Greeting, FYI, we noticed a -4.3% regression of vm-scalability.throughput due to commit: commit: 4f09feb8bf083be3834080ddf3782aee12a7c3f7 ("mm: Force update of mem cgroup soft limit tree on usage excess") url: https://github.com/0day-ci/linux/commits/Tim-Chen/Soft-limit-memory-management-bug-fixes/20210218-054228 in testcase: vm-scalability on test machine: 192 threads Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory with following parameters: runtime: 300s test: lru-file-readonce cpufreq_governor: performance ucode: 0x5003006 test-description: The motivation behind this suite is to exercise functions and regions of the mm/ of the Linux kernel which are of interest to us. test-url: https://git.kernel.org/cgit/linux/kernel/git/wfg/vm-scalability.git/ 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 bin/lkp install job.yaml # job file is attached in this email bin/lkp split-job --compatible job.yaml bin/lkp run compatible-job.yaml ========================================================================================= compiler/cpufreq_governor/kconfig/rootfs/runtime/tbox_group/test/testcase/ucode: gcc-9/performance/x86_64-rhel-8.3/debian-10.4-x86_64-20200603.cgz/300s/lkp-csl-2ap4/lru-file-readonce/vm-scalability/0x5003006 commit: f0812bba8b ("mm: Fix dropped memcg from mem cgroup soft limit tree") 4f09feb8bf ("mm: Force update of mem cgroup soft limit tree on usage excess") f0812bba8bbd02bf 4f09feb8bf083be3834080ddf37 ---------------- --------------------------- %stddev %change %stddev \ | \ 131114 -4.2% 125626 vm-scalability.median 25345052 -4.3% 24265993 vm-scalability.throughput 224.30 +3.6% 232.43 vm-scalability.time.elapsed_time 224.30 +3.6% 232.43 vm-scalability.time.elapsed_time.max 31611 +4.5% 33046 vm-scalability.time.system_time 5225 -2.5% 5095 vmstat.system.cs 95143 ± 4% +20.7% 114842 ± 15% meminfo.Active 94109 ± 4% +20.9% 113806 ± 15% meminfo.Active(anon) 7347 ± 8% +17.9% 8665 ± 11% softirqs.CPU73.RCU 6628 ± 7% +21.3% 8040 ± 11% softirqs.CPU77.RCU 7378 ± 7% +35.9% 10027 ± 20% softirqs.CPU90.RCU 6976 ± 4% +31.5% 9173 ± 25% softirqs.CPU95.RCU 8983 ± 24% +19.5% 10737 ± 4% softirqs.CPU96.SCHED 220617 ± 33% -68.4% 69618 ±106% numa-meminfo.node0.Inactive(anon) 4442 ± 24% -31.2% 3055 ± 18% numa-meminfo.node0.PageTables 2964 ± 45% +153.1% 7504 ± 23% numa-meminfo.node1.Active 2620 ± 39% +179.8% 7332 ± 25% numa-meminfo.node1.Active(anon) 37702 ±130% +248.6% 131446 ± 60% numa-meminfo.node1.AnonPages 39649 ±131% +368.8% 185879 ± 43% numa-meminfo.node1.Inactive(anon) 3078 ± 11% +50.9% 4647 ± 21% numa-meminfo.node1.PageTables 4610 ± 59% +1241.2% 61840 ± 78% numa-meminfo.node1.Shmem 23809 ± 4% +20.6% 28704 ± 15% proc-vmstat.nr_active_anon 939.33 ± 4% -8.1% 863.33 ± 3% proc-vmstat.nr_isolated_file 56262 +9.7% 61735 ± 8% proc-vmstat.nr_shmem 23811 ± 4% +20.6% 28705 ± 15% proc-vmstat.nr_zone_active_anon 75883 ± 2% +23.1% 93446 ± 19% proc-vmstat.pgactivate 