* [bytedance:5.15-velinux] [mm/page_alloc] 70a0ba8eec: netperf.Throughput_Mbps 157.8% improvement
@ 2025-10-24 5:38 kernel test robot
0 siblings, 0 replies; only message in thread
From: kernel test robot @ 2025-10-24 5:38 UTC (permalink / raw)
To: jason.zeng; +Cc: oe-lkp, lkp, oliver.sang
Hello,
kernel test robot noticed a 157.8% improvement of netperf.Throughput_Mbps on:
commit: 70a0ba8eec7fecbbdf4d0aab45233a10ec83ab31 ("mm/page_alloc: limit number of high-order pages on PCP during bulk free")
https://github.com/bytedance/kernel.git 5.15-velinux
testcase: netperf
config: x86_64-bytedance-5.15-velinux
compiler: gcc-12
test machine: 256 threads 2 sockets Intel(R) Xeon(R) 6767P CPU @ 2.4GHz (Granite Rapids) with 256G memory
parameters:
ip: ipv4
runtime: 300s
nr_threads: 50%
cluster: cs-localhost
test: UDP_STREAM
cpufreq_governor: performance
Details are as below:
-------------------------------------------------------------------------------------------------->
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20251024/202510241319.250f4fef-lkp@intel.com
=========================================================================================
cluster/compiler/cpufreq_governor/ip/kconfig/nr_threads/rootfs/runtime/tbox_group/test/testcase:
cs-localhost/gcc-12/performance/ipv4/x86_64-bytedance-5.15-velinux/50%/debian-13-x86_64-20250902.cgz/300s/lkp-gnr-2sp3/UDP_STREAM/netperf
commit:
55140b0d44 ("mm/page_alloc: free pages in a single pass during bulk free")
70a0ba8eec ("mm/page_alloc: limit number of high-order pages on PCP during bulk free")
55140b0d4411f37e 70a0ba8eec7fecbbdf4d0aab452
---------------- ---------------------------
%stddev %change %stddev
\ | \
86953 -21.7% 68097 uptime.idle
2.148e+10 ± 3% -88.9% 2.383e+09 ± 2% cpuidle..time
3.947e+08 ± 3% -99.8% 927032 ± 8% cpuidle..usage
840.50 ± 13% -72.2% 233.83 ± 13% perf-c2c.DRAM.local
132713 -73.1% 35759 ± 11% perf-c2c.HITM.local
2944 ± 2% +92.3% 5663 ± 16% perf-c2c.HITM.remote
135657 -69.5% 41423 ± 8% perf-c2c.HITM.total
27.73 ± 2% -24.8 2.90 ± 7% mpstat.cpu.all.idle%
15.95 ± 8% -14.7 1.26 ± 4% mpstat.cpu.all.soft%
55.64 +39.9 95.49 mpstat.cpu.all.sys%
0.68 -0.3 0.35 ± 5% mpstat.cpu.all.usr%
264.67 ± 6% -51.6% 128.17 ± 60% mpstat.max_utilization.seconds
75.80 +31.5% 99.71 mpstat.max_utilization_pct
2640 +32.0% 3484 turbostat.Avg_MHz
73.34 +23.4 96.79 turbostat.Busy%
27.42 ± 2% -24.4 2.98 ± 2% turbostat.C1%
0.12 ± 4% -14.3% 0.10 turbostat.IPC
4.817e+08 ± 2% -78.7% 1.025e+08 turbostat.IRQ
11472272 +10.7% 12705147 turbostat.NMI
0.02 ± 17% -0.0 0.00 turbostat.POLL%
645.37 -7.6% 596.50 turbostat.PkgWatt
21473 +41.6% 30408 ± 2% netperf.ThroughputBoth_Mbps
2748559 +41.6% 3892308 ± 2% netperf.ThroughputBoth_total_Mbps
10543 -78.8% 2230 ± 4% netperf.ThroughputRecv_Mbps
1349560 -78.8% 285540 ± 4% netperf.ThroughputRecv_total_Mbps
10929 ± 3% +157.8% 28177 ± 2% netperf.Throughput_Mbps
1398999 ± 3% +157.8% 3606768 ± 2% netperf.Throughput_total_Mbps
114930 ± 19% +19082.1% 22046105 ± 19% netperf.time.involuntary_context_switches
88.44 ± 3% +81.4% 160.43 ± 3% netperf.time.user_time
1.573e+09 +41.6% 2.228e+09 ± 2% netperf.workload
0.75 ± 5% +669.2% 5.78 ± 53% perf-sched.total_wait_and_delay.average.ms
3961911 -85.2% 587653 ± 65% perf-sched.total_wait_and_delay.count.ms
7677 ± 14% -48.0% 3989 ± 13% perf-sched.total_wait_and_delay.max.ms
0.66 ± 4% +763.7% 5.74 ± 52% perf-sched.total_wait_time.average.ms
4976 -24.1% 3776 ± 7% perf-sched.total_wait_time.max.ms
0.75 ± 5% +669.2% 5.78 ± 53% perf-sched.wait_and_delay.avg.ms.[unknown].[unknown].[unknown].[unknown].[unknown]
3961911 -85.2% 587653 ± 65% perf-sched.wait_and_delay.count.[unknown].[unknown].[unknown].[unknown].[unknown]
7677 ± 14% -48.0% 3989 ± 13% perf-sched.wait_and_delay.max.ms.[unknown].[unknown].[unknown].[unknown].[unknown]
0.66 ± 4% +763.7% 5.74 ± 52% perf-sched.wait_time.avg.ms.[unknown].[unknown].[unknown].[unknown].[unknown]
4976 -24.1% 3776 ± 7% perf-sched.wait_time.max.ms.[unknown].[unknown].[unknown].[unknown].[unknown]
1013810 ± 2% +72.7% 1751139 ± 12% meminfo.Active
1013235 ± 2% +72.8% 1750563 ± 12% meminfo.Active(anon)
