[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 2 Feb 2016 11:55:57 -0800
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Will Deacon <will.deacon@....com>
Cc: Boqun Feng <boqun.feng@...il.com>,
Paul McKenney <paulmck@...ux.vnet.ibm.com>,
Peter Zijlstra <peterz@...radead.org>,
"Maciej W. Rozycki" <macro@...tec.com>,
David Daney <ddaney@...iumnetworks.com>,
Måns Rullgård <mans@...sr.com>,
Ralf Baechle <ralf@...ux-mips.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [RFC][PATCH] mips: Fix arch_spin_unlock()
On Tue, Feb 2, 2016 at 11:30 AM, Will Deacon <will.deacon@....com> wrote:
>
> FWIW, and this is by no means conclusive, I hacked that up quickly and
> ran hackbench a few times on the nearest idle arm64 system. The results
> were consistently ~4% slower using acquire for rcu_dereference.
Ok, that's *much* more noticeable than I would have expected. I take
it that load-acquire is really really slow on current arm64
implementations.
That, btw, is one reason why I despise weak memory ordering. In most
cases I've ever seen, the hardware designers have said "barriers don't
much matter" and just made them crazy slow. My old alpha was the worst
of the worst.
It's shades of my least favorite x86 microarchitecture ever: netburst.
Make common cases fast, but make exceptional cases _so_ slow that it's
actually a noticeable pain.
Just out of interest, is store-release slow too? Because that should
be easy to make fast.
Linus
Powered by blists - more mailing lists