| 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: <20190930084637.GD17687@shao2-debian>
Date: Mon, 30 Sep 2019 16:46:37 +0800
From: kernel test robot <rong.a.chen@...el.com>
To: Al Viro <viro@...iv.linux.org.uk>
Cc: "zhengbin (A)" <zhengbin13@...wei.com>,
LKML <linux-kernel@...r.kernel.org>,
Stephen Rothwell <sfr@...b.auug.org.au>, lkp@...org
Subject: d4f4de5e5e: aim9.dir_rtns_1.ops_per_sec -26.1% regression
Greeting,
FYI, we noticed a -26.1% regression of aim9.dir_rtns_1.ops_per_sec due to commit:
commit: d4f4de5e5ef8efde85febb6876cd3c8ab1631999 ("Fix the locking in dcache_readdir() and friends")
https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
in testcase: aim9
on test machine: 4 threads Intel(R) Core(TM) i3-3220 CPU @ 3.30GHz with 4G memory
with following parameters:
testtime: 5s
test: all
cpufreq_governor: performance
ucode: 0x21
test-description: Suite IX is the "AIM Independent Resource Benchmark:" the famous synthetic benchmark.
test-url: https://sourceforge.net/projects/aimbench/files/aim-suite9/
In addition to that, the commit also has significant impact on the following tests:
+------------------+------------------------------------------------------------------+
| testcase: change | aim9: aim9.dir_rtns_1.ops_per_sec -26.3% regression |
| test machine | 4 threads Intel(R) Core(TM) i3-3220 CPU @ 3.30GHz with 4G memory |
| test parameters | cpufreq_governor=performance |
| | test=dir_rtns_1 |
| | testtime=300s |
| | ucode=0x21 |
+------------------+------------------------------------------------------------------+
If you fix the issue, kindly add following tag
Reported-by: kernel test robot <rong.a.chen@...el.com>
Details are as below:
-------------------------------------------------------------------------------------------------->
To reproduce:
git clone https://github.com/intel/lkp-tests.git
cd lkp-tests
bin/lkp install job.yaml # job file is attached in this email
bin/lkp run job.yaml
=========================================================================================
compiler/cpufreq_governor/kconfig/rootfs/tbox_group/test/testcase/testtime/ucode:
gcc-7/performance/x86_64-rhel-7.6/debian-x86_64-2019-05-14.cgz/lkp-ivb-d03/all/aim9/5s/0x21
commit:
mainline-tracking-v5.3-190917T112302Z
d4f4de5e5e ("Fix the locking in dcache_readdir() and friends")
mainline-trackin d4f4de5e5ef8efde85febb6876c
---------------- ---------------------------
%stddev %change %stddev
\ | \
10976333 -26.1% 8116500 aim9.dir_rtns_1.ops_per_sec
916821 +1.6% 931164 aim9.sync_disk_wrt.ops_per_sec
145728 -1.1% 144085 aim9.tcp_test.ops_per_sec
276050 +1.3% 279655 aim9.udp_test.ops_per_sec
2307 ± 59% -53.1% 1083 ± 45% interrupts.CPU3.RES:Rescheduling_interrupts
3476 -2.0% 3406 proc-vmstat.nr_kernel_stack
1.205e+08 ± 47% -73.8% 31527642 ± 73% cpuidle.C3.time
403015 ± 70% -77.9% 89237 ± 65% cpuidle.C3.usage
1085 ± 2% +16.1% 1260 ± 4% slabinfo.kmalloc-96.active_objs
1085 ± 2% +16.1% 1260 ± 4% slabinfo.kmalloc-96.num_objs
176352 ±167% +294.7% 696134 ± 90% softirqs.CPU3.NET_RX
73740 ± 51% +136.7% 174570 ± 52% softirqs.CPU3.RCU
36381 ± 23% +33.7% 48655 ± 13% sched_debug.cfs_rq:/.min_vruntime.min
6.86 ± 59% +85.8% 12.75 ± 37% sched_debug.cfs_rq:/.runnable_load_avg.min
-57832 -85.8% -8235 sched_debug.cfs_rq:/.spread0.min
403011 ± 70% -77.9% 89237 ± 65% turbostat.C3
9.99 ± 47% -7.4 2.61 ± 73% turbostat.C3%
3.64 ±214% +323.7% 15.41 ± 99% turbostat.CPU%c6
39.33 -7.8% 36.25 turbostat.CoreTmp
