| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20250811112537.GA7448@lst.de> Date: Mon, 11 Aug 2025 13:25:37 +0200 From: Christoph Hellwig <hch@....de> To: kernel test robot <oliver.sang@...el.com> Cc: Damien Le Moal <dlemoal@...nel.org>, oe-lkp@...ts.linux.dev, lkp@...el.com, linux-kernel@...r.kernel.org, Jens Axboe <axboe@...nel.dk>, "Martin K . Petersen" <martin.petersen@...cle.com>, Christoph Hellwig <hch@....de>, John Garry <john.g.garry@...cle.com>, Johannes Thumshirn <johannes.thumshirn@....com>, Hannes Reinecke <hare@...e.de>, linux-block@...r.kernel.org Subject: Re: [linus:master] [block] 9b8b84879d: fio.read_iops 59.4% regression On Mon, Aug 11, 2025 at 01:10:07PM +0800, kernel test robot wrote: > > > Hello, > > kernel test robot noticed a 59.4% regression of fio.read_iops on: >From looking at the changes below, this applies to the iops reported by vmstat. Assuming that's the ones reported by vmstat -d, it counts the disk IOPS. Reducing them for the same large I/O size workload would indeed by the intent of this change. > v6.16-rc4 9b8b84879d4adc506b0d3944e20 > ---------------- --------------------------- > %stddev %change %stddev > \ | \ > 794920 -24.2% 602314 cpuidle..usage > 19205 -19.2% 15517 ± 3% meminfo.Shmem > 1.96 -31.8% 1.34 iostat.cpu.system > 1.20 -4.1% 1.15 iostat.cpu.user > 0.46 -0.1 0.39 mpstat.cpu.all.irq% > 0.41 -0.2 0.19 ± 2% mpstat.cpu.all.soft% > 1.07 -0.3 0.74 mpstat.cpu.all.sys% > 537011 -59.3% 218357 ± 2% vmstat.io.bi > 3506 -44.0% 1963 vmstat.system.cs > 2045 -21.3% 1609 vmstat.system.in > 20391442 -71.1% 5903000 ± 7% proc-vmstat.nr_foll_pin_acquired > 20371384 -71.2% 5868357 ± 7% proc-vmstat.nr_foll_pin_released > 4810 -19.2% 3886 ± 3% proc-vmstat.nr_shmem > 1.633e+08 -59.3% 66380544 ± 2% proc-vmstat.pgpgin > 3372 ± 7% -75.0% 842.18 ± 12% sched_debug.cpu.clock_task.stddev > 362239 -53.4% 168976 sched_debug.cpu.nr_switches.avg > 573921 -62.6% 214563 ± 3% sched_debug.cpu.nr_switches.max > 219534 ± 9% -40.2% 131216 ± 7% sched_debug.cpu.nr_switches.min > 142166 ± 8% -76.5% 33405 ± 19% sched_debug.cpu.nr_switches.stddev > 252.27 ± 12% -40.1% 151.04 ± 24% sched_debug.cpu.nr_uninterruptible.max > -603.88 -51.9% -290.65 sched_debug.cpu.nr_uninterruptible.min > 352.77 ± 10% -50.7% 173.92 ± 11% sched_debug.cpu.nr_uninterruptible.stddev > 79.78 -69.7 10.03 ± 17% fio.latency_1000ms% > 19.05 ± 4% +63.5 82.55 fio.latency_2000ms% > 0.09 ± 11% +0.2 0.32 ± 22% fio.latency_500ms% > 1.03 ± 15% +1.0 2.05 ± 16% fio.latency_750ms% > 531.17 -59.4% 215.81 ± 2% fio.read_bw_MBps > 1.032e+09 +74.2% 1.797e+09 ± 2% fio.read_clat_90%_ns > 1.059e+09 +88.5% 1.996e+09 ± 2% fio.read_clat_95%_ns > 1.111e+09 +1439.6% 1.711e+10 fio.read_clat_99%_ns > 9.325e+08 +146.3% 2.296e+09 ± 2% fio.read_clat_mean_ns > 81374983 ± 4% +6906.8% 5.702e+09 ± 7% fio.read_clat_stddev > 265.59 -59.4% 107.90 ± 2% fio.read_iops > 30096187 +146.2% 74087641 ± 2% fio.read_slat_mean_us > 8137305 ± 5% +12002.2% 9.848e+08 fio.read_slat_stddev > 3.265e+08 -59.3% 1.328e+08 ± 2% fio.time.file_system_inputs > 2.00 -50.0% 1.00 fio.time.percent_of_cpu_this_job_got > 192892 -65.9% 65682 fio.time.voluntary_context_switches > 79712 -59.3% 32412 ± 2% fio.workload > 0.04 ± 12% +58.8% 0.06 ± 