965784 -16.9% 802110 meminfo.AnonPages
9128042 -40.8% 5407071 ± 4% meminfo.Cached
6772256 -57.5% 2880600 ± 8% meminfo.Committed_AS
5551502 -83.4% 919631 ± 3% meminfo.Inactive
5551354 -83.4% 919483 ± 3% meminfo.Inactive(anon)
3549531 -89.1% 386805 ± 12% meminfo.Mapped
13247319 -35.6% 8534127 ± 3% meminfo.Memused
22057 -29.8% 15491 meminfo.PageTables
5602172 -66.4% 1881201 ± 12% meminfo.Shmem
13376418 -33.9% 8846841 ± 2% meminfo.max_used_kB
931.16 ± 23% +4921.5% 46758 ±146% numa-vmstat.node0.nr_active_anon
124950 -72.8% 34006 ± 14% numa-vmstat.node0.nr_mapped
1449 ± 19% +3512.1% 52351 ±144% numa-vmstat.node0.nr_shmem
931.16 ± 23% +4921.4% 46757 ±146% numa-vmstat.node0.nr_zone_active_anon
252137 ± 2% +55.1% 391082 ± 10% numa-vmstat.node1.nr_active_anon
1412738 ± 2% -67.5% 458528 ± 14% numa-vmstat.node1.nr_file_pages
1234368 ± 4% -89.9% 125143 ± 61% numa-vmstat.node1.nr_inactive_anon
764116 -91.8% 62578 ± 18% numa-vmstat.node1.nr_mapped
3359 ± 8% -41.6% 1960 ± 14% numa-vmstat.node1.nr_page_table_pages
1399125 -70.1% 418018 ± 10% numa-vmstat.node1.nr_shmem
252137 ± 2% +55.1% 391081 ± 10% numa-vmstat.node1.nr_zone_active_anon
1234367 ± 4% -89.9% 125143 ± 61% numa-vmstat.node1.nr_zone_inactive_anon
4204 ± 23% +4347.7% 186988 ±146% numa-meminfo.node0.Active
3724 ± 23% +4913.0% 186700 ±146% numa-meminfo.node0.Active(anon)
498738 -72.7% 136034 ± 14% numa-meminfo.node0.Mapped
5798 ± 19% +3511.2% 209412 ±144% numa-meminfo.node0.Shmem
1009339 ± 2% +55.0% 1564631 ± 10% numa-meminfo.node1.Active
1009243 ± 2% +55.0% 1564344 ± 10% numa-meminfo.node1.Active(anon)
5650898 ± 2% -67.5% 1834433 ± 14% numa-meminfo.node1.FilePages
4936718 ± 4% -89.9% 500899 ± 61% numa-meminfo.node1.Inactive
4936693 ± 4% -89.9% 500825 ± 61% numa-meminfo.node1.Inactive(anon)
3053490 -91.7% 252553 ± 17% numa-meminfo.node1.Mapped
7486356 ± 4% -55.7% 3316775 ± 16% numa-meminfo.node1.MemUsed
13431 ± 8% -41.6% 7845 ± 14% numa-meminfo.node1.PageTables
5596446 -70.1% 1672391 ± 10% numa-meminfo.node1.Shmem
253460 ± 2% +72.7% 437631 ± 12% proc-vmstat.nr_active_anon
241472 -16.9% 200542 proc-vmstat.nr_anon_pages
6120877 +1.9% 6238602 proc-vmstat.nr_dirty_background_threshold
12256721 +1.9% 12492458 proc-vmstat.nr_dirty_threshold
2282483 -40.8% 1351838 ± 4% proc-vmstat.nr_file_pages
61581025 +1.9% 62759954 proc-vmstat.nr_free_pages
1388160 -83.4% 229959 ± 3% proc-vmstat.nr_inactive_anon
889178 -89.0% 97383 ± 12% proc-vmstat.nr_mapped
5523 -29.9% 3872 proc-vmstat.nr_page_table_pages
1401014 -66.4% 470370 ± 12% proc-vmstat.nr_shmem
49237 -3.8% 47374 proc-vmstat.nr_slab_reclaimable
125948 -1.3% 124255 proc-vmstat.nr_slab_unreclaimable
253460 ± 2% +72.7% 437631 ± 12% proc-vmstat.nr_zone_active_anon
1388160 -83.4% 229959 ± 3% proc-vmstat.nr_zone_inactive_anon
1.576e+09 +40.2% 2.21e+09 ± 2% proc-vmstat.numa_hit
1.576e+09 +40.2% 2.21e+09 ± 2% proc-vmstat.numa_local
162736 ± 14% +85.7% 302272 ± 18% proc-vmstat.pgactivate
1.573e+09 +40.4% 2.208e+09 ± 2% proc-vmstat.pgalloc_normal
1585512 -16.9% 1318078 proc-vmstat.pgfault
1.57e+09 +40.6% 2.208e+09 ± 2% proc-vmstat.pgfree
38581 +16.9% 45092 proc-vmstat.pgreuse
1.36 ± 4% -20.2% 1.08 ± 14% perf-stat.i.MPKI
1.716e+10 ± 2% +7.4% 1.843e+10 ± 2% perf-stat.i.branch-instructions
0.31 -0.2 0.09 perf-stat.i.branch-miss-rate%
53096999 -72.6% 14544729 ± 2% perf-stat.i.branch-misses
1.83 ± 6% +6.9 8.69 ± 22% perf-stat.i.cache-miss-rate%
6.645e+09 -81.8% 1.207e+09 ± 8% perf-stat.i.cache-references
2618496 ± 3% -94.5% 143522 ± 18% perf-stat.i.context-switches
8.58 ± 2% +14.4% 9.82 ± 2% perf-stat.i.cpi
6.869e+11 +30.1% 8.935e+11 perf-stat.i.cpu-cycles
46520 ± 26% -99.5% 255.07 perf-stat.i.cpu-migrations
6419 ± 3% +67.7% 10764 ± 16% perf-stat.i.cycles-between-cache-misses
8e+10 ± 2% +12.5% 8.997e+10 ± 2% perf-stat.i.instructions
0.12 ± 2% -9.8% 0.11 ± 2% perf-stat.i.ipc
10.23 ± 3% -98.6% 0.14 ± 58% perf-stat.i.metric.K/sec
4878 -18.2% 3988 perf-stat.i.minor-faults
4881 -18.2% 3991 perf-stat.i.page-faults
1.34 ± 4% -19.4% 1.08 ± 14% perf-stat.overall.MPKI
0.31 -0.2 0.08 perf-stat.overall.branch-miss-rate%
1.61 ± 2% +6.6 8.22 ± 22% perf-stat.overall.cache-miss-rate%
