Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758998AbcDMFas (ORCPT ); Wed, 13 Apr 2016 01:30:48 -0400 Received: from mail-pa0-f47.google.com ([209.85.220.47]:32974 "EHLO mail-pa0-f47.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758962AbcDMFaq (ORCPT ); Wed, 13 Apr 2016 01:30:46 -0400 Date: Tue, 12 Apr 2016 22:30:36 -0700 (PDT) From: Hugh Dickins X-X-Sender: hugh@eggly.anvils To: kernel test robot cc: Hugh Dickins , Stephen Rothwell , "Kirill A. Shutemov" , Yang Shi , Ning Qu , David Rientjes , LKML , lkp@01.org Subject: Re: [lkp] [huge tmpfs] d7c7d56ca6: vm-scalability.throughput -5.5% regression In-Reply-To: <20160413015427.GC659@yexl-desktop> Message-ID: References: <20160413015427.GC659@yexl-desktop> User-Agent: Alpine 2.11 (LSU 23 2013-08-11) MIME-Version: 1.0 Content-Type: MULTIPART/MIXED; BOUNDARY="0-198338652-1460525443=:13068" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 34879 Lines: 706 This message is in MIME format. The first part should be readable text, while the remaining parts are likely unreadable without MIME-aware tools. --0-198338652-1460525443=:13068 Content-Type: TEXT/PLAIN; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE On Wed, 13 Apr 2016, kernel test robot wrote: > FYI, we noticed that vm-scalability.throughput -5.5% regression on >=20 > https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git maste= r > commit d7c7d56ca61aec18e5e0cb3a64e50073c42195f7 ("huge tmpfs: avoid prema= ture exposure of new pagetable") Very useful info, thank you. I presume it confirms exactly what Kirill warned me of, that doing the map_pages after instead of before the fault, comes with a performance disadvantage. I shall look into it, but not immediately (and we know other reasons why that patch has to be revisited). Hugh >=20 >=20 > =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 > compiler/cpufreq_governor/kconfig/rootfs/runtime/tbox_group/test/testcase= : > gcc-4.9/performance/x86_64-rhel/debian-x86_64-2015-02-07.cgz/300s/lkp-h= sw01/lru-file-mmap-read-rand/vm-scalability >=20 > commit:=20 > 517348161d2725b8b596feb10c813bf596dc6a47 > d7c7d56ca61aec18e5e0cb3a64e50073c42195f7 >=20 > 517348161d2725b8 d7c7d56ca61aec18e5e0cb3a64=20 > ---------------- --------------------------=20 > fail:runs %reproduction fail:runs > | | | =20 > 1801726 =B1 0% -5.5% 1702808 =B1 0% vm-scalability.throughp= ut > 317.89 =B1 0% +2.9% 327.15 =B1 0% vm-scalability.time.ela= psed_time > 317.89 =B1 0% +2.9% 327.15 =B1 0% vm-scalability.time.ela= psed_time.max > 872240 =B1 4% +8.5% 946467 =B1 1% vm-scalability.time.inv= oluntary_context_switches > 6.73e+08 =B1 0% -92.5% 50568722 =B1 0% vm-scalability.time.maj= or_page_faults > 2109093 =B1 9% -25.8% 1564815 =B1 7% vm-scalability.time.max= imum_resident_set_size > 37881 =B1 0% +586.9% 260194 =B1 0% vm-scalability.time.min= or_page_faults > 5087 =B1 0% +3.7% 5277 =B1 0% vm-scalability.time.per= cent_of_cpu_this_job_got > 16047 =B1 0% +7.5% 17252 =B1 0% vm-scalability.time.sys= tem_time > 127.19 =B1 0% -88.3% 14.93 =B1 1% vm-scalability.time.use= r_time > 72572 =B1 7% +56.0% 113203 =B1 3% cpuidle.C1-HSW.usage > 9.879e+08 =B1 4% -32.5% 6.67e+08 =B1 8% cpuidle.C6-HSW.time > 605545 =B1 3% -12.9% 527295 =B1 1% softirqs.RCU > 164170 =B1 7% +20.5% 197881 =B1 6% softirqs.SCHED > 2584429 =B1 3% -25.5% 1925241 =B1 2% vmstat.memory.free > 252507 =B1 0% +36.2% 343994 =B1 0% vmstat.system.in > 2.852e+08 =B1 5% +163.9% 7.527e+08 =B1 1% numa-numastat.node0.loc= al_node > 2.852e+08 =B1 5% +163.9% 7.527e+08 =B1 1% numa-numastat.node0.num= a_hit > 2.876e+08 =B1 6% +162.8% 7.559e+08 =B1 0% numa-numastat.node1.loc= al_node > 2.876e+08 =B1 6% +162.8% 7.559e+08 =B1 0% numa-numastat.node1.num= a_hit > 6.73e+08 =B1 0% -92.5% 50568722 =B1 0% time.major_page_faults > 2109093 =B1 9% -25.8% 1564815 =B1 7% time.maximum_resident_s= et_size > 37881 =B1 0% +586.9% 260194 =B1 0% time.minor_page_faults > 127.19 =B1 0% -88.3% 14.93 =B1 1% time.user_time > 94.37 =B1 0% +2.0% 96.27 =B1 0% turbostat.