8% perf-sched.sch_delay.avg.ms.rcu_gp_kthread.kthread.ret_from_fork.ret_from_fork_asm > 0.04 ± 6% +26.3% 0.05 ± 7% perf-sched.sch_delay.avg.ms.schedule_timeout.rcu_gp_fqs_loop.rcu_gp_kthread.kthread > 0.02 ± 9% +1688.7% 0.28 ±215% perf-sched.sch_delay.avg.ms.worker_thread.kthread.ret_from_fork.ret_from_fork_asm > 20.28 +79.9% 36.48 ± 5% perf-sched.total_wait_and_delay.average.ms > 10366 -57.3% 4429 ± 20% perf-sched.total_wait_and_delay.count.ms > 2608 ± 10% +73.2% 4517 ± 16% perf-sched.total_wait_and_delay.max.ms > 20.26 +79.8% 36.42 ± 5% perf-sched.total_wait_time.average.ms > 2608 ± 10% +73.2% 4517 ± 16% perf-sched.total_wait_time.max.ms > 12.62 ± 5% -12.2% 11.08 ± 8% perf-sched.wait_and_delay.avg.ms.anon_pipe_read.vfs_read.ksys_read.do_syscall_64 > 10.19 ± 3% +2433.3% 258.14 ± 46% perf-sched.wait_and_delay.avg.ms.worker_thread.kthread.ret_from_fork.ret_from_fork_asm > 1251 ± 3% +12.0% 1401 ± 5% perf-sched.wait_and_delay.count.anon_pipe_read.vfs_read.ksys_read.do_syscall_64 > 2688 -93.4% 178.00 ±237% perf-sched.wait_and_delay.count.io_schedule.blk_mq_get_tag.__blk_mq_alloc_requests.blk_mq_submit_bio > 26.88 -10.7% 24.00 ± 2% perf-sched.wait_and_delay.count.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop.0.do_poll > 820.75 ± 7% +12.7% 924.62 ± 5% perf-sched.wait_and_delay.count.schedule_timeout.rcu_gp_fqs_loop.rcu_gp_kthread.kthread > 85.75 ± 12% -30.0% 60.00 ± 21% perf-sched.wait_and_delay.count.smpboot_thread_fn.kthread.ret_from_fork.ret_from_fork_asm > 3976 -93.5% 260.25 ±170% perf-sched.wait_and_delay.count.worker_thread.kthread.ret_from_fork.ret_from_fork_asm > 12.62 ± 5% -12.2% 11.08 ± 8% perf-sched.wait_time.avg.ms.anon_pipe_read.vfs_read.ksys_read.do_syscall_64 > 10.17 ± 3% +2434.4% 257.87 ± 46% perf-sched.wait_time.avg.ms.worker_thread.kthread.ret_from_fork.ret_from_fork_asm > 51939177 -9.4% 47072722 ± 2% perf-stat.i.branch-instructions > 9.12 -1.5 7.62 perf-stat.i.branch-miss-rate% > 3695435 -16.8% 3076023 ± 2% perf-stat.i.branch-misses > 36.34 ± 3% +4.6 40.96 perf-stat.i.cache-miss-rate% > 8932098 ± 2% -29.9% 6263038 perf-stat.i.cache-misses > 24645568 -37.9% 15296925 perf-stat.i.cache-references > 3527 -44.1% 1971 perf-stat.i.context-switches > 3.78 +7.3% 4.06 perf-stat.i.cpi > 4.492e+08 -12.4% 3.934e+08 perf-stat.i.cpu-cycles > 37.18 ± 8% -35.4% 24.01 ± 16% perf-stat.i.cpu-migrations > 47.65 ± 2% +39.6% 66.51 ± 5% perf-stat.i.cycles-between-cache-misses > 2.559e+08 -9.8% 2.309e+08 ± 2% perf-stat.i.instructions > 0.17 -16.3% 0.14 ± 3% perf-stat.i.metric.K/sec > 34.91 ± 2% -22.6% 27.02 perf-stat.overall.MPKI > 7.12 -0.6 6.54 perf-stat.overall.branch-miss-rate% > 36.25 ± 3% +4.6 40.89 perf-stat.overall.cache-miss-rate% > 1.76 -3.1% 1.70 perf-stat.overall.cpi > 50.32 ± 2% +25.2% 62.99 perf-stat.overall.cycles-between-cache-misses > 0.57 +3.2% 0.59 perf-stat.overall.ipc > 964879 +122.7% 2148375 ± 2% perf-stat.overall.path-length > 51775109 -9.1% 47059552 perf-stat.ps.branch-instructions > 3683872 -16.4% 3079080 perf-stat.ps.branch-misses > 8902465 ± 2% -30.0% 6234709 perf-stat.ps.cache-misses > 24563872 -37.9% 15247155 perf-stat.ps.cache-references > 3515 -44.2% 1963 perf-stat.ps.context-switches > 4.477e+08 -12.3% 3.927e+08 perf-stat.ps.cpu-cycles > 37.05 ± 8% -35.2% 23.99 ± 16% perf-stat.ps.cpu-migrations > 2.551e+08 -9.5% 2.308e+08 perf-stat.ps.instructions > 7.691e+10 -9.5% 6.959e+10 perf-stat.total.instructions > 14.40 ± 13% -11.2 3.15 ± 97% perf-profile.calltrace.cycles-pp.asm_common_interrupt.cpuidle_enter_state.cpuidle_enter.cpuidle_idle_call.do_idle > 14.28 ± 14% -11.1 3.14 ± 98% perf-profile.calltrace.cycles-pp.common_interrupt.asm_common_interrupt.cpuidle_enter_state.cpuidle_enter.cpuidle_idle_call > 9.53 ± 18% -7.6 1.95 ±104% perf-profile.calltrace.cycles-pp.__irq_exit_rcu.common_interrupt.asm_common_interrupt.cpuidle_enter_state.cpuidle_enter > 9.40 ± 18% -7.5 1.94 ±104% perf-profile.calltrace.cycles-pp.handle_softirqs.__irq_exit_rcu.common_interrupt.asm_common_interrupt.cpuidle_enter_state > 11.38 ± 10% -7.4 3.93 ± 81% perf-profile.calltrace.cycles-pp.syscall > 10.63 ± 11% -7.0 3.63 ± 80% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.syscall > 10.43 ± 11% -6.9 3.56 ± 80% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.syscall > 8.43 ± 12% -6.5 1.92 ±104% perf-profile.calltrace.cycles-pp.blk_done_softirq.handle_softirqs.__irq_exit_rcu.common_interrupt.asm_common_interrupt > 9.51 ± 12% -6.2 3.27 ± 79% perf-profile.calltrace.cycles-pp.__x64_sys_io_submit.do_syscall_64.entry_SYSCALL_64_after_hwframe.syscall > 9.40 ± 12% -6.2 3.22 ± 79% perf-profile.calltrace.cycles-pp.io_submit_one.__x64_sys_io_submit.do_syscall_64.entry_SYSCALL_64_after_hwframe.syscall > 9.14 ± 12% -6.0 3.10 ± 79% perf-profile.calltrace.cycles-pp.aio_read.io_submit_one.__x64_sys_io_submit.do_syscall_64.entry_SYSCALL_64_after_hwframe > 9.04 ± 13% -6.0 3.06 ± 79% perf-profile.calltrace.cycles-pp.xfs_file_read_iter.aio_read.io_submit_one.__x64_sys_io_submit.do_syscall_64 > 8.92 ± 12% -5.9 3.02 ± 78% perf-profile.calltrace.cycles-pp.xfs_file_dio_read.xfs_file_read_iter.aio_read.io_submit_one.__x64_sys_io_submit > 8.74 ± 12% -5.8 2.90 ± 78% perf-profile.calltrace.cycles-pp.iomap_dio_rw.xfs_file_dio_read.xfs_file_read_iter.aio_read.io_submit_one > 8.73 ± 12% -5.8 2.90 ± 78% perf-profile.calltrace.cycles-pp.__iomap_dio_rw.iomap_dio_rw.xfs_file_dio_read.xfs_file_read_iter.aio_read > 7.43 ± 10% -5.7 1.70 ±106% perf-profile.calltrace.cycles-pp.scsi_io_completion.blk_done_softirq.handle_softirqs.__irq_exit_rcu.common_interrupt > 7.52 ± 10% -5.7 1.80 ±107% perf-profile.calltrace.cycles-pp.scsi_end_request.scsi_io_completion.blk_done_softirq.handle_softirqs.__irq_exit_rcu > 7.58 ± 8% -5.2 2.39 ± 79% perf-profile.calltrace.cycles-pp.iomap_dio_bio_iter.__iomap_dio_rw.iomap_dio_rw.xfs_file_dio_read.xfs_file_read_iter > 7.56 ± 13% -4.9 2.62 ± 70% perf-profile.calltrace.cycles-pp.worker_thread.kthread.ret_from_fork.ret_from_fork_asm > 5.26 ± 10% -3.9 1.41 ± 86% perf-profile.calltrace.cycles-pp.submit_bio_noacct_nocheck.iomap_dio_bio_iter.__iomap_dio_rw.iomap_dio_rw.xfs_file_dio_read > 5.18 ± 10% -3.8 1.38 ± 87% perf-profile.calltrace.cycles-pp.__submit_bio.submit_bio_noacct_nocheck.iomap_dio_bio_iter.__iomap_dio_rw.iomap_dio_rw > 5.13 ± 10% -3.8 1.37 ± 86% perf-profile.calltrace.cycles-pp.blk_mq_submit_bio.__submit_bio.submit_bio_noacct_nocheck.iomap_dio_bio_iter.__iomap_dio_rw > 3.73 ± 6% -2.9 0.78 ±108% perf-profile.calltrace.cycles-pp.