Hi Yonghong, > -----Original Message----- > From: Yonghong Song <yhs@xxxxxx> > Sent: Friday, August 20, 2021 2:29 PM > To: Sang, Oliver <oliver.sang@xxxxxxxxx>; Andrii Nakryiko <andrii@xxxxxxxxxx> > Cc: Daniel Borkmann <daniel@xxxxxxxxxxxxx>; Peter Zijlstra > <peterz@xxxxxxxxxxxxx>; LKML <linux-kernel@xxxxxxxxxxxxxxx>; Linux Memory > Management List <linux-mm@xxxxxxxxx>; lkp@xxxxxxxxxxxx; lkp <lkp@xxxxxxxxx> > Subject: [LKP] Re: [bpf] b89fbfbb85: netperf.Throughput_tps -21.3% regression > > > > On 8/19/21 10:50 PM, kernel test robot wrote: > > > > > > Greeting, > > > > FYI, we noticed a -21.3% regression of netperf.Throughput_tps due to commit: > > > > > > commit: b89fbfbb854c9afc3047e8273cc3a694650b802e ("bpf: Implement > > minimal BPF perf link") > > https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git > > master > > Could you do a test again to double check? > The above patch is related to bpf perf_event area. > It should not have any impact with netperf. We first tested again and confirmed the regression still exists (-21.4%), the result is stable [1], more details are in [2]. 652c1b17b85b9c19 b89fbfbb854c9afc3047e8273cc ---------------- --------------------------- %stddev %change %stddev \ | \ 382134 -21.4% 300300 netperf.Throughput_total_tps 23958 -21.4% 18828 netperf.Throughput_tps Then we tried another platform, Ice Lake, it has 128 threads 2 sockets Intel(R) Xeon(R) Gold 6338 CPU @ 2.00GHz with 256G, but we found the regression is still existing (-5.4%), the result is stable [3], more details are in [4]. 652c1b17b85b9c19 b89fbfbb854c9afc3047e8273cc ---------------- --------------------------- %stddev %change %stddev \ | \ 522472 -5.3% 494763 netperf.Throughput_total_tps 32654 -5.3% 30922 netperf.Throughput_tps [1] b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json: "netperf.Throughput_tps": [ b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18860.396249999998, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18808.879375, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18905.214375000003, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18750.746874999997, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18803.823125000006, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 19201.846, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18819.47, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18979.38, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18837.719375000004, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18809.4725, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18776.799375, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18773.092500000002, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18743.978125, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18794.814374999998, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18810.783124999998, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18821.65625, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18680.149375, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18747.203125, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18819.728750000002, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18830.020625, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 18656.501874999998 b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- ], b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json: "netperf.Throughput_total_tps": [ b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 301766.33999999997, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 300942.07, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 302483.43000000005, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 300011.94999999995, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 300861.1700000001, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 288027.69, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 301111.52, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 303670.08, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 301403.51000000007, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 300951.56, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 300428.79, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 300369.48000000004, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 299903.65, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 300717.02999999997, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 300972.52999999997, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 301146.5, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 298882.39, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 299955.25, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 301115.66000000003, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 301280.33, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 298504.02999999997 b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- ], [2] ========================================================================================= cluster/compiler/cpufreq_governor/ip/kconfig/nr_threads/rootfs/runtime/tbox_group/test/testcase/ucode: cs-localhost/gcc-9/performance/ipv4/x86_64-rhel-8.3/16/debian-10.4-x86_64-20200603.cgz/300s/lkp-csl-2ap3/TCP_CRR/netperf/0x5003006 commit: 652c1b17b8 ("bpf: Refactor perf_event_set_bpf_prog() to use struct bpf_prog input") b89fbfbb85 ("bpf: Implement minimal BPF perf link") 652c1b17b85b9c19 b89fbfbb854c9afc3047e8273cc ---------------- --------------------------- %stddev %change %stddev \ | \ 381418 -21.3% 300214 netperf.Throughput_total_tps 23906 -21.3% 18820 netperf.Throughput_tps 872.95 -5.2% 827.95 netperf.time.percent_of_cpu_this_job_got 2586 -4.3% 2474 netperf.time.system_time 2.25e+08 -22.8% 1.736e+08 netperf.time.voluntary_context_switches 1.144e+08 -21.3% 90064356 netperf.workload 2346775 -8.7% 2143411 vmstat.system.cs 872.95 -5.2% 827.95 time.percent_of_cpu_this_job_got 2586 -4.3% 2474 time.system_time 2.25e+08 -22.8% 1.736e+08 time.voluntary_context_switches 0.01 ± 7% +68.4% 0.01 ± 7% perf-sched.wait_and_delay.avg.ms.schedule_timeout.wait_woken.sk_wait_data.tcp_recvmsg_locked 3491773 ± 2% -22.6% 2702148 perf-sched.wait_and_delay.count.schedule_timeout.inet_csk_accept.inet_accept.