| 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: <202508251419.d8117ed7-lkp@intel.com>
Date: Mon, 25 Aug 2025 15:11:32 +0800
From: kernel test robot <oliver.sang@...el.com>
To: "Rafael J. Wysocki" <rafael.j.wysocki@...el.com>
CC: <oe-lkp@...ts.linux.dev>, <lkp@...el.com>, <linux-kernel@...r.kernel.org>,
Christian Loehle <christian.loehle@....com>, <linux-pm@...r.kernel.org>,
<oliver.sang@...el.com>
Subject: [linus:master] [cpuidle] 779b1a1cb1:
perf-bench-sched-pipe.ops_per_sec 12.9% regression
Hello,
kernel test robot noticed a 12.9% regression of perf-bench-sched-pipe.ops_per_sec on:
commit: 779b1a1cb13ae17028aeddb2fbbdba97357a1e15 ("cpuidle: governors: menu: Avoid selecting states with too much latency")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
[still regression on linus/master 1b237f190eb3d36f52dffe07a40b5eb210280e00]
[still regression on linux-next/master 0f4c93f7eb861acab537dbe94441817a270537bf]
testcase: perf-bench-sched-pipe
config: x86_64-rhel-9.4
compiler: gcc-12
test machine: 192 threads 4 sockets Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz (Cascade Lake) with 176G memory
parameters:
loops: 10000000ops
mode: threads
cpufreq_governor: performance
In addition to that, the commit also has significant impact on the following tests:
+------------------+-----------------------------------------------------------------------------+
| testcase: change | qperf: qperf.udp.latency 8.1% regression |
| test machine | 8 threads Intel(R) Core(TM) i7-6700 CPU @ 3.40GHz (Skylake) with 16G memory |
| test parameters | cluster=cs-localhost |
| | cpufreq_governor=performance |
| | runtime=600s |
+------------------+-----------------------------------------------------------------------------+
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <oliver.sang@...el.com>
| Closes: https://lore.kernel.org/oe-lkp/202508251419.d8117ed7-lkp@intel.com
Details are as below:
-------------------------------------------------------------------------------------------------->
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20250825/202508251419.d8117ed7-lkp@intel.com
=========================================================================================
compiler/cpufreq_governor/kconfig/loops/mode/rootfs/tbox_group/testcase:
gcc-12/performance/x86_64-rhel-9.4/10000000ops/threads/debian-12-x86_64-20240206.cgz/lkp-csl-2sp10/perf-bench-sched-pipe
commit:
v6.17-rc2
779b1a1cb1 ("cpuidle: governors: menu: Avoid selecting states with too much latency")
v6.17-rc2 779b1a1cb13ae17028aeddb2fbb
---------------- ---------------------------
%stddev %change %stddev
\ | \
9.124e+09 +14.6% 1.045e+10 cpuidle..time
32.00 ± 23% +37.5% 44.00 ± 10% perf-c2c.DRAM.remote
0.01 ± 22% +55.3% 0.01 ± 18% perf-sched.sch_delay.avg.ms.schedule_timeout.__wait_for_common.wait_for_completion_state.kernel_clone
805462 -12.0% 708615 vmstat.system.cs
47.14 ± 4% +8.9% 51.35 ± 5% boot-time.boot
8181 ± 5% +9.4% 8955 ± 5% boot-time.idle