8.57 ± 2% +15.7% 9.91 ± 2% perf-stat.overall.cpi
6420 ± 3% +46.8% 9428 ± 16% perf-stat.overall.cycles-between-cache-misses
0.12 ± 2% -13.6% 0.10 ± 2% perf-stat.overall.ipc
15502 -20.0% 12401 perf-stat.overall.path-length
1.717e+10 ± 2% +7.6% 1.848e+10 ± 2% perf-stat.ps.branch-instructions
53004191 -72.4% 14638673 ± 2% perf-stat.ps.branch-misses
6.631e+09 -81.9% 1.203e+09 ± 8% perf-stat.ps.cache-references
2614035 ± 3% -94.4% 146411 ± 18% perf-stat.ps.context-switches
6.853e+11 +30.4% 8.935e+11 perf-stat.ps.cpu-cycles
46475 ± 25% -99.5% 247.55 perf-stat.ps.cpu-migrations
8.003e+10 +12.7% 9.022e+10 ± 2% perf-stat.ps.instructions
4755 -18.9% 3859 perf-stat.ps.minor-faults
4758 -18.8% 3862 perf-stat.ps.page-faults
2.439e+13 +13.3% 2.763e+13 ± 2% perf-stat.total.instructions
228146 ± 27% -96.7% 7597 ±141% sched_debug.cfs_rq:/.MIN_vruntime.avg
22331084 ± 10% -91.3% 1944137 ±141% sched_debug.cfs_rq:/.MIN_vruntime.max
2069150 ± 17% -94.1% 121288 ±141% sched_debug.cfs_rq:/.MIN_vruntime.stddev
0.62 +34.6% 0.84 sched_debug.cfs_rq:/.h_nr_running.avg
0.40 -79.2% 0.08 ± 9% sched_debug.cfs_rq:/.h_nr_running.stddev
228146 ± 27% -96.7% 7597 ±141% sched_debug.cfs_rq:/.max_vruntime.avg
22331084 ± 10% -91.3% 1944137 ±141% sched_debug.cfs_rq:/.max_vruntime.max
2069150 ± 17% -94.1% 121288 ±141% sched_debug.cfs_rq:/.max_vruntime.stddev
21297524 ± 7% +83.4% 39059246 sched_debug.cfs_rq:/.min_vruntime.avg
29255285 ± 7% +35.0% 39484619 sched_debug.cfs_rq:/.min_vruntime.max
12149520 ± 8% +90.3% 23117001 ± 11% sched_debug.cfs_rq:/.min_vruntime.min
3965031 ± 13% -60.4% 1568773 ± 11% sched_debug.cfs_rq:/.min_vruntime.stddev
0.62 +34.7% 0.84 sched_debug.cfs_rq:/.nr_running.avg
0.40 -82.8% 0.07 ± 9% sched_debug.cfs_rq:/.nr_running.stddev
642.64 +35.1% 868.05 sched_debug.cfs_rq:/.runnable_avg.avg
12.72 ±196% +1646.5% 222.19 ± 33% sched_debug.cfs_rq:/.runnable_avg.min
272.63 -72.3% 75.64 ± 6% sched_debug.cfs_rq:/.runnable_avg.stddev
1510202 ± 86% -109.2% -139483 sched_debug.cfs_rq:/.spread0.avg
9468217 ± 17% -96.9% 288852 ±192% sched_debug.cfs_rq:/.spread0.max
-7640223 +110.5% -16085770 sched_debug.cfs_rq:/.spread0.min
3965570 ± 13% -60.4% 1568865 ± 11% sched_debug.cfs_rq:/.spread0.stddev
641.82 +34.9% 865.99 sched_debug.cfs_rq:/.util_avg.avg
1107 ± 5% -9.7% 999.89 sched_debug.cfs_rq:/.util_avg.max
11.22 ±215% +1672.3% 198.89 ± 33% sched_debug.cfs_rq:/.util_avg.min
271.63 -72.1% 75.66 ± 7% sched_debug.cfs_rq:/.util_avg.stddev
282.84 ± 2% +83.2% 518.08 ± 3% sched_debug.cfs_rq:/.util_est_enqueued.avg
196.82 +10.1% 216.79 ± 2% sched_debug.cfs_rq:/.util_est_enqueued.stddev
538212 +42.0% 764052 ± 2% sched_debug.cpu.avg_idle.avg
424676 -32.0% 288955 ± 3% sched_debug.cpu.avg_idle.stddev
6049 +36.3% 8247 sched_debug.cpu.curr->pid.avg
3855 -80.9% 736.09 ± 5% sched_debug.cpu.curr->pid.stddev
0.00 ± 49% +117.2% 0.00 ± 50% sched_debug.cpu.next_balance.stddev
0.62 +35.0% 0.84 sched_debug.cpu.nr_running.avg
0.41 -78.0% 0.09 ± 9% sched_debug.cpu.nr_running.stddev
1506613 ± 10% -94.2% 88111 ± 18% sched_debug.cpu.nr_switches.avg
726658 ± 14% -52.6% 344629 ± 24% sched_debug.cpu.nr_switches.stddev
46.92 ± 4% -33.6 13.30 ± 2% perf-profile.calltrace.cycles-pp.__ip_append_data.ip_make_skb.udp_sendmsg.__sock_sendmsg.__sys_sendto
40.10 ± 4% -32.5 7.55 perf-profile.calltrace.cycles-pp.sk_page_frag_refill.__ip_append_data.ip_make_skb.udp_sendmsg.__sock_sendmsg
40.08 ± 4% -32.5 7.53 perf-profile.calltrace.cycles-pp.skb_page_frag_refill.sk_page_frag_refill.__ip_append_data.ip_make_skb.udp_sendmsg
39.59 ± 4% -32.3 7.31 perf-profile.calltrace.cycles-pp.__alloc_pages.skb_page_frag_refill.sk_page_frag_refill.__ip_append_data.ip_make_skb
39.52 ± 4% -32.3 7.26 perf-profile.calltrace.cycles-pp.get_page_from_freelist.__alloc_pages.skb_page_frag_refill.sk_page_frag_refill.__ip_append_data
39.09 ± 4% -32.1 6.98 perf-profile.calltrace.cycles-pp.rmqueue.get_page_from_freelist.__alloc_pages.skb_page_frag_refill.sk_page_frag_refill
38.76 ± 4% -32.1 6.70 perf-profile.calltrace.cycles-pp.rmqueue_bulk.rmqueue.get_page_from_freelist.__alloc_pages.skb_page_frag_refill