__common_interrupt.common_interrupt.asm_common_interrupt.cpuidle_enter_state.cpuidle_enter > 3.93 ± 13% -2.9 1.00 ± 74% perf-profile.calltrace.cycles-pp.__blk_mq_alloc_requests.blk_mq_submit_bio.__submit_bio.submit_bio_noacct_nocheck.iomap_dio_bio_iter > 3.65 ± 7% -2.9 0.77 ±108% perf-profile.calltrace.cycles-pp.handle_edge_irq.__common_interrupt.common_interrupt.asm_common_interrupt.cpuidle_enter_state > 3.81 ± 14% -2.8 0.99 ± 73% perf-profile.calltrace.cycles-pp.blk_mq_get_tag.__blk_mq_alloc_requests.blk_mq_submit_bio.__submit_bio.submit_bio_noacct_nocheck > 4.21 ± 22% -2.7 1.48 ± 71% perf-profile.calltrace.cycles-pp.blk_mq_sched_dispatch_requests.blk_mq_run_work_fn.process_one_work.worker_thread.kthread > 4.24 ± 22% -2.7 1.53 ± 71% perf-profile.calltrace.cycles-pp.blk_mq_run_work_fn.process_one_work.worker_thread.kthread.ret_from_fork > 4.56 ± 20% -2.7 1.91 ± 68% perf-profile.calltrace.cycles-pp.process_one_work.worker_thread.kthread.ret_from_fork.ret_from_fork_asm > 4.04 ± 22% -2.6 1.48 ± 71% perf-profile.calltrace.cycles-pp.__blk_mq_sched_dispatch_requests.blk_mq_sched_dispatch_requests.blk_mq_run_work_fn.process_one_work.worker_thread > 3.14 ± 9% -2.5 0.66 ±109% perf-profile.calltrace.cycles-pp.handle_irq_event.handle_edge_irq.__common_interrupt.common_interrupt.asm_common_interrupt > 2.91 ± 8% -2.3 0.59 ±109% perf-profile.calltrace.cycles-pp.__handle_irq_event_percpu.handle_irq_event.handle_edge_irq.__common_interrupt.common_interrupt > 2.72 ± 9% -2.2 0.51 ±107% perf-profile.calltrace.cycles-pp.schedule.worker_thread.kthread.ret_from_fork.ret_from_fork_asm > 2.64 ± 9% -2.1 0.50 ±108% perf-profile.calltrace.cycles-pp.__schedule.schedule.worker_thread.kthread.ret_from_fork > 2.76 ± 27% -1.1 1.71 ± 26% perf-profile.calltrace.cycles-pp.schedule_idle.do_idle.cpu_startup_entry.start_secondary.common_startup_64 > 2.64 ± 28% -1.0 1.64 ± 27% perf-profile.calltrace.cycles-pp.__schedule.schedule_idle.do_idle.cpu_startup_entry.start_secondary > 1.28 ± 20% -0.8 0.48 ± 85% perf-profile.calltrace.cycles-pp.tick_nohz_idle_exit.do_idle.cpu_startup_entry.start_secondary.common_startup_64 > 0.98 ± 36% -0.7 0.31 ±101% perf-profile.calltrace.cycles-pp.blk_mq_run_hw_queue.blk_mq_get_tag.__blk_mq_alloc_requests.blk_mq_submit_bio.__submit_bio > 0.97 ± 22% -0.7 0.32 ±103% perf-profile.calltrace.cycles-pp.tick_nohz_restart_sched_tick.tick_nohz_idle_exit.do_idle.cpu_startup_entry.start_secondary > 1.29 ± 11% -0.3 0.95 ± 20% perf-profile.calltrace.cycles-pp.flush_smp_call_function_queue.do_idle.cpu_startup_entry.start_secondary.common_startup_64 > 0.28 ±100% +0.7 0.97 ± 35% perf-profile.calltrace.cycles-pp.schedule_timeout.rcu_gp_fqs_loop.rcu_gp_kthread.kthread.ret_from_fork > 0.26 ±173% +1.9 2.18 ± 45% perf-profile.calltrace.cycles-pp.asm_sysvec_irq_work.handle_softirqs.__irq_exit_rcu.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt > 0.26 ±173% +1.9 2.18 ± 45% perf-profile.calltrace.cycles-pp.sysvec_irq_work.asm_sysvec_irq_work.handle_softirqs.