| 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: <20160727015425.GL21454@yexl-desktop>
Date: Wed, 27 Jul 2016 09:54:26 +0800
From: kernel test robot <xiaolong.ye@...el.com>
To: Xin Long <lucien.xin@...il.com>
Cc: "David S. Miller" <davem@...emloft.net>,
LKML <linux-kernel@...r.kernel.org>,
Stephen Rothwell <sfr@...b.auug.org.au>, lkp@...org
Subject: [lkp] [sctp] a6c2f79287: netperf.Throughput_Mbps -37.2% regression
FYI, we noticed a -37.2% regression of netperf.Throughput_Mbps due to commit:
commit a6c2f792873aff332a4689717c3cd6104f46684c ("sctp: implement prsctp TTL policy")
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
in testcase: netperf
on test machine: 4 threads Ivy Bridge with 8G memory
with following parameters:
ip: ipv4
runtime: 300s
nr_threads: 200%
cluster: cs-localhost
send_size: 10K
test: SCTP_STREAM_MANY
cpufreq_governor: performance
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.
Details are as below:
-------------------------------------------------------------------------------------------------->
To reproduce:
git clone git://git.kernel.org/pub/scm/linux/kernel/git/wfg/lkp-tests.git
cd lkp-tests
bin/lkp install job.yaml # job file is attached in this email
bin/lkp run job.yaml
=========================================================================================
cluster/compiler/cpufreq_governor/ip/kconfig/nr_threads/rootfs/runtime/send_size/tbox_group/test/testcase:
cs-localhost/gcc-6/performance/ipv4/x86_64-rhel-7.2/200%/debian-x86_64-2015-02-07.cgz/300s/10K/lkp-ivb-d02/SCTP_STREAM_MANY/netperf
commit:
826d253d57 ("sctp: add SCTP_PR_ASSOC_STATUS on sctp sockopt")
a6c2f79287 ("sctp: implement prsctp TTL policy")
826d253d57b11f69 a6c2f792873aff332a4689717c
---------------- --------------------------
%stddev %change %stddev
\ | \
3923 ± 0% -37.2% 2462 ± 0% netperf.Throughput_Mbps
11701 ± 14% -21.9% 9139 ± 11% meminfo.AnonHugePages
642.75 ± 7% +15.8% 744.00 ± 6% slabinfo.kmalloc-512.active_objs
4993 ± 2% +14.6% 5724 ± 2% slabinfo.kmalloc-64.active_objs
4993 ± 2% +14.6% 5724 ± 2% slabinfo.kmalloc-64.num_objs
9.00 ± 0% -77.8% 2.00 ± 0% vmstat.procs.r
112616 ± 1% +19.0% 133959 ± 0% vmstat.system.cs
4053 ± 0% +6.3% 4309 ± 0% vmstat.system.in
16466114 ± 0% -37.4% 10307354 ± 0% softirqs.NET_RX
72067 ± 10% -63.5% 26326 ± 4% softirqs.RCU
8598 ± 4% +954.6% 90677 ± 1% softirqs.SCHED
605899 ± 0% -45.7% 329111 ± 0% softirqs.TIMER
2529 ± 4% -15.1% 2147 ± 1% proc-vmstat.nr_alloc_batch
1.48e+08 ± 0% -37.1% 93059843 ± 0% proc-vmstat.numa_hit
1.48e+08 ± 0% -37.1% 93059842 ± 0% proc-vmstat.numa_local
3.742e+08 ± 1% -36.9% 2.362e+08 ± 0% proc-vmstat.pgalloc_dma32