38.32 ± 5% -31.7 6.64 perf-profile.calltrace.cycles-pp._raw_spin_lock.rmqueue_bulk.rmqueue.get_page_from_freelist.__alloc_pages
38.23 ± 5% -31.6 6.63 perf-profile.calltrace.cycles-pp.native_queued_spin_lock_slowpath._raw_spin_lock.rmqueue_bulk.rmqueue.get_page_from_freelist
18.60 ± 7% -18.6 0.00 perf-profile.calltrace.cycles-pp.__udp_enqueue_schedule_skb.udp_queue_rcv_one_skb.udp_unicast_rcv_skb.__udp4_lib_rcv.ip_protocol_deliver_rcu
18.72 ± 7% -17.8 0.92 ± 6% perf-profile.calltrace.cycles-pp.udp_unicast_rcv_skb.__udp4_lib_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish.__netif_receive_skb_one_core
18.70 ± 7% -17.8 0.90 ± 6% perf-profile.calltrace.cycles-pp.udp_queue_rcv_one_skb.udp_unicast_rcv_skb.__udp4_lib_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish
18.78 ± 7% -17.8 1.03 ± 5% perf-profile.calltrace.cycles-pp.__udp4_lib_rcv.ip_protocol_deliver_rcu.ip_local_deliver_finish.__netif_receive_skb_one_core.process_backlog
18.80 ± 7% -17.8 1.04 ± 6% perf-profile.calltrace.cycles-pp.ip_protocol_deliver_rcu.ip_local_deliver_finish.__netif_receive_skb_one_core.process_backlog.__napi_poll
18.80 ± 7% -17.8 1.05 ± 6% perf-profile.calltrace.cycles-pp.ip_local_deliver_finish.__netif_receive_skb_one_core.process_backlog.__napi_poll.net_rx_action
18.87 ± 7% -17.7 1.16 ± 5% perf-profile.calltrace.cycles-pp.__netif_receive_skb_one_core.process_backlog.__napi_poll.net_rx_action.__softirqentry_text_start
18.91 ± 7% -17.7 1.23 ± 5% perf-profile.calltrace.cycles-pp.process_backlog.__napi_poll.net_rx_action.__softirqentry_text_start.do_softirq
18.92 ± 7% -17.7 1.24 ± 5% perf-profile.calltrace.cycles-pp.__napi_poll.net_rx_action.__softirqentry_text_start.do_softirq.__local_bh_enable_ip
18.93 ± 7% -17.7 1.27 ± 5% perf-profile.calltrace.cycles-pp.net_rx_action.__softirqentry_text_start.do_softirq.__local_bh_enable_ip.ip_finish_output2
18.95 ± 7% -17.6 1.30 ± 5% perf-profile.calltrace.cycles-pp.__softirqentry_text_start.do_softirq.__local_bh_enable_ip.ip_finish_output2.ip_send_skb
18.97 ± 7% -17.6 1.32 ± 5% perf-profile.calltrace.cycles-pp.__local_bh_enable_ip.ip_finish_output2.ip_send_skb.udp_send_skb.udp_sendmsg
18.97 ± 7% -17.6 1.32 ± 5% perf-profile.calltrace.cycles-pp.do_softirq.__local_bh_enable_ip.ip_finish_output2.ip_send_skb.udp_send_skb
66.78 -17.4 49.36 perf-profile.calltrace.cycles-pp.udp_sendmsg.__sock_sendmsg.__sys_sendto.__x64_sys_sendto.do_syscall_64
66.80 -17.4 49.40 perf-profile.calltrace.cycles-pp.__sock_sendmsg.__sys_sendto.__x64_sys_sendto.do_syscall_64.entry_SYSCALL_64_after_hwframe
19.14 ± 7% -17.4 1.78 ± 7% perf-profile.calltrace.cycles-pp.ip_finish_output2.ip_send_skb.udp_send_skb.udp_sendmsg.__sock_sendmsg
66.91 -17.3 49.58 perf-profile.calltrace.cycles-pp.__sys_sendto.__x64_sys_sendto.do_syscall_64.entry_SYSCALL_64_after_hwframe.send_omni_inner
66.91 -17.3 49.59 perf-profile.calltrace.cycles-pp.__x64_sys_sendto.do_syscall_64.entry_SYSCALL_64_after_hwframe.send_omni_inner.send_udp_stream
66.95 -17.3 49.68 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.send_omni_inner.send_udp_stream.main
66.93 -17.3 49.66 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.send_omni_inner.send_udp_stream.main
67.06 -17.2 49.84 perf-profile.calltrace.cycles-pp.send_omni_inner.send_udp_stream.main
67.06 -17.2 49.84 perf-profile.calltrace.cycles-pp.send_udp_stream.main
19.20 ± 7% -17.1 2.05 ± 8% perf-profile.calltrace.cycles-pp.ip_send_skb.udp_send_skb.udp_sendmsg.__sock_sendmsg.__sys_sendto
19.22 ± 7% -17.0 2.19 ± 10% perf-profile.calltrace.cycles-pp.udp_send_skb.udp_sendmsg.__sock_sendmsg.__sys_sendto.__x64_sys_sendto
14.66 ± 7% -14.7 0.00 perf-profile.calltrace.cycles-pp.__sk_mem_raise_allocated.__udp_enqueue_schedule_skb.udp_queue_rcv_one_skb.udp_unicast_rcv_skb.__udp4_lib_rcv
14.29 ± 7% -14.3 0.00 perf-profile.calltrace.cycles-pp.mem_cgroup_charge_skmem.__sk_mem_raise_allocated.__udp_enqueue_schedule_skb.udp_queue_rcv_one_skb.udp_unicast_rcv_skb
13.68 ± 7% -13.7 0.00 perf-profile.calltrace.cycles-pp.try_charge_memcg.mem_cgroup_charge_skmem.__sk_mem_raise_allocated.__udp_enqueue_schedule_skb.udp_queue_rcv_one_skb