__sys_accept4_file 4512682 ± 4% +21.9% 5501355 ± 2% perf-sched.wait_and_delay.count.schedule_timeout.wait_woken.sk_wait_data.tcp_recvmsg_locked 0.01 ± 8% +86.5% 0.01 ± 8% perf-sched.wait_time.avg.ms.schedule_timeout.wait_woken.sk_wait_data.tcp_recvmsg_locked 0.03 ± 8% +171.3% 0.08 ±231% perf-sched.wait_time.max.ms.preempt_schedule_common.__cond_resched.apparmor_file_alloc_security.security_file_alloc.__alloc_file 113308 -2.8% 110156 proc-vmstat.nr_slab_unreclaimable 4057280 ± 2% -8.6% 3709088 ± 2% proc-vmstat.numa_hit 3847678 ± 2% -8.9% 3506977 ± 2% proc-vmstat.numa_local 4055392 ± 2% -8.5% 3709694 ± 2% proc-vmstat.pgalloc_normal 1770294 ± 2% -23.5% 1355122 ± 2% proc-vmstat.pgfree 2456040 ± 16% -34.0% 1620559 ± 20% softirqs.CPU121.NET_RX 1737990 ± 24% -36.1% 1109923 ± 27% softirqs.CPU163.NET_RX 3.519e+08 -22.1% 2.742e+08 softirqs.NET_RX 26128869 ± 2% -15.7% 22015818 softirqs.RCU 237092 ± 2% +84.0% 436186 softirqs.TIMER 170292 -12.0% 149884 slabinfo.filp.active_objs 2663 -12.0% 2343 slabinfo.filp.active_slabs 170459 -12.0% 150036 slabinfo.filp.num_objs 2663 -12.0% 2343 slabinfo.filp.num_slabs 49886 ± 5% -11.6% 44079 ± 4% slabinfo.kmalloc-512.active_objs 50080 ± 5% -11.6% 44269 ± 4% slabinfo.kmalloc-512.num_objs 12.06 ± 4% +26.2% 15.22 ± 2% perf-stat.i.MPKI 1.004e+10 -18.0% 8.233e+09 perf-stat.i.branch-instructions 1.562e+08 -19.6% 1.256e+08 perf-stat.i.branch-misses 10.92 +7.2 18.09 perf-stat.i.cache-miss-rate% 64015758 ± 2% +75.1% 1.121e+08 perf-stat.i.cache-misses 5.888e+08 +5.2% 6.196e+08 perf-stat.i.cache-references 2379887 -8.7% 2173629 perf-stat.i.context-switches 1.57 ± 2% +17.2% 1.84 perf-stat.i.cpi 7.739e+10 -3.5% 7.467e+10 perf-stat.i.cpu-cycles 1272 ± 2% -43.8% 715.53 ± 3% perf-stat.i.cycles-between-cache-misses 903463 ± 18% -27.4% 655855 ± 5% perf-stat.i.dTLB-load-misses 1.518e+10 -18.3% 1.24e+10 perf-stat.i.dTLB-loads 228225 ± 31% -52.8% 107778 ± 23% perf-stat.i.dTLB-store-misses 8.879e+09 -18.5% 7.234e+09 perf-stat.i.dTLB-stores 89.92 -4.7 85.18 perf-stat.i.iTLB-load-miss-rate% 1.075e+08 -21.1% 84797404 perf-stat.i.iTLB-load-misses 11549994 ± 4% +23.9% 14312532 ± 2% perf-stat.i.iTLB-loads 5.079e+10 -18.2% 4.155e+10 perf-stat.i.instructions 0.65 -15.2% 0.55 perf-stat.i.ipc 0.40 -3.5% 0.39 perf-stat.i.metric.GHz 180.81 -17.8% 148.59 perf-stat.i.metric.M/sec 13211450 +59.2% 21027780 perf-stat.i.node-load-misses 1191177 ± 5% +68.5% 2006794 ± 9% perf-stat.i.node-loads 10841458 -22.2% 8430280 perf-stat.i.node-store-misses 11.59 +28.7% 14.91 perf-stat.overall.MPKI 1.56 -0.0 1.53 perf-stat.overall.branch-miss-rate% 10.87 +7.2 18.09 perf-stat.overall.cache-miss-rate% 1.52 +18.0% 1.80 perf-stat.overall.cpi 1209 ± 2% -44.9% 666.15 perf-stat.overall.cycles-between-cache-misses 90.30 -4.7 85.56 perf-stat.overall.iTLB-load-miss-rate% 472.39 +3.8% 490.15 perf-stat.overall.instructions-per-iTLB-miss 0.66 -15.2% 0.56 perf-stat.overall.ipc 1e+10 -18.0% 8.205e+09 perf-stat.ps.branch-instructions 1.557e+08 -19.6% 1.251e+08 perf-stat.ps.branch-misses 63792201 ± 2% +75.1% 1.117e+08 perf-stat.ps.cache-misses 5.867e+08 +5.2% 6.174e+08 perf-stat.ps.cache-references 2371522 -8.7% 2165977 perf-stat.ps.context-switches 7.712e+10 -3.5% 7.441e+10 perf-stat.ps.cpu-cycles 900712 ± 18% -27.4% 653896 ± 5% perf-stat.ps.dTLB-load-misses 1.513e+10 -18.3% 1.236e+10 perf-stat.ps.dTLB-loads 227485 ± 31% -52.8% 107457 ± 23% perf-stat.ps.dTLB-store-misses 8.848e+09 -18.5% 7.208e+09 perf-stat.ps.dTLB-stores 1.072e+08 -21.1% 84499344 perf-stat.ps.iTLB-load-misses 11509852 ± 4% +23.9% 14262519 ± 2% perf-stat.ps.iTLB-loads 5.062e+10 -18.2% 4.14e+10 perf-stat.ps.instructions 13165220 +59.2% 20953789 perf-stat.ps.node-load-misses 1187095 ± 5% +68.5% 1999816 ± 9% perf-stat.ps.node-loads 10803432 -22.2% 8400688 perf-stat.ps.node-store-misses 1.543e+13 -18.2% 1.262e+13 perf-stat.total.instructions 5.24 ± 7% -1.2 4.05 ± 7% perf-profile.calltrace.cycles-pp.inet_recvmsg.__sys_recvfrom.__x64_sys_recvfrom.do_syscall_64.entry_SYSCALL_64_after_hwframe 5.22 ± 7% -1.2 4.03 ± 7% perf-profile.calltrace.cycles-pp.tcp_recvmsg.inet_recvmsg.__sys_recvfrom.__x64_sys_recvfrom.do_syscall_64 4.13 ± 21% -1.0 3.17 ± 6% perf-profile.calltrace.cycles-pp.poll_idle.cpuidle_enter_state.cpuidle_enter.do_idle.cpu_startup_entry 2.75 ± 6% -0.6 2.15 ± 8% perf-profile.calltrace.cycles-pp.__sys_accept4_file.__sys_accept4.__x64_sys_accept.do_syscall_64.entry_SYSCALL_64_after_hwframe 1.71 ± 9% -0.5 1.25 ± 9% perf-profile.calltrace.cycles-pp.ret_from_fork 1.71 ± 9% -0.5 1.25 ± 9% perf-profile.calltrace.cycles-pp.kthread.ret_from_fork 1.67 ± 9% -0.5 1.22 ± 9% perf-profile.calltrace.cycles-pp.__softirqentry_text_start.run_ksoftirqd.smpboot_thread_fn.kthread.ret_from_fork 1.66 ± 9% -0.5 1.21 ± 9% perf-profile.calltrace.cycles-pp.rcu_core.