4.733e+08 ± 0% -36.6% 3e+08 ± 0% proc-vmstat.pgalloc_normal
8.476e+08 ± 0% -36.7% 5.362e+08 ± 0% proc-vmstat.pgfree
99.31 ± 0% -44.6% 55.03 ± 0% turbostat.%Busy
3269 ± 0% -44.8% 1804 ± 1% turbostat.Avg_MHz
0.05 ± 17% +51942.1% 24.72 ± 1% turbostat.CPU%c1
0.64 ± 0% +3067.3% 20.11 ± 3% turbostat.CPU%c6
20.20 ± 0% -24.7% 15.21 ± 0% turbostat.CorWatt
0.12 ± 39% +1906.4% 2.36 ± 3% turbostat.Pkg%pc2
0.46 ± 10% +1696.7% 8.27 ± 6% turbostat.Pkg%pc6
37.54 ± 0% -14.5% 32.11 ± 0% turbostat.PkgWatt
76510 ± 46% +2.6e+05% 1.954e+08 ± 1% cpuidle.C1-IVB.time
6742 ± 76% +2.9e+05% 19860856 ± 1% cpuidle.C1-IVB.usage
19769 ± 17% +5414.4% 1090172 ± 3% cpuidle.C1E-IVB.time
151.00 ± 11% +4024.8% 6228 ± 3% cpuidle.C1E-IVB.usage
33074 ± 14% +5086.6% 1715442 ± 2% cpuidle.C3-IVB.time
114.50 ± 14% +6075.1% 7070 ± 3% cpuidle.C3-IVB.usage
8006184 ± 0% +4066.3% 3.336e+08 ± 2% cpuidle.C6-IVB.time
8874 ± 0% +4197.0% 381350 ± 2% cpuidle.C6-IVB.usage
476.75 ± 92% +77822.9% 371497 ± 12% cpuidle.POLL.time
46.25 ±144% +30824.3% 14302 ± 3% cpuidle.POLL.usage
1.758e+11 ± 2% -38.4% 1.083e+11 ± 1% perf-stat.L1-dcache-load-misses
6.405e+11 ± 0% -29.7% 4.5e+11 ± 0% perf-stat.L1-dcache-loads
6.28e+09 ± 3% -30.0% 4.398e+09 ± 0% perf-stat.L1-dcache-prefetch-misses
8.142e+10 ± 1% -39.3% 4.941e+10 ± 0% perf-stat.L1-dcache-store-misses
5.32e+11 ± 0% -33.6% 3.533e+11 ± 1% perf-stat.L1-dcache-stores
5.59e+10 ± 1% -19.3% 4.512e+10 ± 1% perf-stat.L1-icache-load-misses
5.199e+10 ± 0% -42.1% 3.01e+10 ± 1% perf-stat.LLC-loads
3.066e+10 ± 1% -42.4% 1.765e+10 ± 1% perf-stat.LLC-prefetches
5.282e+10 ± 0% -40.0% 3.168e+10 ± 1% perf-stat.LLC-stores
2.776e+11 ± 0% -27.9% 2.002e+11 ± 0% perf-stat.branch-instructions
3.861e+09 ± 1% -50.8% 1.9e+09 ± 0% perf-stat.branch-load-misses
2.772e+11 ± 0% -27.8% 2e+11 ± 0% perf-stat.branch-loads
3.864e+09 ± 1% -50.8% 1.9e+09 ± 0% perf-stat.branch-misses
1.179e+11 ± 0% -43.9% 6.61e+10 ± 0% perf-stat.bus-cycles
7.126e+09 ± 16% -96.8% 2.272e+08 ± 4% perf-stat.cache-misses
1.173e+11 ± 0% -38.0% 7.278e+10 ± 0% perf-stat.cache-references
34232822 ± 1% +19.1% 40787311 ± 0% perf-stat.context-switches
3.897e+12 ± 0% -44.2% 2.174e+12 ± 1% perf-stat.cpu-cycles
12019 ± 35% +306.6% 48867 ± 1% perf-stat.cpu-migrations
4.069e+09 ± 20% -58.4% 1.694e+09 ± 46% perf-stat.dTLB-load-misses
6.421e+11 ± 0% -30.3% 4.476e+11 ± 1% perf-stat.dTLB-loads
5.285e+08 ± 22% -71.0% 1.531e+08 ± 17% perf-stat.dTLB-store-misses
5.32e+11 ± 0% -33.4% 3.544e+11 ± 1% perf-stat.dTLB-stores
3.735e+08 ± 5% -48.5% 1.923e+08 ± 3% perf-stat.iTLB-load-misses
1.803e+08 ± 52% +662.4% 1.374e+09 ± 2% perf-stat.iTLB-loads
1.505e+12 ± 0% -29.3% 1.064e+12 ± 0% perf-stat.instructions
339045 ± 0% +4.4% 354068 ± 0% perf-stat.minor-faults