__irq_exit_rcu.sysvec_apic_timer_interrupt > 0.07 ±264% +2.0 2.02 ± 69% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe._Fork > 0.07 ±264% +2.0 2.02 ± 69% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe._Fork > 0.07 ±264% +2.0 2.11 ± 66% perf-profile.calltrace.cycles-pp._Fork > 1.23 ± 41% +6.8 7.99 ± 59% perf-profile.calltrace.cycles-pp.delay_tsc.wait_for_lsr.serial8250_console_write.console_flush_all.console_unlock > 3.68 ± 13% +16.8 20.52 ± 57% perf-profile.calltrace.cycles-pp.io_serial_in.wait_for_lsr.serial8250_console_write.console_flush_all.console_unlock > 6.33 ± 5% +23.0 29.34 ± 55% perf-profile.calltrace.cycles-pp.wait_for_lsr.serial8250_console_write.console_flush_all.console_unlock.vprintk_emit > 6.47 ± 4% +23.4 29.91 ± 55% perf-profile.calltrace.cycles-pp.serial8250_console_write.console_flush_all.console_unlock.vprintk_emit._printk > 14.58 ± 13% -11.2 3.34 ± 90% perf-profile.children.cycles-pp.asm_common_interrupt > 14.45 ± 13% -11.1 3.33 ± 90% perf-profile.children.cycles-pp.common_interrupt > 11.46 ± 9% -7.5 3.96 ± 81% perf-profile.children.cycles-pp.syscall > 8.70 ± 11% -6.5 2.18 ± 94% perf-profile.children.cycles-pp.blk_done_softirq > 9.51 ± 12% -6.2 3.27 ± 79% perf-profile.children.cycles-pp.__x64_sys_io_submit > 9.40 ± 12% -6.2 3.22 ± 79% perf-profile.children.cycles-pp.io_submit_one > 9.14 ± 12% -6.0 3.10 ± 79% perf-profile.children.cycles-pp.aio_read > 9.04 ± 13% -6.0 3.06 ± 79% perf-profile.children.cycles-pp.xfs_file_read_iter > 8.92 ± 12% -5.9 3.02 ± 79% perf-profile.children.cycles-pp.xfs_file_dio_read > 8.74 ± 12% -5.8 2.90 ± 78% perf-profile.children.cycles-pp.iomap_dio_rw > 8.73 ± 12% -5.8 2.90 ± 78% perf-profile.children.cycles-pp.__iomap_dio_rw > 7.67 ± 9% -5.7 1.95 ± 95% perf-profile.children.cycles-pp.scsi_io_completion > 7.66 ± 9% -5.7 1.95 ± 95% perf-profile.children.cycles-pp.scsi_end_request > 7.59 ± 8% -5.1 2.45 ± 74% perf-profile.children.cycles-pp.iomap_dio_bio_iter > 7.57 ± 13% -4.9 2.63 ± 70% perf-profile.children.cycles-pp.worker_thread > 8.51 ± 11% -3.8 4.66 ± 30% perf-profile.children.cycles-pp.__schedule > 5.28 ± 10% -3.8 1.52 ± 72% perf-profile.children.cycles-pp.submit_bio_noacct_nocheck > 4.57 ± 21% -3.7 0.85 ± 46% perf-profile.children.cycles-pp.__blk_mq_do_dispatch_sched > 5.18 ± 10% -3.7 1.49 ± 73% perf-profile.children.cycles-pp.__submit_bio > 5.13 ± 10% -3.7 1.48 ± 72% perf-profile.children.cycles-pp.blk_mq_submit_bio > 3.45 ± 15% -3.4 0.08 ±113% perf-profile.children.cycles-pp.blk_mq_run_hw_queues > 3.87 ± 6% -2.9 0.93 ± 84% perf-profile.children.cycles-pp.__common_interrupt > 3.78 ± 7% -2.9 0.92 ± 83% perf-profile.children.cycles-pp.handle_edge_irq > 3.93 ± 13% -2.8 1.11 ± 54% perf-profile.children.cycles-pp.__blk_mq_alloc_requests > 4.82 ± 21% -2.8 2.06 ± 52% perf-profile.children.cycles-pp.blk_mq_sched_dispatch_requests > 5.42 ± 7% -2.7 2.69 ± 36% perf-profile.children.cycles-pp.schedule > 3.82 ± 13% -2.7 1.10 ± 54% perf-profile.children.cycles-pp.blk_mq_get_tag > 4.57 ± 20% -2.7 1.91 ± 68% perf-profile.children.cycles-pp.process_one_work > 4.24 ± 22% -2.6 1.59 ± 64% perf-profile.children.cycles-pp.blk_mq_run_work_fn > 4.64 ± 20% -2.6 2.06 ± 52% perf-profile.children.cycles-pp.__blk_mq_sched_dispatch_requests > 3.32 ± 18% -2.5 0.78 ± 90% perf-profile.children.cycles-pp.mod_delayed_work_on > 3.33 ± 18% -2.5 0.80 ± 90% perf-profile.children.cycles-pp.kblockd_mod_delayed_work_on > 3.24 ± 9% -2.5 0.78 ± 85% perf-profile.children.cycles-pp.handle_irq_event > 2.99 ± 8% -2.3 0.70 ± 85% perf-profile.children.cycles-pp.