Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755622AbcJGFkD (ORCPT ); Fri, 7 Oct 2016 01:40:03 -0400 Received: from mail-qt0-f194.google.com ([209.85.216.194]:34901 "EHLO mail-qt0-f194.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752154AbcJGFjw (ORCPT ); Fri, 7 Oct 2016 01:39:52 -0400 MIME-Version: 1.0 Reply-To: doug@easyco.com In-Reply-To: <20161007003850.GA7197@kernel.org> References: <20161007003850.GA7197@kernel.org> From: Doug Dumitru Date: Thu, 6 Oct 2016 22:39:21 -0700 X-Google-Sender-Auth: yjbsWwPhBb4GtkibVu3DPcGbuos Message-ID: Subject: Re: [GIT PULL] MD update for 4.9 To: Shaohua Li Cc: torvalds@linux-foundation.org, linux-kernel@vger.kernel.org, linux-raid , neilb@suse.de Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 3761 Lines: 97 Mr. Li, There is another thread in [linux-raid] discussing pre-fetches in the raid-6 AVX2 code. My testing implies that the prefetch distance is too short. In your new AVX512 code, it looks like there are 24 instructions, each with latencies of 1, between the prefetch and the actual memory load. I don't have a AVX512 CPU to try this on, but the prefetch might do better at a bigger distance. If I am not mistaken, it takes a lot longer than 24 clocks to fetch 4 cache lines. Just a comment while the code is still fluid. Doug Dumitru EasyCo LLC On Thu, Oct 6, 2016 at 5:38 PM, Shaohua Li wrote: > Hi Linus, > Please pull MD update for 4.9. This update includes: > - new AVX512 instruction based raid6 gen/recovery algorithm > - A couple of md-cluster related bug fixes > - Fix a potential deadlock > - Set nonrotational bit for raid array with SSD > - Set correct max_hw_sectors for raid5/6, which hopefuly can improve > performance a little bit > - Other minor fixes > > Thanks, > Shaohua > > The following changes since commit 7d1e042314619115153a0f6f06e4552c09a50e13: > > Merge tag 'usercopy-v4.8-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux (2016-09-20 17:11:19 -0700) > > are available in the git repository at: > > git://git.kernel.org/pub/scm/linux/kernel/git/shli/md.git tags/md/4.9-rc1 > > for you to fetch changes up to bb086a89a406b5d877ee616f1490fcc81f8e1b2b: > > md: set rotational bit (2016-10-03 10:20:27 -0700) > > ---------------------------------------------------------------- > Chao Yu (1): > raid5: fix to detect failure of register_shrinker > > Gayatri Kammela (5): > lib/raid6: Add AVX512 optimized gen_syndrome functions > lib/raid6: Add AVX512 optimized recovery functions > lib/raid6/test/Makefile: Add avx512 gen_syndrome and recovery functions > lib/raid6: Add AVX512 optimized xor_syndrome functions > raid6/test/test.c: bug fix: Specify aligned(alignment) attributes to the char arrays > > Guoqing Jiang (9): > md-cluster: call md_kick_rdev_from_array once ack failed > md-cluster: use FORCEUNLOCK in lockres_free > md-cluster: remove some unnecessary dlm_unlock_sync > md: changes for MD_STILL_CLOSED flag > md-cluster: clean related infos of cluster > md-cluster: protect md_find_rdev_nr_rcu with rcu lock > md-cluster: convert the completion to wait queue > md-cluster: introduce dlm_lock_sync_interruptible to fix tasks hang > md-cluster: make resync lock also could be interruptted > > Shaohua Li (5): > raid5: allow arbitrary max_hw_sectors > md/bitmap: fix wrong cleanup > md: fix a potential deadlock > raid5: handle register_shrinker failure > md: set rotational bit > > arch/x86/Makefile | 5 +- > drivers/md/bitmap.c | 4 +- > drivers/md/md-cluster.c | 99 ++++++--- > drivers/md/md.c | 44 +++- > drivers/md/md.h | 5 +- > drivers/md/raid5.c | 11 +- > include/linux/raid/pq.h | 4 + > lib/raid6/Makefile | 2 +- > lib/raid6/algos.c | 12 + > lib/raid6/avx512.c | 569 +++++++++++++++++++++++++++++++++++++++++++++++ > lib/raid6/recov_avx512.c | 388 ++++++++++++++++++++++++++++++++ > lib/raid6/test/Makefile | 5 +- > lib/raid6/test/test.c | 7 +- > lib/raid6/x86.h | 10 + > 14 files changed, 1111 insertions(+), 54 deletions(-) > create mode 100644 lib/raid6/avx512.c > create mode 100644 lib/raid6/recov_avx512.c > -- > To unsubscribe from this list: send the line "unsubscribe linux-raid" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- Doug Dumitru EasyCo LLC