__softirqentry_text_start.run_ksoftirqd.smpboot_thread_fn.kthread 1.67 ± 9% -0.5 1.22 ± 9% perf-profile.calltrace.cycles-pp.run_ksoftirqd.smpboot_thread_fn.kthread.ret_from_fork 1.68 ± 9% -0.5 1.23 ± 9% perf-profile.calltrace.cycles-pp.smpboot_thread_fn.kthread.ret_from_fork 1.65 ± 9% -0.5 1.20 ± 9% perf-profile.calltrace.cycles-pp.rcu_do_batch.rcu_core.__softirqentry_text_start.run_ksoftirqd.smpboot_thread_fn 1.58 ± 9% -0.4 1.14 ± 9% perf-profile.calltrace.cycles-pp.kmem_cache_free.rcu_do_batch.rcu_core.__softirqentry_text_start.run_ksoftirqd 1.40 ± 9% -0.4 1.02 ± 9% perf-profile.calltrace.cycles-pp.obj_cgroup_uncharge_pages.kmem_cache_free.rcu_do_batch.rcu_core.__softirqentry_text_start 1.36 ± 9% -0.4 0.99 ± 9% perf-profile.calltrace.cycles-pp.page_counter_uncharge.obj_cgroup_uncharge_pages.kmem_cache_free.rcu_do_batch.rcu_core 1.39 ± 8% -0.3 1.09 ± 9% perf-profile.calltrace.cycles-pp.inet_accept.__sys_accept4_file.__sys_accept4.__x64_sys_accept.do_syscall_64 1.07 ± 10% -0.3 0.78 ± 9% perf-profile.calltrace.cycles-pp.page_counter_cancel.page_counter_uncharge.obj_cgroup_uncharge_pages.kmem_cache_free.rcu_do_batch 1.10 ± 6% -0.3 0.83 ± 7% perf-profile.calltrace.cycles-pp.sock_def_readable.tcp_child_process.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish 1.03 ± 6% -0.3 0.78 ± 7% perf-profile.calltrace.cycles-pp.__wake_up_common_lock.sock_def_readable.tcp_child_process.tcp_v4_rcv.ip_protocol_deliver_rcu 1.04 ± 8% -0.2 0.79 ± 8% perf-profile.calltrace.cycles-pp.sock_def_readable.tcp_rcv_established.tcp_v4_do_rcv.tcp_v4_rcv.ip_protocol_deliver_rcu 0.96 ± 6% -0.2 0.72 ± 7% perf-profile.calltrace.cycles-pp.__wake_up_common.__wake_up_common_lock.sock_def_readable.tcp_child_process.tcp_v4_rcv 0.95 ± 6% -0.2 0.71 ± 7% perf-profile.calltrace.cycles-pp.autoremove_wake_function.__wake_up_common.__wake_up_common_lock.sock_def_readable.tcp_child_process 0.97 ± 8% -0.2 0.74 ± 8% perf-profile.calltrace.cycles-pp.__wake_up_common_lock.sock_def_readable.tcp_rcv_established.tcp_v4_do_rcv.tcp_v4_rcv 0.92 ± 6% -0.2 0.68 ± 7% perf-profile.calltrace.cycles-pp.try_to_wake_up.autoremove_wake_function.__wake_up_common.__wake_up_common_lock.sock_def_readable 0.90 ± 8% -0.2 0.68 ± 8% perf-profile.calltrace.cycles-pp.__wake_up_common.__wake_up_common_lock.sock_def_readable.tcp_rcv_established.tcp_v4_do_rcv 0.88 ± 8% -0.2 0.67 ± 8% perf-profile.calltrace.cycles-pp.try_to_wake_up.__wake_up_common.__wake_up_common_lock.sock_def_readable.tcp_rcv_established 0.75 ± 7% -0.2 0.57 ± 23% perf-profile.calltrace.cycles-pp.alloc_file_pseudo.sock_alloc_file.__sys_accept4_file.__sys_accept4.__x64_sys_accept 0.67 ± 7% -0.2 0.49 ± 41% perf-profile.calltrace.cycles-pp.inet_csk_get_port.__inet_bind.inet_bind.__sys_bind.__x64_sys_bind 0.55 ± 22% +0.2 0.72 ± 7% perf-profile.calltrace.cycles-pp.tcp_v4_send_synack.tcp_conn_request.tcp_rcv_state_process.tcp_v4_do_rcv.tcp_v4_rcv 0.49 ± 40% +0.3 0.75 ± 8% perf-profile.calltrace.cycles-pp.tcp_rcv_state_process.tcp_child_process.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish 1.65 ± 6% +0.3 1.97 ± 7% perf-profile.calltrace.cycles-pp.tcp_v4_syn_recv_sock.tcp_check_req.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish 1.87 ± 6% +0.4 2.24 ± 7% perf-profile.calltrace.cycles-pp.tcp_check_req.tcp_v4_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish.ip_local_deliver 1.56 ± 6% +0.5 2.02 ± 6% perf-profile.calltrace.cycles-pp.tcp_conn_request.tcp_rcv_state_process.tcp_v4_do_rcv.tcp_v4_rcv.ip_protocol_deliver_rcu 0.00 +0.6 0.57 ± 6% perf-profile.calltrace.cycles-pp.sock_def_readable.tcp_data_queue.tcp_rcv_established.tcp_v4_do_rcv.tcp_v4_rcv 3.08 ± 6% +0.7 3.74 ± 7% perf-profile.calltrace.cycles-pp.ip_output.__ip_queue_xmit.__tcp_transmit_skb.tcp_connect.tcp_v4_connect 3.05 ± 6% +0.7 3.73 ± 7% perf-profile.calltrace.cycles-pp.ip_finish_output2.ip_output.__ip_queue_xmit.__tcp_transmit_skb.tcp_connect 3.42 ± 6% +0.8 4.18 ± 7% perf-profile.calltrace.cycles-pp.__ip_queue_xmit.__tcp_transmit_skb.tcp_connect.tcp_v4_connect.__inet_stream_connect 3.64 ± 6% +0.9 4.50 ± 6% perf-profile.calltrace.cycles-pp.__tcp_transmit_skb.tcp_connect.tcp_v4_connect.__inet_stream_connect.inet_stream_connect 4.42 ± 6% +1.0 5.44 ± 7% perf-profile.calltrace.cycles-pp.tcp_connect.tcp_v4_connect.__inet_stream_connect.inet_stream_connect.__sys_connect 5.36 ± 6% +1.3 6.68 ± 7% perf-profile.calltrace.cycles-pp.tcp_v4_connect.__inet_stream_connect.inet_stream_connect.__sys_connect.__x64_sys_connect 11.59 ± 5% +2.1 13.65 ± 6% perf-profile.calltrace.cycles-pp.inet_stream_connect.__sys_connect.__x64_sys_connect.do_syscall_64.entry_SYSCALL_64_after_hwframe 11.54 ± 5% +2.1 13.61 ± 6% perf-profile.calltrace.cycles-pp.__inet_stream_connect.inet_stream_connect.__sys_connect.__x64_sys_connect.do_syscall_64 5.94 ± 6% -1.3 4.60 ± 6% perf-profile.children.cycles-pp.__x64_sys_recvfrom 5.90 ± 6% -1.3 4.58 ± 6% perf-profile.children.cycles-pp.__sys_recvfrom 5.32 ± 6% -1.2 4.08 ± 6% perf-profile.children.cycles-pp.inet_recvmsg 5.30 ± 6% -1.2 4.07 ± 6% perf-profile.children.cycles-pp.tcp_recvmsg 4.17 ± 22% -1.0 3.19 ± 6% perf-profile.children.cycles-pp.poll_idle 4.48 ± 6% -0.8 3.71 ± 6% perf-profile.children.cycles-pp.tcp_recvmsg_locked 3.06 ± 8% -0.7 2.39 ± 7% perf-profile.children.cycles-pp.kmem_cache_free 1.66 ± 6% -0.6 1.06 ± 6% perf-profile.children.cycles-pp._raw_spin_lock_bh 2.80 ± 6% -0.6 2.21 ± 7% perf-profile.children.cycles-pp.__x64_sys_accept 2.78 ± 6% -0.6 2.20 ± 7% perf-profile.children.cycles-pp.__sys_accept4 2.75 ± 6% -0.6 2.17 ± 7% perf-profile.children.cycles-pp.