Greeting, FYI, we noticed a 6.9% improvement of stress-ng.fanotify.ops_per_sec due to commit: commit: dbb16df6443c59e8a1ef21c2272fcf387d600ddf ("Revert "memcg: cleanup racy sum avoidance code"") https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master in testcase: stress-ng on test machine: 96 threads 2 sockets Intel(R) Xeon(R) Gold 6252 CPU @ 2.10GHz (Cascade Lake) with 512G memory with following parameters: nr_threads: 10% disk: 1HDD testtime: 60s fs: ext4 class: os test: fanotify cpufreq_governor: performance Details are as below: ========================================================================================= class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime: os/gcc-11/performance/1HDD/ext4/x86_64-rhel-8.3/10%/debian-11.1-x86_64-20220510.cgz/lkp-csl-2sp7/fanotify/stress-ng/60s commit: a5d2172180 ("mm/zsmalloc: do not attempt to free IS_ERR handle") dbb16df644 ("Revert "memcg: cleanup racy sum avoidance code"") a5d2172180e8f94a dbb16df6443c59e8a1ef21c2272 ---------------- --------------------------- %stddev %change %stddev \ | \ 74446473 ± 3% +7.0% 79659479 ± 2% stress-ng.fanotify.ops 1233866 ± 3% +6.9% 1318614 ± 2% stress-ng.fanotify.ops_per_sec 19099410 ± 2% +6.9% 20408525 stress-ng.time.file_system_outputs 12033 +7.3% 12907 stress-ng.time.involuntary_context_switches 1619 +4.2% 1687 stress-ng.time.percent_of_cpu_this_job_got 6716 ± 5% -12.6% 5871 ± 2% stress-ng.time.voluntary_context_switches 0.71 ± 29% -95.5% 0.03 ± 14% iostat.cpu.iowait 0.72 ± 29% -0.7 0.03 ± 16% mpstat.cpu.all.iowait% 4322 ± 16% -21.1% 3409 ± 3% meminfo.Active(file) 3447 ± 21% -25.7% 2560 ± 4% meminfo.Buffers 2278 ± 21% -71.3% 653.67 ± 62% numa-vmstat.node0.nr_written 2667 ± 31% -81.1% 503.67 ± 74% numa-vmstat.node1.nr_written 3464 ± 20% -26.0% 2563 ± 4% vmstat.memory.buff 3559 -4.0% 3417 vmstat.system.cs 0.37 ± 8% +0.1 0.50 ± 15% perf-profile.children.cycles-pp.fanotify_should_merge 1.08 ± 9% +0.2 1.28 ± 10% perf-profile.children.cycles-pp.fanotify_merge 0.08 ± 88% +0.5 0.56 ± 42% perf-profile.children.cycles-pp.start_kernel 0.08 ± 88% +0.5 0.56 ± 42% perf-profile.children.cycles-pp.arch_call_rest_init 0.08 ± 88% +0.5 0.56 ± 42% perf-profile.children.cycles-pp.rest_init 0.36 ± 8% +0.1 0.49 ± 15% perf-profile.self.cycles-pp.fanotify_should_merge 1080 ± 16% -21.2% 851.50 ± 3% proc-vmstat.nr_active_file 2388563 ± 2% +6.8% 2551331 proc-vmstat.nr_dirtied 4950 ± 26% -76.6% 1157 ± 67% proc-vmstat.nr_written 1080 ± 16% -21.2% 851.50 ± 3% proc-vmstat.nr_zone_active_file 6102953 +3.4% 6309059 proc-vmstat.numa_hit 6016520 +3.4% 6222216 proc-vmstat.numa_local 6102815 +3.4% 6308704 proc-vmstat.pgalloc_normal 5988882 +3.4% 6192473 proc-vmstat.pgfree 6.114e+09 +3.2% 6.307e+09 perf-stat.i.branch-instructions 3040 -4.8% 2895 perf-stat.i.context-switches 4.852e+10 +3.6% 5.027e+10 perf-stat.i.cpu-cycles 4.547e+09 +2.7% 4.668e+09 perf-stat.i.dTLB-stores 2.965e+10 +3.1% 3.057e+10 perf-stat.i.instructions 2151 +4.2% 2243 perf-stat.i.instructions-per-iTLB-miss 0.51 +3.6% 0.52 perf-stat.i.metric.GHz 653.24 +3.6% 676.88 perf-stat.i.metric.K/sec 198.34 +3.0% 204.22 perf-stat.i.metric.M/sec 36433263 +5.1% 38301409 perf-stat.i.node-load-misses 2010 +5.0% 2111 perf-stat.overall.instructions-per-iTLB-miss 6.021e+09 +3.2% 6.211e+09 perf-stat.ps.branch-instructions 2986 -4.6% 2847 perf-stat.ps.context-switches 4.779e+10 +3.6% 4.951e+10 perf-stat.ps.cpu-cycles 4.479e+09 +2.7% 4.598e+09 perf-stat.ps.dTLB-stores 2.92e+10 +3.1% 3.011e+10 perf-stat.ps.instructions 35888487 +5.1% 37728073 perf-stat.ps.node-load-misses 2922671 +3.8% 3032931 ± 2% perf-stat.ps.node-loads 1.873e+12 +3.4% 1.937e+12 perf-stat.total.instructions 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. 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. -- 0-DAY CI Kernel Test Service https://01.org/lkp