11.15 ± 8% -11.1 0.00 perf-profile.calltrace.cycles-pp.__skb_recv_udp.udp_recvmsg.inet_recvmsg.__sys_recvfrom.__x64_sys_recvfrom
8.46 ± 9% -8.5 0.00 perf-profile.calltrace.cycles-pp.udp_rmem_release.__skb_recv_udp.udp_recvmsg.inet_recvmsg.__sys_recvfrom
8.22 ± 7% -8.2 0.00 perf-profile.calltrace.cycles-pp.page_counter_try_charge.try_charge_memcg.mem_cgroup_charge_skmem.__sk_mem_raise_allocated.__udp_enqueue_schedule_skb
7.52 ± 7% -7.5 0.00 perf-profile.calltrace.cycles-pp.__sk_mem_reduce_allocated.udp_rmem_release.__skb_recv_udp.udp_recvmsg.inet_recvmsg
6.82 ± 8% -6.8 0.00 perf-profile.calltrace.cycles-pp.refill_stock.__sk_mem_reduce_allocated.udp_rmem_release.__skb_recv_udp.udp_recvmsg
6.74 ± 8% -6.7 0.00 perf-profile.calltrace.cycles-pp.drain_stock.refill_stock.__sk_mem_reduce_allocated.udp_rmem_release.__skb_recv_udp
6.68 ± 8% -6.7 0.00 perf-profile.calltrace.cycles-pp.page_counter_uncharge.drain_stock.refill_stock.__sk_mem_reduce_allocated.udp_rmem_release
7.52 -6.4 1.14 ± 6% perf-profile.calltrace.cycles-pp.skb_copy_datagram_iter.udp_recvmsg.inet_recvmsg.__sys_recvfrom.__x64_sys_recvfrom
7.51 -6.4 1.14 ± 6% perf-profile.calltrace.cycles-pp.__skb_datagram_iter.skb_copy_datagram_iter.udp_recvmsg.inet_recvmsg.__sys_recvfrom
6.80 -5.8 1.05 ± 6% perf-profile.calltrace.cycles-pp._copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter.udp_recvmsg.inet_recvmsg
6.70 -5.7 1.03 ± 6% perf-profile.calltrace.cycles-pp.copy_user_enhanced_fast_string._copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter.udp_recvmsg
5.59 ± 2% -0.7 4.94 ± 4% perf-profile.calltrace.cycles-pp._copy_from_iter.ip_generic_getfrag.__ip_append_data.ip_make_skb.udp_sendmsg
5.37 ± 2% -0.6 4.73 ± 4% perf-profile.calltrace.cycles-pp.copy_user_enhanced_fast_string._copy_from_iter.ip_generic_getfrag.__ip_append_data.ip_make_skb
5.69 -0.6 5.08 ± 4% perf-profile.calltrace.cycles-pp.ip_generic_getfrag.__ip_append_data.ip_make_skb.udp_sendmsg.__sock_sendmsg
96.28 +3.3 99.62 perf-profile.calltrace.cycles-pp.main
28.41 ± 2% +21.4 49.78 perf-profile.calltrace.cycles-pp.accept_connection.accept_connections.main
28.41 ± 2% +21.4 49.78 perf-profile.calltrace.cycles-pp.accept_connections.main
28.41 ± 2% +21.4 49.78 perf-profile.calltrace.cycles-pp.process_requests.spawn_child.accept_connection.accept_connections.main
28.41 ± 2% +21.4 49.78 perf-profile.calltrace.cycles-pp.spawn_child.accept_connection.accept_connections.main
28.41 ± 2% +21.4 49.78 perf-profile.calltrace.cycles-pp.recv_omni.process_requests.spawn_child.accept_connection.accept_connections
28.14 ± 2% +21.6 49.76 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.recv_omni.process_requests.spawn_child.accept_connection
28.12 ± 2% +21.6 49.76 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.recv_omni.process_requests.spawn_child
27.97 ± 2% +21.8 49.74 perf-profile.calltrace.cycles-pp.__x64_sys_recvfrom.do_syscall_64.entry_SYSCALL_64_after_hwframe.recv_omni.process_requests
27.96 ± 2% +21.8 49.74 perf-profile.calltrace.cycles-pp.__sys_recvfrom.__x64_sys_recvfrom.do_syscall_64.entry_SYSCALL_64_after_hwframe.recv_omni
27.71 ± 2% +22.0 49.71 perf-profile.calltrace.cycles-pp.inet_recvmsg.__sys_recvfrom.__x64_sys_recvfrom.do_syscall_64.entry_SYSCALL_64_after_hwframe
27.70 ± 2% +22.0 49.71 perf-profile.calltrace.cycles-pp.udp_recvmsg.inet_recvmsg.__sys_recvfrom.__x64_sys_recvfrom.do_syscall_64
0.00 +33.2 33.16 perf-profile.calltrace.cycles-pp.ip_idents_reserve.__ip_select_ident.__ip_make_skb.ip_make_skb.udp_sendmsg
0.00 +33.4 33.36 perf-profile.calltrace.cycles-pp.__ip_select_ident.__ip_make_skb.ip_make_skb.udp_sendmsg.__sock_sendmsg
0.00 +33.5 33.46 perf-profile.calltrace.cycles-pp.__ip_make_skb.ip_make_skb.udp_sendmsg.__sock_sendmsg.__sys_sendto
8.89 ± 4% +39.4 48.33 perf-profile.calltrace.cycles-pp.skb_release_data.__consume_stateless_skb.udp_recvmsg.inet_recvmsg.__sys_recvfrom
8.89 ± 4% +39.4 48.33 perf-profile.calltrace.cycles-pp.__consume_stateless_skb.udp_recvmsg.inet_recvmsg.__sys_recvfrom.__x64_sys_recvfrom
8.44 ± 4% +39.5 47.99 perf-profile.calltrace.cycles-pp.free_unref_page.skb_release_data.__consume_stateless_skb.udp_recvmsg.inet_recvmsg