339041 ± 0% +4.4% 354062 ± 0% perf-stat.page-faults
3.892e+12 ± 0% -43.9% 2.182e+12 ± 0% perf-stat.ref-cycles
3.052e+12 ± 0% -25.9% 2.261e+12 ± 0% perf-stat.stalled-cycles-frontend
34082 ± 16% -65.5% 11746 ± 83% sched_debug.cfs_rq:/.MIN_vruntime.avg
75047 ± 3% -58.2% 31366 ± 63% sched_debug.cfs_rq:/.MIN_vruntime.max
34069 ± 3% -59.9% 13660 ± 61% sched_debug.cfs_rq:/.MIN_vruntime.stddev
1403286 ± 4% -34.9% 913870 ± 14% sched_debug.cfs_rq:/.load.avg
1874082 ± 7% -27.9% 1352098 ± 19% sched_debug.cfs_rq:/.load.max
1041057 ± 0% -66.8% 345620 ± 35% sched_debug.cfs_rq:/.load.min
1063 ± 0% -54.6% 482.93 ± 15% sched_debug.cfs_rq:/.load_avg.avg
1204 ± 3% -20.6% 956.50 ± 4% sched_debug.cfs_rq:/.load_avg.max
939.38 ± 0% -86.4% 127.33 ± 31% sched_debug.cfs_rq:/.load_avg.min
104.33 ± 19% +238.6% 353.26 ± 6% sched_debug.cfs_rq:/.load_avg.stddev
34082 ± 16% -65.5% 11746 ± 83% sched_debug.cfs_rq:/.max_vruntime.avg
75047 ± 3% -58.2% 31366 ± 63% sched_debug.cfs_rq:/.max_vruntime.max
34069 ± 3% -59.9% 13660 ± 61% sched_debug.cfs_rq:/.max_vruntime.stddev
74820 ± 0% -10.4% 67031 ± 1% sched_debug.cfs_rq:/.min_vruntime.min
1091 ± 7% +378.3% 5221 ± 13% sched_debug.cfs_rq:/.min_vruntime.stddev
1.35 ± 3% -35.4% 0.88 ± 14% sched_debug.cfs_rq:/.nr_running.avg
1.83 ± 6% -29.5% 1.29 ± 19% sched_debug.cfs_rq:/.nr_running.max
1.00 ± 0% -66.7% 0.33 ± 35% sched_debug.cfs_rq:/.nr_running.min
943.28 ± 0% -61.0% 367.80 ± 20% sched_debug.cfs_rq:/.runnable_load_avg.avg
991.42 ± 1% -21.2% 781.00 ± 14% sched_debug.cfs_rq:/.runnable_load_avg.max
894.88 ± 1% -90.9% 81.46 ± 20% sched_debug.cfs_rq:/.runnable_load_avg.min
39.07 ± 31% +667.7% 299.97 ± 16% sched_debug.cfs_rq:/.runnable_load_avg.stddev
-794.96 ±-98% +505.2% -4810 ±-18% sched_debug.cfs_rq:/.spread0.avg
-1935 ±-42% +442.8% -10508 ±-13% sched_debug.cfs_rq:/.spread0.min
1092 ± 7% +378.0% 5221 ± 13% sched_debug.cfs_rq:/.spread0.stddev
949.60 ± 0% -41.6% 554.85 ± 12% sched_debug.cfs_rq:/.util_avg.avg
976.08 ± 1% -9.9% 879.71 ± 1% sched_debug.cfs_rq:/.util_avg.max
925.83 ± 0% -67.0% 305.25 ± 20% sched_debug.cfs_rq:/.util_avg.min
18.71 ± 30% +1151.4% 234.18 ± 6% sched_debug.cfs_rq:/.util_avg.stddev
2.60 ± 9% -66.3% 0.88 ± 62% sched_debug.cpu.clock.stddev
2.60 ± 9% -66.3% 0.88 ± 62% sched_debug.cpu.clock_task.stddev
941.57 ± 1% -65.8% 322.38 ± 25% sched_debug.cpu.cpu_load[0].avg
992.46 ± 1% -29.5% 700.08 ± 25% sched_debug.cpu.cpu_load[0].max
880.67 ± 4% -93.3% 58.58 ± 64% sched_debug.cpu.cpu_load[0].min
44.14 ± 46% +526.2% 276.40 ± 28% sched_debug.cpu.cpu_load[0].stddev
944.76 ± 0% -59.6% 381.23 ± 17% sched_debug.cpu.cpu_load[1].avg
980.25 ± 1% -17.5% 808.25 ± 5% sched_debug.cpu.cpu_load[1].max