__sys_accept4_file 2.37 ± 9% -0.5 1.85 ± 8% perf-profile.children.cycles-pp.rcu_core 2.33 ± 9% -0.5 1.81 ± 8% perf-profile.children.cycles-pp.rcu_do_batch 2.35 ± 8% -0.5 1.84 ± 8% perf-profile.children.cycles-pp.obj_cgroup_uncharge_pages 2.31 ± 8% -0.5 1.81 ± 8% perf-profile.children.cycles-pp.page_counter_uncharge 1.12 ± 7% -0.5 0.63 ± 7% perf-profile.children.cycles-pp.native_queued_spin_lock_slowpath 1.37 ± 6% -0.5 0.89 ± 6% perf-profile.children.cycles-pp.lock_sock_nested 1.71 ± 9% -0.5 1.25 ± 9% perf-profile.children.cycles-pp.ret_from_fork 1.71 ± 9% -0.5 1.25 ± 9% perf-profile.children.cycles-pp.kthread 1.67 ± 9% -0.5 1.22 ± 9% perf-profile.children.cycles-pp.run_ksoftirqd 1.68 ± 9% -0.5 1.23 ± 9% perf-profile.children.cycles-pp.smpboot_thread_fn 2.84 ± 6% -0.4 2.43 ± 6% perf-profile.children.cycles-pp.sk_wait_data 1.78 ± 6% -0.4 1.36 ± 6% perf-profile.children.cycles-pp.__dev_queue_xmit 1.83 ± 9% -0.4 1.43 ± 8% perf-profile.children.cycles-pp.page_counter_cancel 2.14 ± 6% -0.4 1.75 ± 6% perf-profile.children.cycles-pp._raw_spin_lock 1.44 ± 7% -0.3 1.13 ± 6% perf-profile.children.cycles-pp.kmem_cache_alloc 1.40 ± 7% -0.3 1.10 ± 8% perf-profile.children.cycles-pp.inet_accept 1.56 ± 6% -0.3 1.26 ± 6% perf-profile.children.cycles-pp.alloc_file_pseudo 1.56 ± 6% -0.3 1.27 ± 6% perf-profile.children.cycles-pp.sock_alloc_file 1.03 ± 6% -0.3 0.73 ± 7% perf-profile.children.cycles-pp.percpu_counter_add_batch 1.20 ± 6% -0.3 0.92 ± 6% perf-profile.children.cycles-pp.__alloc_skb 0.62 ± 6% -0.3 0.37 ± 9% perf-profile.children.cycles-pp.__tcp_get_metrics 0.97 ± 5% -0.2 0.73 ± 6% perf-profile.children.cycles-pp.autoremove_wake_function 1.15 ± 6% -0.2 0.92 ± 8% perf-profile.children.cycles-pp.inet_csk_accept 0.59 ± 5% -0.2 0.36 ± 7% perf-profile.children.cycles-pp.__cgroup_bpf_run_filter_skb 0.80 ± 6% -0.2 0.58 ± 6% perf-profile.children.cycles-pp.sk_forced_mem_schedule 1.12 ± 6% -0.2 0.90 ± 6% perf-profile.children.cycles-pp.alloc_file 1.10 ± 6% -0.2 0.89 ± 6% perf-profile.children.cycles-pp.alloc_empty_file 1.03 ± 6% -0.2 0.82 ± 7% perf-profile.children.cycles-pp.sock_def_wakeup 1.08 ± 6% -0.2 0.88 ± 6% perf-profile.children.cycles-pp.__alloc_file 1.21 ± 6% -0.2 1.01 ± 6% perf-profile.children.cycles-pp.inet_csk_destroy_sock 0.78 ± 6% -0.2 0.58 ± 6% perf-profile.children.cycles-pp.sk_stream_alloc_skb 0.89 ± 6% -0.2 0.70 ± 7% perf-profile.children.cycles-pp.dev_hard_start_xmit 0.36 ± 6% -0.2 0.18 ± 6% perf-profile.children.cycles-pp.ip_finish_output 0.85 ± 6% -0.2 0.68 ± 7% perf-profile.children.cycles-pp.__sk_mem_reduce_allocated 0.78 ± 6% -0.2 0.61 ± 7% perf-profile.children.cycles-pp.loopback_xmit 1.00 ± 6% -0.2 0.83 ± 6% perf-profile.children.cycles-pp.__kfree_skb 1.15 ± 6% -0.2 0.98 ± 6% perf-profile.children.cycles-pp._raw_spin_lock_irqsave 0.68 ± 6% -0.2 0.51 ± 7% perf-profile.children.cycles-pp.tcp_get_metrics 0.79 ± 6% -0.2 0.63 ± 6% perf-profile.children.cycles-pp.tcp_send_fin 0.92 ± 6% -0.2 0.76 ± 6% perf-profile.children.cycles-pp.inet_csk_clone_lock 0.67 ± 6% -0.2 0.51 ± 8% perf-profile.children.cycles-pp.sk_reset_timer 0.68 ± 6% -0.2 0.52 ± 7% perf-profile.children.cycles-pp.sock_alloc 0.67 ± 6% -0.2 0.52 ± 7% perf-profile.children.cycles-pp.new_inode_pseudo 0.89 ± 5% -0.1 0.74 ± 6% perf-profile.children.cycles-pp.sk_clone_lock 0.77 ± 7% -0.1 0.63 ± 7% perf-profile.children.cycles-pp.tcp_done 0.76 ± 6% -0.1 0.62 ± 6% perf-profile.children.cycles-pp.security_file_alloc 0.58 ± 7% -0.1 0.45 ± 7% perf-profile.children.cycles-pp.alloc_inode 0.59 ± 6% -0.1 0.46 ± 7% perf-profile.children.cycles-pp.mod_timer 0.67 ± 6% -0.1 0.54 ± 6% perf-profile.children.cycles-pp.__dentry_kill 0.69 ± 6% -0.1 0.57 ± 6% perf-profile.children.cycles-pp.apparmor_file_alloc_security 0.45 ± 6% -0.1 0.33 ± 7% perf-profile.children.cycles-pp.lock_timer_base 0.58 ± 7% -0.1 0.47 ± 6% perf-profile.children.cycles-pp.apparmor_file_free_security 0.59 ± 7% -0.1 0.47 ± 6% perf-profile.children.cycles-pp.security_file_free 0.68 ± 7% -0.1 0.56 ± 7% perf-profile.children.cycles-pp.inet_csk_get_port 0.44 ± 5% -0.1 0.33 ± 7% perf-profile.children.cycles-pp.sk_filter_trim_cap 0.47 ± 7% -0.1 0.36 ± 7% perf-profile.children.cycles-pp.sock_alloc_inode 0.54 ± 8% -0.1 0.44 ± 10% perf-profile.children.cycles-pp.propagate_protected_usage 0.68 ± 6% -0.1 0.58 ± 6% perf-profile.children.cycles-pp.skb_release_data 0.43 ± 6% -0.1 0.33 ± 7% perf-profile.children.cycles-pp.kmalloc_reserve 0.45 ± 10% -0.1 0.36 ± 10% perf-profile.children.cycles-pp.inet_csk_bind_conflict 0.38 ± 6% -0.1 0.30 ± 7% perf-profile.children.cycles-pp.__kmalloc_node_track_caller 0.40 ± 6% -0.1 0.32 ± 7% perf-profile.children.cycles-pp.raw_local_deliver 0.46 ± 6% -0.1 0.38 ± 7% perf-profile.children.cycles-pp.read_tsc 0.40 ± 6% -0.1 0.32 ± 5% perf-profile.children.cycles-pp.skb_release_all 0.44 ± 6% -0.1 0.36 ± 7% perf-profile.children.cycles-pp.sock_recvmsg 0.37 ± 7% -0.1 0.29 ± 8% perf-profile.children.cycles-pp.__tcp_send_ack 0.46 ± 7% -0.1 0.38 ± 7% perf-profile.children.cycles-pp.__skb_datagram_iter 0.48 ± 6% -0.1 0.40 ± 7% perf-profile.children.cycles-pp.___might_sleep 0.43 ± 7% -0.1 0.36 ± 8% perf-profile.children.cycles-pp.security_socket_recvmsg 0.38 ± 7% -0.1 0.30 ± 5% perf-profile.children.cycles-pp.skb_release_head_state 0.37 ± 7% -0.1 0.29 ± 7% perf-profile.children.cycles-pp.apparmor_sk_clone_security 0.37 ± 7% -0.1 0.30 ± 6% perf-profile.children.cycles-pp.security_sk_clone 0.29 ± 6% -0.1 0.22 ± 8% perf-profile.children.cycles-pp.