PREEMPT, if that wakeup sets NEED_RESCHED, the return to
userspace after futex_wake() should reschedule.


So I'm really not getting it.