1038398 +5.0% 1089871 ± 2% proc-vmstat.pgfault 65866 +3.5% 68179 proc-vmstat.pgreuse 18338900 ± 6% -8.5% 16783260 ± 4% proc-vmstat.slabs_scanned 1216 ± 9% -8.2% 1116 ± 3% interrupts.CPU145.CAL:Function_call_interrupts 27.33 ± 16% +195.1% 80.67 ± 72% interrupts.CPU172.RES:Rescheduling_interrupts 162.00 ± 4% +21.0% 196.00 ± 17% interrupts.CPU33.RES:Rescheduling_interrupts 163.83 ± 3% +15.8% 189.67 ± 14% interrupts.CPU34.RES:Rescheduling_interrupts 129.83 ± 2% +19.8% 155.50 ± 9% interrupts.CPU56.RES:Rescheduling_interrupts 97.50 ± 10% +40.0% 136.50 ± 20% interrupts.CPU65.RES:Rescheduling_interrupts 261.17 ± 44% -44.2% 145.67 ± 18% interrupts.CPU73.RES:Rescheduling_interrupts 49.17 ± 48% +133.9% 115.00 ± 36% interrupts.CPU85.RES:Rescheduling_interrupts 41.83 ± 24% +144.2% 102.17 ± 52% interrupts.CPU88.RES:Rescheduling_interrupts 48.17 ± 38% +215.2% 151.83 ± 54% interrupts.CPU89.RES:Rescheduling_interrupts 38.17 ± 15% +106.1% 78.67 ± 39% interrupts.CPU90.RES:Rescheduling_interrupts 55160 ± 33% -68.5% 17396 ±106% numa-vmstat.node0.nr_inactive_anon 3614 ± 8% -17.7% 2974 ± 11% numa-vmstat.node0.nr_mapped 1105 ± 24% -31.0% 762.67 ± 17% numa-vmstat.node0.nr_page_table_pages 55168 ± 33% -68.5% 17402 ±106% numa-vmstat.node0.nr_zone_inactive_anon 663.00 ± 39% +179.7% 1854 ± 25% numa-vmstat.node1.nr_active_anon 9426 ±130% +248.2% 32821 ± 60% numa-vmstat.node1.nr_anon_pages 9914 ±131% +368.5% 46447 ± 43% numa-vmstat.node1.nr_inactive_anon 764.00 ± 11% +51.7% 1159 ± 20% numa-vmstat.node1.nr_page_table_pages 1162 ± 58% +1233.7% 15500 ± 77% numa-vmstat.node1.nr_shmem 663.17 ± 39% +179.6% 1854 ± 25% numa-vmstat.node1.nr_zone_active_anon 9920 ±131% +368.3% 46454 ± 43% numa-vmstat.node1.nr_zone_inactive_anon 9.08 ± 2% +22.4% 11.12 ± 12% perf-stat.i.MPKI 1.303e+10 -4.7% 1.242e+10 perf-stat.i.branch-instructions 0.39 +0.1 0.50 ± 28% perf-stat.i.branch-miss-rate% 1.873e+08 +8.2% 2.027e+08 perf-stat.i.cache-misses 5.924e+08 +7.4% 6.365e+08 perf-stat.i.cache-references 5087 -2.6% 4957 perf-stat.i.context-switches 6.02 +5.2% 6.33 perf-stat.i.cpi 2001 -6.8% 1865 perf-stat.i.cycles-between-cache-misses 0.05 ± 19% +0.0 0.07 ± 24% perf-stat.i.dTLB-load-miss-rate% 1.544e+10 -4.2% 1.479e+10 perf-stat.i.dTLB-loads 5.401e+09 -2.9% 5.247e+09 perf-stat.i.dTLB-stores 24692490 +6.3% 26255880 perf-stat.i.iTLB-load-misses 5.933e+10 -4.4% 5.67e+10 perf-stat.i.instructions 1.89 ± 3% -9.9% 1.70 ± 5% perf-stat.i.major-faults 179.72 -4.0% 172.60 perf-stat.i.metric.M/sec 21954317 +7.3% 23554489 perf-stat.i.node-load-misses 12269278 -3.4% 11855032 perf-stat.i.node-store-misses 14458012 -3.3% 13976704 perf-stat.i.node-stores 10.01 +12.0% 11.21 perf-stat.overall.MPKI 0.32 +0.0 0.34 ± 4% perf-stat.overall.branch-miss-rate% 7.28 +5.3% 7.67 perf-stat.overall.cpi 2307 -7.0% 2146 perf-stat.overall.cycles-between-cache-misses 2408 -10.0% 2167 perf-stat.overall.instructions-per-iTLB-miss 0.14 -5.0% 0.13 