%Busy > 2919 =B1 0% +2.0% 2977 =B1 0% turbostat.Avg_MHz > 5.12 =B1 4% -38.7% 3.14 =B1 5% turbostat.CPU%c6 > 2.00 =B1 13% -44.8% 1.10 =B1 22% turbostat.Pkg%pc2 > 240.00 =B1 0% +4.2% 250.14 =B1 0% turbostat.PkgWatt > 55.36 =B1 3% +16.3% 64.40 =B1 2% turbostat.RAMWatt > 17609 =B1103% -59.4% 7148 =B1 72% latency_stats.avg.pipe_= write.__vfs_write.vfs_write.SyS_write.entry_SYSCALL_64_fastpath > 63966 =B1152% -68.4% 20204 =B1 64% latency_stats.max.pipe_= read.__vfs_read.vfs_read.SyS_read.entry_SYSCALL_64_fastpath > 299681 =B1123% -89.7% 30889 =B1 13% latency_stats.max.pipe_= wait.pipe_read.__vfs_read.vfs_read.SyS_read.entry_SYSCALL_64_fastpath > 0.00 =B1 -1% +Inf% 35893 =B1 10% latency_stats.max.wait_= on_page_bit_killable.__lock_page_or_retry.filemap_fault.__do_fault.do_fault= =2Ehandle_mm_fault.__do_page_fault.do_page_fault.page_fault > 90871 =B1125% -56.2% 39835 =B1 74% latency_stats.sum.pipe_= read.__vfs_read.vfs_read.SyS_read.entry_SYSCALL_64_fastpath > 61821 =B1 22% -86.6% 8254 =B1 62% latency_stats.sum.sigsu= spend.SyS_rt_sigsuspend.entry_SYSCALL_64_fastpath > 0.00 =B1 -1% +Inf% 59392 =B1118% latency_stats.sum.throt= tle_direct_reclaim.try_to_free_pages.__alloc_pages_nodemask.alloc_pages_cur= rent.__page_cache_alloc.__do_page_cache_readahead.filemap_fault.xfs_filemap= _fault.__do_fault.do_fault.handle_mm_fault.__do_page_fault > 0.00 =B1 -1% +Inf% 1549096 =B1 24% latency_stats.sum.wait_= on_page_bit_killable.__lock_page_or_retry.filemap_fault.__do_fault.do_fault= =2Ehandle_mm_fault.__do_page_fault.do_page_fault.page_fault > 639.30 =B1 8% -38.8% 391.40 =B1 6% slabinfo.RAW.active_obj= s > 639.30 =B1 8% -38.8% 391.40 =B1 6% slabinfo.RAW.num_objs > 555.90 =B1 14% -50.7% 274.10 =B1 36% slabinfo.nfs_commit_dat= a.active_objs > 555.90 =B1 14% -50.7% 274.10 =B1 36% slabinfo.nfs_commit_dat= a.num_objs > 10651978 =B1 0% -80.0% 2126718 =B1 0% slabinfo.radix_tree_nod= e.active_objs > 218915 =B1 0% -81.9% 39535 =B1 0% slabinfo.radix_tree_nod= e.active_slabs > 12259274 =B1 0% -81.9% 2213762 =B1 0% slabinfo.radix_tree_nod= e.num_objs > 218915 =B1 0% -81.9% 39535 =B1 0% slabinfo.radix_tree_nod= e.num_slabs > 8503640 =B1 1% -87.8% 1038681 =B1 0% meminfo.Active > 8155208 =B1 1% -91.5% 692744 =B1 0% meminfo.Active(file) > 47732497 =B1 0% +13.9% 54365008 =B1 0% meminfo.Cached > 38794624 =B1 0% +36.4% 52899738 =B1 0% meminfo.Inactive > 38748440 =B1 0% +36.4% 52853183 =B1 0% meminfo.Inactive(file) > 45315491 =B1 0% -24.0% 34459599 =B1 0% meminfo.Mapped > 2693407 =B1 5% -30.7% 1867438 =B1 3% meminfo.MemFree > 7048370 =B1 0% -81.5% 1303216 =B1 0% meminfo.SReclaimable > 7145508 =B1 0% -80.4% 1400313 =B1 0% meminfo.Slab > 4168849 =B1 2% -88.1% 496040 =B1 27% numa-meminfo.node0.Acti= ve > 3987391 =B1 1% -91.3% 346768 =B1 0% numa-meminfo.node0.Acti= ve(file) > 23809283 =B1 0% +13.8% 27087077 =B1 0% numa-meminfo.node0.File= Pages > 19423374 =B1 0% +35.8% 26379857 =B1 0% numa-meminfo.node0.Inac= tive > 19402281 =B1 0% +35.8% 26356354 =B1 0% numa-meminfo.node0.Inac= tive(file) > 22594121 =B1 0% -24.1% 17153129 =B1 0% numa-meminfo.node0.Mapp= ed > 1430871 =B1 5% -31.2% 984861 =B1 2% numa-meminfo.node0.MemF= ree > 3457483 =B1 1% -81.4% 642147 =B1 0% numa-meminfo.node0.SRec= laimable > 3507005 =B1 1% -80.3% 692577 =B1 0% numa-meminfo.node0.Slab > 4349443 =B1 3% -87.5% 543711 =B1 24% numa-meminfo.node1.Acti= ve > 4181422 =B1 3% -91.7% 346861 =B1 1% numa-meminfo.node1.Acti= ve(file) > 23896184 =B1 0% +14.2% 27287954 =B1 0% numa-meminfo.node1.File= Pages > 19329324 =B1 0% +37.2% 26528591 =B1 