__handle_irq_event_percpu > 2.89 ± 7% -2.2 0.66 ± 84% perf-profile.children.cycles-pp.ahci_single_level_irq_intr > 2.13 ± 12% -1.8 0.38 ± 95% perf-profile.children.cycles-pp.io_schedule > 2.16 ± 4% -1.6 0.55 ± 82% perf-profile.children.cycles-pp.ahci_handle_port_intr > 2.71 ± 13% -1.6 1.13 ± 89% perf-profile.children.cycles-pp.__queue_work > 2.76 ± 10% -1.5 1.30 ± 25% perf-profile.children.cycles-pp.__pick_next_task > 2.41 ± 11% -1.4 1.06 ± 26% perf-profile.children.cycles-pp.pick_next_task_fair > 1.90 ± 16% -1.3 0.58 ± 94% perf-profile.children.cycles-pp.blk_update_request > 2.25 ± 15% -1.3 0.93 ± 92% perf-profile.children.cycles-pp.kick_pool > 3.62 ± 10% -1.3 2.32 ± 35% perf-profile.children.cycles-pp.enqueue_task_fair > 1.71 ± 8% -1.3 0.43 ± 87% perf-profile.children.cycles-pp.ahci_qc_complete > 3.28 ± 23% -1.2 2.07 ± 27% perf-profile.children.cycles-pp.schedule_idle > 2.40 ± 28% -1.2 1.22 ± 28% perf-profile.children.cycles-pp.scsi_queue_rq > 2.44 ± 11% -1.1 1.33 ± 41% perf-profile.children.cycles-pp.dequeue_entities > 2.54 ± 12% -1.1 1.45 ± 39% perf-profile.children.cycles-pp.try_to_block_task > 2.13 ± 9% -1.0 1.08 ± 40% perf-profile.children.cycles-pp.hrtimer_start_range_ns > 1.19 ± 40% -1.0 0.21 ±123% perf-profile.children.cycles-pp.__blk_flush_plug > 1.58 ± 8% -1.0 0.60 ± 83% perf-profile.children.cycles-pp.bio_iov_iter_get_pages > 1.56 ± 9% -1.0 0.60 ± 84% perf-profile.children.cycles-pp.__bio_iov_iter_get_pages > 1.26 ± 14% -1.0 0.31 ± 44% perf-profile.children.cycles-pp.sched_balance_newidle > 1.15 ± 40% -1.0 0.20 ±122% perf-profile.children.cycles-pp.blk_mq_flush_plug_list > 1.27 ± 11% -0.9 0.35 ± 83% perf-profile.children.cycles-pp.ata_qc_complete_multiple > 1.32 ± 20% -0.9 0.42 ± 71% perf-profile.children.cycles-pp.__wake_up_common > 1.06 ± 42% -0.9 0.20 ±122% perf-profile.children.cycles-pp.blk_mq_dispatch_list > 1.18 ± 21% -0.9 0.33 ± 77% perf-profile.children.cycles-pp.autoremove_wake_function > 1.04 ± 16% -0.8 0.21 ± 78% perf-profile.children.cycles-pp.sbitmap_get > 1.05 ± 16% -0.8 0.23 ± 93% perf-profile.children.cycles-pp.scsi_mq_get_budget > 1.66 ± 18% -0.8 0.88 ± 43% perf-profile.children.cycles-pp.tick_nohz_idle_exit > 1.01 ± 52% -0.7 0.26 ±104% perf-profile.children.cycles-pp.sd_setup_read_write_cmnd > 1.48 ± 16% -0.7 0.74 ± 56% perf-profile.children.cycles-pp.tick_nohz_stop_tick > 1.50 ± 16% -0.7 0.75 ± 55% perf-profile.children.cycles-pp.tick_nohz_idle_stop_tick > 1.29 ± 9% -0.7 0.59 ± 29% perf-profile.children.cycles-pp.sched_balance_update_blocked_averages > 1.55 ± 11% -0.7 0.86 ± 29% perf-profile.children.cycles-pp.enqueue_dl_entity > 1.56 ± 11% -0.7 0.88 ± 28% perf-profile.children.cycles-pp.dl_server_start > 1.25 ± 26% -0.7 0.59 ± 36% perf-profile.children.cycles-pp.blk_mq_run_hw_queue > 0.77 ± 21% -0.7 0.12 ± 80% perf-profile.children.cycles-pp.sbitmap_find_bit > 1.07 ± 9% -0.6 0.46 ± 37% perf-profile.children.cycles-pp.