perf-stat.overall.ipc 36.46 +2.1 38.51 perf-stat.overall.node-load-miss-rate% 1.346e+10 -4.6% 1.285e+10 perf-stat.ps.branch-instructions 1.936e+08 +8.2% 2.095e+08 perf-stat.ps.cache-misses 6.14e+08 +7.2% 6.579e+08 perf-stat.ps.cache-references 5172 -2.6% 5038 perf-stat.ps.context-switches 1.597e+10 -4.1% 1.531e+10 perf-stat.ps.dTLB-loads 5.575e+09 -2.9% 5.411e+09 perf-stat.ps.dTLB-stores 25461461 +6.3% 27074171 perf-stat.ps.iTLB-load-misses 6.131e+10 -4.3% 5.867e+10 perf-stat.ps.instructions 1.76 ± 2% -5.0% 1.68 perf-stat.ps.major-faults 22688997 +7.3% 24343963 perf-stat.ps.node-load-misses 12711574 -3.5% 12262917 perf-stat.ps.node-store-misses 14988174 -3.4% 14480039 perf-stat.ps.node-stores 0.01 ± 11% +87.2% 0.01 ± 18% perf-sched.sch_delay.avg.ms.__sched_text_start.__sched_text_start.smpboot_thread_fn.kthread.ret_from_fork 0.00 ±114% +353.6% 0.02 ± 47% perf-sched.sch_delay.max.ms.__sched_text_start.__sched_text_start.exit_to_user_mode_prepare.syscall_exit_to_user_mode.entry_SYSCALL_64_after_hwframe 0.02 ± 17% +338.9% 0.08 ± 62% perf-sched.sch_delay.max.ms.__sched_text_start.__sched_text_start.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop 0.02 ± 49% +3.1e+05% 76.43 ±185% perf-sched.sch_delay.max.ms.__sched_text_start.__sched_text_start.schedule_timeout.kcompactd.kthread 0.68 ±212% +983.9% 7.35 ± 37% perf-sched.sch_delay.max.ms.__sched_text_start.__sched_text_start.smpboot_thread_fn.kthread.ret_from_fork 3.16 ± 75% +8842.7% 282.23 ±218% perf-sched.sch_delay.max.ms.__sched_text_start.__sched_text_start.worker_thread.kthread.ret_from_fork 197.84 ± 7% -23.7% 150.97 ± 18% perf-sched.total_wait_and_delay.average.ms 13780 ± 6% +75.6% 24203 ± 27% perf-sched.total_wait_and_delay.count.ms 197.81 ± 7% -23.7% 150.88 ± 18% perf-sched.total_wait_time.average.ms 2.31 ±100% +286.1% 8.90 ± 38% perf-sched.wait_and_delay.avg.ms.__sched_text_start.__sched_text_start.devkmsg_read.vfs_read.ksys_read 2.32 ±100% +284.4% 8.91 ± 38% perf-sched.wait_and_delay.avg.ms.__sched_text_start.__sched_text_start.do_syslog.part.0 215.21 ± 4% -62.0% 81.72 ± 21% perf-sched.wait_and_delay.avg.ms.__sched_text_start.__sched_text_start.do_task_dead.do_exit.do_group_exit 0.91 ± 4% +145.7% 2.25 ± 5% perf-sched.wait_and_delay.avg.ms.__sched_text_start.__sched_text_start.do_wait.kernel_wait4.__do_sys_wait4 344.42 -92.3% 26.39 ± 33% perf-sched.wait_and_delay.avg.ms.__sched_text_start.__sched_text_start.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop 478.60 -36.5% 304.11 ± 31% perf-sched.wait_and_delay.avg.ms.__sched_text_start.__sched_text_start.schedule_timeout.kcompactd.kthread 655.59 ± 3% +9.7% 719.00 ± 2% perf-sched.wait_and_delay.avg.ms.__sched_text_start.__sched_text_start.smpboot_thread_fn.kthread.ret_from_fork 167.67 ±124% +426.7% 883.17 ± 12% perf-sched.wait_and_delay.count.__sched_text_start.__sched_text_start.devkmsg_read.vfs_read.ksys_read 11.17 ± 3% +11.9% 12.50 ± 4% perf-sched.wait_and_delay.count.