0% numa-meminfo.node1.Inac= tive > 19304364 =B1 0% +37.3% 26505692 =B1 0% numa-meminfo.node1.Inac= tive(file) > 22671758 =B1 0% -23.7% 17303673 =B1 0% numa-meminfo.node1.Mapp= ed > 1299430 =B1 7% -32.8% 873435 =B1 6% numa-meminfo.node1.MemF= ree > 3589265 =B1 1% -81.6% 661650 =B1 0% numa-meminfo.node1.SRec= laimable > 3636880 =B1 1% -80.5% 708315 =B1 0% numa-meminfo.node1.Slab > 994864 =B1 1% -91.3% 86711 =B1 0% numa-vmstat.node0.nr_ac= tive_file > 5952715 =B1 0% +13.8% 6773427 =B1 0% numa-vmstat.node0.nr_fi= le_pages > 356982 =B1 5% -31.5% 244513 =B1 3% numa-vmstat.node0.nr_fr= ee_pages > 4853127 =B1 0% +35.8% 6590709 =B1 0% numa-vmstat.node0.nr_in= active_file > 394.70 =B1 15% -62.9% 146.60 =B1 32% numa-vmstat.node0.nr_is= olated_file > 5649360 =B1 0% -24.1% 4288873 =B1 0% numa-vmstat.node0.nr_ma= pped > 28030 =B1 53% -97.7% 648.30 =B1 10% numa-vmstat.node0.nr_pa= ges_scanned > 864516 =B1 1% -81.4% 160512 =B1 0% numa-vmstat.node0.nr_sl= ab_reclaimable > 1.522e+08 =B1 4% +155.9% 3.893e+08 =B1 1% numa-vmstat.node0.numa_= hit > 1.521e+08 =B1 4% +155.9% 3.893e+08 =B1 1% numa-vmstat.node0.numa_= local > 217926 =B1 3% -84.4% 33949 =B1 2% numa-vmstat.node0.worki= ngset_activate > 60138428 =B1 2% -72.5% 16533446 =B1 0% numa-vmstat.node0.worki= ngset_nodereclaim > 4367580 =B1 3% +158.4% 11285489 =B1 1% numa-vmstat.node0.worki= ngset_refault > 1043245 =B1 3% -91.7% 86749 =B1 1% numa-vmstat.node1.nr_ac= tive_file > 5974941 =B1 0% +14.2% 6823255 =B1 0% numa-vmstat.node1.nr_fi= le_pages > 323798 =B1 7% -33.0% 216945 =B1 5% numa-vmstat.node1.nr_fr= ee_pages > 4829122 =B1 1% +37.2% 6627644 =B1 0% numa-vmstat.node1.nr_in= active_file > 395.80 =B1 8% -68.5% 124.80 =B1 46% numa-vmstat.node1.nr_is= olated_file > 5669082 =B1 0% -23.7% 4326551 =B1 0% numa-vmstat.node1.nr_ma= pped > 32004 =B1 60% -99.9% 47.00 =B1 9% numa-vmstat.node1.nr_pa= ges_scanned > 897351 =B1 1% -81.6% 165406 =B1 0% numa-vmstat.node1.nr_sl= ab_reclaimable > 1.535e+08 =B1 4% +154.6% 3.909e+08 =B1 0% numa-vmstat.node1.numa_= hit > 1.535e+08 =B1 4% +154.7% 3.909e+08 =B1 0% numa-vmstat.node1.numa_= local > 235134 =B1 5% -85.7% 33507 =B1 2% numa-vmstat.node1.worki= ngset_activate > 59647268 =B1 1% -72.1% 16626347 =B1 0% numa-vmstat.node1.worki= ngset_nodereclaim > 4535102 =B1 4% +151.1% 11389137 =B1 0% numa-vmstat.node1.worki= ngset_refault > 347641 =B1 13% +97.0% 684832 =B1 0% proc-vmstat.allocstall > 7738 =B1 9% +236.5% 26042 =B1 0% proc-vmstat.kswapd_low_= wmark_hit_quickly > 2041367 =B1 1% -91.5% 173206 =B1 0% proc-vmstat.nr_active_f= ile > 1233230 =B1 0% +11.7% 1378011 =B1 0% proc-vmstat.nr_dirty_ba= ckground_threshold > 2466460 =B1 0% +11.7% 2756024 =B1 0% proc-vmstat.nr_dirty_th= reshold > 11933740 =B1 0% +13.9% 13594909 =B1 0% proc-vmstat.nr_file_pag= es > 671934 =B1 5% -31.1% 463093 =B1 3% proc-vmstat.nr_free_pag= es > 9685062 =B1 0% +36.5% 13216819 =B1 0% proc-vmstat.nr_inactive= _file > 792.80 =B1 10% -67.9% 254.20 =B1 34% proc-vmstat.nr_isolated= _file > 11327952 =B1 0% -23.9% 8616859 =B1 0% proc-vmstat.nr_mapped > 73994 =B1 51% -99.1% 657.00 =B1 7% proc-vmstat.nr_pages_sc= anned > 1762423 =B1 0% -81.5% 325807 =B1 0% proc-vmstat.nr_slab_rec= laimable > 72.30 =B1 23% +852.4% 688.60 =B1 58% proc-vmstat.nr_vmscan_i= mmediate_reclaim > 5392 =B1 2% -11.9% 4750 =B1 2% proc-vmstat.numa_hint_f= aults > 5.728e+08 =B1 5% +163.4% 1.509e+09 =B1 0% proc-vmstat.numa_hit > 5.728e+08 =B1 5% +163.4% 1.509e+09 =B1 0% proc-vmstat.numa_local > 5638 =B1 4% -12.5% 4935 =B1 3% proc-vmstat.numa_pte_up= dates > 8684 =B1 8% +215.8% 27427 =B1 0% proc-vmstat.pageoutrun > 3220941 =B1 0% -90.2% 315751 =B1 0% proc-vmstat.pgactivate > 17739240 =B1 1% +143.6% 43217427 =B1 0% proc-vmstat.pgalloc_dma= 32 > 6.6e+08 =B1 0% +138.1% 1.572e+09 =B1 0% proc-vmstat.pgalloc_nor= mal > 6.737e+08 =B1 0% -92.4% 51517407 =B1 0% proc-vmstat.pgfault > 6.767e+08 =B1 0% +138.5% 1.614e+09 =B1 0% proc-vmstat.pgfree > 6.73e+08 =B1 0% -92.5% 50568722 =B1 0% proc-vmstat.pgmajfault > 31567471 =B1 1% +91.6% 60472288 =B1 0% proc-vmstat.pgscan_dire= ct_dma32 > 1.192e+09 =B1 2% +84.5% 2.199e+09 =B1 0% proc-vmstat.pgscan_dire= ct_normal > 16309661 =B1 0% +150.4% 40841573 =B1 0% proc-vmstat.pgsteal_dir= ect_dma32 > 6.151e+08 =B1 0% +140.8% 1.481e+09 =B1 0% proc-vmstat.pgsteal_dir= ect_normal > 939746 =B1 18% +101.3% 1891322 =B1 6% proc-vmstat.pgsteal_ksw= apd_dma32 > 27432476 =B1 4% +162.4% 71970660 =B1 2% proc-vmstat.pgsteal_ksw= apd_normal > 4.802e+08 =B1 5% -81.5% 88655347 =B1 0% proc-vmstat.slabs_scann= ed > 452671 =B1 2% -85.1% 67360 =B1 1% proc-vmstat.workingset_= activate > 1.198e+08 =B1 1% -72.4% 33135682 =B1 0% proc-vmstat.workingset_= nodereclaim > 8898128 =B1 1% +154.6% 22657102 =B1 0% proc-vmstat.workingset_= refault > 613962 =B1 12% -18.6% 499880 =B1 9% sched_debug.cfs_rq:/.mi= n_vruntime.stddev > 31.47 =B1 38% +203.5% 95.52 =B1 29% sched_debug.cfs_rq:/.nr= _spread_over.max > 6.19 =B1 32% +150.9% 15.53 =B1 24% sched_debug.cfs_rq:/.nr= _spread_over.stddev > 41.71 =B1 51% -42.3% 24.07 =B1 12% sched_debug.cfs_rq:/.ru= nnable_load_avg.avg > 1094 =B1106% -60.9% 427.95 =B1 25% sched_debug.cfs_rq:/.ru= nnable_load_avg.max > 163.22 =B1 92% -63.2% 60.09 =B1 28% sched_debug.cfs_rq:/.ru= nnable_load_avg.stddev > 613932 =B1 12% -18.6% 499833 =B1 9% sched_debug.cfs_rq:/.sp= read0.stddev > 35.20 =B1 8% -29.1% 24.97 =B1 11% sched_debug.cpu.cpu_loa= d[0].avg > 731.80 =B1 11% -36.1% 467.45 =B1 21% sched_debug.cpu.cpu_loa= d[0].max > 116.23 =B1 10% -43.5% 65.72 =B1 23% sched_debug.cpu.cpu_loa= d[0].stddev > 35.25 =B1 8% -25.6% 26.23 =B1 10% sched_debug.cpu.cpu_loa= d[1].avg > 722.47 =B1 10% -30.2% 504.05 =B1 18% sched_debug.cpu.cpu_loa= d[1].max > 115.25 =B1 10% -38.5% 70.82 =B1 19% sched_debug.cpu.cpu_loa= d[1].stddev > 35.37 =B1 8% -22.4% 27.45 =B1 8% sched_debug.cpu.cpu_loa= d[2].avg > 721.90 =B1 9% -27.7% 521.60 =B1 16% sched_debug.cpu.cpu_loa= d[2].max > 10.85 =B1 14% +16.9% 12.68 =B1 6% sched_debug.cpu.cpu_loa= d[2].min > 114.93 =B1 9% -35.1% 74.62 =B1 16% sched_debug.cpu.cpu_loa= d[2].stddev > 35.20 =B1 8% -21.3% 27.70 =B1 5% sched_debug.cpu.cpu_loa= d[3].avg > 705.73 =B1 9% -29.6% 496.57 =B1 13% sched_debug.cpu.cpu_loa= d[3].max > 10.95 =B1 13% +18.7% 13.00 =B1 4% sched_debug.cpu.cpu_loa= d[3].min > 112.58 =B1 9% -34.8% 73.35 =B1 12% sched_debug.cpu.cpu_loa= d[3].stddev > 34.96 =B1 8% -21.7% 27.39 =B1 5% sched_debug.cpu.cpu_loa= d[4].avg > 684.63 =B1 10% -32.0% 465.83 =B1 11% sched_debug.cpu.cpu_loa= d[4].max > 11.10 =B1 12% +17.7% 13.07 =B1 3% sched_debug.cpu.cpu_loa= d[4].min > 110.03 =B1 9% -36.1% 70.28 =B1 10% sched_debug.cpu.cpu_loa= d[4].stddev > 293.58 =B1 28% +110.8% 618.85 =B1 32% sched_debug.cpu.curr->p= id.min > 18739 =B1 3% +10.5% 20713 =B1 1% sched_debug.cpu.nr_swit= ches.avg > 33332 =B1 10% +21.0% 40337 =B1 6% sched_debug.cpu.nr_swit= ches.max > 4343 =B1 10% +34.8% 5852 =B1 8% sched_debug.cpu.nr_swit= ches.stddev > 19363 =B1 3% +9.2% 21136 =B1 1% sched_debug.cpu.sched_c= ount.avg > 20.35 =B1 17% -31.5% 13.93 =B1 22% sched_debug.cpu.sched_g= oidle.min > 9245 =B1 3% +12.5% 10398 =B1 0% sched_debug.cpu.ttwu_co= unt.avg > 16837 =B1 10% +27.0% 21390 =B1 8% sched_debug.cpu.ttwu_co= unt.max > 2254 =B1 8% +39.5% 3143 =B1 8% sched_debug.cpu.ttwu_co= unt.stddev > 8052 =B1 4% +16.2% 9353 =B1 0% sched_debug.cpu.ttwu_lo= cal.avg > 5846 =B1 4% +11.0% 6491 =B1 2% sched_debug.cpu.ttwu_lo= cal.min > 1847 =B1 11% +39.8% 2582 =B1 8% sched_debug.cpu.ttwu_lo= cal.stddev > 3.66 =B1 4% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= __add_to_page_cache_locked.add_to_page_cache_lru.filemap_fault.xfs_filemap_= fault.