From: "Darrick J. Wong" Subject: Re: [PATCH v5 0/4] crc32c: Add faster algorithm and self-test code Date: Thu, 6 Oct 2011 13:20:42 -0700 Message-ID: <20111006202042.GD12447@tux1.beaverton.ibm.com> References: <20111004235357.1560.12602.stgit@elm3c44.beaverton.ibm.com> Reply-To: djwong@us.ibm.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Joakim Tjernlund , Bob Pearson , linux-kernel , Mingming Cao , linux-crypto , linux-fsdevel , linux-ext4@vger.kernel.org To: Andreas Dilger , Herbert Xu , Theodore Tso , David Miller Return-path: Received: from e4.ny.us.ibm.com ([32.97.182.144]:46767 "EHLO e4.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756845Ab1JFUUh (ORCPT ); Thu, 6 Oct 2011 16:20:37 -0400 Received: from /spool/local by e4.ny.us.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Thu, 6 Oct 2011 16:20:36 -0400 Content-Disposition: inline In-Reply-To: <20111004235357.1560.12602.stgit@elm3c44.beaverton.ibm.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Tue, Oct 04, 2011 at 04:53:57PM -0700, Darrick J. Wong wrote: > Hi all, > > This patchset (re)uses Bob Pearson's crc32 slice-by-8 code to stamp out a > software crc32c implementation. It requires that all ten of his patches (at > least the ones dated 31 Aug 2011) be applied. It removes the crc32c > implementation in crypto/ in favor of using the stamped-out one in lib/. There > is also a change to Kconfig so that the kernel builder can pick an > implementation best suited for the hardware. > > The motivation for this patchset is that I am working on adding full metadata > checksumming to ext4. As far as performance impact of adding checksumming > goes, I see nearly no change with a standard mail server ffsb simulation. On a > test that involves only file creation and deletion and extent tree writes, I > see a drop of about 50 pcercent with the current kernel crc32c implementation; > this improves to a drop of about 20 percent with the enclosed crc32c code. > > When metadata is usually a small fraction of total IO, this new implementation > doesn't help much because metadata is usually a small fraction of total IO. > However, when we are doing IO that is almost all metadata (such as rm -rf'ing a > tree), then this patch speeds up the operation substantially. > > Incidentally, given that iscsi, sctp, and btrfs also use crc32c, this patchset > should improve their speed as well. I have not yet quantified that, however. As for Mr. Tjernlund's unresolved questions regarding the v4 patch, I have tested this new code on x64/x32/ppc32/ppc64 and it seems to work fine, both with the crc32c selftest and also on a practical level with ext4 metadata checksumming enabled. Updating to Bob's newest calculation code brings about a 10-15% speedup on the ppc64 box. I also see that slice-by-8 is about 20% faster than slice-by-4 on my ppc32 box. I did _not_ see any failures on ppc32 when running an extended ext4+checksum test suite. Details of the ppc32 box: root@dyn9047029101:~# cat /proc/cpuinfo processor : 0 cpu : 740/750 temperature : 45 C (uncalibrated) clock : 500.000000MHz revision : 131.0 (pvr 0008 8300) bogomips : 49.86 total bogomips : 49.86 timebase : 24934966 platform : PowerMac model : PowerMac1,1 machine : PowerMac1,1 motherboard : PowerMac1,1 MacRISC Power Macintosh detected as : 66 (Blue&White G3) pmac flags : 00000000 L2 cache : 1024K unified pmac-generation : NewWorld Memory : 896 MB root@dyn9047029101:~# gcc --version gcc-4.4.real (Ubuntu 4.4.3-4ubuntu5) 4.4.3 Copyright (C) 2009 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. root@dyn9047029101:~# for