7.98 ± 5% +39.9 47.86 perf-profile.calltrace.cycles-pp.free_pcppages_bulk.free_unref_page.skb_release_data.__consume_stateless_skb.udp_recvmsg
7.64 ± 5% +40.1 47.71 perf-profile.calltrace.cycles-pp.native_queued_spin_lock_slowpath._raw_spin_lock.free_pcppages_bulk.free_unref_page.skb_release_data
7.65 ± 5% +40.1 47.79 perf-profile.calltrace.cycles-pp._raw_spin_lock.free_pcppages_bulk.free_unref_page.skb_release_data.__consume_stateless_skb
46.92 ± 4% -33.6 13.31 ± 2% perf-profile.children.cycles-pp.__ip_append_data
40.10 ± 4% -32.5 7.55 perf-profile.children.cycles-pp.sk_page_frag_refill
40.08 ± 4% -32.5 7.54 perf-profile.children.cycles-pp.skb_page_frag_refill
39.63 ± 4% -32.3 7.37 perf-profile.children.cycles-pp.__alloc_pages
39.56 ± 4% -32.2 7.32 perf-profile.children.cycles-pp.get_page_from_freelist
39.14 ± 4% -32.1 7.03 perf-profile.children.cycles-pp.rmqueue
38.79 ± 4% -32.0 6.75 perf-profile.children.cycles-pp.rmqueue_bulk
18.61 ± 7% -18.4 0.17 ± 9% perf-profile.children.cycles-pp.__udp_enqueue_schedule_skb
18.72 ± 7% -17.8 0.92 ± 6% perf-profile.children.cycles-pp.udp_unicast_rcv_skb
18.70 ± 7% -17.8 0.91 ± 6% perf-profile.children.cycles-pp.udp_queue_rcv_one_skb
18.79 ± 7% -17.8 1.03 ± 6% perf-profile.children.cycles-pp.__udp4_lib_rcv
18.80 ± 7% -17.8 1.05 ± 5% perf-profile.children.cycles-pp.ip_protocol_deliver_rcu
18.80 ± 7% -17.8 1.05 ± 6% perf-profile.children.cycles-pp.ip_local_deliver_finish
19.03 ± 7% -17.7 1.31 ± 5% perf-profile.children.cycles-pp.__softirqentry_text_start
18.87 ± 7% -17.7 1.16 ± 5% perf-profile.children.cycles-pp.__netif_receive_skb_one_core
18.92 ± 7% -17.7 1.24 ± 5% perf-profile.children.cycles-pp.__napi_poll
18.92 ± 7% -17.7 1.24 ± 5% perf-profile.children.cycles-pp.process_backlog
18.93 ± 7% -17.7 1.27 ± 5% perf-profile.children.cycles-pp.net_rx_action
18.98 ± 7% -17.6 1.33 ± 5% perf-profile.children.cycles-pp.__local_bh_enable_ip
18.97 ± 7% -17.6 1.32 ± 5% perf-profile.children.cycles-pp.do_softirq
66.79 -17.4 49.37 perf-profile.children.cycles-pp.udp_sendmsg
66.80 -17.4 49.40 perf-profile.children.cycles-pp.__sock_sendmsg
19.14 ± 7% -17.4 1.78 ± 7% perf-profile.children.cycles-pp.ip_finish_output2
66.91 -17.3 49.59 perf-profile.children.cycles-pp.__sys_sendto
66.91 -17.3 49.59 perf-profile.children.cycles-pp.__x64_sys_sendto
67.06 -17.2 49.84 perf-profile.children.cycles-pp.send_omni_inner
67.06 -17.2 49.84 perf-profile.children.cycles-pp.send_udp_stream
19.20 ± 7% -17.1 2.05 ± 8% perf-profile.children.cycles-pp.ip_send_skb
19.22 ± 7% -17.0 2.20 ± 10% perf-profile.children.cycles-pp.udp_send_skb
14.66 ± 7% -14.6 0.10 ± 10% perf-profile.children.cycles-pp.__sk_mem_raise_allocated
14.30 ± 7% -14.2 0.07 ± 10% perf-profile.children.cycles-pp.mem_cgroup_charge_skmem
13.70 ± 7% -13.7 0.04 ± 45% perf-profile.children.cycles-pp.try_charge_memcg
11.15 ± 8% -11.0 0.20 ± 21% perf-profile.children.cycles-pp.__skb_recv_udp
8.46 ± 9% -8.4 0.10 ± 7% perf-profile.children.cycles-pp.udp_rmem_release
8.24 ± 7% -8.2 0.00 perf-profile.children.cycles-pp.page_counter_try_charge
7.52 ± 7% -7.4 0.09 ± 5% perf-profile.children.cycles-pp.__sk_mem_reduce_allocated
6.82 ± 8% -6.8 0.04 ± 45% perf-profile.children.cycles-pp.refill_stock
6.74 ± 8% -6.7 0.01 ±223% perf-profile.children.cycles-pp.drain_stock
6.68 ± 8% -6.7 0.00 perf-profile.children.cycles-pp.page_counter_uncharge
7.52 -6.4 1.14 ± 6% perf-profile.children.cycles-pp.skb_copy_datagram_iter
7.51 -6.4 1.14 ± 6% perf-profile.children.cycles-pp.__skb_datagram_iter
12.24 -6.3 5.92 ± 4% perf-profile.children.cycles-pp.copy_user_enhanced_fast_string
6.81 -5.8 1.05 ± 6% perf-profile.children.cycles-pp._copy_to_iter
1.12 ± 7% -1.1 0.04 ± 44% perf-profile.children.cycles-pp.__mod_memcg_state
0.88 -0.8 0.09 ± 61% perf-profile.children.cycles-pp.__schedule
0.82 -0.8 0.05 ±102% perf-profile.children.cycles-pp.__skb_wait_for_more_packets
0.88 ± 4% -0.7 0.13 ± 3% perf-profile.children.cycles-pp.sysvec_apic_timer_interrupt
0.82 ± 4% -0.7 0.13 ± 3% perf-profile.children.cycles-pp.__sysvec_apic_timer_interrupt
0.81 ± 4% -0.7 0.12 ± 3% perf-profile.children.cycles-pp.hrtimer_interrupt