__sched_text_start.__sched_text_start.do_nanosleep.hrtimer_nanosleep.__x64_sys_nanosleep 167.50 ±124% +427.1% 882.83 ± 12% perf-sched.wait_and_delay.count.__sched_text_start.__sched_text_start.do_syslog.part.0 303.50 +169.0% 816.50 ± 14% perf-sched.wait_and_delay.count.__sched_text_start.__sched_text_start.do_task_dead.do_exit.do_group_exit 302.17 +128.4% 690.17 perf-sched.wait_and_delay.count.__sched_text_start.__sched_text_start.do_wait.kernel_wait4.__do_sys_wait4 323.33 ± 28% -87.0% 42.00 ±141% perf-sched.wait_and_delay.count.__sched_text_start.__sched_text_start.preempt_schedule_common._cond_resched.truncate_inode_pages_range 185.67 ±115% +385.9% 902.17 ± 12% perf-sched.wait_and_delay.count.__sched_text_start.__sched_text_start.schedule_hrtimeout_range_clock.ep_poll.do_epoll_wait 26.17 +2201.9% 602.33 perf-sched.wait_and_delay.count.__sched_text_start.__sched_text_start.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop 79.33 +71.2% 135.83 ± 32% perf-sched.wait_and_delay.count.__sched_text_start.__sched_text_start.schedule_timeout.kcompactd.kthread 974.50 ± 19% +390.9% 4783 ± 70% perf-sched.wait_and_delay.count.__sched_text_start.__sched_text_start.worker_thread.kthread.ret_from_fork 5.23 ±104% +63694.8% 3339 ± 66% perf-sched.wait_and_delay.max.ms.__sched_text_start.__sched_text_start.devkmsg_read.vfs_read.ksys_read 5.23 ±104% +63704.2% 3339 ± 66% perf-sched.wait_and_delay.max.ms.__sched_text_start.__sched_text_start.do_syslog.part.0 21.62 ± 3% +4530.4% 1001 perf-sched.wait_and_delay.max.ms.__sched_text_start.__sched_text_start.do_wait.kernel_wait4.__do_sys_wait4 1019 +227.9% 3342 ± 66% perf-sched.wait_and_delay.max.ms.__sched_text_start.__sched_text_start.pipe_read.new_sync_read.vfs_read 333.34 ± 70% +963.4% 3544 ± 61% perf-sched.wait_and_delay.max.ms.__sched_text_start.__sched_text_start.schedule_hrtimeout_range_clock.ep_poll.do_epoll_wait 1000 +365.0% 4654 ± 71% perf-sched.wait_and_delay.max.ms.__sched_text_start.__sched_text_start.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop 5.03 +4104.3% 211.64 ± 41% perf-sched.wait_and_delay.max.ms.__sched_text_start.__sched_text_start.schedule_timeout.rcu_gp_kthread.kthread 3.46 ± 44% +157.3% 8.90 ± 38% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.devkmsg_read.vfs_read.ksys_read 3.47 ± 44% +156.2% 8.90 ± 38% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.do_syslog.part.0 215.20 ± 4% -62.0% 81.71 ± 21% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.do_task_dead.do_exit.do_group_exit 0.91 ± 4% +146.2% 2.24 ± 5% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.do_wait.kernel_wait4.