__update_blocked_fair > 1.38 ± 17% -0.6 0.79 ± 28% perf-profile.children.cycles-pp.timerqueue_add > 0.99 ± 17% -0.6 0.43 ± 29% perf-profile.children.cycles-pp.start_dl_timer > 1.25 ± 18% -0.6 0.69 ± 43% perf-profile.children.cycles-pp.tick_nohz_restart_sched_tick > 0.73 ± 60% -0.5 0.22 ±100% perf-profile.children.cycles-pp.scsi_alloc_sgtables > 0.58 ± 52% -0.5 0.08 ±101% perf-profile.children.cycles-pp.dd_dispatch_request > 0.56 ± 55% -0.5 0.08 ± 96% perf-profile.children.cycles-pp.__dd_dispatch_request > 0.75 ± 20% -0.5 0.29 ± 96% perf-profile.children.cycles-pp.iov_iter_extract_pages > 0.71 ± 19% -0.4 0.27 ± 93% perf-profile.children.cycles-pp.pin_user_pages_fast > 0.66 ± 18% -0.4 0.24 ± 97% perf-profile.children.cycles-pp.gup_fast_fallback > 0.63 ± 18% -0.4 0.23 ± 98% perf-profile.children.cycles-pp.gup_fast > 0.52 ± 77% -0.4 0.14 ±116% perf-profile.children.cycles-pp.dd_insert_requests > 0.50 ± 16% -0.4 0.13 ± 96% perf-profile.children.cycles-pp.__x64_sys_io_getevents > 0.58 ± 20% -0.4 0.22 ± 95% perf-profile.children.cycles-pp.gup_fast_pgd_range > 0.46 ± 41% -0.4 0.09 ± 91% perf-profile.children.cycles-pp.scsi_finish_command > 0.41 ± 17% -0.4 0.06 ±121% perf-profile.children.cycles-pp.bio_endio > 0.46 ± 23% -0.3 0.11 ±116% perf-profile.children.cycles-pp.__ata_qc_complete > 0.80 ± 16% -0.3 0.47 ± 36% perf-profile.children.cycles-pp.update_curr > 0.51 ± 18% -0.3 0.18 ± 52% perf-profile.children.cycles-pp.__switch_to_asm > 0.45 ± 19% -0.3 0.12 ±104% perf-profile.children.cycles-pp.do_io_getevents > 0.46 ± 71% -0.3 0.14 ± 95% perf-profile.children.cycles-pp.iomap_iter > 0.40 ± 20% -0.3 0.10 ± 93% perf-profile.children.cycles-pp.blk_mq_complete_request > 0.38 ± 25% -0.3 0.10 ±100% perf-profile.children.cycles-pp.blk_mq_complete_request_remote > 0.59 ± 14% -0.3 0.33 ± 45% perf-profile.children.cycles-pp.__get_next_timer_interrupt > 0.35 ± 27% -0.3 0.10 ±104% perf-profile.children.cycles-pp.gup_fast_pmd_leaf > 0.33 ± 22% -0.2 0.08 ±126% perf-profile.children.cycles-pp.dma_direct_unmap_sg > 0.36 ± 20% -0.2 0.11 ±117% perf-profile.children.cycles-pp.bio_alloc_bioset > 0.35 ± 25% -0.2 0.10 ±118% perf-profile.children.cycles-pp.read_events > 0.33 ± 31% -0.2 0.09 ±113% perf-profile.children.cycles-pp.aio_read_events_ring > 0.34 ± 28% -0.2 0.10 ±118% perf-profile.children.cycles-pp.aio_read_events > 0.28 ± 28% -0.2 0.06 ±110% perf-profile.children.cycles-pp.__sg_free_table > 0.46 ± 23% -0.2 0.24 ± 26% perf-profile.children.cycles-pp.set_next_entity > 0.27 ± 30% -0.2 0.05 ±119% perf-profile.children.cycles-pp.__blk_mq_end_request > 0.34 ± 17% -0.2 0.13 ± 75% perf-profile.children.cycles-pp.__update_load_avg_cfs_rq > 0.42 ± 14% -0.2 0.21 ± 35% perf-profile.children.cycles-pp.switch_hrtimer_base > 0.27 ± 37% -0.2 0.07 ±100% perf-profile.children.cycles-pp.__sg_alloc_table > 0.47 ± 21% -0.2 0.29 ± 23% perf-profile.children.cycles-pp.set_next_task_fair > 0.29 ± 33% -0.2 0.11 ± 85% perf-profile.children.cycles-pp.try_grab_folio_fast > 0.41 ± 15% -0.2 0.23 ± 27% perf-profile.children.cycles-pp.switch_mm_irqs_off > 0.34 ± 14% -0.2 0.18 ± 36% perf-profile.children.cycles-pp.get_nohz_timer_target > 0.24 ± 89% -0.2 0.08 ± 61% perf-profile.children.cycles-pp.xas_find > 0.20 ± 42% -0.2 0.04 ±107% perf-profile.children.cycles-pp._find_next_zero_bit > 0.18 ± 40% -0.1 0.05 ±117% perf-profile.children.cycles-pp.aio_complete_rw > 0.18 ± 33% -0.1 0.06 ±101% perf-profile.children.cycles-pp.refresh_cpu_vm_stats > 0.24 ± 19% -0.1 0.15 ± 42% perf-profile.children.cycles-pp.perf_event_task_tick > 0.12 ± 30% -0.1 0.04 ±111% perf-profile.children.cycles-pp.