Greeting, FYI, we noticed a 5.6% improvement of will-it-scale.per_process_ops due to commit: commit: f030aade9165080f3539fb86fc2ce9ffc391813c ("x86/xen: Move pv specific parts of arch/x86/xen/mmu.c to mmu_pv.c") https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master in testcase: will-it-scale on test machine: 8 threads Ivy Bridge with 16G memory with following parameters: nr_task: 100% mode: process test: poll2 ucode: 0x20 cpufreq_governor: performance test-description: Will It Scale takes a testcase and runs it from 1 through to n parallel copies to see if the testcase will scale. It builds both a process and threads based test in order to see any differences between the two. test-url: https://github.com/antonblanchard/will-it-scale 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 ========================================================================================= compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase/ucode: gcc-7/performance/x86_64-rhel-7.2/process/100%/debian-x86_64-2018-04-03.cgz/lkp-ivb-d01/poll2/will-it-scale/0x20 commit: 28c11b0f79 ("x86/xen: Move pv irq related functions under CONFIG_XEN_PV umbrella") f030aade91 ("x86/xen: Move pv specific parts of arch/x86/xen/mmu.c to mmu_pv.c") 28c11b0f798c6727 f030aade9165080f3539fb86fc ---------------- -------------------------- fail:runs %reproduction fail:runs | | | :4 25% 1:4 dmesg.RIP:__do_page_fault 1:4 -25% :4 dmesg.RIP:__pte_alloc :4 25% 1:4 dmesg.RIP:skb_copy_datagram_iter 1:4 -25% :4 kmsg.c91>]usb_hcd_irq 1:4 -25% :4 kmsg.d7641b>]usb_hcd_irq :4 25% 1:4 kmsg.dea1619>]usb_hcd_irq :4 25% 1:4 kmsg.ef1c654>]usb_hcd_irq 1:4 -25% :4 kmsg.effc28>]usb_hcd_irq :4 25% 1:4 kmsg.f222649>]usb_hcd_irq :4 25% 1:4 kmsg.f7cf435>]usb_hcd_irq 0:4 0% 0:4 perf-profile.children.cycles-pp.poll_select_set_timeout 0:4 0% 0:4 perf-profile.self.cycles-pp.poll_select_set_timeout %stddev %change %stddev \ | \ 401137 +5.6% 423557 will-it-scale.per_process_ops 3209103 +5.6% 3388457 will-it-scale.workload 5.70 -1.7% 5.61 boot-time.dhcp 0.00 ± 17% -0.0 0.00 ± 38% mpstat.cpu.soft% 3839 ± 90% +304.9% 15546 ± 55% cpuidle.C3.usage 7549 ± 3% +124.5% 16951 ± 51% cpuidle.C6.usage 0.82 -7.4% 0.76 ± 2% perf-stat.cpi 8419 +7.3% 9036 ± 4% perf-stat.instructions-per-iTLB-miss 1.21 +8.0% 1.31 ± 2% perf-stat.ipc 42479 ± 6% +10.9% 47103 ± 5% sched_debug.cfs_rq:/.load.stddev 731344 ± 12% -32.9% 490912 ± 10% sched_debug.cpu.avg_idle.avg 42994 ± 6% +9.8% 47202 ± 5% sched_debug.cpu.load.stddev 3830 ± 91% +305.6% 15535 ± 55% turbostat.C3 7544 ± 3% +124.6% 16944 ± 51% turbostat.C6 0.08 ±101% +500.0% 0.51 ± 61% turbostat.CPU%c3 46.98 +1.6% 47.76 turbostat.CorWatt 0.02 ±100% +1375.0% 0.29 ± 80% turbostat.Pkg%pc3 51.87 +1.5% 52.65 turbostat.PkgWatt 46.78 -4.4 42.40 perf-profile.calltrace.cycles-pp.__fget_light.do_sys_poll.__x64_sys_poll.do_syscall_64.entry_SYSCALL_64_after_hwframe 86.64 -0.9 85.77 perf-profile.calltrace.cycles-pp.do_sys_poll.__x64_sys_poll.do_syscall_64.entry_SYSCALL_64_after_hwframe 89.84 -0.7 89.16 perf-profile.calltrace.cycles-pp.__x64_sys_poll.do_syscall_64.entry_SYSCALL_64_after_hwframe 91.35 -0.6 90.75 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe 91.68 -0.6 91.12 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe 0.70 ± 5% +0.1 0.77 ± 4% perf-profile.calltrace.cycles-pp.__might_fault._copy_from_user.do_sys_poll.__x64_sys_poll.do_syscall_64 3.27 ± 2% +0.2 3.43 ± 3% perf-profile.calltrace.cycles-pp.copy_user_enhanced_fast_string._copy_from_user.do_sys_poll.__x64_sys_poll.do_syscall_64 3.66 +0.2 3.87 perf-profile.calltrace.cycles-pp.__entry_SYSCALL_64_trampoline 3.67 +0.3 3.94 perf-profile.calltrace.cycles-pp.syscall_return_via_sysret 4.25 ± 2% +0.3 4.52 ± 2% perf-profile.calltrace.cycles-pp._copy_from_user.do_sys_poll.__x64_sys_poll.do_syscall_64.entry_SYSCALL_64_after_hwframe 44.61 -4.6 39.99 perf-profile.children.cycles-pp.__fget_light 89.08 -0.7 88.36 perf-profile.children.cycles-pp.do_sys_poll 89.92 -0.7 89.23 perf-profile.children.cycles-pp.__x64_sys_poll 91.42 -0.6 90.82 perf-profile.children.cycles-pp.do_syscall_64 91.76 -0.5 91.21 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe 3.29 ± 2% +0.2 3.45 ± 3% perf-profile.children.cycles-pp.copy_user_enhanced_fast_string 3.75 +0.2 3.95 perf-profile.children.cycles-pp.__entry_SYSCALL_64_trampoline 4.32 ± 2% +0.3 4.59 perf-profile.children.cycles-pp._copy_from_user 4.18 +0.3 4.48 perf-profile.children.cycles-pp.syscall_return_via_sysret 4.54 ± 2% +0.4 4.94 perf-profile.children.cycles-pp.__fdget 42.30 -4.9 37.43 perf-profile.self.cycles-pp.__fget_light 0.42 ± 4% +0.0 0.46 ± 4% perf-profile.self.cycles-pp.___might_sleep 0.35 +0.1 0.42 ± 3% perf-profile.self.cycles-pp.entry_SYSCALL_64_after_hwframe 3.24 ± 2% +0.2 3.40 ± 3% perf-profile.self.cycles-pp.copy_user_enhanced_fast_string 3.75 +0.2 3.95 perf-profile.self.cycles-pp.__entry_SYSCALL_64_trampoline 2.24 +0.3 2.50 ± 3% perf-profile.self.cycles-pp.__fdget 4.17 ± 2% +0.3 4.47 perf-profile.self.cycles-pp.syscall_return_via_sysret 36.22 +3.5 39.71 perf-profile.self.cycles-pp.do_sys_poll will-it-scale.per_process_ops 440000 +-+----------------------------------------------------------------+ | | 430000 +-+ O O O O O O | 420000 O-+O O O O O O O O O O O O O O O O | O O O O | 410000 +-+ | | .+. +.+.. + | 400000 +-+ .+.. .+.. .+..+. +.. : + | |..+ + + +.+ + : + | 390000 +-+ : :: : | 380000 +-+ : : : : | | : : : : | 370000 +-+ : : : | | + + | 360000 +-+----------------------------------------------------------------+ [*] 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