43305 ± 9% +17.8% 51027 ± 6% meminfo.AnonHugePages
284445 +20.6% 343101 meminfo.Shmem
0.50 -0.1 0.43 mpstat.cpu.all.sys%
0.45 ± 2% -0.1 0.40 ± 2% mpstat.cpu.all.usr%
70114 ± 40% -74.6% 17817 ±146% numa-vmstat.node3.nr_anon_pages
68448 ± 3% +22.0% 83539 numa-vmstat.node3.nr_shmem
96.76 ± 2% +11.5% 107.92 ± 2% uptime.boot
17588 ± 2% +12.0% 19695 ± 3% uptime.idle
280480 ± 40% -74.6% 71234 ±146% numa-meminfo.node3.AnonPages
323960 ± 34% -69.7% 98297 ±102% numa-meminfo.node3.AnonPages.max
49659 ± 55% -35.6% 32001 ± 3% numa-meminfo.node3.Mapped
273725 ± 3% +21.4% 332416 numa-meminfo.node3.Shmem
217049 -12.9% 189123 perf-bench-sched-pipe.ops_per_sec
46.09 +14.8% 52.90 perf-bench-sched-pipe.time.elapsed_time
46.09 +14.8% 52.90 perf-bench-sched-pipe.time.elapsed_time.max
91.83 -12.2% 80.67 perf-bench-sched-pipe.time.percent_of_cpu_this_job_got
243570 +5.6% 257111 proc-vmstat.nr_active_anon
956368 +1.5% 970579 proc-vmstat.nr_file_pages
28172 -5.2% 26711 proc-vmstat.nr_mapped
71358 +19.9% 85579 proc-vmstat.nr_shmem
243570 +5.6% 257111 proc-vmstat.nr_zone_active_anon
454916 +4.5% 475605 proc-vmstat.pgfault
49587 ± 4% +8.5% 53826 ± 4% sched_debug.cpu.clock.avg
49598 ± 4% +8.5% 53836 ± 4% sched_debug.cpu.clock.max
49568 ± 4% +8.6% 53812 ± 4% sched_debug.cpu.clock.min
49456 ± 4% +8.6% 53686 ± 4% sched_debug.cpu.clock_task.avg
49580 ± 4% +8.5% 53811 ± 4% sched_debug.cpu.clock_task.max
49577 ± 4% +8.5% 53816 ± 4% sched_debug.cpu_clk
48580 ± 4% +8.7% 52811 ± 4% sched_debug.ktime
50297 ± 4% +9.3% 54960 ± 5% sched_debug.sched_clk
2.211e+09 -15.9% 1.859e+09 ± 3% perf-stat.i.branch-instructions
60344331 ± 2% -13.6% 52146391 ± 2% perf-stat.i.branch-misses
4974491 ± 5% -15.9% 4185197 ± 7% perf-stat.i.cache-misses
80956268 ± 3% -12.0% 71234171 ± 7% perf-stat.i.cache-references
854948 -12.6% 747062 perf-stat.i.context-switches
1.19 +3.9% 1.24 ± 2% perf-stat.i.cpi
1.148e+10 ± 2% -13.9% 9.886e+09 perf-stat.i.cpu-cycles
1.031e+10 -14.7% 8.79e+09 ± 3% perf-stat.i.instructions
4.45 -12.7% 3.89 perf-stat.i.metric.K/sec
7605 ± 2% -7.4% 7042 ± 2% perf-stat.i.minor-faults
7605 ± 2% -7.4% 7042 ± 2% perf-stat.i.page-faults
2.73 +0.1 2.80 perf-stat.overall.branch-miss-rate%
2.162e+09 -15.7% 1.823e+09 ± 3% perf-stat.ps.branch-instructions
58974968 ± 2% -13.3% 51107382 ± 2% perf-stat.ps.branch-misses
4857873 ± 5% -15.6% 4100205 ± 7% perf-stat.ps.cache-misses
79195784 ± 3% -11.8% 69883377 ± 7% perf-stat.ps.cache-references
836709 -12.4% 733164 perf-stat.ps.context-switches
1.123e+10 ± 2% -13.7% 9.698e+09 perf-stat.ps.cpu-cycles
1.008e+10 -14.5% 8.62e+09 ± 3% perf-stat.ps.instructions
7359 ± 2% -7.1% 6839 ± 2% perf-stat.ps.minor-faults
7359 ± 2% -7.1% 6839 ± 2% perf-stat.ps.page-faults
3.97 ±107% -3.2 0.75 ±223% perf-profile.calltrace.cycles-pp.console_flush_all.console_unlock.vprintk_emit.devkmsg_emit.devkmsg_write
3.97 ±107% -3.2 0.75 ±223% perf-profile.calltrace.cycles-pp.console_unlock.vprintk_emit.devkmsg_emit.devkmsg_write.vfs_write
