From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id DD5CDC25B07 for ; Tue, 9 Aug 2022 09:59:06 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S239923AbiHIJ7E (ORCPT ); Tue, 9 Aug 2022 05:59:04 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:49594 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S239078AbiHIJ7E (ORCPT ); Tue, 9 Aug 2022 05:59:04 -0400 Received: from frasgout.his.huawei.com (frasgout.his.huawei.com [185.176.79.56]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id B08AE23BC5; Tue, 9 Aug 2022 02:59:02 -0700 (PDT) Received: from fraeml710-chm.china.huawei.com (unknown [172.18.147.207]) by frasgout.his.huawei.com (SkyGuard) with ESMTP id 4M27hR6XZ5z67xh9; Tue, 9 Aug 2022 17:54:31 +0800 (CST) Received: from lhrpeml500003.china.huawei.com (7.191.162.67) by fraeml710-chm.china.huawei.com (10.206.15.59) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2375.24; Tue, 9 Aug 2022 11:59:00 +0200 Received: from [10.195.245.222] (10.195.245.222) by lhrpeml500003.china.huawei.com (7.191.162.67) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2375.24; Tue, 9 Aug 2022 10:58:59 +0100 Message-ID: <3451fa5a-6229-073f-ae18-0c232cd48ed5@huawei.com> Date: Tue, 9 Aug 2022 10:58:58 +0100 MIME-Version: 1.0 User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101 Thunderbird/91.6.1 Subject: Re: [ata] 0568e61225: stress-ng.copy-file.ops_per_sec -15.0% regression To: Damien Le Moal , kernel test robot CC: Christoph Hellwig , "Martin K. Petersen" , LKML , "Linux Memory Management List" , , , , , , , References: <1f498d4a-f93f-ceb4-b713-753196e5e08d@opensource.wdc.com> From: John Garry In-Reply-To: <1f498d4a-f93f-ceb4-b713-753196e5e08d@opensource.wdc.com> Content-Type: text/plain; charset="UTF-8"; format=flowed Content-Transfer-Encoding: 8bit X-Originating-IP: [10.195.245.222] X-ClientProxiedBy: lhrpeml500004.china.huawei.com (7.191.163.9) To lhrpeml500003.china.huawei.com (7.191.162.67) X-CFilter-Loop: Reflected Precedence: bulk List-ID: X-Mailing-List: linux-ide@vger.kernel.org On 08/08/2022 15:52, Damien Le Moal wrote: > On 2022/08/05 1:05, kernel test robot wrote: >> >> >> Greeting, >> >> FYI, we noticed a -15.0% regression of stress-ng.copy-file.ops_per_sec due to commit: >> >> >> commit: 0568e6122574dcc1aded2979cd0245038efe22b6 ("ata: libata-scsi: cap ata_device->max_sectors according to shost->max_sectors") >> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master >> >> in testcase: stress-ng >> on test machine: 96 threads 2 sockets Ice Lake with 256G memory >> with following parameters: >> >> nr_threads: 10% >> disk: 1HDD >> testtime: 60s >> fs: f2fs >> class: filesystem >> test: copy-file >> cpufreq_governor: performance >> ucode: 0xb000280 > > Without knowing what the device adapter is, hard to say where the problem is. I > suspect that with the patch applied, we may be ending up with a small default > max_sectors value, causing overhead due to more commands than necessary. > > Will check what I see with my test rig. As far as I can see, this patch should not make a difference unless the ATA shost driver is