Greeting, FYI, we noticed a 11.1% improvement of stress-ng.eventfd.ops_per_sec due to commit: commit: 763802b53a427ed3cbd419dbba255c414fdd9e7c ("x86/mm: split vmalloc_sync_all()") https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master in testcase: stress-ng on test machine: 96 threads Intel(R) Xeon(R) Gold 6252 CPU @ 2.10GHz with 192G memory with following parameters: nr_threads: 100% disk: 1HDD testtime: 1s class: os cpufreq_governor: performance ucode: 0x500002c fs: ext4 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 run job.yaml ========================================================================================= class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_group/testcase/testtime/ucode: os/gcc-7/performance/1HDD/ext4/x86_64-rhel-7.6/100%/debian-x86_64-20191114.cgz/lkp-csl-2sp5/stress-ng/1s/0x500002c commit: 0715e6c516 ("mm, slub: prevent kmalloc_node crashes and memory leaks") 763802b53a ("x86/mm: split vmalloc_sync_all()") 0715e6c516f106ed 763802b53a427ed3cbd419dbba2 ---------------- --------------------------- fail:runs %reproduction fail:runs | | | 1:4 -25% :4 dmesg.WARNING:at_ip__fsnotify_parent/0x :4 25% 1:4 dmesg.WARNING:at_ip__slab_free/0x 1:4 -25% :4 dmesg.WARNING:at_ip_perf_event_mmap_output/0x 0:4 0% 0:4 perf-profile.children.cycles-pp.error_entry %stddev %change %stddev \ | \ 676172 +11.1% 751446 stress-ng.eventfd.ops 676084 +11.1% 751365 stress-ng.eventfd.ops_per_sec 10015 ± 9% +67.8% 16808 ± 9% stress-ng.opcode.ops 9972 ± 9% +68.1% 16762 ± 9% stress-ng.opcode.ops_per_sec 8718 +69.5% 14779 ± 2% stress-ng.seccomp.ops 8717 +69.5% 14779 ± 2% stress-ng.seccomp.ops_per_sec 890754 ± 3% +17.2% 1043917 ± 3% stress-ng.sigrt.ops 884276 ± 3% +17.2% 1036699 ± 3% stress-ng.sigrt.ops_per_sec 828196 ± 3% +11.8% 925775 ± 3% stress-ng.sigsuspend.ops 826776 ± 3% +11.8% 924425 ± 3% stress-ng.sigsuspend.ops_per_sec 2130447 ± 7% -27.7% 1541253 stress-ng.switch.ops 2130009 ± 7% -27.7% 1540958 stress-ng.switch.ops_per_sec 38755106 +3.6% 40159267 stress-ng.time.voluntary_context_switches 662944 +11.1% 736437 ± 2% stress-ng.userfaultfd.ops 662942 +11.1% 736447 ± 2% stress-ng.userfaultfd.ops_per_sec 7663774 ± 9% -13.0% 6666649 ± 3% cpuidle.POLL.time 197406 ± 5% +80.3% 355925 ± 41% numa-meminfo.node1.Mapped 6446 ± 9% -10.0% 5803 ± 8% numa-vmstat.node0.nr_page_table_pages 50338 ± 7% +77.9% 89534 ± 42% numa-vmstat.node1.nr_mapped 140973 ± 10% -16.8% 117339 ± 13% softirqs.CPU2.NET_RX 145464 ± 8% -14.6% 124231 ± 11% softirqs.CPU22.NET_RX 3847 +9.6% 4218 ± 2% perf-stat.i.major-faults 43.17 +1.5 44.72 perf-stat.overall.node-load-miss-rate% 3655 +10.1% 4022 perf-stat.ps.major-faults 5694 ± 5% -31.5% 3902 ± 5% slabinfo.DCCP.active_objs 322.50 ± 6% -31.8% 220.00 ± 5% slabinfo.DCCP.active_slabs 5814 ± 6% -31.8% 3968 ± 6% slabinfo.DCCP.num_objs 322.50 ± 6% -31.8% 220.00 ± 5% slabinfo.DCCP.num_slabs 6748 ± 7% -28.3% 4836 ± 2% slabinfo.ccid2_hc_tx_sock.active_objs 6821 ± 8% -28.4% 4881 ± 2% slabinfo.ccid2_hc_tx_sock.num_objs 4.10 ± 39% -2.7 1.45 ±173% perf-profile.calltrace.cycles-pp.__sched_text_start.schedule.__x64_sys_sched_yield.do_syscall_64.entry_SYSCALL_64_after_hwframe 4.00 ±104% -1.6 2.40 ±173% perf-profile.calltrace.cycles-pp.__sched_yield 3.23 ±103% -1.3 1.89 ±173% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__sched_yield 3.11 ±103% -1.3 1.82 ±173% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__sched_yield 4.87 ± 39% -3.1 1.72 ±173% perf-profile.children.cycles-pp.__x64_sys_sched_yield 4.11 ±103% -1.7 2.40 ±173% perf-profile.children.cycles-pp.__sched_yield 0.16 ± 18% +0.1 0.22 ± 8% perf-profile.children.cycles-pp.prep_new_page 0.05 ± 58% +0.1 0.13 ± 32% perf-profile.children.cycles-pp.native_flush_tlb_one_user 0.09 ± 23% +0.1 0.17 ± 37% perf-profile.children.cycles-pp.__pte_alloc 0.57 ± 28% +0.4 0.94 ± 23% perf-profile.children.cycles-pp.get_page_from_freelist 0.63 ± 27% +0.4 1.06 ± 23% perf-profile.children.cycles-pp.__alloc_pages_nodemask 0.05 ± 58% +0.1 0.13 ± 32% perf-profile.self.cycles-pp.native_flush_tlb_one_user 597911 ± 37% -73.5% 158203 ±132% sched_debug.cfs_rq:/.MIN_vruntime.avg 5478739 ± 25% -61.0% 2138711 ± 99% sched_debug.cfs_rq:/.MIN_vruntime.max 1513560 ± 12% -66.9% 501143 ±108% sched_debug.cfs_rq:/.MIN_vruntime.stddev 597912 ± 37% -73.5% 158203 ±132% sched_debug.cfs_rq:/.max_vruntime.avg 5478808 ± 25% -61.0% 2138713 ± 99% sched_debug.cfs_rq:/.max_vruntime.max 1513567 ± 12% -66.9% 501143 ±108% sched_debug.cfs_rq:/.max_vruntime.stddev 0.65 ± 12% +19.6% 0.78 ± 3% sched_debug.cfs_rq:/.nr_running.avg 681.06 ± 11% +23.1% 838.62 ± 4% sched_debug.cfs_rq:/.util_avg.avg 11.96 ± 18% +503.1% 72.12 ± 60% sched_debug.cpu.clock.stddev 11.96 ± 18% +503.1% 72.12 ± 60% sched_debug.cpu.clock_task.stddev 28699 ± 15% +56.3% 44844 ± 30% sched_debug.cpu.nr_switches.stddev 1799168 ± 3% +14.6% 2061693 ± 8% sched_debug.cpu.sched_count.max 128733 ± 5% +16.5% 149939 ± 5% sched_debug.cpu.ttwu_local.max 12286 ± 17% +52.3% 18712 ± 9% sched_debug.cpu.ttwu_local.stddev 1477973 ± 4% +16.5% 1722271 ± 10% sched_debug.cpu.yld_count.max 199.75 ± 20% +32.2% 264.00 ± 16% interrupts.100:PCI-MSI.31981633-edge.i40e-eth0-TxRx-64 186.75 ± 19% +45.2% 271.25 ± 15% interrupts.102:PCI-MSI.31981635-edge.i40e-eth0-TxRx-66 