i in /sys/devices/system/cpu/cpu0/cache/*/*; do echo $i $(cat $i); done /sys/devices/system/cpu/cpu0/cache/index0/coherency_line_size 32 /sys/devices/system/cpu/cpu0/cache/index0/level 1 /sys/devices/system/cpu/cpu0/cache/index0/number_of_sets 128 /sys/devices/system/cpu/cpu0/cache/index0/shared_cpu_map 00000000,00000000,00000000,00000001 /sys/devices/system/cpu/cpu0/cache/index0/size 32K /sys/devices/system/cpu/cpu0/cache/index0/type Data /sys/devices/system/cpu/cpu0/cache/index0/ways_of_associativity 8 /sys/devices/system/cpu/cpu0/cache/index1/coherency_line_size 32 /sys/devices/system/cpu/cpu0/cache/index1/level 1 /sys/devices/system/cpu/cpu0/cache/index1/number_of_sets 128 /sys/devices/system/cpu/cpu0/cache/index1/shared_cpu_map 00000000,00000000,00000000,00000001 /sys/devices/system/cpu/cpu0/cache/index1/size 32K /sys/devices/system/cpu/cpu0/cache/index1/type Instruction /sys/devices/system/cpu/cpu0/cache/index1/ways_of_associativity 8 /sys/devices/system/cpu/cpu0/cache/index2/coherency_line_size 128 /sys/devices/system/cpu/cpu0/cache/index2/level 2 /sys/devices/system/cpu/cpu0/cache/index2/number_of_sets 4096 /sys/devices/system/cpu/cpu0/cache/index2/shared_cpu_map 00000000,00000000,00000000,00000001 /sys/devices/system/cpu/cpu0/cache/index2/size 1024K /sys/devices/system/cpu/cpu0/cache/index2/type Unified /sys/devices/system/cpu/cpu0/cache/index2/ways_of_associativity 2 The ppc64 box: root@elm3c7:~# cat /proc/cpuinfo processor : 0 cpu : POWER5+ (gs) clock : 1900.098000MHz revision : 2.0 (pvr 003b 0200) (the rest is omitted for brevity) root@elm3c7:~# gcc --version gcc-4.4.real (Ubuntu 4.4.3-4ubuntu5) 4.4.3 Copyright (C) 2009 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. root@elm3c7:~# for i in /sys/devices/system/cpu/cpu0/cache/*/*; do echo $i $(cat $i); done /sys/devices/system/cpu/cpu0/cache/index0/coherency_line_size 128 /sys/devices/system/cpu/cpu0/cache/index0/level 1 /sys/devices/system/cpu/cpu0/cache/index0/number_of_sets 64 /sys/devices/system/cpu/cpu0/cache/index0/shared_cpu_map 00000000,00000000,00000000,00000001 /sys/devices/system/cpu/cpu0/cache/index0/size 32K /sys/devices/system/cpu/cpu0/cache/index0/type Data /sys/devices/system/cpu/cpu0/cache/index0/ways_of_associativity 4 /sys/devices/system/cpu/cpu0/cache/index1/coherency_line_size 128 /sys/devices/system/cpu/cpu0/cache/index1/level 1 /sys/devices/system/cpu/cpu0/cache/index1/number_of_sets 256 /sys/devices/system/cpu/cpu0/cache/index1/shared_cpu_map 00000000,00000000,00000000,00000001 /sys/devices/system/cpu/cpu0/cache/index1/size 64K /sys/devices/system/cpu/cpu0/cache/index1/type Instruction /sys/devices/system/cpu/cpu0/cache/index1/ways_of_associativity 2 /sys/devices/system/cpu/cpu0/cache/index2/coherency_line_size 128 /sys/devices/system/cpu/cpu0/cache/index2/level 2 /sys/devices/system/cpu/cpu0/cache/index2/number_of_sets 1536 /sys/devices/system/cpu/cpu0/cache/index2/shared_cpu_map 00000000,00000000,00000000,00000005 /sys/devices/system/cpu/cpu0/cache/index2/size 1920K /sys/devices/system/cpu/cpu0/cache/index2/type Unified /sys/devices/system/cpu/cpu0/cache/index2/ways_of_associativity 10 /sys/devices/system/cpu/cpu0/cache/index3/coherency_line_size 128 /sys/devices/system/cpu/cpu0/cache/index3/level 3 /sys/devices/system/cpu/cpu0/cache/index3/number_of_sets 1 /sys/devices/system/cpu/cpu0/cache/index3/shared_cpu_map 00000000,00000000,00000000,00000005 /sys/devices/system/cpu/cpu0/cache/index3/size 36864K /sys/devices/system/cpu/cpu0/cache/index3/type Unified /sys/devices/system/cpu/cpu0/cache/index3/ways_of_associativity 0 --D