From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from e2.ny.us.ibm.com (e2.ny.us.ibm.com [32.97.182.142]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "e2.ny.us.ibm.com", Issuer "Equifax" (verified OK)) by ozlabs.org (Postfix) with ESMTPS id 88C4ADE8F1 for ; Thu, 10 Jul 2008 02:39:59 +1000 (EST) Received: from d01relay02.pok.ibm.com (d01relay02.pok.ibm.com [9.56.227.234]) by e2.ny.us.ibm.com (8.13.8/8.13.8) with ESMTP id m69GdukZ022873 for ; Wed, 9 Jul 2008 12:39:56 -0400 Received: from d01av01.pok.ibm.com (d01av01.pok.ibm.com [9.56.224.215]) by d01relay02.pok.ibm.com (8.13.8/8.13.8/NCO v9.0) with ESMTP id m69GduaK212948 for ; Wed, 9 Jul 2008 12:39:56 -0400 Received: from d01av01.pok.ibm.com (loopback [127.0.0.1]) by d01av01.pok.ibm.com (8.12.11.20060308/8.13.3) with ESMTP id m69Gdtuc006419 for ; Wed, 9 Jul 2008 12:39:55 -0400 Date: Wed, 9 Jul 2008 12:36:33 -0400 From: Josh Boyer To: Benjamin Herrenschmidt Subject: Re: [PATCH] powerpc: rework 4xx PTE access and TLB miss Message-ID: <20080709123633.72758580@zod.rchland.ibm.com> In-Reply-To: <20080709102308.4735d48a@zod.rchland.ibm.com> References: <20080708055449.EDFDADDEF5@ozlabs.org> <20080709102308.4735d48a@zod.rchland.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Cc: linuxppc-dev@ozlabs.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Wed, 9 Jul 2008 10:23:08 -0400 Josh Boyer wrote: > On Tue, 08 Jul 2008 15:54:40 +1000 > Benjamin Herrenschmidt wrote: > > > This is some preliminary work to improve TLB management on SW loaded > > TLB powerpc platforms. This introduce support for non-atomic PTE > > operations in pgtable-ppc32.h and removes write back to the PTE from > > the TLB miss handlers. In addition, the DSI interrupt code no longer > > tries to fixup write permission, this is left to generic code, and > > _PAGE_HWWRITE is gone. > > > > Signed-off-by: Benjamin Herrenschmidt > > Boots for me now. So far it's surviving a dbench run, and I'll try > hackbench/kernbench shortly. For those interested, here are the results of my stupidly simple hackbench runs: Setup: - AMCC PowerPC Sequoia 440EPx board, 128MiB of DRAM, 533 MHz CPU clock - NFS rootfs over 100MBit ethernet - soft-float userspace Times were gathered by booting to a prompt, running five 'hackbench 50' tests, and averaging the results. sequoia_defconfig: Avg time - 48.1118 seconds sequoia_defconfig+patch: Avg time - 47.8996 seconds I consider that difference to be in the noise range, but the important thing is that it doesn't show a regression in performance. Feel free to flame away at my lame testing. I didn't have the setup or time to do much else at the moment. josh