__do_sys_wait4 0.01 ±149% +1185.7% 0.09 ± 53% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.exit_to_user_mode_prepare.syscall_exit_to_user_mode.entry_SYSCALL_64_after_hwframe 0.51 ± 20% +111.1% 1.08 ± 21% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.preempt_schedule_common._cond_resched.truncate_inode_pages_range 0.01 ±146% +9.6e+05% 86.17 ± 70% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.preempt_schedule_common._cond_resched.ww_mutex_lock 1.73 ± 16% +95.0% 3.37 ± 29% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.rcu_gp_kthread.kthread.ret_from_fork 344.42 -92.3% 26.38 ± 33% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop 478.59 -36.6% 303.45 ± 31% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.schedule_timeout.kcompactd.kthread 655.58 ± 3% +9.7% 718.99 ± 2% perf-sched.wait_time.avg.ms.__sched_text_start.__sched_text_start.smpboot_thread_fn.kthread.ret_from_fork 7.49 ± 52% +44482.5% 3339 ± 66% perf-sched.wait_time.max.ms.__sched_text_start.__sched_text_start.devkmsg_read.vfs_read.ksys_read 7.49 ± 52% +44481.2% 3339 ± 66% perf-sched.wait_time.max.ms.__sched_text_start.__sched_text_start.do_syslog.part.0 21.62 ± 3% +4531.3% 1001 perf-sched.wait_time.max.ms.__sched_text_start.__sched_text_start.do_wait.kernel_wait4.__do_sys_wait4 0.02 ±145% +12809.8% 1.98 ±134% perf-sched.wait_time.max.ms.__sched_text_start.__sched_text_start.exit_to_user_mode_prepare.syscall_exit_to_user_mode.entry_SYSCALL_64_after_hwframe 1019 +227.9% 3342 ± 66% perf-sched.wait_time.max.ms.__sched_text_start.__sched_text_start.pipe_read.new_sync_read.vfs_read 0.02 ±142% +2.9e+06% 436.52 ± 67% perf-sched.wait_time.max.ms.__sched_text_start.__sched_text_start.preempt_schedule_common._cond_resched.ww_mutex_lock 413.83 ± 43% +756.5% 3544 ± 61% perf-sched.wait_time.max.ms.__sched_text_start.__sched_text_start.schedule_hrtimeout_range_clock.ep_poll.do_epoll_wait 1000 +365.1% 4654 ± 71% perf-sched.wait_time.max.ms.__sched_text_start.__sched_text_start.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop 5.02 +4118.6% 211.60 ± 41% perf-sched.wait_time.max.ms.__sched_text_start.__sched_text_start.schedule_timeout.rcu_gp_kthread.kthread 30.68 ± 3% -2.6 28.08 ± 2% perf-profile.calltrace.cycles-pp.lock_page_lruvec_irqsave.__pagevec_lru_add.lru_cache_add.add_to_page_cache_lru.page_cache_ra_unbounded 30.67 ± 3% -2.6 28.07 ± 2% perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.lock_page_lruvec_irqsave.__pagevec_lru_add.lru_cache_add.add_to_page_cache_lru 30.63 ± 3% -2.6 28.03 ± 2% perf-profile.calltrace.cycles-pp.native_queued_spin_lock_slowpath._raw_spin_lock_irqsave.lock_page_lruvec_irqsave.__pagevec_lru_add.lru_cache_add 31.63 ± 3% -2.6 29.07 ± 2% perf-profile.calltrace.cycles-pp.__pagevec_lru_add.lru_cache_add.add_to_page_cache_lru.page_cache_ra_unbounded.generic_file_buffered_read_get_pages 31.68 ± 3% -2.6 29.11 ± 2% perf-profile.calltrace.cycles-pp.lru_cache_add.add_to_page_cache_lru.page_cache_ra_unbounded.generic_file_buffered_read_get_pages.generic_file_buffered_read 49.97 ± 2% -2.6 47.42 perf-profile.calltrace.cycles-pp.