38.83 -7.3% 36.00 turbostat.PkgTmp
1.57 ±141% +3.2 4.78 ± 15% perf-profile.calltrace.cycles-pp.div_double
0.02 ±149% +0.1 0.08 ± 26% perf-profile.children.cycles-pp.native_sched_clock
0.02 ±149% +0.1 0.08 ± 26% perf-profile.children.cycles-pp.sched_clock
0.02 ±146% +0.1 0.09 ± 27% perf-profile.children.cycles-pp.arch_stack_walk
0.02 ±146% +0.1 0.09 ± 20% perf-profile.children.cycles-pp.swake_up_one
0.02 ±149% +0.1 0.09 ± 35% perf-profile.children.cycles-pp.sched_clock_cpu
0.01 ±223% +0.1 0.09 ± 24% perf-profile.children.cycles-pp.swake_up_locked
0.03 ±152% +0.1 0.10 ± 30% perf-profile.children.cycles-pp.stack_trace_save_tsk
0.01 ±223% +0.1 0.09 ± 34% perf-profile.children.cycles-pp.__fxstat64
0.05 ±114% +0.1 0.14 ± 33% perf-profile.children.cycles-pp.security_task_getsecid
0.03 ±152% +0.1 0.11 ± 32% perf-profile.children.cycles-pp.__account_scheduler_latency
0.05 ±104% +0.1 0.14 ± 27% perf-profile.children.cycles-pp.selinux_inode_free_security
0.05 ±141% +0.1 0.14 ± 39% perf-profile.children.cycles-pp.__x64_sys_chdir
0.05 ±141% +0.1 0.14 ± 39% perf-profile.children.cycles-pp.ksys_chdir
0.05 ±120% +0.1 0.14 ± 29% perf-profile.children.cycles-pp.enqueue_entity
0.01 ±223% +0.1 0.11 ± 36% perf-profile.children.cycles-pp.d_set_d_op
0.05 ± 94% +0.1 0.15 ± 38% perf-profile.children.cycles-pp.ttwu_do_activate
0.05 ± 94% +0.1 0.15 ± 38% perf-profile.children.cycles-pp.activate_task
0.05 ±120% +0.1 0.15 ± 36% perf-profile.children.cycles-pp.enqueue_task_fair
0.00 +0.1 0.13 ± 27% perf-profile.children.cycles-pp.__srcu_read_lock
0.06 ±116% +0.2 0.22 ± 38% perf-profile.children.cycles-pp.d_lookup
0.08 ±103% +0.2 0.24 ± 32% perf-profile.children.cycles-pp.may_open
0.07 ±119% +0.2 0.25 ± 38% perf-profile.children.cycles-pp.inode_doinit_with_dentry
0.06 ±101% +0.2 0.24 ± 33% perf-profile.children.cycles-pp.fsnotify_grab_connector
0.08 ±107% +0.2 0.28 ± 35% perf-profile.children.cycles-pp.fsnotify_destroy_marks
0.10 ±104% +0.2 0.30 ± 27% perf-profile.children.cycles-pp.selinux_inode_init_security
0.11 ±115% +0.2 0.33 ± 39% perf-profile.children.cycles-pp.lockref_put_or_lock
0.12 ±116% +0.2 0.36 ± 29% perf-profile.children.cycles-pp._IO_fgets
0.12 ±102% +0.3 0.38 ± 36% perf-profile.children.cycles-pp.vfprintf
0.21 ±106% +0.3 0.56 ± 29% perf-profile.children.cycles-pp.destroy_inode
0.37 ± 79% +0.5 0.86 ± 44% perf-profile.children.cycles-pp.kthread
0.37 ± 79% +0.5 0.86 ± 45% perf-profile.children.cycles-pp.ret_from_fork
0.27 ±112% +0.5 0.80 ± 29% perf-profile.children.cycles-pp.user_path_at_empty
1.57 ±141% +3.2 4.79 ± 15% perf-profile.children.cycles-pp.div_double
0.03 ±141% +0.1 0.09 ± 37% perf-profile.self.cycles-pp.selinux_inode_free_security
0.01 ±223% +0.1 0.07 ± 31% perf-profile.self.cycles-pp.may_open
0.03 ±142% +0.1 0.11 ± 41% perf-profile.self.cycles-pp.shmem_getattr
0.01 ±223% +0.1 0.10 ± 27% perf-profile.self.cycles-pp.may_create
0.01 ±223% +0.1 0.11 ± 36% perf-profile.self.cycles-pp.d_set_d_op
0.04 ±142% +0.1 0.14 ± 36% perf-profile.self.cycles-pp.inode_doinit_with_dentry
0.01 ±223% +0.1 0.12 ± 29% perf-profile.self.cycles-pp.__d_lookup_done
0.03 ±143% +0.1 0.15 ± 16% perf-profile.self.cycles-pp.getname_flags
0.02 ±144% +0.1 0.15 ± 36% perf-profile.self.cycles-pp.__alloc_fd
0.00 +0.1 0.13 ± 22% perf-profile.self.cycles-pp.__srcu_read_lock
0.07 ±110% +0.1 0.20 ± 26% perf-profile.self.cycles-pp.selinux_inode_init_security
0.09 ±114% +0.2 0.28 ± 36% perf-profile.self.cycles-pp.lockref_put_or_lock