0.76 ± 4% -0.7 0.09 perf-profile.children.cycles-pp.__hrtimer_run_queues
0.75 ± 4% -0.7 0.09 ± 4% perf-profile.children.cycles-pp.tick_sched_timer
5.60 ± 2% -0.7 4.94 ± 4% perf-profile.children.cycles-pp._copy_from_iter
0.71 -0.7 0.05 ±102% perf-profile.children.cycles-pp.schedule_timeout
0.73 ± 4% -0.7 0.08 perf-profile.children.cycles-pp.tick_sched_handle
0.73 ± 4% -0.7 0.08 perf-profile.children.cycles-pp.update_process_times
0.97 ± 4% -0.6 0.34 ± 4% perf-profile.children.cycles-pp.asm_sysvec_apic_timer_interrupt
5.70 ± 2% -0.6 5.09 ± 4% perf-profile.children.cycles-pp.ip_generic_getfrag
0.66 -0.6 0.09 ± 61% perf-profile.children.cycles-pp.schedule
0.54 -0.5 0.06 ± 11% perf-profile.children.cycles-pp.simple_copy_to_iter
0.81 ± 4% -0.5 0.32 ± 10% perf-profile.children.cycles-pp.handle_internal_command
0.81 ± 4% -0.5 0.32 ± 10% perf-profile.children.cycles-pp.run_builtin
0.80 ± 4% -0.5 0.32 ± 10% perf-profile.children.cycles-pp.cmd_record
0.63 -0.4 0.20 ± 3% perf-profile.children.cycles-pp.__check_object_size
0.28 -0.2 0.05 perf-profile.children.cycles-pp.__free_one_page
0.28 ± 4% -0.1 0.14 ± 12% perf-profile.children.cycles-pp.__zone_watermark_ok
0.38 -0.1 0.25 ± 5% perf-profile.children.cycles-pp.sock_alloc_send_pskb
0.12 ± 3% -0.1 0.02 ± 99% perf-profile.children.cycles-pp.__ksize
0.19 ± 2% -0.1 0.10 ± 3% perf-profile.children.cycles-pp.free_pcp_prepare
0.15 ± 2% -0.1 0.06 ± 49% perf-profile.children.cycles-pp.syscall_exit_to_user_mode
0.13 -0.1 0.04 ± 74% perf-profile.children.cycles-pp.exit_to_user_mode_prepare
0.26 -0.1 0.20 ± 6% perf-profile.children.cycles-pp.__alloc_skb
0.27 -0.1 0.20 ± 6% perf-profile.children.cycles-pp.alloc_skb_with_frags
0.12 -0.1 0.06 perf-profile.children.cycles-pp.__slab_free
0.12 ± 6% -0.1 0.06 perf-profile.children.cycles-pp.scheduler_tick
0.12 ± 3% -0.1 0.07 ± 7% perf-profile.children.cycles-pp.sockfd_lookup_light
0.11 ± 3% -0.0 0.06 perf-profile.children.cycles-pp.__fget_light
0.18 -0.0 0.15 ± 3% perf-profile.children.cycles-pp.kfree
0.09 -0.0 0.06 ± 6% perf-profile.children.cycles-pp.__virt_addr_valid
0.06 ± 6% -0.0 0.05 perf-profile.children.cycles-pp.syscall_return_via_sysret
0.05 +0.0 0.06 ± 7% perf-profile.children.cycles-pp.__udp4_lib_lookup
0.18 ± 11% +0.0 0.22 ± 9% perf-profile.children.cycles-pp.generic_perform_write
0.18 ± 12% +0.0 0.22 ± 9% perf-profile.children.cycles-pp.generic_file_write_iter
0.19 ± 12% +0.0 0.23 ± 10% perf-profile.children.cycles-pp.vfs_write
0.20 ± 11% +0.0 0.24 ± 10% perf-profile.children.cycles-pp.ksys_write
0.06 ± 6% +0.0 0.11 ± 4% perf-profile.children.cycles-pp.kmem_cache_free
0.05 ± 7% +0.1 0.10 ± 3% perf-profile.children.cycles-pp.move_addr_to_kernel
0.00 +0.1 0.05 perf-profile.children.cycles-pp.__mkroute_output
0.00 +0.1 0.05 perf-profile.children.cycles-pp.ip_finish_output
0.21 ± 15% +0.1 0.26 ± 11% perf-profile.children.cycles-pp.record__pushfn
0.00 +0.1 0.06 ± 9% perf-profile.children.cycles-pp.sk_filter_trim_cap
0.00 +0.1 0.06 perf-profile.children.cycles-pp.enqueue_to_backlog
0.00 +0.1 0.06 perf-profile.children.cycles-pp.ip_rcv
0.00 +0.1 0.06 ± 6% perf-profile.children.cycles-pp.netif_rx_internal
0.11 ± 6% +0.1 0.17 ± 5% perf-profile.children.cycles-pp.fib_table_lookup
0.00 +0.1 0.07 ± 7% perf-profile.children.cycles-pp.netif_rx
0.00 +0.1 0.07 ± 5% perf-profile.children.cycles-pp._copy_from_user
0.00 +0.1 0.08 perf-profile.children.cycles-pp.__cgroup_bpf_run_filter_skb
0.12 ± 4% +0.1 0.20 ± 4% perf-profile.children.cycles-pp.ip_route_output_key_hash_rcu
0.17 ± 4% +0.1 0.28 ± 5% perf-profile.children.cycles-pp.ip_route_output_flow
0.00 +0.1 0.12 ± 28% perf-profile.children.cycles-pp.udp4_hwcsum
0.18 ± 4% +0.1 0.33 ± 13% perf-profile.children.cycles-pp.free_unref_page_commit
0.00 +0.1 0.15 ± 29% perf-profile.children.cycles-pp.ip_send_check
0.00 +0.2 0.17 ± 25% perf-profile.children.cycles-pp.__ip_local_out
0.09 ± 8% +0.2 0.28 ± 22% perf-profile.children.cycles-pp.loopback_xmit
0.00 +0.2 0.20 ± 30% perf-profile.children.cycles-pp.sock_wfree
0.09 ± 4% +0.2 0.30 ± 22% perf-profile.children.cycles-pp.xmit_one
