Date: Thu, 14 Dec 2017 20:48:53 +0100
From: Peter Zijlstra <peterz@infradead.org>
To: Christopher Lameter <cl@linux.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
        "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
        Boqun Feng <boqun.feng@gmail.com>,
        Andy Lutomirski <luto@amacapital.net>,
        Dave Watson <davejwatson@fb.com>,
        linux-kernel <linux-kernel@vger.kernel.org>,
        linux-api <linux-api@vger.kernel.org>, Paul Turner <pjt@google.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Russell King <linux@arm.linux.org.uk>,
        Thomas Gleixner <tglx@linutronix.de>, Ingo Molnar <mingo@redhat.com>,
        "H. Peter Anvin" <hpa@zytor.com>, Andrew Hunter <ahh@google.com>,
        Andi Kleen <andi@firstfloor.org>, Ben Maurer <bmaurer@fb.com>,
        rostedt <rostedt@goodmis.org>, Josh Triplett <josh@joshtriplett.org>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Will Deacon <will.deacon@arm.com>,
        Michael Kerrisk <mtk.manpages@gmail.com>,
        Alexander Viro <viro@zeniv.linux.org.uk>
Subject: Re: [RFC PATCH for 4.16 02/21] rseq: Introduce restartable sequences
 system call (v12)
Message-ID: <20171214194853.GE3326@worktop>
References: <20171214161403.30643-1-mathieu.desnoyers@efficios.com>
 <20171214161403.30643-3-mathieu.desnoyers@efficios.com>
 <alpine.DEB.2.20.1712141033100.6249@nuc-kabylake>
 <12046460.34426.1513275177081.JavaMail.zimbra@efficios.com>
 <alpine.DEB.2.20.1712141238180.8340@nuc-kabylake>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <alpine.DEB.2.20.1712141238180.8340@nuc-kabylake>
User-Agent: Mutt/1.5.22.1 (2013-10-16)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 611
Lines: 12

On Thu, Dec 14, 2017 at 12:50:13PM -0600, Christopher Lameter wrote:
> Ultimately I wish fast increments like done by this_cpu_inc() could be
> implemented in an efficient way on non x86 platforms that do not have
> cheap instructions like that.

So the problem isn't migration; for that we could wrap the operation in
preempt_disable() which is not more expensive than rseq would be. And a
lot more deterministic.

The problem instead is interrupts, which can result in nested load-store
operations, and that comes apart. This then means having to disable
interrupts over these things and _that_ is expensive.