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 S932388AbcLIGi7 (ORCPT <rfc822;w@1wt.eu>);
        Fri, 9 Dec 2016 01:38:59 -0500
Received: from bombadil.infradead.org ([198.137.202.9]:42765 "EHLO
        bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1752847AbcLIGi6 (ORCPT
        <rfc822;linux-kernel@vger.kernel.org>);
        Fri, 9 Dec 2016 01:38:58 -0500
Date: Fri, 9 Dec 2016 07:38:47 +0100
From: Peter Zijlstra <peterz@infradead.org>
To: Thomas Gleixner <tglx@linutronix.de>
Cc: LKML <linux-kernel@vger.kernel.org>,
        John Stultz <john.stultz@linaro.org>, Ingo Molnar <mingo@kernel.org>,
        David Gibson <david@gibson.dropbear.id.au>,
        Liav Rehana <liavr@mellanox.com>,
        Chris Metcalf <cmetcalf@mellanox.com>,
        Richard Cochran <richardcochran@gmail.com>,
        Parit Bhargava <prarit@redhat.com>,
        Laurent Vivier <lvivier@redhat.com>,
        "Christopher S. Hall" <christopher.s.hall@intel.com>
Subject: Re: [patch 5/6] [RFD] timekeeping: Provide optional 128bit math
Message-ID: <20161209063847.GC15765@worktop.programming.kicks-ass.net>
References: <20161208202623.883855034@linutronix.de>
 <20161208204229.005418487@linutronix.de>
 <20161209052638.GC3061@worktop.programming.kicks-ass.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20161209052638.GC3061@worktop.programming.kicks-ass.net>
User-Agent: Mutt/1.5.22.1 (2013-10-16)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Fri, Dec 09, 2016 at 06:26:38AM +0100, Peter Zijlstra wrote:
> On Thu, Dec 08, 2016 at 08:49:39PM -0000, Thomas Gleixner wrote:
> 
> > +static inline u64 timekeeping_delta_to_ns(struct tk_read_base *tkr, u64 delta)
> > +{
> > +	u32 dh, dl;
> > +	u64 nsec;
> > +
> > +	dl = delta;
> > +	dh = delta >> 32;
> > +
> > +	nsec = ((u64)dl * tkr->mult) + tkr->xtime_nsec;
> > +	nsec >>= tkr->shift;
> > +	if (unlikely(dh))
> > +		nsec += ((u64)dh * tkr->mult) << (32 - tkr->shift);
> > +	return nsec;
> > +}
> 
> Just for giggles, on tilegx the branch is actually slower than doing the
> mult unconditionally.
> 
> The problem is that the two multiplies would otherwise completely
> pipeline, whereas with the conditional you serialize them.

On my Haswell laptop the unconditional version is faster too.

> (came to light while talking about why the mul_u64_u32_shr() fallback
> didn't work right for them, which was a combination of the above issue
> and the fact that their compiler 'lost' the fact that these are
> 32x32->64 mults and did 64x64 ones instead).

Turns out using GCC-6.2.1 we have the same problem on i386, GCC doesn't
recognise the 32x32 mults and generates crap.

This used to work :/