0.10 ± 6% +0.2 0.30 ± 20% perf-profile.children.cycles-pp.dev_hard_start_xmit
0.14 ± 2% +0.3 0.40 ± 16% perf-profile.children.cycles-pp.__dev_queue_xmit
0.00 +0.6 0.55 ± 8% perf-profile.children.cycles-pp.kfree_skb_reason
51.74 ± 3% +2.9 54.59 perf-profile.children.cycles-pp._raw_spin_lock
51.51 ± 3% +2.9 54.39 perf-profile.children.cycles-pp.native_queued_spin_lock_slowpath
96.28 +3.7 99.95 perf-profile.children.cycles-pp.main
95.33 +4.4 99.71 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
95.30 +4.4 99.70 perf-profile.children.cycles-pp.do_syscall_64
28.41 ± 2% +21.4 49.78 perf-profile.children.cycles-pp.accept_connection
28.41 ± 2% +21.4 49.78 perf-profile.children.cycles-pp.accept_connections
28.41 ± 2% +21.4 49.78 perf-profile.children.cycles-pp.process_requests
28.41 ± 2% +21.4 49.78 perf-profile.children.cycles-pp.recv_omni
28.41 ± 2% +21.4 49.78 perf-profile.children.cycles-pp.spawn_child
27.98 ± 2% +21.8 49.76 perf-profile.children.cycles-pp.__x64_sys_recvfrom
27.97 ± 2% +21.8 49.76 perf-profile.children.cycles-pp.__sys_recvfrom
27.71 ± 2% +22.0 49.72 perf-profile.children.cycles-pp.inet_recvmsg
27.70 ± 2% +22.0 49.72 perf-profile.children.cycles-pp.udp_recvmsg
0.36 ± 6% +33.0 33.32 perf-profile.children.cycles-pp.ip_idents_reserve
0.38 ± 5% +33.0 33.37 perf-profile.children.cycles-pp.__ip_select_ident
0.42 ± 5% +33.0 33.46 perf-profile.children.cycles-pp.__ip_make_skb
8.89 ± 4% +39.4 48.33 perf-profile.children.cycles-pp.__consume_stateless_skb
8.00 ± 5% +39.9 47.86 perf-profile.children.cycles-pp.free_pcppages_bulk
8.45 ± 4% +40.0 48.42 perf-profile.children.cycles-pp.free_unref_page
8.90 ± 4% +40.0 48.88 perf-profile.children.cycles-pp.skb_release_data
7.36 ± 6% -7.4 0.00 perf-profile.self.cycles-pp.page_counter_try_charge
12.06 -6.2 5.91 ± 4% perf-profile.self.cycles-pp.copy_user_enhanced_fast_string
5.58 ± 6% -5.6 0.00 perf-profile.self.cycles-pp.page_counter_uncharge
5.39 ± 7% -5.4 0.00 perf-profile.self.cycles-pp.try_charge_memcg
1.10 ± 7% -1.1 0.04 ± 44% perf-profile.self.cycles-pp.__mod_memcg_state
0.50 -0.4 0.11 ± 4% perf-profile.self.cycles-pp.__check_object_size
0.38 ± 2% -0.3 0.05 ± 8% perf-profile.self.cycles-pp.rmqueue_bulk
0.74 ± 2% -0.3 0.42 ± 6% perf-profile.self.cycles-pp.__ip_append_data
0.47 ± 2% -0.3 0.20 ± 8% perf-profile.self.cycles-pp.skb_page_frag_refill
0.28 ± 3% -0.1 0.14 ± 11% perf-profile.self.cycles-pp.__zone_watermark_ok
0.19 ± 3% -0.1 0.10 ± 3% perf-profile.self.cycles-pp.free_pcp_prepare
0.12 -0.1 0.06 perf-profile.self.cycles-pp.__slab_free
0.10 -0.0 0.05 ± 8% perf-profile.self.cycles-pp.__fget_light
0.22 -0.0 0.20 ± 5% perf-profile.self.cycles-pp._raw_spin_lock
0.06 ± 6% -0.0 0.05 perf-profile.self.cycles-pp.syscall_return_via_sysret
0.06 -0.0 0.05 perf-profile.self.cycles-pp.__alloc_pages
0.00 +0.1 0.05 perf-profile.self.cycles-pp.__cgroup_bpf_run_filter_skb
0.00 +0.1 0.05 ± 8% perf-profile.self.cycles-pp.kmem_cache_free
0.00 +0.1 0.06 ± 9% perf-profile.self.cycles-pp.kfree
0.23 ± 5% +0.1 0.29 ± 4% perf-profile.self.cycles-pp.skb_release_data
0.00 +0.1 0.06 ± 6% perf-profile.self.cycles-pp.free_unref_page
0.09 ± 6% +0.1 0.15 ± 5% perf-profile.self.cycles-pp.fib_table_lookup
0.00 +0.1 0.06 perf-profile.self.cycles-pp.send_omni_inner
0.01 ±223% +0.1 0.08 ± 6% perf-profile.self.cycles-pp.udp_sendmsg
0.00 +0.1 0.08 ± 6% perf-profile.self.cycles-pp.__ip_make_skb
0.00 +0.1 0.12 ± 29% perf-profile.self.cycles-pp.udp4_hwcsum
0.18 ± 4% +0.1 0.33 ± 13% perf-profile.self.cycles-pp.free_unref_page_commit
0.00 +0.1 0.15 ± 28% perf-profile.self.cycles-pp.ip_send_check
0.00 +0.2 0.18 ± 34% perf-profile.self.cycles-pp.sock_wfree
51.44 ± 3% +2.9 54.39 perf-profile.self.cycles-pp.native_queued_spin_lock_slowpath
0.35 ± 5% +32.9 33.27 perf-profile.self.cycles-pp.ip_idents_reserve
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
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2025-10-24 5:39 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-10-24 5:38 [bytedance:5.15-velinux] [mm/page_alloc] 70a0ba8eec: netperf.Throughput_Mbps 157.8% improvement kernel test robot
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.