setting the max_sectors value unnecessarily low. > >> >> >> >> >> If you fix the issue, kindly add following tag >> Reported-by: kernel test robot >> >> >> Details are as below: >> --------------------------------------------------------------------------------------------------> >> >> >> To reproduce: >> >> git clone https://github.com/intel/lkp-tests.git >> cd lkp-tests >> sudo bin/lkp install job.yaml # job file is attached in this email >> bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run >> sudo bin/lkp run generated-yaml-file >> >> # if come across any failure that blocks the test, >> # please remove ~/.lkp and /lkp dir to run from a clean state. >> >> ========================================================================================= >> class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime/ucode: >> filesystem/gcc-11/performance/1HDD/f2fs/x86_64-rhel-8.3/10%/debian-11.1-x86_64-20220510.cgz/lkp-icl-2sp1/copy-file/stress-ng/60s/0xb000280 >> >> commit: >> 4cbfca5f77 ("scsi: scsi_transport_sas: cap shost opt_sectors according to DMA optimal limit") >> 0568e61225 ("ata: libata-scsi: cap ata_device->max_sectors according to shost->max_sectors") >> >> 4cbfca5f7750520f 0568e6122574dcc1aded2979cd0 >> ---------------- --------------------------- >> %stddev %change %stddev >> \ | \ >> 1627 -14.9% 1385 stress-ng.copy-file.ops >> 27.01 -15.0% 22.96 stress-ng.copy-file.ops_per_sec >> 8935079 -11.9% 7870629 stress-ng.time.file_system_outputs >> 14.88 ± 5% -31.8% 10.14 ± 3% stress-ng.time.percent_of_cpu_this_job_got >> 50912 -14.7% 43413 vmstat.io.bo >> 93.78 +1.4% 95.10 iostat.cpu.idle >> 3.89 -31.6% 2.66 iostat.cpu.iowait >> 4.01 -1.3 2.74 mpstat.cpu.all.iowait% >> 0.23 ± 9% -0.1 0.17 ± 11% mpstat.cpu.all.sys% >> 1.66 ± 37% -1.2 0.51 ± 55% perf-profile.calltrace.cycles-pp.f2fs_write_end.generic_perform_write.f2fs_buffered_write_iter.f2fs_file_write_iter.do_iter_readv_writev >> 1.66 ± 37% -1.1 0.59 ± 25% perf-profile.children.cycles-pp.f2fs_write_end >> 1.51 ± 40% -1.1 0.45 ± 26% perf-profile.children.cycles-pp.f2fs_dirty_data_folio >> 1.21 ± 49% -1.0 0.23 ± 33% perf-profile.children.cycles-pp.f2fs_update_dirty_folio >> 0.88 ± 56% -0.8 0.04 ±111% perf-profile.children.cycles-pp.native_queued_spin_lock_slowpath >> 0.14 ± 26% +0.1 0.25 ± 28% perf-profile.children.cycles-pp.page_cache_ra_unbounded >> 0.88 ± 56% -0.8 0.04 ±112% perf-profile.self.cycles-pp.native_queued_spin_lock_slowpath >> 3164876 ± 9% -20.2% 2524713 ± 7% perf-stat.i.cache-misses >> 4.087e+08 -4.6% 3.899e+08 perf-stat.i.dTLB-loads >> 313050 ± 10% -18.4% 255410 ± 6% perf-stat.i.node-loads >> 972573 ± 9% -16.4% 812873 ± 6% perf-stat.i.node-stores >> 3114748 ± 9% -20.2% 2484807 ± 7% perf-stat.ps.cache-misses >> 4.022e+08 -4.6% 3.837e+08 perf-stat.ps.dTLB-loads >> 308178 ± 10% -18.4% 251418 ± 6% perf-stat.ps.node-loads >> 956996 ± 9% -16.4% 799948 ± 6% perf-stat.ps.node-stores >> 358486 -8.3% 328694 proc-vmstat.nr_active_file >> 1121620 -11.9% 987816 proc-vmstat.nr_dirtied >> 179906 -6.7% 167912 proc-vmstat.nr_dirty >> 1151201 -1.7% 1131322 proc-vmstat.nr_file_pages >> 