__do_fault > 0.00 =B1 -1% +Inf% 1.12 =B1 0% perf-profile.cycles-pp.= __add_to_page_cache_locked.add_to_page_cache_lru.mpage_readpages.xfs_vm_rea= dpages.__do_page_cache_readahead > 0.00 =B1 -1% +Inf% 77.72 =B1 0% perf-profile.cycles-pp.= __alloc_pages_nodemask.alloc_pages_current.__page_cache_alloc.__do_page_cac= he_readahead.filemap_fault > 79.28 =B1 0% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= __alloc_pages_nodemask.alloc_pages_current.__page_cache_alloc.filemap_fault= =2Exfs_filemap_fault > 11.43 =B1 5% -89.4% 1.21 =B1 4% perf-profile.cycles-pp.= __delete_from_page_cache.__remove_mapping.shrink_page_list.shrink_inactive_= list.shrink_zone_memcg > 0.00 =B1 -1% +Inf% 96.93 =B1 0% perf-profile.cycles-pp.= __do_fault.do_fault.handle_mm_fault.__do_page_fault.do_page_fault > 91.04 =B1 0% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= __do_fault.handle_pte_fault.handle_mm_fault.__do_page_fault.do_page_fault > 0.00 =B1 -1% +Inf% 96.66 =B1 0% perf-profile.cycles-pp.= __do_page_cache_readahead.filemap_fault.xfs_filemap_fault.__do_fault.do_fau= lt > 29.86 =B1 3% -96.9% 0.92 =B1 19% perf-profile.cycles-pp.= __list_lru_walk_one.isra.3.list_lru_walk_one.scan_shadow_nodes.shrink_slab.= shrink_zone > 1.59 =B1 14% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= __lru_cache_add.lru_cache_add.add_to_page_cache_lru.filemap_fault.xfs_filem= ap_fault > 0.00 =B1 -1% +Inf% 5.67 =B1 5% perf-profile.cycles-pp.= __lru_cache_add.lru_cache_add.add_to_page_cache_lru.mpage_readpages.xfs_vm_= readpages > 0.00 =B1 -1% +Inf% 78.11 =B1 0% perf-profile.cycles-pp.= __page_cache_alloc.__do_page_cache_readahead.filemap_fault.xfs_filemap_faul= t.__do_fault > 79.40 =B1 0% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= __page_cache_alloc.filemap_fault.xfs_filemap_fault.__do_fault.handle_pte_fa= ult > 1.28 =B1 4% -38.7% 0.78 =B1 1% perf-profile.cycles-pp.= __radix_tree_lookup.__delete_from_page_cache.__remove_mapping.shrink_page_l= ist.shrink_inactive_list > 25.30 =B1 6% -84.2% 3.99 =B1 5% perf-profile.cycles-pp.= __remove_mapping.shrink_page_list.shrink_inactive_list.shrink_zone_memcg.sh= rink_zone > 0.56 =B1 0% +98.2% 1.11 =B1 0% perf-profile.cycles-pp.= __rmqueue.get_page_from_freelist.__alloc_pages_nodemask.alloc_pages_current= =2E__page_cache_alloc > 0.00 =B1 -1% +Inf% 1.11 =B1 0% perf-profile.cycles-pp.= __xfs_get_blocks.xfs_get_blocks.do_mpage_readpage.mpage_readpages.xfs_vm_re= adpages > 0.01 =B1133% +30254.3% 2.66 =B1 8% perf-profile.cycles-pp.= _raw_spin_lock.free_pcppages_bulk.free_hot_cold_page.free_hot_cold_page_lis= t.shrink_page_list > 5.07 =B1 25% +268.7% 18.71 =B1 3% perf-profile.cycles-pp.= _raw_spin_lock.get_page_from_freelist.__alloc_pages_nodemask.alloc_pages_cu= rrent.__page_cache_alloc > 9.16 =B1 6% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= _raw_spin_lock.list_lru_add.__delete_from_page_cache.__remove_mapping.shrin= k_page_list > 0.69 =B1 64% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= _raw_spin_lock.list_lru_del.__add_to_page_cache_locked.add_to_page_cache_lr= u.filemap_fault > 27.69 =B1 3% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= _raw_spin_lock.shadow_lru_isolate.__list_lru_walk_one.list_lru_walk_one.sca= n_shadow_nodes > 10.77 =B1 10% +238.5% 36.45 =B1 1% perf-profile.cycles-pp.= _raw_spin_lock_irq.shrink_inactive_list.shrink_zone_memcg.shrink_zone.do_tr= y_to_free_pages > 0.35 =B1 9% +193.4% 1.02 =B1 13% perf-profile.cycles-pp.= _raw_spin_lock_irq.shrink_inactive_list.shrink_zone_memcg.shrink_zone.kswap= d > 12.86 =B1 9% -89.4% 1.36 =B1 17% perf-profile.cycles-pp.= _raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_inactive_li= st.shrink_zone_memcg > 1.11 =B1 18% +333.5% 4.83 =B1 6% perf-profile.cycles-pp.