__alloc_pages_nodemask.page_cache_ra_unbounded.generic_file_buffered_read_get_pages.generic_file_buffered_read.xfs_file_buffered_aio_read 5.62 ± 5% -0.5 5.10 ± 3% perf-profile.calltrace.cycles-pp.read_pages.page_cache_ra_unbounded.generic_file_buffered_read_get_pages.generic_file_buffered_read.xfs_file_buffered_aio_read 5.57 ± 5% -0.5 5.05 ± 3% perf-profile.calltrace.cycles-pp.iomap_readahead_actor.iomap_apply.iomap_readahead.read_pages.page_cache_ra_unbounded 5.61 ± 5% -0.5 5.09 ± 3% perf-profile.calltrace.cycles-pp.iomap_readahead.read_pages.page_cache_ra_unbounded.generic_file_buffered_read_get_pages.generic_file_buffered_read 5.60 ± 5% -0.5 5.09 ± 3% perf-profile.calltrace.cycles-pp.iomap_apply.iomap_readahead.read_pages.page_cache_ra_unbounded.generic_file_buffered_read_get_pages 5.34 ± 5% -0.5 4.83 ± 3% perf-profile.calltrace.cycles-pp.iomap_readpage_actor.iomap_readahead_actor.iomap_apply.iomap_readahead.read_pages 0.70 ± 14% +0.3 0.95 ± 9% perf-profile.calltrace.cycles-pp.try_charge.mem_cgroup_charge.__add_to_page_cache_locked.add_to_page_cache_lru.page_cache_ra_unbounded 0.38 ± 71% +0.3 0.70 ± 9% perf-profile.calltrace.cycles-pp.page_counter_try_charge.try_charge.mem_cgroup_charge.__add_to_page_cache_locked.add_to_page_cache_lru 0.00 +3.5 3.52 ± 9% perf-profile.calltrace.cycles-pp.memcg_check_events.mem_cgroup_charge.__add_to_page_cache_locked.add_to_page_cache_lru.page_cache_ra_unbounded 38.12 ± 2% +3.8 41.90 ± 2% perf-profile.calltrace.cycles-pp.add_to_page_cache_lru.page_cache_ra_unbounded.generic_file_buffered_read_get_pages.generic_file_buffered_read.xfs_file_buffered_aio_read 5.32 ± 19% +6.3 11.60 ± 6% perf-profile.calltrace.cycles-pp.mem_cgroup_charge.__add_to_page_cache_locked.add_to_page_cache_lru.page_cache_ra_unbounded.generic_file_buffered_read_get_pages 6.41 ± 17% +6.3 12.76 ± 5% perf-profile.calltrace.cycles-pp.__add_to_page_cache_locked.add_to_page_cache_lru.page_cache_ra_unbounded.generic_file_buffered_read_get_pages.generic_file_buffered_read 76.16 ± 2% -5.1 71.05 perf-profile.children.cycles-pp.native_queued_spin_lock_slowpath 38.16 ± 5% -3.5 34.66 ± 3% perf-profile.children.cycles-pp._raw_spin_lock_irqsave 50.11 ± 2% -2.6 47.51 perf-profile.children.cycles-pp.__alloc_pages_nodemask 30.76 ± 3% -2.6 28.17 ± 2% perf-profile.children.cycles-pp.lock_page_lruvec_irqsave 31.70 ± 3% -2.6 29.14 ± 2% perf-profile.children.cycles-pp.__pagevec_lru_add 31.72 ± 3% -2.6 29.16 ± 2% perf-profile.children.cycles-pp.lru_cache_add 4.13 ± 16% -1.4 2.69 ± 19% perf-profile.children.cycles-pp.get_page_from_freelist 3.98 ± 17% -1.4 2.55 ± 20% perf-profile.children.cycles-pp.rmqueue 3.73 ± 18% -1.4 2.32 ± 22% perf-profile.children.cycles-pp.rmqueue_bulk 3.60 ± 18% -1.4 2.19 ± 22% perf-profile.children.cycles-pp._raw_spin_lock 5.62 ± 5% -0.5 5.09 ± 2% perf-profile.children.cycles-pp.read_pages 5.57 ± 5% -0.5 5.05 ± 3% perf-profile.children.cycles-pp.iomap_readahead_actor 5.61 ± 5% -0.5 5.09 ± 3% perf-profile.children.cycles-pp.iomap_readahead 5.60 ± 5% -0.5 5.08 ± 3% perf-profile.children.cycles-pp.iomap_apply 5.36 ± 5% -0.5 4.85 ± 3% perf-profile.children.cycles-pp.iomap_readpage_actor 2.93 ± 5% -0.3 2.65 ± 3% perf-profile.children.cycles-pp.memset_erms 2.23 ± 5% -0.2 2.01 ± 3% perf-profile.children.cycles-pp.iomap_set_range_uptodate 0.73 ± 2% -0.1 0.67 perf-profile.children.cycles-pp.