3.97 ±107% -3.2 0.75 ±223% perf-profile.calltrace.cycles-pp.devkmsg_emit.devkmsg_write.vfs_write.ksys_write.do_syscall_64
3.97 ±107% -3.2 0.75 ±223% perf-profile.calltrace.cycles-pp.devkmsg_write.vfs_write.ksys_write.do_syscall_64.entry_SYSCALL_64_after_hwframe
3.97 ±107% -3.2 0.75 ±223% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.write
3.97 ±107% -3.2 0.75 ±223% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.write
3.97 ±107% -3.2 0.75 ±223% perf-profile.calltrace.cycles-pp.ksys_write.do_syscall_64.entry_SYSCALL_64_after_hwframe.write
3.97 ±107% -3.2 0.75 ±223% perf-profile.calltrace.cycles-pp.vprintk_emit.devkmsg_emit.devkmsg_write.vfs_write.ksys_write
3.97 ±107% -3.2 0.75 ±223% perf-profile.calltrace.cycles-pp.write
3.61 ±106% -3.0 0.62 ±223% perf-profile.calltrace.cycles-pp.serial8250_console_write.console_flush_all.console_unlock.vprintk_emit.devkmsg_emit
3.45 ±105% -2.9 0.56 ±223% perf-profile.calltrace.cycles-pp.wait_for_lsr.serial8250_console_write.console_flush_all.console_unlock.vprintk_emit
1.80 ± 53% +1.0 2.84 ± 44% perf-profile.calltrace.cycles-pp.__x64_sys_exit_group.x64_sys_call.do_syscall_64.entry_SYSCALL_64_after_hwframe
1.80 ± 53% +1.0 2.84 ± 44% perf-profile.calltrace.cycles-pp.do_exit.do_group_exit.__x64_sys_exit_group.x64_sys_call.do_syscall_64
1.80 ± 53% +1.0 2.84 ± 44% perf-profile.calltrace.cycles-pp.do_group_exit.__x64_sys_exit_group.x64_sys_call.do_syscall_64.entry_SYSCALL_64_after_hwframe
1.80 ± 53% +1.0 2.84 ± 44% perf-profile.calltrace.cycles-pp.x64_sys_call.do_syscall_64.entry_SYSCALL_64_after_hwframe
3.97 ±107% -3.2 0.75 ±223% perf-profile.children.cycles-pp.console_flush_all
3.97 ±107% -3.2 0.75 ±223% perf-profile.children.cycles-pp.console_unlock
3.97 ±107% -3.2 0.75 ±223% perf-profile.children.cycles-pp.devkmsg_emit
3.97 ±107% -3.2 0.75 ±223% perf-profile.children.cycles-pp.devkmsg_write
3.97 ±107% -3.2 0.75 ±223% perf-profile.children.cycles-pp.vprintk_emit
3.61 ±106% -3.0 0.62 ±223% perf-profile.children.cycles-pp.serial8250_console_write
3.45 ±105% -2.9 0.56 ±223% perf-profile.children.cycles-pp.wait_for_lsr
1.80 ± 53% +1.0 2.84 ± 44% perf-profile.children.cycles-pp.__x64_sys_exit_group
***************************************************************************************************
lkp-skl-d07: 8 threads Intel(R) Core(TM) i7-6700 CPU @ 3.40GHz (Skylake) with 16G memory
=========================================================================================
cluster/compiler/cpufreq_governor/kconfig/rootfs/runtime/tbox_group/testcase:
cs-localhost/gcc-12/performance/x86_64-rhel-9.4/debian-12-x86_64-20240206.cgz/600s/lkp-skl-d07/qperf
commit:
v6.17-rc2
779b1a1cb1 ("cpuidle: governors: menu: Avoid selecting states with too much latency")
v6.17-rc2 779b1a1cb13ae17028aeddb2fbb
---------------- ---------------------------
%stddev %change %stddev
\ | \
861.00 ± 26% -29.1% 610.40 ± 19% perf-sched.wait_and_delay.count.smpboot_thread_fn.kthread.ret_from_fork.ret_from_fork_asm