= _raw_spin_lock_irqsave.pagevec_lru_move_fn.__lru_cache_add.lru_cache_add.ad= d_to_page_cache_lru > 5.38 =B1 5% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= add_to_page_cache_lru.filemap_fault.xfs_filemap_fault.__do_fault.handle_pte= _fault > 0.00 =B1 -1% +Inf% 7.15 =B1 4% perf-profile.cycles-pp.= add_to_page_cache_lru.mpage_readpages.xfs_vm_readpages.__do_page_cache_read= ahead.filemap_fault > 0.00 =B1 -1% +Inf% 78.06 =B1 0% perf-profile.cycles-pp.= alloc_pages_current.__page_cache_alloc.__do_page_cache_readahead.filemap_fa= ult.xfs_filemap_fault > 79.38 =B1 0% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= alloc_pages_current.__page_cache_alloc.filemap_fault.xfs_filemap_fault.__do= _fault > 0.00 =B1 -1% +Inf% 97.32 =B1 0% perf-profile.cycles-pp.= do_fault.handle_mm_fault.__do_page_fault.do_page_fault.page_fault > 5.19 =B1 2% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= do_mpage_readpage.mpage_readpage.xfs_vm_readpage.filemap_fault.xfs_filemap_= fault > 0.00 =B1 -1% +Inf% 10.68 =B1 1% perf-profile.cycles-pp.= do_mpage_readpage.mpage_readpages.xfs_vm_readpages.__do_page_cache_readahea= d.filemap_fault > 0.72 =B1 67% -98.3% 0.01 =B1 87% perf-profile.cycles-pp.= do_syscall_64.return_from_SYSCALL_64.__libc_fork > 72.75 =B1 1% -23.2% 55.88 =B1 1% perf-profile.cycles-pp.= do_try_to_free_pages.try_to_free_pages.__alloc_pages_nodemask.alloc_pages_c= urrent.__page_cache_alloc > 0.00 =B1 -1% +Inf% 96.86 =B1 0% perf-profile.cycles-pp.= filemap_fault.xfs_filemap_fault.__do_fault.do_fault.handle_mm_fault > 90.80 =B1 0% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= filemap_fault.xfs_filemap_fault.__do_fault.handle_pte_fault.handle_mm_fault > 2.39 =B1 3% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= filemap_map_pages.handle_pte_fault.handle_mm_fault.__do_page_fault.do_page_= fault > 0.97 =B1 12% +321.3% 4.07 =B1 6% perf-profile.cycles-pp.= free_hot_cold_page.free_hot_cold_page_list.shrink_page_list.shrink_inactive= _list.shrink_zone_memcg > 1.03 =B1 9% +303.4% 4.17 =B1 5% perf-profile.cycles-pp.= free_hot_cold_page_list.shrink_page_list.shrink_inactive_list.shrink_zone_m= emcg.shrink_zone > 0.65 =B1 23% +451.9% 3.58 =B1 6% perf-profile.cycles-pp.= free_pcppages_bulk.free_hot_cold_page.free_hot_cold_page_list.shrink_page_l= ist.shrink_inactive_list > 0.00 =B1 -1% +Inf% 21.18 =B1 2% perf-profile.cycles-pp.= get_page_from_freelist.__alloc_pages_nodemask.alloc_pages_current.__page_ca= che_alloc.__do_page_cache_readahead > 6.22 =B1 21% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= get_page_from_freelist.__alloc_pages_nodemask.alloc_pages_current.__page_ca= che_alloc.filemap_fault > 94.07 =B1 0% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= handle_pte_fault.handle_mm_fault.__do_page_fault.do_page_fault.page_fault > 0.57 =B1 1% +104.6% 1.16 =B1 2% perf-profile.cycles-pp.= isolate_lru_pages.isra.47.shrink_inactive_list.shrink_zone_memcg.shrink_zon= e.do_try_to_free_pages > 2.96 =B1 7% -30.5% 2.05 =B1 9% perf-profile.cycles-pp.= kthread.ret_from_fork > 9.58 =B1 6% -100.0% 0.00 =B1229% perf-profile.cycles-pp.= list_lru_add.__delete_from_page_cache.__remove_mapping.shrink_page_list.shr= ink_inactive_list > 1.88 =B1 6% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= list_lru_del.__add_to_page_cache_locked.add_to_page_cache_lru.filemap_fault= =2Exfs_filemap_fault > 29.08 =B1 3% -97.0% 0.89 =B1 19% perf-profile.cycles-pp.= list_lru_walk_one.scan_shadow_nodes.shrink_slab.shrink_zone.do_try_to_free_= pages > 1.59 =B1 14% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= lru_cache_add.add_to_page_cache_lru.filemap_fault.xfs_filemap_fault.__do_fa= ult > 0.00 =B1 -1% +Inf% 5.68 =B1 5% perf-profile.cycles-pp.= lru_cache_add.add_to_page_cache_lru.mpage_readpages.xfs_vm_readpages.