100181 +9.9% 110078 ± 2% proc-vmstat.nr_inactive_file >> 846362 -14.6% 722471 proc-vmstat.nr_written >> 358486 -8.3% 328694 proc-vmstat.nr_zone_active_file >> 100181 +9.9% 110078 ± 2% proc-vmstat.nr_zone_inactive_file >> 180668 -6.8% 168456 proc-vmstat.nr_zone_write_pending >> 556469 -3.5% 536985 proc-vmstat.pgactivate >> 3385454 -14.6% 2889953 proc-vmstat.pgpgout >> >> >> >> >> Disclaimer: >> Results have been estimated based on internal Intel analysis and are provided >> for informational purposes only. Any difference in system hardware or software >> design or configuration may affect actual performance. >> >> > > From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============7551581913653157619==" MIME-Version: 1.0 From: John Garry To: lkp@lists.01.org Subject: Re: [ata] 0568e61225: stress-ng.copy-file.ops_per_sec -15.0% regression Date: Tue, 09 Aug 2022 10:58:58 +0100 Message-ID: <3451fa5a-6229-073f-ae18-0c232cd48ed5@huawei.com> In-Reply-To: <1f498d4a-f93f-ceb4-b713-753196e5e08d@opensource.wdc.com> List-Id: --===============7551581913653157619== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable On 08/08/2022 15:52, Damien Le Moal wrote: > On 2022/08/05 1:05, kernel test robot wrote: >> >> >> Greeting, >> >> FYI, we noticed a -15.0% regression of stress-ng.copy-file.ops_per_sec d= ue to commit: >> >> >> commit: 0568e6122574dcc1aded2979cd0245038efe22b6 ("ata: libata-scsi: cap= ata_device->max_sectors according to shost->max_sectors") >> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master >> >> in testcase: stress-ng >> on test machine: 96 threads 2 sockets Ice Lake with 256G memory >> with following parameters: >> >> nr_threads: 10% >> disk: 1HDD >> testtime: 60s >> fs: f2fs >> class: filesystem >> test: copy-file >> cpufreq_governor: performance >> ucode: 0xb000280 > = > Without knowing what the device adapter is, hard to say where the problem= is. I > suspect that with the patch applied, we may be ending up with a small def= ault > max_sectors value, causing overhead due to more commands than necessary. > = > Will check what I see with my test rig. As far as I can see, this patch should not make a difference unless the = ATA shost driver is setting the max_sectors value unnecessarily low. > = >> >> >> >> >> If you fix the issue, kindly add following tag >> Reported-by: kernel test robot >> >> >> Details are as below: >> ------------------------------------------------------------------------= --------------------------> >> >> >> To reproduce: >> >> git clone https://github.com/intel/lkp-tests.git >> cd lkp-tests >> sudo bin/lkp install job.yaml # job file is attached = in this email >> bin/lkp split-job --compatible job.yaml # generate the yaml fil= e for lkp run >> sudo bin/lkp run generated-yaml-file >> >> # if come across any failure that blocks the test, >> # please remove ~/.lkp and /lkp dir to run from a clean state. >> >> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >> class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_g= roup/test/testcase/testtime/ucode: >> filesystem/gcc-11/performance/1HDD/f2fs/x86_64-rhel-8.3/10%/debian-11= .1-x86_64-20220510.cgz/lkp-icl-2sp1/copy-file/stress-ng/60s/0xb000280 >> >> commit: >> 4cbfca5f77 ("scsi: scsi_transport_sas: cap shost opt_sectors accordin= g to DMA optimal limit") >> 0568e61225 ("ata: libata-scsi: cap ata_device->max_sectors according = to shost->max_sectors") >> >> 4cbfca5f7750520f 0568e6122574dcc1aded2979cd0 >> ---------------- --------------------------- >> %stddev %change %stddev >> \ | \ >> 1627 -14.9% 1385 stress-ng.copy-file.ops >> 27.01 -15.0% 22.96 stress-ng.copy-file.ops_p= er_sec >> 8935079 -11.9% 7870629 stress-ng.time.file_syste= m_outputs >> 14.88 =C2=B1 5% -31.8% 10.14 =C2=B1 3% stress-ng.time.= percent_of_cpu_this_job_got >> 50912 -14.7% 43413 vmstat.io.bo >> 93.78 +1.4% 95.10 iostat.cpu.idle >> 3.89 -31.6% 2.66 iostat.cpu.iowait >> 4.01 -1.3 2.74 mpstat.cpu.all.iowait% >> 0.23 =C2=B1 9% -0.1 0.17 =C2=B1 11% mpstat.cpu.all.= sys% >> 1.66 =C2=B1 37% -1.2 0.51 =C2=B1 55% perf-profile.ca= lltrace.cycles-pp.f2fs_write_end.generic_perform_write.f2fs_buffered_write_= iter.f2fs_file_write_iter.do_iter_readv_writev >> 1.66 =C2=B1 37% -1.1 0.59 =C2=B1 25% perf-profile.ch= ildren.cycles-pp.f2fs_write_end >> 1.51 =C2=B1 40% -1.1 0.45 =C2=B1 26% perf-profile.ch= ildren.cycles-pp.f2fs_dirty_data_folio >> 1.21 =C2=B1 49% -1.0 0.23 =C2=B1 33% perf-profile.ch= ildren.cycles-pp.f2fs_update_dirty_folio >> 0.88 =C2=B1 56% -0.8 0.04 =C2=B1111% perf-profile.ch= ildren.cycles-pp.native_queued_spin_lock_slowpath >> 0.14 =C2=B1 26% +0.1 0.25 =C2=B1 28% perf-profile.ch= ildren.cycles-pp.page_cache_ra_unbounded >> 0.88 =C2=B1 56% -0.8 0.04 =C2=B1112% perf-profile.se= lf.cycles-pp.native_queued_spin_lock_slowpath >> 3164876 =C2=B1 9% -20.2% 2524713 =C2=B1 7% perf-stat.i.cac= he-misses >> 4.087e+08 -4.6% 3.899e+08 perf-stat.i.dTLB-loads >> 313050 =C2=B1 10% -18.4% 255410 =C2=B1 6% perf-stat.i.nod= e-loads >> 972573 =C2=B1 9% -16.4% 812873 =C2=B1 6% perf-stat.i.nod= e-stores >> 3114748 =C2=B1 9% -20.2% 2484807 =C2=B1 7% perf-stat.ps.ca= che-misses >> 4.022e+08 -4.6% 3.837e+08 perf-stat.ps.dTLB-loads >> 308178 =C2=B1 10% -18.4% 251418 =C2=B1 6% perf-stat.ps.no= de-loads >> 956996 =C2=B1 9% -16.4% 799948 =C2=B1 6% perf-stat.ps.no= de-stores >> 358486 -8.3% 328694 proc-vmstat.nr_active_file >> 1121620 -11.9% 987816 proc-vmstat.nr_dirtied >> 179906 -6.7% 167912 proc-vmstat.nr_dirty >> 1151201 -1.7% 1131322 proc-vmstat.nr_file_pages >> 100181 +9.9% 110078 =C2=B1 2% proc-vmstat.nr_inact= ive_file >> 846362 -14.6% 722471 proc-vmstat.nr_written >> 358486 -8.3% 328694 proc-vmstat.nr_zone_activ= e_file >> 100181 +9.9% 110078 =C2=B1 2% proc-vmstat.nr_zone_= inactive_file >> 180668 -6.8% 168456 proc-vmstat.nr_zone_write= _pending >> 556469 -3.5% 536985 proc-vmstat.pgactivate >> 3385454 -14.6% 2889953 proc-vmstat.pgpgout >> >> >> >> >> Disclaimer: >> Results have been estimated based on internal Intel analysis and are pro= vided >> for informational purposes only. Any difference in system hardware or so= ftware >> design or configuration may affect actual performance. >> >> > = >=20 --===============7551581913653157619==--