__ksize 0.31 ± 6% -0.1 0.24 ± 9% perf-profile.children.cycles-pp.netif_rx 0.37 ± 7% -0.1 0.29 ± 7% perf-profile.children.cycles-pp.select_task_rq_fair 0.32 ± 6% -0.1 0.25 ± 8% perf-profile.children.cycles-pp.tcp_write_queue_purge 0.30 ± 7% -0.1 0.23 ± 9% perf-profile.children.cycles-pp.netif_rx_internal 0.24 ± 5% -0.1 0.17 ± 7% perf-profile.children.cycles-pp.ip_local_out 0.30 ± 6% -0.1 0.23 ± 8% perf-profile.children.cycles-pp.irqtime_account_irq 0.40 ± 6% -0.1 0.33 ± 7% perf-profile.children.cycles-pp.apparmor_socket_post_create 0.40 ± 6% -0.1 0.33 ± 6% perf-profile.children.cycles-pp.security_socket_post_create 0.25 ± 8% -0.1 0.18 ± 6% perf-profile.children.cycles-pp.__skb_clone 0.32 ± 6% -0.1 0.25 ± 7% perf-profile.children.cycles-pp.tcp_mstamp_refresh 0.22 ± 4% -0.1 0.16 ± 6% perf-profile.children.cycles-pp.__ip_local_out 0.26 ± 7% -0.1 0.20 ± 9% perf-profile.children.cycles-pp.validate_xmit_skb 0.27 ± 7% -0.1 0.21 ± 6% perf-profile.children.cycles-pp.sk_stop_timer 0.26 ± 6% -0.1 0.20 ± 7% perf-profile.children.cycles-pp.enqueue_to_backlog 0.21 ± 8% -0.1 0.15 ± 8% perf-profile.children.cycles-pp.kmem_cache_alloc_node 0.14 ± 7% -0.1 0.08 ± 10% perf-profile.children.cycles-pp.migrate_enable 0.20 ± 7% -0.1 0.15 ± 8% perf-profile.children.cycles-pp.__check_object_size 0.18 ± 6% -0.1 0.12 ± 6% perf-profile.children.cycles-pp.ip_send_check 0.24 ± 8% -0.1 0.19 ± 7% perf-profile.children.cycles-pp.__slab_free 0.24 ± 6% -0.1 0.19 ± 6% perf-profile.children.cycles-pp.inet_csk_clear_xmit_timers 0.26 ± 8% -0.0 0.22 ± 6% perf-profile.children.cycles-pp.d_alloc_pseudo 0.21 ± 6% -0.0 0.17 ± 9% perf-profile.children.cycles-pp.copy_user_generic_unrolled 0.12 ± 8% -0.0 0.07 ± 8% perf-profile.children.cycles-pp.sock_put 0.28 ± 6% -0.0 0.24 ± 7% perf-profile.children.cycles-pp.__switch_to 0.18 ± 7% -0.0 0.14 ± 7% perf-profile.children.cycles-pp.del_timer 0.21 ± 5% -0.0 0.17 ± 9% perf-profile.children.cycles-pp.kfree 0.15 ± 9% -0.0 0.11 ± 7% perf-profile.children.cycles-pp.get_obj_cgroup_from_current 0.12 ± 10% -0.0 0.08 ± 10% perf-profile.children.cycles-pp.tcp_queue_rcv 0.08 ± 9% -0.0 0.05 ± 41% perf-profile.children.cycles-pp.netdev_core_pick_tx 0.13 ± 7% -0.0 0.10 ± 11% perf-profile.children.cycles-pp.netif_skb_features 0.08 ± 9% -0.0 0.05 ± 34% perf-profile.children.cycles-pp.nf_hook_slow 0.22 ± 6% -0.0 0.18 ± 9% perf-profile.children.cycles-pp.__might_sleep 0.10 ± 9% -0.0 0.07 ± 10% perf-profile.children.cycles-pp.__destroy_inode 0.14 ± 7% -0.0 0.11 ± 8% perf-profile.children.cycles-pp.tcp_rcv_space_adjust 0.10 ± 8% -0.0 0.07 ± 9% perf-profile.children.cycles-pp.destroy_inode 0.09 ± 8% -0.0 0.06 ± 11% perf-profile.children.cycles-pp.inet_csk_complete_hashdance 0.14 ± 7% -0.0 0.11 ± 8% perf-profile.children.cycles-pp.copyout 0.22 ± 7% -0.0 0.19 ± 8% perf-profile.children.cycles-pp.__update_load_avg_se 0.18 ± 6% -0.0 0.15 ± 8% perf-profile.children.cycles-pp.__fget_light 0.14 ± 8% -0.0 0.11 ± 9% perf-profile.children.cycles-pp.simple_copy_to_iter 0.15 ± 8% -0.0 0.13 ± 7% perf-profile.children.cycles-pp.__cond_resched 0.09 ± 9% -0.0 0.06 ± 10% perf-profile.children.cycles-pp.inet_reqsk_alloc 0.13 ± 6% -0.0 0.11 ± 8% perf-profile.children.cycles-pp.d_instantiate 0.09 ± 12% -0.0 0.06 ± 10% perf-profile.children.cycles-pp.security_sock_rcv_skb 0.08 ± 11% -0.0 0.05 ± 24% perf-profile.children.cycles-pp.migrate_disable 0.11 ± 10% -0.0 0.08 ± 10% perf-profile.children.cycles-pp.tcp_validate_incoming 0.14 ± 8% -0.0 0.11 ± 8% perf-profile.children.cycles-pp.tcp_wfree 0.12 ± 7% -0.0 0.09 ± 7% perf-profile.children.cycles-pp.tcp_add_backlog 0.07 ± 10% -0.0 0.05 ± 33% perf-profile.children.cycles-pp.fsnotify_destroy_marks 0.09 ± 8% -0.0 0.07 ± 10% perf-profile.children.cycles-pp.sock_setsockopt 0.10 ± 7% -0.0 0.08 ± 8% perf-profile.children.cycles-pp.tcp_init_xmit_timers 0.09 ± 9% +0.0 0.13 ± 9% perf-profile.children.cycles-pp.inet_twsk_alloc 0.06 ± 8% +0.0 0.09 ± 7% perf-profile.children.cycles-pp.tcp_mtup_init 0.23 ± 6% +0.0 0.27 ± 5% perf-profile.children.cycles-pp.tcp_make_synack 0.06 ± 9% +0.1 0.12 ± 8% perf-profile.children.cycles-pp.tcp_synack_rtt_meas 0.06 ± 10% +0.1 0.11 ± 8% perf-profile.children.cycles-pp.tcp_ca_openreq_child 0.00 +0.1 0.06 ± 10% perf-profile.children.cycles-pp.tcp_rack_update_reo_wnd 0.07 ± 9% +0.1 0.14 ± 9% perf-profile.children.cycles-pp.__inet_lookup_listener 0.12 ± 7% +0.1 0.20 ± 7% perf-profile.children.cycles-pp.__ip_finish_output 0.06 ± 13% +0.1 0.14 ± 8% perf-profile.children.cycles-pp.__ip_dev_find 0.01 ±212% +0.1 0.09 ± 8% perf-profile.children.cycles-pp.inet_lhash2_lookup 0.00 +0.1 0.08 ± 8% perf-profile.children.cycles-pp.sock_prot_inuse_add 0.00 +0.1 0.09 ± 10% perf-profile.children.cycles-pp.tcp_push 0.07 ± 10% +0.1 0.16 ± 8% perf-profile.children.cycles-pp.tcp_select_initial_window 0.07 ± 10% +0.1 0.16 ± 7% perf-profile.children.cycles-pp.tcp_sync_mss 0.07 ± 9% +0.1 0.16 ± 8% perf-profile.children.cycles-pp.secure_tcp_ts_off 0.36 ± 6% +0.1 0.47 ± 6% perf-profile.children.cycles-pp.tcp_finish_connect 0.02 ±120% +0.1 0.13 ± 10% perf-profile.children.cycles-pp.inet_lookup_ifaddr_rcu 0.08 ± 11% +0.1 0.19 ± 8% perf-profile.children.cycles-pp.skb_entail 0.17 ± 5% +0.1 0.29 ± 7% perf-profile.children.cycles-pp.tcp_connect_init 0.10 ± 9% +0.1 0.21 ± 7% perf-profile.children.cycles-pp.tcp_openreq_init_rwin 0.00 +0.1 0.13 ± 9% perf-profile.children.cycles-pp.tcp_stream_memory_free 0.14 ± 6% +0.1 0.26 ± 7% perf-profile.children.cycles-pp.tcp_parse_options 0.13 ± 8% +0.1 0.26 ± 8% perf-profile.children.cycles-pp.tcp_update_pacing_rate 0.58 ± 5% +0.2 0.73 ± 5% perf-profile.children.cycles-pp.tcp_v4_send_synack 0.08 ± 7% +0.2 0.24 ± 6% perf-profile.children.cycles-pp.tcp_assign_congestion_control 0.34 ± 7% +0.2 0.50 ± 6% perf-profile.children.cycles-pp.tcp_v4_init_sock 0.33 ± 6% +0.2 0.49 ± 6% perf-profile.children.cycles-pp.tcp_init_sock 0.23 ± 6% +0.2 0.40 ± 6% perf-profile.children.cycles-pp.fib_table_lookup 0.00 +0.2 0.17 ± 8% perf-profile.children.cycles-pp.