280319 -4.7% 267006 vmstat.system.cs
4510 +4.3% 4706 qperf.tcp.latency
13138 ± 17% -28.1% 9451 ± 17% qperf.time.involuntary_context_switches
34417677 -4.8% 32782157 qperf.time.voluntary_context_switches
3362 +8.1% 3634 qperf.udp.latency
8.011e+08 -4.6% 7.642e+08 perf-stat.i.branch-instructions
12882352 -3.6% 12424168 perf-stat.i.branch-misses
281595 -4.7% 268383 perf-stat.i.context-switches
1.54 -5.3% 1.46 perf-stat.i.cpi
5.769e+09 -10.5% 5.161e+09 perf-stat.i.cpu-cycles
44.97 ± 2% -8.6% 41.11 ± 3% perf-stat.i.cpu-migrations
16208 ± 2% -13.4% 14038 ± 3% perf-stat.i.cycles-between-cache-misses
4.094e+09 -4.1% 3.927e+09 perf-stat.i.instructions
0.68 +7.6% 0.73 perf-stat.i.ipc
35.20 -4.7% 33.55 perf-stat.i.metric.K/sec
1.41 -16.0% 1.18 ± 33% perf-stat.overall.cpi
7.999e+08 -13.9% 6.888e+08 ± 33% perf-stat.ps.branch-instructions
12865498 -13.0% 11190853 ± 33% perf-stat.ps.branch-misses
281062 -14.2% 241224 ± 33% perf-stat.ps.context-switches
5.76e+09 -19.2% 4.652e+09 ± 33% perf-stat.ps.cpu-cycles
44.89 ± 2% -17.5% 37.01 ± 33% perf-stat.ps.cpu-migrations
4.088e+09 -13.4% 3.539e+09 ± 33% perf-stat.ps.instructions
2.458e+12 -13.4% 2.128e+12 ± 33% perf-stat.total.instructions
24.74 ± 5% -24.7 0.00 perf-profile.calltrace.cycles-pp.poll_idle.cpuidle_enter_state.cpuidle_enter.cpuidle_idle_call.do_idle
5.80 ± 4% -0.5 5.30 ± 3% perf-profile.calltrace.cycles-pp.__schedule.schedule.schedule_timeout.__skb_wait_for_more_packets.__skb_recv_udp
1.01 ± 11% +0.3 1.27 ± 12% perf-profile.calltrace.cycles-pp.entry_SYSRETQ_unsafe_stack.read
0.22 ±122% +0.4 0.66 ± 8% perf-profile.calltrace.cycles-pp.asm_sysvec_apic_timer_interrupt.cpuidle_enter_state.cpuidle_enter.cpuidle_idle_call.do_idle
0.16 ±152% +0.5 0.62 ± 8% perf-profile.calltrace.cycles-pp.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt.cpuidle_enter_state.cpuidle_enter.cpuidle_idle_call
6.05 ± 3% +0.6 6.64 ± 4% perf-profile.calltrace.cycles-pp.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.read
5.92 ± 3% +0.6 6.51 ± 4% perf-profile.calltrace.cycles-pp.vfs_read.ksys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe.read
6.73 ± 3% +0.8 7.51 ± 4% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.read
6.69 ± 3% +0.8 7.48 ± 4% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.read
22.08 ± 5% +19.4 41.52 ± 7% perf-profile.calltrace.cycles-pp.intel_idle.cpuidle_enter_state.cpuidle_enter.cpuidle_idle_call.do_idle
24.80 ± 6% -24.8 0.00 perf-profile.children.cycles-pp.poll_idle
0.38 ± 9% -0.2 0.19 ± 19% perf-profile.children.cycles-pp.local_clock_noinstr
0.93 ± 6% -0.2 0.76 ± 5% perf-profile.children.cycles-pp.native_sched_clock
0.35 ± 10% -0.1 0.28 ± 7% perf-profile.children.cycles-pp.cpuidle_governor_latency_req
0.14 ± 11% -0.0 0.09 ± 14% perf-profile.children.cycles-pp.call_function_single_prep_ipi
0.22 ± 12% -0.0 0.17 ± 12% perf-profile.children.cycles-pp.get_cpu_device
0.19 ± 8% +0.1 0.25 ± 9% perf-profile.children.cycles-pp.rseq_update_cpu_node_id