185.75 ± 24% +43.2% 266.00 ± 9% interrupts.105:PCI-MSI.31981638-edge.i40e-eth0-TxRx-69 514.00 ± 53% -52.1% 246.25 ± 17% interrupts.121:PCI-MSI.31981654-edge.i40e-eth0-TxRx-85 165.75 ± 14% +50.1% 248.75 ± 16% interrupts.37:PCI-MSI.31981570-edge.i40e-eth0-TxRx-1 219.75 ± 5% +19.3% 262.25 ± 8% interrupts.74:PCI-MSI.31981607-edge.i40e-eth0-TxRx-38 163.50 ± 28% +52.8% 249.75 ± 26% interrupts.88:PCI-MSI.31981621-edge.i40e-eth0-TxRx-52 353.75 ± 30% -30.1% 247.25 ± 9% interrupts.92:PCI-MSI.31981625-edge.i40e-eth0-TxRx-56 5023300 +38.4% 6952194 ± 4% interrupts.CAL:Function_call_interrupts 49997 ± 3% +26.4% 63199 ± 5% interrupts.CPU0.CAL:Function_call_interrupts 165.00 ± 14% +50.3% 248.00 ± 16% interrupts.CPU1.37:PCI-MSI.31981570-edge.i40e-eth0-TxRx-1 50367 ± 5% +23.5% 62178 ± 5% interrupts.CPU1.CAL:Function_call_interrupts 49188 ± 4% +28.5% 63218 ± 4% interrupts.CPU10.CAL:Function_call_interrupts 49113 ± 2% +29.9% 63780 ± 4% interrupts.CPU11.CAL:Function_call_interrupts 9008 ± 16% +15.6% 10417 ± 13% interrupts.CPU11.TLB:TLB_shootdowns 48444 +31.1% 63517 ± 4% interrupts.CPU12.CAL:Function_call_interrupts 7854 ± 4% +25.6% 9867 ± 7% interrupts.CPU12.TLB:TLB_shootdowns 49604 ± 4% +28.2% 63574 ± 5% interrupts.CPU13.CAL:Function_call_interrupts 49028 ± 2% +29.2% 63359 ± 5% interrupts.CPU14.CAL:Function_call_interrupts 48957 ± 3% +30.2% 63747 ± 5% interrupts.CPU15.CAL:Function_call_interrupts 49046 ± 2% +29.2% 63385 ± 4% interrupts.CPU16.CAL:Function_call_interrupts 50026 ± 3% +27.6% 63819 ± 5% interrupts.CPU17.CAL:Function_call_interrupts 49931 ± 3% +28.5% 64178 ± 5% interrupts.CPU18.CAL:Function_call_interrupts 49739 ± 2% +28.5% 63939 ± 5% interrupts.CPU19.CAL:Function_call_interrupts 49687 ± 5% +27.1% 63162 ± 4% interrupts.CPU2.CAL:Function_call_interrupts 49652 ± 2% +28.7% 63885 ± 5% interrupts.CPU20.CAL:Function_call_interrupts 50193 ± 2% +28.8% 64670 ± 5% interrupts.CPU21.CAL:Function_call_interrupts 49557 ± 2% +30.0% 64404 ± 5% interrupts.CPU22.CAL:Function_call_interrupts 50152 ± 28% -44.2% 27990 ± 55% interrupts.CPU22.RES:Rescheduling_interrupts 51078 ± 5% +26.9% 64811 ± 4% interrupts.CPU23.CAL:Function_call_interrupts 51801 ± 2% +29.2% 66939 ± 4% interrupts.CPU24.CAL:Function_call_interrupts 50932 ± 3% +30.1% 66261 ± 3% interrupts.CPU25.CAL:Function_call_interrupts 9788 ± 11% +22.0% 11940 ± 7% interrupts.CPU25.TLB:TLB_shootdowns 50882 ± 2% +30.2% 66235 ± 4% interrupts.CPU26.CAL:Function_call_interrupts 51005 +30.6% 66598 ± 6% interrupts.CPU27.CAL:Function_call_interrupts 51409 +27.3% 65447 ± 5% interrupts.CPU28.CAL:Function_call_interrupts 51061 +30.0% 66375 ± 