0.11 ±117% +0.2 0.33 ± 28% perf-profile.self.cycles-pp._IO_fgets
0.15 ±110% +0.2 0.40 ± 30% perf-profile.self.cycles-pp.inode_permission
0.12 ±103% +0.2 0.37 ± 36% perf-profile.self.cycles-pp.vfprintf
1.52 ±141% +3.3 4.78 ± 15% perf-profile.self.cycles-pp.div_double
aim9.dir_rtns_1.ops_per_sec
1.2e+07 +-+---------------------------------------------------------------+
|.+.+ +.+.+ +.+.+.+.+.+.+.+.+ +.+.+ +.+.+ +.+.|
1e+07 +-+ : : : : : : : : : : |
| : : : : : : : : : : |
O O : O O O O O O O O O O O O O O O O : O : O : : : |
8e+06 +-+ : : : : :O:O O:O : : : |
| : : : : : : : : : : |
6e+06 +-+ : : : : : : : : : : |
| : : : : : : : : : : |
4e+06 +-+ : : : : : : : : : : |
| : : : : : : : : : : |
| : : : : : : : |
2e+06 +-+ : : : : : : : |
| : : : : : : : |
0 +-+-O-------------------------------------------------------------+
[*] bisect-good sample
[O] bisect-bad sample
***************************************************************************************************
lkp-ivb-d03: 4 threads Intel(R) Core(TM) i3-3220 CPU @ 3.30GHz with 4G memory
=========================================================================================
compiler/cpufreq_governor/kconfig/rootfs/tbox_group/test/testcase/testtime/ucode:
gcc-7/performance/x86_64-rhel-7.6/debian-x86_64-2019-05-14.cgz/lkp-ivb-d03/dir_rtns_1/aim9/300s/0x21
commit:
mainline-tracking-v5.3-190917T112302Z
d4f4de5e5e ("Fix the locking in dcache_readdir() and friends")
mainline-trackin d4f4de5e5ef8efde85febb6876c
---------------- ---------------------------
%stddev %change %stddev
\ | \
10956895 -26.3% 8074108 aim9.dir_rtns_1.ops_per_sec
235.88 ± 3% +8.0% 254.66 aim9.time.system_time
64.12 ± 14% -29.3% 45.34 ± 5% aim9.time.user_time
20.04 ± 3% +1.6 21.67 mpstat.cpu.all.sys%
0.00 +0.1 0.07 ± 33% perf-profile.children.cycles-pp.schedule
0.00 +13.6 13.63 ± 81% perf-profile.children.cycles-pp.scan_positives
395.86 +2.6% 406.00 proc-vmstat.nr_active_file
395.86 +2.6% 406.00 proc-vmstat.nr_zone_active_file
1092 ± 2% +18.0% 1288 ± 3% slabinfo.kmalloc-96.active_objs
1092 ± 2% +18.0% 1288 ± 3% slabinfo.kmalloc-96.num_objs
376.29 ± 96% +1150.0% 4703 ±111% interrupts.CPU3.NMI:Non-maskable_interrupts
376.29 ± 96% +1150.0% 4703 ±111% interrupts.CPU3.PMI:Performance_monitoring_interrupts
318.00 ± 15% -32.2% 215.50 ± 6% interrupts.TLB:TLB_shootdowns
58972 ± 13% -23.9% 44848 ± 9% softirqs.CPU1.RCU
61482 ± 9% -20.1% 49110 ± 3% softirqs.CPU2.RCU
122988 ± 14% +25.8% 154740 ± 8% softirqs.CPU3.TIMER
243766 ± 8% -20.7% 193320 ± 3% softirqs.RCU
20.60 ± 54% +67.9% 34.59 ± 23% perf-stat.i.MPKI
2.08 ± 27% +1.1 3.20 ± 39% perf-stat.i.branch-miss-rate%
21252312 ± 50% -57.6% 9012007 ± 56% perf-stat.i.branch-misses
6.94 ± 57% +5.5 12.42 ± 27% perf-stat.i.cache-miss-rate%
5870887 ± 40% -35.5% 3788398 ± 5% perf-stat.i.cache-references
0.45 ± 40% +0.2 0.69 ± 21% perf-stat.i.dTLB-load-miss-rate%
5.18 ± 63% +5.3 10.45 ± 28% perf-stat.overall.cache-miss-rate%
0.63 ± 4% +19.4% 0.75 ± 15% perf-stat.overall.cpi
21183509 ± 50% -57.5% 8992621 ± 56% perf-stat.ps.branch-misses
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.
Thanks,
Rong Chen
View attachment "config-5.3.0-00001-gd4f4de5e5ef8e" of type "text/plain" (199448 bytes)
View attachment "job-script" of type "text/plain" (7161 bytes)
View attachment "job.yaml" of type "text/plain" (4741 bytes)
View attachment "reproduce" of type "text/plain" (254 bytes)
Powered by blists - more mailing lists