__sk_free 0.22 ± 7% +0.2 0.41 ± 6% perf-profile.children.cycles-pp.tcp_schedule_loss_probe 0.14 ± 7% +0.2 0.33 ± 7% perf-profile.children.cycles-pp.ip_rcv_finish 0.63 ± 5% +0.2 0.83 ± 7% perf-profile.children.cycles-pp.tcp_init_transfer 0.10 ± 8% +0.2 0.31 ± 7% perf-profile.children.cycles-pp.ip_rcv_finish_core 0.04 ± 40% +0.2 0.25 ± 6% perf-profile.children.cycles-pp.tcp_newly_delivered 0.07 ± 10% +0.2 0.28 ± 7% perf-profile.children.cycles-pp.fib_lookup_good_nhc 0.19 ± 7% +0.2 0.41 ± 6% perf-profile.children.cycles-pp.inet_csk_route_child_sock 0.18 ± 6% +0.2 0.40 ± 6% perf-profile.children.cycles-pp.inet_csk_route_req 0.11 ± 7% +0.2 0.34 ± 8% perf-profile.children.cycles-pp.inet_sk_rebuild_header 0.11 ± 8% +0.2 0.35 ± 7% perf-profile.children.cycles-pp.inet_sock_destruct 0.10 ± 9% +0.3 0.37 ± 8% perf-profile.children.cycles-pp.__tcp_select_window 1.69 ± 5% +0.3 2.00 ± 6% perf-profile.children.cycles-pp.tcp_v4_syn_recv_sock 0.20 ± 6% +0.3 0.52 ± 8% perf-profile.children.cycles-pp.tcp_ack_update_rtt 0.15 ± 8% +0.3 0.49 ± 7% perf-profile.children.cycles-pp.inet_ehashfn 1.91 ± 5% +0.4 2.27 ± 6% perf-profile.children.cycles-pp.tcp_check_req 0.17 ± 6% +0.4 0.54 ± 7% perf-profile.children.cycles-pp.ip_rcv_core 1.59 ± 5% +0.5 2.05 ± 6% perf-profile.children.cycles-pp.tcp_conn_request 0.43 ± 5% +0.6 0.99 ± 5% perf-profile.children.cycles-pp.ip_route_output_flow 2.02 ± 6% +0.6 2.67 ± 6% perf-profile.children.cycles-pp.tcp_ack 0.26 ± 7% +0.8 1.02 ± 7% perf-profile.children.cycles-pp.__sk_dst_check 0.60 ± 5% +0.8 1.37 ± 6% perf-profile.children.cycles-pp.ip_route_output_key_hash 0.55 ± 5% +0.8 1.33 ± 6% perf-profile.children.cycles-pp.ip_route_output_key_hash_rcu 5.73 ± 5% +0.9 6.60 ± 6% perf-profile.children.cycles-pp.tcp_rcv_synsent_state_process 0.24 ± 6% +0.9 1.19 ± 7% perf-profile.children.cycles-pp.ipv4_dst_check 4.49 ± 5% +1.0 5.47 ± 6% perf-profile.children.cycles-pp.tcp_connect 5.45 ± 5% +1.3 6.72 ± 6% perf-profile.children.cycles-pp.tcp_v4_connect 11.81 ± 5% +2.0 13.85 ± 6% perf-profile.children.cycles-pp.__x64_sys_connect 11.79 ± 5% +2.1 13.84 ± 6% perf-profile.children.cycles-pp.__sys_connect 11.59 ± 5% +2.1 13.65 ± 6% perf-profile.children.cycles-pp.inet_stream_connect 11.54 ± 5% +2.1 13.61 ± 6% perf-profile.children.cycles-pp.__inet_stream_connect 4.00 ± 20% -0.9 3.08 ± 6% perf-profile.self.cycles-pp.poll_idle 1.12 ± 7% -0.5 0.63 ± 7% perf-profile.self.cycles-pp.native_queued_spin_lock_slowpath 1.82 ± 9% -0.4 1.42 ± 8% perf-profile.self.cycles-pp.page_counter_cancel 1.01 ± 6% -0.3 0.72 ± 6% perf-profile.self.cycles-pp.percpu_counter_add_batch 1.65 ± 6% -0.3 1.36 ± 5% perf-profile.self.cycles-pp._raw_spin_lock 0.62 ± 6% -0.3 0.36 ± 8% perf-profile.self.cycles-pp.__tcp_get_metrics 1.09 ± 6% -0.3 0.84 ± 6% perf-profile.self.cycles-pp.__inet_lookup_established 0.47 ± 9% -0.2 0.25 ± 7% perf-profile.self.cycles-pp.__sk_destruct 0.79 ± 6% -0.2 0.57 ± 6% perf-profile.self.cycles-pp.sk_forced_mem_schedule 0.99 ± 6% -0.2 0.78 ± 7% perf-profile.self.cycles-pp._raw_spin_lock_bh 0.84 ± 6% -0.2 0.67 ± 7% perf-profile.self.cycles-pp.__sk_mem_reduce_allocated 1.14 ± 6% -0.2 0.97 ± 6% perf-profile.self.cycles-pp._raw_spin_lock_irqsave 0.36 ± 6% -0.1 0.22 ± 8% perf-profile.self.cycles-pp.__cgroup_bpf_run_filter_skb 0.67 ± 6% -0.1 0.55 ± 6% perf-profile.self.cycles-pp.apparmor_file_alloc_security 0.40 ± 7% -0.1 0.28 ± 7% perf-profile.self.cycles-pp.tcp_recvmsg_locked 0.58 ± 7% -0.1 0.46 ± 5% perf-profile.self.cycles-pp.apparmor_file_free_security 0.50 ± 6% -0.1 0.38 ± 7% perf-profile.self.cycles-pp.__dev_queue_xmit 0.54 ± 9% -0.1 0.43 ± 10% perf-profile.self.cycles-pp.propagate_protected_usage 0.40 ± 6% -0.1 0.30 ± 7% perf-profile.self.cycles-pp.kmem_cache_free 0.33 ± 6% -0.1 0.24 ± 7% perf-profile.self.cycles-pp.tcp_clean_rtx_queue 0.45 ± 10% -0.1 0.35 ± 10% perf-profile.self.cycles-pp.inet_csk_bind_conflict 0.28 ± 8% -0.1 0.20 ± 9% perf-profile.self.cycles-pp.ip_finish_output2 0.39 ± 6% -0.1 0.31 ± 7% perf-profile.self.cycles-pp.raw_local_deliver 0.45 ± 6% -0.1 0.37 ± 6% perf-profile.self.cycles-pp.read_tsc 0.47 ± 6% -0.1 0.39 ± 7% perf-profile.self.cycles-pp.___might_sleep 0.37 ± 7% -0.1 0.29 ± 7% perf-profile.self.cycles-pp.apparmor_sk_clone_security 0.26 ± 7% -0.1 0.19 ± 8% perf-profile.self.cycles-pp.tcp_sendmsg_locked 0.28 ± 6% -0.1 0.21 ± 7% perf-profile.self.cycles-pp.__ksize 0.39 ± 6% -0.1 0.32 ± 7% perf-profile.self.cycles-pp.do_idle 0.39 ± 6% -0.1 0.33 ± 7% perf-profile.self.cycles-pp.apparmor_socket_post_create 0.20 ± 8% -0.1 0.14 ± 7% perf-profile.self.cycles-pp.__skb_clone 0.23 ± 8% -0.1 0.17 ± 9% perf-profile.self.cycles-pp.exit_to_user_mode_prepare 0.13 ± 7% -0.1 0.07 ± 11% perf-profile.self.cycles-pp.migrate_enable 0.17 ± 6% -0.1 0.12 ± 6% perf-profile.self.cycles-pp.ip_send_check 0.24 ± 8% -0.1 0.19 ± 8% perf-profile.self.cycles-pp.__slab_free 0.24 ± 7% -0.1 0.19 ± 7% perf-profile.self.cycles-pp.__kmalloc_node_track_caller 0.24 ± 7% -0.1 0.19 ± 8% perf-profile.self.cycles-pp.__alloc_skb 0.17 ± 7% -0.0 0.13 ± 7% perf-profile.self.cycles-pp.net_rx_action 0.18 ± 7% -0.0 0.13 ± 10% perf-profile.self.cycles-pp.process_backlog 0.16 ± 7% -0.0 0.12 ± 6% perf-profile.self.cycles-pp.fib_table_lookup 0.12 ± 9% -0.0 0.07 ± 8% perf-profile.self.cycles-pp.sock_put 0.27 ± 5% -0.0 0.23 ± 8% perf-profile.self.cycles-pp.