4% interrupts.CPU29.CAL:Function_call_interrupts 49434 ± 3% +26.7% 62656 ± 4% interrupts.CPU3.CAL:Function_call_interrupts 50720 +31.4% 66654 ± 4% interrupts.CPU30.CAL:Function_call_interrupts 51178 +31.3% 67202 ± 5% interrupts.CPU31.CAL:Function_call_interrupts 50693 ± 2% +32.1% 66987 ± 6% interrupts.CPU32.CAL:Function_call_interrupts 50574 +33.1% 67302 ± 4% interrupts.CPU33.CAL:Function_call_interrupts 50294 ± 2% +37.0% 68886 ± 3% interrupts.CPU34.CAL:Function_call_interrupts 33348 ± 15% -53.5% 15505 ± 20% interrupts.CPU34.RES:Rescheduling_interrupts 9363 ± 14% +54.9% 14504 ± 7% interrupts.CPU34.TLB:TLB_shootdowns 51489 +31.2% 67572 ± 5% interrupts.CPU35.CAL:Function_call_interrupts 50404 +34.9% 67973 ± 4% interrupts.CPU36.CAL:Function_call_interrupts 9037 ± 6% +34.4% 12142 ± 13% interrupts.CPU36.TLB:TLB_shootdowns 52030 +29.1% 67166 ± 5% interrupts.CPU37.CAL:Function_call_interrupts 219.00 ± 6% +19.3% 261.25 ± 8% interrupts.CPU38.74:PCI-MSI.31981607-edge.i40e-eth0-TxRx-38 51469 +31.6% 67738 ± 4% interrupts.CPU38.CAL:Function_call_interrupts 52479 +32.2% 69376 ± 6% interrupts.CPU39.CAL:Function_call_interrupts 49292 ± 2% +27.5% 62828 ± 4% interrupts.CPU4.CAL:Function_call_interrupts 52398 ± 3% +31.1% 68717 ± 5% interrupts.CPU40.CAL:Function_call_interrupts 51368 +34.3% 68992 ± 5% interrupts.CPU41.CAL:Function_call_interrupts 51143 +34.3% 68692 ± 5% interrupts.CPU42.CAL:Function_call_interrupts 50781 +35.9% 69023 ± 5% interrupts.CPU43.CAL:Function_call_interrupts 51994 +34.8% 70068 ± 5% interrupts.CPU44.CAL:Function_call_interrupts 52166 +34.2% 70010 ± 5% interrupts.CPU45.CAL:Function_call_interrupts 63778 ± 10% -51.8% 30741 ± 62% interrupts.CPU45.RES:Rescheduling_interrupts 50979 +37.4% 70043 ± 4% interrupts.CPU46.CAL:Function_call_interrupts 8717 ± 10% +34.8% 11751 ± 9% interrupts.CPU46.TLB:TLB_shootdowns 52047 +34.2% 69854 ± 5% interrupts.CPU47.CAL:Function_call_interrupts 38043 ± 34% -57.5% 16167 ± 19% interrupts.CPU47.RES:Rescheduling_interrupts 52734 ± 5% +33.2% 70262 ± 5% interrupts.CPU48.CAL:Function_call_interrupts 52017 ± 4% +35.4% 70428 ± 5% interrupts.CPU49.CAL:Function_call_interrupts 49061 ± 3% +27.7% 62666 ± 5% interrupts.CPU5.CAL:Function_call_interrupts 52911 ± 4% +33.4% 70572 ± 4% interrupts.CPU50.CAL:Function_call_interrupts 52317 ± 3% +35.3% 70796 ± 4% interrupts.CPU51.CAL:Function_call_interrupts 162.75 ± 28% +53.0% 249.00 ± 26% interrupts.CPU52.88:PCI-MSI.31981621-edge.i40e-eth0-TxRx-52 51758 ± 2% +37.8% 71344 ± 4% interrupts.CPU52.CAL:Function_call_interrupts 53500 ± 7% +33.4% 71350 ± 4% interrupts.CPU53.CAL:Function_call_interrupts 53055 ± 4% +34.6% 71431 ± 4% interrupts.CPU54.CAL:Function_call_interrupts 52696 ± 2% +37.0% 72214 ± 4% interrupts.CPU55.CAL:Function_call_interrupts 353.50 ± 30% -30.1% 247.00 ± 10% interrupts.CPU56.92:PCI-MSI.31981625-edge.i40e-eth0-TxRx-56 52234 ± 2% +39.0% 72622 ± 5% interrupts.CPU56.CAL:Function_call_interrupts 52386 ± 2% +38.6% 72629 ± 5% interrupts.CPU57.CAL:Function_call_interrupts 52690 ± 5% +38.8% 73139 ± 4% interrupts.CPU58.CAL:Function_call_interrupts 52458 ± 4% +41.5% 74240 ± 3% interrupts.CPU59.CAL:Function_call_interrupts 49151 ± 2% +27.6% 62729 ± 5% interrupts.CPU6.CAL:Function_call_interrupts 53260 ± 5% +39.9% 74485 ± 4% interrupts.CPU60.CAL:Function_call_interrupts 52369 ± 3% +42.5% 74637 ± 4% interrupts.CPU61.CAL:Function_call_interrupts 52702 ± 2% +41.8% 74741 ± 4% interrupts.CPU62.CAL:Function_call_interrupts 52506 ± 3% +44.0% 75587 ± 4% interrupts.CPU63.CAL:Function_call_interrupts 64174 ± 77% -73.4% 17058 ± 22% interrupts.CPU63.RES:Rescheduling_interrupts 199.00 ± 20% +32.2% 263.00 ± 16% interrupts.CPU64.100:PCI-MSI.31981633-edge.i40e-eth0-TxRx-64 52978 +41.9% 75180 ± 5% interrupts.CPU64.CAL:Function_call_interrupts 53057 ± 2% +43.1% 75930 ± 4% interrupts.CPU65.CAL:Function_call_interrupts 186.00 ± 20% +45.3% 270.25 ± 15% interrupts.CPU66.102:PCI-MSI.31981635-edge.i40e-eth0-TxRx-66 52732 +45.7% 76822 ± 4% interrupts.CPU66.CAL:Function_call_interrupts 8489 ± 4% +18.5% 10063 ± 6% interrupts.CPU66.TLB:TLB_shootdowns 52828 +45.1% 76660 ± 4% interrupts.CPU67.CAL:Function_call_interrupts 55361 ± 18% -51.5% 26858 ± 58% interrupts.CPU67.RES:Rescheduling_interrupts 53993 ± 4% +43.1% 77270 ± 4% interrupts.CPU68.CAL:Function_call_interrupts 185.00 ± 25% +43.5% 265.50 ± 9% interrupts.CPU69.105:PCI-MSI.31981638-edge.i40e-eth0-TxRx-69 55097 ± 5% +43.1% 78872 ± 3% interrupts.CPU69.CAL:Function_call_interrupts 48734 ± 2% +30.5% 63619 ± 5% interrupts.CPU7.CAL:Function_call_interrupts 10300 ± 18% +33.5% 13748 ± 14% interrupts.CPU7.TLB:TLB_shootdowns 53283 ± 2% +47.1% 78406 ± 4% interrupts.CPU70.CAL:Function_call_interrupts 54946 ± 6% +43.2% 78707 ± 4% interrupts.CPU71.CAL:Function_call_interrupts 56696 ± 2% +41.4% 80153 ± 4% interrupts.CPU72.CAL:Function_call_interrupts 54954 +48.3% 81488 ± 5% interrupts.CPU73.CAL:Function_call_interrupts 54304 +48.1% 80419 ± 4% interrupts.CPU74.CAL:Function_call_interrupts 55555 ± 2% +48.1% 82290 ± 6% interrupts.CPU75.CAL:Function_call_interrupts 55435 +45.8% 80815 ± 5% interrupts.CPU76.CAL:Function_call_interrupts 54668 +49.2% 81554 ± 4% interrupts.CPU77.CAL:Function_call_interrupts 54432 +50.5% 81924 ± 5% interrupts.CPU78.CAL:Function_call_interrupts 