From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753520AbcFBQfj (ORCPT ); Thu, 2 Jun 2016 12:35:39 -0400 Received: from mail-oi0-f49.google.com ([209.85.218.49]:36738 "EHLO mail-oi0-f49.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752701AbcFBQfi (ORCPT ); Thu, 2 Jun 2016 12:35:38 -0400 MIME-Version: 1.0 In-Reply-To: <20160602023101.7364.qmail@ns.sciencehorizons.net> References: <20160602023101.7364.qmail@ns.sciencehorizons.net> Date: Thu, 2 Jun 2016 09:35:37 -0700 X-Google-Sender-Auth: 1KVqI05y4mOkccD3ysdhqNm6B1w Message-ID: Subject: Re: [PATCH v3 06/10] fs/namei.c: Improve dcache hash function From: Linus Torvalds To: George Spelvin Cc: "J. Bruce Fields" , Linux Kernel Mailing List , Peter Zijlstra Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Jun 1, 2016 at 7:31 PM, George Spelvin wrote: > > I could use two more iterations of HASH_MIX() or something similar, > then just take the x value, but that's 6 cycles. If a multiply is > 4 or 5 cycles, that's a net loss. Yes. Especially since the multiply will often end up more able to be run in parallel with other things. > But worrying about that is the reason I left provision for arch-specific > hooks, and I'm already working on the first: the PA-RISC doesn't have > an integer multiplier at all, although the FPU can do 32-bit integer > multiplies. Don't worry about pa-risc. There may be a handful of users, where even "users" is more of a "boot up occasionally just for perverse fun" rather than anything else. That's true of at least half the architectures we support - the only ones that really matter and where performance is a real isseu are currently x86, arm and powerpc. Linus