__switch_to 0.14 ± 11% -0.0 0.10 ± 11% perf-profile.self.cycles-pp.do_syscall_64 0.08 ± 13% -0.0 0.04 ± 49% perf-profile.self.cycles-pp.select_task_rq_fair 0.15 ± 9% -0.0 0.11 ± 9% perf-profile.self.cycles-pp.kmem_cache_alloc_node 0.21 ± 5% -0.0 0.17 ± 9% perf-profile.self.cycles-pp.copy_user_generic_unrolled 0.21 ± 7% -0.0 0.17 ± 9% perf-profile.self.cycles-pp.__local_bh_enable_ip 0.20 ± 5% -0.0 0.17 ± 9% perf-profile.self.cycles-pp.kfree 0.08 ± 7% -0.0 0.04 ± 64% perf-profile.self.cycles-pp.netdev_core_pick_tx 0.15 ± 6% -0.0 0.12 ± 9% perf-profile.self.cycles-pp.__tcp_close 0.11 ± 9% -0.0 0.08 ± 10% perf-profile.self.cycles-pp.tcp_queue_rcv 0.07 ± 9% -0.0 0.04 ± 56% perf-profile.self.cycles-pp.migrate_disable 0.07 ± 8% -0.0 0.04 ± 64% perf-profile.self.cycles-pp.inet_reqsk_alloc 0.09 ± 9% -0.0 0.05 ± 10% perf-profile.self.cycles-pp.tcp_data_queue 0.18 ± 8% -0.0 0.15 ± 6% perf-profile.self.cycles-pp.__entry_text_start 0.18 ± 6% -0.0 0.14 ± 8% perf-profile.self.cycles-pp.__fget_light 0.11 ± 9% -0.0 0.08 ± 7% perf-profile.self.cycles-pp.get_obj_cgroup_from_current 0.14 ± 8% -0.0 0.11 ± 9% perf-profile.self.cycles-pp.__check_object_size 0.14 ± 8% -0.0 0.11 ± 9% perf-profile.self.cycles-pp.enqueue_to_backlog 0.14 ± 9% -0.0 0.11 ± 10% perf-profile.self.cycles-pp.sk_reset_timer 0.12 ± 9% -0.0 0.09 ± 8% perf-profile.self.cycles-pp.validate_xmit_skb 0.14 ± 9% -0.0 0.11 ± 9% perf-profile.self.cycles-pp.sk_setup_caps 0.09 ± 7% -0.0 0.06 ± 11% perf-profile.self.cycles-pp.irqtime_account_irq 0.11 ± 10% -0.0 0.08 ± 10% perf-profile.self.cycles-pp.tcp_validate_incoming 0.14 ± 8% -0.0 0.11 ± 8% perf-profile.self.cycles-pp.tcp_wfree 0.07 ± 10% -0.0 0.05 ± 23% perf-profile.self.cycles-pp.tcp_rcv_synsent_state_process 0.08 ± 9% -0.0 0.06 ± 10% perf-profile.self.cycles-pp.sk_filter_trim_cap 0.08 ± 6% -0.0 0.06 ± 11% perf-profile.self.cycles-pp.__netif_receive_skb_one_core 0.07 ± 8% -0.0 0.05 ± 23% perf-profile.self.cycles-pp.tcp_options_write 0.08 ± 9% -0.0 0.06 ± 13% perf-profile.self.cycles-pp.netif_skb_features 0.08 ± 9% -0.0 0.06 ± 9% perf-profile.self.cycles-pp._raw_spin_unlock_irqrestore 0.06 ± 10% +0.0 0.08 ± 8% perf-profile.self.cycles-pp.tcp_init_transfer 0.09 ± 9% +0.0 0.12 ± 9% perf-profile.self.cycles-pp.inet_twsk_alloc 0.07 ± 8% +0.0 0.10 ± 7% perf-profile.self.cycles-pp.inet_csk_get_port 0.06 ± 8% +0.0 0.09 ± 8% perf-profile.self.cycles-pp.tcp_mtup_init 0.11 ± 7% +0.0 0.16 ± 7% perf-profile.self.cycles-pp.sk_clone_lock 0.00 +0.1 0.06 ± 10% perf-profile.self.cycles-pp.tcp_rack_update_reo_wnd 0.00 +0.1 0.06 ± 12% perf-profile.self.cycles-pp.inet_twsk_kill 0.07 ± 7% +0.1 0.13 ± 8% perf-profile.self.cycles-pp.tcp_connect 0.07 ± 7% +0.1 0.14 ± 7% perf-profile.self.cycles-pp.__inet_bind 0.12 ± 7% +0.1 0.19 ± 7% perf-profile.self.cycles-pp.__ip_finish_output 0.06 ± 11% +0.1 0.14 ± 5% perf-profile.self.cycles-pp.tcp_make_synack 0.06 ± 6% +0.1 0.14 ± 8% perf-profile.self.cycles-pp.tcp_v4_syn_recv_sock 0.01 ±251% +0.1 0.09 ± 8% perf-profile.self.cycles-pp.inet_lhash2_lookup 0.00 +0.1 0.08 ± 8% perf-profile.self.cycles-pp.sock_prot_inuse_add 0.00 +0.1 0.09 ± 9% perf-profile.self.cycles-pp.tcp_push 0.07 ± 10% +0.1 0.16 ± 8% perf-profile.self.cycles-pp.tcp_select_initial_window 0.06 ± 10% +0.1 0.15 ± 8% perf-profile.self.cycles-pp.tcp_sync_mss 0.00 +0.1 0.09 ± 11% perf-profile.self.cycles-pp.tcp_v4_send_synack 0.05 ± 22% +0.1 0.14 ± 8% perf-profile.self.cycles-pp.tcp_create_openreq_child 0.04 ± 55% +0.1 0.14 ± 8% perf-profile.self.cycles-pp.tcp_get_metrics 0.03 ± 91% +0.1 0.13 ± 7% perf-profile.self.cycles-pp.tcp_openreq_init_rwin 0.06 ± 11% +0.1 0.16 ± 8% perf-profile.self.cycles-pp.tcp_connect_init 0.02 ±120% +0.1 0.13 ± 10% perf-profile.self.cycles-pp.inet_lookup_ifaddr_rcu 0.08 ± 15% +0.1 0.19 ± 9% perf-profile.self.cycles-pp.tcp_update_metrics 0.06 ± 13% +0.1 0.19 ± 7% perf-profile.self.cycles-pp.skb_entail 0.14 ± 6% +0.1 0.26 ± 7% perf-profile.self.cycles-pp.tcp_parse_options 0.00 +0.1 0.12 ± 8% perf-profile.self.cycles-pp.tcp_stream_memory_free 0.04 ± 55% +0.1 0.17 ± 8% perf-profile.self.cycles-pp.tcp_init_metrics 0.46 ± 6% +0.1 0.59 ± 6% perf-profile.self.cycles-pp.tcp_ack 0.11 ± 7% +0.1 0.24 ± 8% perf-profile.self.cycles-pp.tcp_event_new_data_sent 0.12 ± 8% +0.1 0.25 ± 8% perf-profile.self.cycles-pp.tcp_update_pacing_rate 0.00 ±458% +0.1 0.13 ± 8% perf-profile.self.cycles-pp.secure_tcp_ts_off 0.44 ± 4% +0.1 0.58 ± 8% perf-profile.self.cycles-pp.tcp_v4_rcv 0.00 +0.1 0.15 ± 9% perf-profile.self.cycles-pp.__sk_free 0.06 ± 8% +0.2 0.21 ± 8% perf-profile.self.cycles-pp.tcp_set_state 0.08 ± 6% +0.2 0.23 ± 6% perf-profile.self.cycles-pp.tcp_assign_congestion_control 0.02 ±146% +0.2 0.21 ± 7% perf-profile.self.cycles-pp.tcp_init_sock 0.10 ± 7% +0.2 0.30 ± 7% perf-profile.self.cycles-pp.ip_rcv_finish_core 0.16 ± 7% +0.2 0.37 ± 7% perf-profile.self.cycles-pp.tcp_schedule_loss_probe 0.07 ± 9% +0.2 0.28 ± 7% perf-profile.self.cycles-pp.fib_lookup_good_nhc 0.03 ± 83% +0.2 0.25 ± 6% perf-profile.self.cycles-pp.tcp_newly_delivered 0.10 ± 10% +0.3 0.37 ± 8% perf-profile.self.cycles-pp.__tcp_select_window 0.28 ± 6% +0.3 0.58 ± 8% perf-profile.self.cycles-pp.tcp_write_xmit 0.11 ± 6% +0.3 0.46 ± 8% perf-profile.self.cycles-pp.tcp_ack_update_rtt 0.14 ± 9% +0.3 0.49 ± 7% perf-profile.self.cycles-pp.inet_ehashfn 0.16 ± 8% +0.4 0.53 ± 7% perf-profile.self.cycles-pp.ip_rcv_core 0.75 ± 5% +0.4 1.17 ± 7% perf-profile.self.cycles-pp.