0.48 ± 6% +0.1 0.55 ± 4% perf-profile.children.cycles-pp.hrtimer_interrupt
0.49 ± 7% +0.1 0.56 ± 4% perf-profile.children.cycles-pp.__sysvec_apic_timer_interrupt
0.16 ± 15% +0.1 0.24 ± 9% perf-profile.children.cycles-pp.run_client_udp_lat
0.28 ± 12% +0.1 0.36 ± 9% perf-profile.children.cycles-pp.__softirqentry_text_end
0.00 +0.1 0.09 ± 23% perf-profile.children.cycles-pp.write@plt
0.25 ± 13% +0.1 0.34 ± 13% perf-profile.children.cycles-pp.__netif_rx
0.46 ± 7% +0.1 0.56 ± 8% perf-profile.children.cycles-pp.check_heap_object
0.24 ± 14% +0.1 0.35 ± 12% perf-profile.children.cycles-pp.move_addr_to_user
0.09 ± 14% +0.1 0.19 ± 18% perf-profile.children.cycles-pp.siphash_3u32
0.19 ± 19% +0.1 0.32 ± 11% perf-profile.children.cycles-pp.__ip_select_ident
0.59 ± 7% +0.2 0.76 ± 7% perf-profile.children.cycles-pp.__rseq_handle_notify_resume
0.87 ± 9% +0.2 1.04 ± 6% perf-profile.children.cycles-pp.__check_object_size
0.59 ± 11% +0.2 0.78 ± 9% perf-profile.children.cycles-pp.__ip_make_skb
0.88 ± 9% +0.2 1.09 ± 9% perf-profile.children.cycles-pp.exit_to_user_mode_loop
3.18 ± 3% +0.3 3.43 ± 5% perf-profile.children.cycles-pp.schedule_idle
3.59 ± 6% +0.5 4.09 ± 7% perf-profile.children.cycles-pp.__flush_smp_call_function_queue
3.83 ± 6% +0.5 4.37 ± 6% perf-profile.children.cycles-pp.flush_smp_call_function_queue
5.92 ± 3% +0.6 6.51 ± 4% perf-profile.children.cycles-pp.vfs_read
6.05 ± 3% +0.6 6.65 ± 4% perf-profile.children.cycles-pp.ksys_read
22.08 ± 5% +19.5 41.53 ± 7% perf-profile.children.cycles-pp.intel_idle
24.57 ± 6% -24.6 0.00 perf-profile.self.cycles-pp.poll_idle
0.90 ± 7% -0.2 0.73 ± 5% perf-profile.self.cycles-pp.native_sched_clock
0.13 ± 11% -0.0 0.08 ± 18% perf-profile.self.cycles-pp.call_function_single_prep_ipi
0.21 ± 12% -0.0 0.17 ± 11% perf-profile.self.cycles-pp.get_cpu_device
0.08 ± 25% +0.0 0.12 ± 10% perf-profile.self.cycles-pp.write
0.19 ± 7% +0.1 0.25 ± 9% perf-profile.self.cycles-pp.rseq_update_cpu_node_id
0.11 ± 18% +0.1 0.18 ± 10% perf-profile.self.cycles-pp.run_client_udp_lat
0.11 ± 12% +0.1 0.19 ± 13% perf-profile.self.cycles-pp.recvfrom
0.12 ± 17% +0.1 0.20 ± 14% perf-profile.self.cycles-pp.run_server_udp_lat
0.21 ± 15% +0.1 0.29 ± 17% perf-profile.self.cycles-pp.enqueue_entity
0.00 +0.1 0.08 ± 22% perf-profile.self.cycles-pp.write@plt
0.25 ± 9% +0.1 0.33 ± 7% perf-profile.self.cycles-pp.__softirqentry_text_end
0.10 ± 18% +0.1 0.19 ± 11% perf-profile.self.cycles-pp.read
0.09 ± 13% +0.1 0.19 ± 19% perf-profile.self.cycles-pp.siphash_3u32
0.22 ± 15% +0.1 0.34 ± 12% perf-profile.self.cycles-pp.entry_SYSCALL_64_after_hwframe
0.47 ± 13% +0.1 0.60 ± 11% perf-profile.self.cycles-pp.do_syscall_64
0.29 ± 14% +0.2 0.45 ± 12% perf-profile.self.cycles-pp.schedule_timeout
0.31 ± 10% +0.3 0.61 ± 12% perf-profile.self.cycles-pp.__skb_recv_udp
22.08 ± 5% +19.4 41.52 ± 7% perf-profile.self.cycles-pp.intel_idle
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
Powered by blists - more mailing lists