__list_del_entry_valid 0.06 ± 11% +0.0 0.08 ± 7% perf-profile.children.cycles-pp.uncharge_page 0.31 ± 7% +0.0 0.35 ± 4% perf-profile.children.cycles-pp.__count_memcg_events 0.18 ± 10% +0.0 0.23 ± 4% perf-profile.children.cycles-pp.mem_cgroup_charge_statistics 0.41 ± 7% +0.0 0.46 ± 3% perf-profile.children.cycles-pp.uncharge_batch 0.16 ± 14% +0.1 0.24 ± 7% perf-profile.children.cycles-pp.propagate_protected_usage 0.51 ± 7% +0.1 0.60 ± 4% perf-profile.children.cycles-pp.__mod_memcg_state 0.97 ± 8% +0.1 1.10 ± 3% perf-profile.children.cycles-pp.__mod_memcg_lruvec_state 0.68 ± 14% +0.2 0.93 ± 6% perf-profile.children.cycles-pp.page_counter_try_charge 0.91 ± 14% +0.3 1.25 ± 6% perf-profile.children.cycles-pp.try_charge 0.05 ± 8% +3.5 3.55 ± 9% perf-profile.children.cycles-pp.memcg_check_events 38.12 ± 2% +3.8 41.91 ± 2% perf-profile.children.cycles-pp.add_to_page_cache_lru 5.34 ± 19% +6.3 11.62 ± 6% perf-profile.children.cycles-pp.mem_cgroup_charge 6.42 ± 17% +6.4 12.77 ± 5% perf-profile.children.cycles-pp.__add_to_page_cache_locked 76.16 ± 2% -5.1 71.05 perf-profile.self.cycles-pp.native_queued_spin_lock_slowpath 2.91 ± 5% -0.3 2.63 ± 3% perf-profile.self.cycles-pp.memset_erms 2.20 ± 5% -0.2 1.98 ± 3% perf-profile.self.cycles-pp.iomap_set_range_uptodate 0.20 ± 8% -0.0 0.17 ± 4% perf-profile.self.cycles-pp.rmqueue_bulk 0.13 ± 4% -0.0 0.11 ± 6% perf-profile.self.cycles-pp.__remove_mapping 0.06 ± 11% +0.0 0.08 ± 7% perf-profile.self.cycles-pp.uncharge_page 0.31 ± 7% +0.0 0.35 ± 4% perf-profile.self.cycles-pp.__count_memcg_events 0.16 ± 14% +0.1 0.24 ± 9% perf-profile.self.cycles-pp.propagate_protected_usage 0.23 ± 13% +0.1 0.32 ± 6% perf-profile.self.cycles-pp.try_charge 0.51 ± 7% +0.1 0.60 ± 3% perf-profile.self.cycles-pp.__mod_memcg_state 0.58 ± 13% +0.2 0.75 ± 6% perf-profile.self.cycles-pp.page_counter_try_charge 2.25 ± 25% +1.8 4.05 ± 6% perf-profile.self.cycles-pp.mem_cgroup_charge 0.01 ±223% +3.5 3.54 ± 9% perf-profile.self.cycles-pp.memcg_check_events vm-scalability.median 134000 +------------------------------------------------------------------+ | +. | 132000 |-+ .+ .+ + .+ + + + | | + : +.++.+ : + + : + : + :+ : | 130000 |-+: : + : + : + : : : : + : | |.+ ++ O O :.+ +.+ O +.+.+O + O + O | 128000 |-+ O + O | | O O O O O O O O | 126000 |-+ O O O O O O | | O O O O O | 124000 |-+ O O O O O O | | O O | 122000 |-+ | | O O | 120000 +------------------------------------------------------------------+ [*] bisect-good sample [O] bisect-bad sample 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. Thanks, Oliver Sang