__tcp_transmit_skb 0.08 ± 8% +0.5 0.56 ± 5% perf-profile.self.cycles-pp.ip_protocol_deliver_rcu 0.24 ± 5% +0.5 0.77 ± 6% perf-profile.self.cycles-pp.ip_route_output_key_hash_rcu 0.23 ± 5% +0.6 0.84 ± 7% perf-profile.self.cycles-pp.ip_output 0.24 ± 6% +0.9 1.17 ± 7% perf-profile.self.cycles-pp.ipv4_dst_check [3] b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json: "netperf.Throughput_tps": [ b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30212.148125000003, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30112.125625, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30178.713125000002, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30206.963125000002, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30208.091875000002, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30139.37625, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30163.055625, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30238.72875, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30188.946874999994, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30239.25625, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30241.643125, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30208.288125, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30232.425, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30157.838125, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30220.53375, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30256.23625, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30214.021250000005, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30203.740624999995, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30209.22375, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 30245.735624999998 b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- ], b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json: "netperf.Throughput_total_tps": [ b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483394.37000000005, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 481794.01, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 482859.41000000003, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483311.41000000003, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483329.47000000003, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 482230.02, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 482608.89, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483819.66, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483023.1499999999, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483828.1, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483866.29, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483332.61, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483718.8, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 482525.41, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483528.54, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 484099.78, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483424.3400000001, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483259.8499999999, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483347.58, b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- 483931.76999999996 b89fbfbb854c9afc3047e8273cc3a694650b802e/matrix.json- ], [4] ========================================================================================= cluster/compiler/cpufreq_governor/ip/kconfig/nr_threads/rootfs/runtime/tbox_group/test/testcase/ucode: cs-localhost/gcc-9/performance/ipv4/x86_64-rhel-8.3/16/debian-10.4-x86_64-20200603.cgz/300s/lkp-icl-2sp2/TCP_CRR/netperf/0xd000280 commit: 652c1b17b8 ("bpf: Refactor perf_event_set_bpf_prog() to use struct bpf_prog input") b89fbfbb85 ("bpf: Implement minimal BPF perf link") 652c1b17b85b9c19 b89fbfbb854c9afc3047e8273cc ---------------- --------------------------- %stddev %change %stddev \ | \ 522472 -5.3% 494763 netperf.Throughput_total_tps 32654 -5.3% 30922 netperf.Throughput_tps 305.01 -0.0% 305.01 netperf.time.elapsed_time 305.01 -0.0% 305.01 netperf.time.elapsed_time.max 514102 +0.3% 515832 netperf.time.involuntary_context_switches 0.25 ±214% -20.0% 0.20 ±200% netperf.time.major_page_faults 2497 -0.0% 2497 netperf.time.maximum_resident_set_size 7473 ± 5% -0.3% 7451 ± 5% netperf.time.minor_page_faults 4096 +0.0% 4096 netperf.time.page_size 940.65 -1.0% 931.30 netperf.time.percent_of_cpu_this_job_got 2762 -0.7% 2743 netperf.time.system_time 107.63 -8.0% 99.06 netperf.time.user_time 1.601e+08 -1.5% 1.577e+08 netperf.time.voluntary_context_switches 1.567e+08 -5.3% 1.484e+08 netperf.workload 305.01 -0.0% 305.01 time.elapsed_time 305.01 -0.0% 305.01 time.elapsed_time.max 514102 +0.3% 515832 time.involuntary_context_switches 0.25 ±214% -20.0% 0.20 ±200% time.major_page_faults 2497 -0.0% 2497 time.maximum_resident_set_size 7473 ± 5% -0.3% 7451 ± 5% time.minor_page_faults 4096 +0.0% 4096 time.page_size 940.65 -1.0% 931.30 time.percent_of_cpu_this_job_got 2762 -0.7% 2743 time.system_time 107.63 -8.0% 99.06 time.user_time 1.601e+08 -1.5% 1.577e+08 time.voluntary_context_switches > > > > > > > in testcase: netperf > > on test machine: 192 threads 4 sockets Intel(R) Xeon(R) Platinum 9242 > > CPU @ 2.30GHz with 192G memory with following parameters: > > > > ip: ipv4 > > runtime: 300s > > nr_threads: 16 > > cluster: cs-localhost > > test: TCP_CRR > > cpufreq_governor: performance > > ucode: 0x5003006 > > > > test-description: Netperf is a benchmark that can be use to measure various > aspect of networking performance. > > test-url: http://www.netperf.org/netperf/ > > > > > > > [...] > _______________________________________________ > LKP mailing list -- lkp@xxxxxxxxxxxx > To unsubscribe send an email to lkp-leave@xxxxxxxxxxxx