55232 +51.4% 83610 ± 5% interrupts.CPU79.CAL:Function_call_interrupts 49125 ± 2% +28.1% 62928 ± 5% interrupts.CPU8.CAL:Function_call_interrupts 54689 ± 2% +51.6% 82895 ± 5% interrupts.CPU80.CAL:Function_call_interrupts 55099 +51.3% 83373 ± 4% interrupts.CPU81.CAL:Function_call_interrupts 54596 +53.8% 83975 ± 5% interrupts.CPU82.CAL:Function_call_interrupts 9389 ± 13% +28.6% 12071 ± 14% interrupts.CPU82.TLB:TLB_shootdowns 56002 +49.1% 83509 ± 4% interrupts.CPU83.CAL:Function_call_interrupts 54713 +55.2% 84921 ± 5% interrupts.CPU84.CAL:Function_call_interrupts 513.25 ± 53% -52.2% 245.50 ± 18% interrupts.CPU85.121:PCI-MSI.31981654-edge.i40e-eth0-TxRx-85 55910 ± 3% +51.5% 84718 ± 5% interrupts.CPU85.CAL:Function_call_interrupts 55008 +53.7% 84569 ± 4% interrupts.CPU86.CAL:Function_call_interrupts 55946 +52.4% 85236 ± 5% interrupts.CPU87.CAL:Function_call_interrupts 57153 ± 2% +50.0% 85758 ± 5% interrupts.CPU88.CAL:Function_call_interrupts 55969 ± 2% +54.4% 86444 ± 4% interrupts.CPU89.CAL:Function_call_interrupts 49276 ± 2% +27.3% 62729 ± 5% interrupts.CPU9.CAL:Function_call_interrupts 56219 +52.3% 85617 ± 5% interrupts.CPU90.CAL:Function_call_interrupts 56359 +55.2% 87469 ± 4% interrupts.CPU91.CAL:Function_call_interrupts 55887 +55.7% 87012 ± 4% interrupts.CPU92.CAL:Function_call_interrupts 56339 +54.8% 87235 ± 5% interrupts.CPU93.CAL:Function_call_interrupts 55958 +57.2% 87966 ± 4% interrupts.CPU94.CAL:Function_call_interrupts 9443 ± 13% +23.7% 11682 ± 6% interrupts.CPU94.TLB:TLB_shootdowns 56661 ± 2% +54.9% 87793 ± 4% interrupts.CPU95.CAL:Function_call_interrupts 8398 ± 50% -55.1% 3769 ± 58% interrupts.IWI:IRQ_work_interrupts stress-ng.seccomp.ops 16000 +-------------------------------------------------------------------+ | O O OO OO O O OO | 14000 |-O O O O | 12000 |-+ | | | 10000 |-+ | |. +.++.+ ++.++.+.+ +. +.++.+ +.++. +.++.+ ++.+. +.+ .++.++.++.| 8000 |-+ : : +.+ + : : + : : + + | | : : : : : : | 6000 |-+ : : : : : : | 4000 |-+ :: :: :: | | :: :: :: | 2000 |-+ :: :: :: | | : : : | 0 +-------------------------------------------------------------------+ stress-ng.seccomp.ops_per_sec 16000 +-------------------------------------------------------------------+ | O O OO OO O O OO | 14000 |-O O O O | 12000 |-+ | | | 10000 |-+ | |. +.++.+ ++.++.+.+ +. +.++.+ +.++. +.++.+ ++.+. +.+ .+ .++.++.| 8000 |-+ : : +.+ + : : + : : + + + | | : : : : : : | 6000 |-+ : : : : : : | 4000 |-+ :: :: :: | | :: :: :: | 2000 |-+ :: :: :: | | : : : | 0 +-------------------------------------------------------------------+ [*] 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, Rong Chen