__do_p= age_cache_readahead > 5.24 =B1 2% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= mpage_readpage.xfs_vm_readpage.filemap_fault.xfs_filemap_fault.__do_fault > 0.00 =B1 -1% +Inf% 18.20 =B1 1% perf-profile.cycles-pp.= mpage_readpages.xfs_vm_readpages.__do_page_cache_readahead.filemap_fault.xf= s_filemap_fault > 2.37 =B1 14% +79.9% 4.27 =B1 13% perf-profile.cycles-pp.= native_flush_tlb_others.try_to_unmap_flush.shrink_page_list.shrink_inactive= _list.shrink_zone_memcg > 0.01 =B1133% +30322.9% 2.66 =B1 8% perf-profile.cycles-pp.= native_queued_spin_lock_slowpath._raw_spin_lock.free_pcppages_bulk.free_hot= _cold_page.free_hot_cold_page_list > 5.07 =B1 25% +268.8% 18.71 =B1 3% perf-profile.cycles-pp.= native_queued_spin_lock_slowpath._raw_spin_lock.get_page_from_freelist.__al= loc_pages_nodemask.alloc_pages_current > 9.16 =B1 6% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= native_queued_spin_lock_slowpath._raw_spin_lock.list_lru_add.__delete_from_= page_cache.__remove_mapping > 0.75 =B1 57% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= native_queued_spin_lock_slowpath._raw_spin_lock.list_lru_del.__add_to_page_= cache_locked.add_to_page_cache_lru > 27.68 =B1 3% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= native_queued_spin_lock_slowpath._raw_spin_lock.shadow_lru_isolate.__list_l= ru_walk_one.list_lru_walk_one > 11.09 =B1 10% +237.5% 37.44 =B1 0% perf-profile.cycles-pp.= native_queued_spin_lock_slowpath._raw_spin_lock_irq.shrink_inactive_list.sh= rink_zone_memcg.shrink_zone > 12.76 =B1 9% -90.9% 1.17 =B1 22% perf-profile.cycles-pp.= native_queued_spin_lock_slowpath._raw_spin_lock_irqsave.__remove_mapping.sh= rink_page_list.shrink_inactive_list > 1.08 =B1 19% +338.2% 4.75 =B1 7% perf-profile.cycles-pp.= native_queued_spin_lock_slowpath._raw_spin_lock_irqsave.pagevec_lru_move_fn= =2E__lru_cache_add.lru_cache_add > 1.81 =B1 2% -73.7% 0.48 =B1 1% perf-profile.cycles-pp.= page_check_address_transhuge.page_referenced_one.rmap_walk_file.rmap_walk.p= age_referenced > 3.24 =B1 1% -42.5% 1.87 =B1 2% perf-profile.cycles-pp.= page_referenced.shrink_page_list.shrink_inactive_list.shrink_zone_memcg.shr= ink_zone > 2.20 =B1 2% -66.0% 0.75 =B1 5% perf-profile.cycles-pp.= page_referenced_one.rmap_walk_file.rmap_walk.page_referenced.shrink_page_li= st > 1.54 =B1 14% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= pagevec_lru_move_fn.__lru_cache_add.lru_cache_add.add_to_page_cache_lru.fil= emap_fault > 0.00 =B1 -1% +Inf% 5.57 =B1 5% perf-profile.cycles-pp.= pagevec_lru_move_fn.__lru_cache_add.lru_cache_add.add_to_page_cache_lru.mpa= ge_readpages > 2.07 =B1 4% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= radix_tree_next_chunk.filemap_map_pages.handle_pte_fault.handle_mm_fault.__= do_page_fault > 3.01 =B1 6% -31.7% 2.05 =B1 9% perf-profile.cycles-pp.= ret_from_fork > 0.72 =B1 67% -98.5% 0.01 =B1 94% perf-profile.cycles-pp.= return_from_SYSCALL_64.__libc_fork > 3.15 =B1 1% -46.0% 1.70 =B1 1% perf-profile.cycles-pp.= rmap_walk.page_referenced.shrink_page_list.shrink_inactive_list.shrink_zone= _memcg > 3.02 =B1 2% -48.4% 1.56 =B1 1% perf-profile.cycles-pp.= rmap_walk_file.rmap_walk.page_referenced.shrink_page_list.shrink_inactive_l= ist > 29.08 =B1 3% -97.0% 0.89 =B1 19% perf-profile.cycles-pp.= scan_shadow_nodes.shrink_slab.shrink_zone.do_try_to_free_pages.try_to_free_= pages > 28.89 =B1 3% -97.1% 0.84 =B1 22% perf-profile.cycles-pp.= shadow_lru_isolate.__list_lru_walk_one.list_lru_walk_one.scan_shadow_nodes.= shrink_slab > 44.93 =B1 4% +21.9% 54.77 =B1 1% perf-profile.cycles-pp.= shrink_inactive_list.shrink_zone_memcg.shrink_zone.do_try_to_free_pages.try= _to_free_pages > 33.07 =B1 4% -50.8% 16.28 =B1 3% perf-profile.cycles-pp.