__dentry_kill > 0.12 ± 63% +0.4 0.51 ± 86% perf-profile.children.cycles-pp.vms_clear_ptes > 0.53 ± 13% +0.5 0.99 ± 34% perf-profile.children.cycles-pp.schedule_timeout > 0.21 ± 43% +0.9 1.11 ± 50% perf-profile.children.cycles-pp.timer_expire_remote > 0.23 ± 33% +0.9 1.16 ± 49% perf-profile.children.cycles-pp.tmigr_handle_remote_cpu > 0.28 ± 32% +1.0 1.26 ± 49% perf-profile.children.cycles-pp.tmigr_handle_remote_up > 0.30 ± 28% +1.0 1.29 ± 48% perf-profile.children.cycles-pp.tmigr_handle_remote > 0.39 ± 24% +1.1 1.53 ± 50% perf-profile.children.cycles-pp.call_timer_fn > 0.48 ± 21% +1.2 1.64 ± 50% perf-profile.children.cycles-pp.__run_timers > 0.37 ± 25% +1.6 2.01 ± 70% perf-profile.children.cycles-pp.__do_sys_clone > 0.42 ± 20% +1.7 2.11 ± 66% perf-profile.children.cycles-pp._Fork > 2.31 ± 14% +6.9 9.21 ± 54% perf-profile.children.cycles-pp.delay_tsc > 4.38 ± 6% +17.3 21.68 ± 57% perf-profile.children.cycles-pp.io_serial_in > 6.71 ± 4% +24.2 30.93 ± 56% perf-profile.children.cycles-pp.wait_for_lsr > 6.81 ± 4% +24.6 31.41 ± 56% perf-profile.children.cycles-pp._printk > 6.81 ± 4% +24.6 31.41 ± 56% perf-profile.children.cycles-pp.console_flush_all > 6.81 ± 4% +24.6 31.41 ± 56% perf-profile.children.cycles-pp.console_unlock > 6.81 ± 4% +24.6 31.41 ± 56% perf-profile.children.cycles-pp.serial8250_console_write > 6.81 ± 4% +24.6 31.41 ± 56% perf-profile.children.cycles-pp.vprintk_emit > 0.73 ± 23% -0.6 0.11 ±112% perf-profile.self.cycles-pp.ahci_single_level_irq_intr > 1.28 ± 15% -0.6 0.72 ± 30% perf-profile.self.cycles-pp.timerqueue_add > 0.80 ± 15% -0.4 0.36 ± 43% perf-profile.self.cycles-pp.__update_blocked_fair > 0.76 ± 18% -0.3 0.42 ± 35% perf-profile.self.cycles-pp.__schedule > 0.50 ± 20% -0.3 0.18 ± 52% perf-profile.self.cycles-pp.__switch_to_asm > 0.45 ± 28% -0.3 0.12 ± 78% perf-profile.self.cycles-pp.ahci_handle_port_intr > 0.37 ± 28% -0.3 0.05 ± 88% perf-profile.self.cycles-pp.__dd_dispatch_request > 0.88 ± 14% -0.3 0.57 ± 27% perf-profile.self.cycles-pp.menu_select > 0.33 ± 28% -0.3 0.05 ±102% perf-profile.self.cycles-pp.sbitmap_find_bit > 0.31 ± 28% -0.3 0.05 ±132% perf-profile.self.cycles-pp.blk_mq_complete_request_remote > 0.33 ± 22% -0.2 0.08 ±126% perf-profile.self.cycles-pp.dma_direct_unmap_sg > 0.27 ± 22% -0.2 0.04 ±108% perf-profile.self.cycles-pp.blk_mq_submit_bio > 0.50 ± 21% -0.2 0.27 ± 34% perf-profile.self.cycles-pp.update_curr > 0.34 ± 18% -0.2 0.12 ± 76% perf-profile.self.cycles-pp.__update_load_avg_cfs_rq > 0.28 ± 23% -0.2 0.10 ±105% perf-profile.self.cycles-pp.dma_direct_map_sg > 0.26 ± 61% -0.2 0.08 ±100% perf-profile.self.cycles-pp.sbitmap_get > 0.28 ± 27% -0.1 0.14 ± 35% perf-profile.self.cycles-pp.dequeue_entity > 0.22 ± 23% -0.1 0.11 ± 33% perf-profile.self.cycles-pp.finish_task_switch > 2.31 ± 14% +6.9 9.21 ± 54% perf-profile.self.cycles-pp.delay_tsc > 4.38 ± 6% +17.3 21.68 ± 57% perf-profile.self.cycles-pp.io_serial_in > > > > > 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://github.com/intel/lkp-tests/wiki ---end quoted text---
Powered by blists - more mailing lists