900.79 ± 1% -89.8% 91.67 ± 28% sched_debug.cpu.cpu_load[1].min
30.83 ± 30% +904.7% 309.72 ± 7% sched_debug.cpu.cpu_load[1].stddev
941.95 ± 0% -61.6% 361.83 ± 18% sched_debug.cpu.cpu_load[2].avg
974.08 ± 1% -18.0% 798.29 ± 5% sched_debug.cpu.cpu_load[2].max
899.00 ± 1% -90.8% 82.42 ± 24% sched_debug.cpu.cpu_load[2].min
29.59 ± 30% +946.9% 309.73 ± 6% sched_debug.cpu.cpu_load[2].stddev
935.02 ± 0% -63.1% 344.78 ± 19% sched_debug.cpu.cpu_load[3].avg
965.83 ± 1% -18.7% 785.50 ± 5% sched_debug.cpu.cpu_load[3].max
893.04 ± 1% -91.8% 73.21 ± 26% sched_debug.cpu.cpu_load[3].min
28.77 ± 29% +976.4% 309.70 ± 6% sched_debug.cpu.cpu_load[3].stddev
922.55 ± 0% -64.2% 330.33 ± 19% sched_debug.cpu.cpu_load[4].avg
951.75 ± 1% -20.0% 761.71 ± 6% sched_debug.cpu.cpu_load[4].max
880.96 ± 0% -92.8% 63.08 ± 31% sched_debug.cpu.cpu_load[4].min
27.92 ± 26% +988.8% 304.04 ± 6% sched_debug.cpu.cpu_load[4].stddev
1302 ± 1% -15.7% 1097 ± 7% sched_debug.cpu.curr->pid.avg
798.04 ± 5% -70.5% 235.67 ± 39% sched_debug.cpu.curr->pid.min
812.92 ± 1% +22.9% 999.29 ± 4% sched_debug.cpu.curr->pid.stddev
1359793 ± 6% -34.4% 892053 ± 17% sched_debug.cpu.load.avg
1830834 ± 4% -28.5% 1308664 ± 19% sched_debug.cpu.load.max
1041241 ± 0% -62.6% 389138 ± 37% sched_debug.cpu.load.min
0.00 ± 21% +77.1% 0.00 ± 7% sched_debug.cpu.next_balance.stddev
150630 ± 0% -24.2% 114118 ± 0% sched_debug.cpu.nr_load_updates.avg
151833 ± 0% -21.6% 119037 ± 0% sched_debug.cpu.nr_load_updates.max
149812 ± 0% -27.3% 108953 ± 1% sched_debug.cpu.nr_load_updates.min
745.51 ± 1% +502.7% 4492 ± 14% sched_debug.cpu.nr_load_updates.stddev
2.60 ± 1% -52.8% 1.23 ± 17% sched_debug.cpu.nr_running.avg
3.79 ± 3% -39.6% 2.29 ± 39% sched_debug.cpu.nr_running.max
1.62 ± 8% -74.4% 0.42 ± 34% sched_debug.cpu.nr_running.min
4242583 ± 1% +20.1% 5094187 ± 0% sched_debug.cpu.nr_switches.avg
3880089 ± 1% +24.6% 4835277 ± 1% sched_debug.cpu.nr_switches.min
487080 ± 18% -61.6% 186796 ± 37% sched_debug.cpu.nr_switches.stddev
netperf.Throughput_Mbps
4000 ++-----------------------------------*--*-----*--*--*--*--*-*--*-----+
| .*.. .*..*..* *. |
3500 *+.*..* *.*..*..*. *. |
3000 ++ : : |
| : O |
2500 O+ O O: O : O O O O O O O O O O O O O O
| : : O O O O O |
2000 ++ : : |
| : : |
1500 ++ : : |
1000 ++ : : |
| : : |
500 ++ :: |
| : |
0 ++-------*------------------------------------O----------------------+
[*] bisect-good sample
[O] bisect-bad sample
Thanks,
Xiaolong
View attachment "config-4.7.0-rc6-01198-ga6c2f79" of type "text/plain" (151646 bytes)
View attachment "job.yaml" of type "text/plain" (4256 bytes)
View attachment "reproduce" of type "text/plain" (848 bytes)
Powered by blists - more mailing lists