= shrink_page_list.shrink_inactive_list.shrink_zone_memcg.shrink_zone.do_try_= to_free_pages > 1.11 =B1 16% -22.6% 0.86 =B1 6% perf-profile.cycles-pp.= shrink_page_list.shrink_inactive_list.shrink_zone_memcg.shrink_zone.kswapd > 29.15 =B1 3% -96.8% 0.94 =B1 16% perf-profile.cycles-pp.= shrink_slab.shrink_zone.do_try_to_free_pages.try_to_free_pages.__alloc_page= s_nodemask > 73.07 =B1 1% -23.5% 55.91 =B1 1% perf-profile.cycles-pp.= shrink_zone.do_try_to_free_pages.try_to_free_pages.__alloc_pages_nodemask.a= lloc_pages_current > 45.01 =B1 4% +22.1% 54.95 =B1 1% perf-profile.cycles-pp.= shrink_zone_memcg.shrink_zone.do_try_to_free_pages.try_to_free_pages.__allo= c_pages_nodemask > 2.35 =B1 14% +78.9% 4.21 =B1 13% perf-profile.cycles-pp.= smp_call_function_many.native_flush_tlb_others.try_to_unmap_flush.shrink_pa= ge_list.shrink_inactive_list > 0.00 =B1 -1% +Inf% 55.91 =B1 1% perf-profile.cycles-pp.= try_to_free_pages.__alloc_pages_nodemask.alloc_pages_current.__page_cache_a= lloc.__do_page_cache_readahead > 72.76 =B1 1% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= try_to_free_pages.__alloc_pages_nodemask.alloc_pages_current.__page_cache_a= lloc.filemap_fault > 2.38 =B1 14% +79.5% 4.28 =B1 13% perf-profile.cycles-pp.= try_to_unmap_flush.shrink_page_list.shrink_inactive_list.shrink_zone_memcg.= shrink_zone > 0.58 =B1 1% +51.9% 0.88 =B1 14% perf-profile.cycles-pp.= workingset_eviction.__remove_mapping.shrink_page_list.shrink_inactive_list.= shrink_zone_memcg > 0.00 =B1 -1% +Inf% 96.89 =B1 0% perf-profile.cycles-pp.= xfs_filemap_fault.__do_fault.do_fault.handle_mm_fault.__do_page_fault > 91.02 =B1 0% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= xfs_filemap_fault.__do_fault.handle_pte_fault.handle_mm_fault.__do_page_fau= lt > 0.00 =B1 -1% +Inf% 1.11 =B1 0% perf-profile.cycles-pp.= xfs_get_blocks.do_mpage_readpage.mpage_readpages.xfs_vm_readpages.__do_page= _cache_readahead > 5.26 =B1 2% -100.0% 0.00 =B1 -1% perf-profile.cycles-pp.= xfs_vm_readpage.filemap_fault.xfs_filemap_fault.__do_fault.handle_pte_fault > 0.00 =B1 -1% +Inf% 18.21 =B1 1% perf-profile.cycles-pp.= xfs_vm_readpages.__do_page_cache_readahead.filemap_fault.xfs_filemap_fault.= __do_fault >=20 >=20 >=20 > lkp-hsw01: Grantley Haswell-EP > Memory: 64G >=20 >=20 >=20 >=20 > vm-scalability.time.user_time >=20 > 140 ++-----------------------------------------------------------------= ---+ > |******* ****** *************************************** **** ******= *** > 120 *+ * * * * * * * = * | > | = | > 100 ++ = | > | = | > 80 ++ = | > | = | > 60 ++ = | > | = | > 40 ++ = | > | = | > 20 OOOO O O OOO O OO OO OO = | > | O OOOOOO OOO O OOOO OO OO = | > 0 ++-----------------------------------------------------------------= ---+ >=20 >=20 > vm-scalability.time.major_page_faults >=20 > 7e+08 ++*-***-----*-----**--*-----*---****-*--*-***------------*---***-= -*-+ > ** * * ***** ****** ** ********* * **** * ***************** **= **** > 6e+08 ++ = | > | = | > 5e+08 ++ = | > | = | > 4e+08 ++ = | > | = | > 3e+08 ++ = | > | = | > 2e+08 ++ = | > | = | > 1e+08 ++ = | > OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO = | > 0 ++---------------------------------------------------------------= ---+ >=20 >=20 > =09[*] bisect-good sample > =09[O] bisect-bad sample >=20 > To reproduce: >=20 > git clone git://git.kernel.org/pub/scm/linux/kernel/git/wfg/lkp-t= ests.git > cd lkp-tests > bin/lkp install job.yaml # job file is attached in this email > bin/lkp run job.yaml >=20 >=20 > Disclaimer: > Results have been estimated based on internal Intel analysis and are prov= ided > for informational purposes only. Any difference in system hardware or sof= tware > design or configuration may affect actual performance. >=20 >=20 > Thanks, > Xiaolong Ye --0-198338652-1460525443=:13068--