Greeting,
FYI, we noticed a -73.5% regression of stress-ng.rename.ops_per_sec due to commit:
commit: 6191cf3ad59fda5901160633fef8e41b064a5246 ("xfs: flush inodegc workqueue tasks before cancel")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
in testcase: stress-ng
on test machine: 96 threads 2 sockets Ice Lake with 256G memory
with following parameters:
nr_threads: 10%
disk: 1HDD
testtime: 60s
fs: xfs
class: filesystem
test: rename
cpufreq_governor: performance
ucode: 0xb000280
In addition to that, the commit also has significant impact on the following tests:
If you fix the issue, kindly add following tag
Reported-by: kernel test robot <[email protected]>
Details are as below:
-------------------------------------------------------------------------------------------------->
To reproduce:
git clone https://github.com/intel/lkp-tests.git
cd lkp-tests
sudo bin/lkp install job.yaml # job file is attached in this email
bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run
sudo bin/lkp run generated-yaml-file
# if come across any failure that blocks the test,
# please remove ~/.lkp and /lkp dir to run from a clean state.
=========================================================================================
class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime/ucode:
filesystem/gcc-9/performance/1HDD/xfs/x86_64-rhel-8.3/10%/debian-10.4-x86_64-20200603.cgz/lkp-icl-2sp1/rename/stress-ng/60s/0xb000280
commit:
a8e422af69 ("xfs: remove unused xfs_ioctl32.h declarations")
6191cf3ad5 ("xfs: flush inodegc workqueue tasks before cancel")
a8e422af69613300 6191cf3ad59fda5901160633fef
---------------- ---------------------------
%stddev %change %stddev
\ | \
13809323 -73.5% 3655053 ? 2% stress-ng.rename.ops
230150 -73.5% 60915 ? 2% stress-ng.rename.ops_per_sec
68641 ? 6% -71.5% 19530 ? 18% stress-ng.time.involuntary_context_switches
762.40 +1.7% 775.00 stress-ng.time.percent_of_cpu_this_job_got
463.07 +3.3% 478.34 stress-ng.time.system_time
10.61 ? 5% -71.3% 3.04 ? 3% stress-ng.time.user_time
152587 ? 6% -72.6% 41861 ? 14% stress-ng.time.voluntary_context_switches
280117 ? 13% -22.7% 216667 ? 19% numa-numastat.node0.numa_hit
10272 ? 4% -48.7% 5273 ? 6% vmstat.system.cs
0.06 ? 21% -0.0 0.04 ? 4% mpstat.cpu.all.soft%
0.20 ? 3% -0.1 0.07 ? 2% mpstat.cpu.all.usr%
2703 ? 4% +5.6% 2854 ? 3% proc-vmstat.nr_active_anon
7988 ? 2% +2.5% 8185 proc-vmstat.nr_shmem
2703 ? 4% +5.6% 2854 ? 3% proc-vmstat.nr_zone_active_anon
522094 -10.7% 466068 proc-vmstat.numa_hit
436762 ? 2% -13.2% 379232 proc-vmstat.numa_local
4853 ? 6% +12.1% 5438 ? 3% proc-vmstat.pgactivate
525596 -10.4% 470724 proc-vmstat.pgalloc_normal
476799 -11.2% 423300 proc-vmstat.pgfree
61932 -2.6% 60328 proc-vmstat.pgpgout
3.60 ? 40% +102.9% 7.30 perf-stat.i.MPKI
4.774e+09 -71.1% 1.381e+09 perf-stat.i.branch-instructions
16114311 ? 32% -72.6% 4418605 ? 4% perf-stat.i.branch-misses
30.00 ? 24% +8.4 38.44 ? 4% perf-stat.i.cache-miss-rate%
20101026 ? 3% -8.7% 18358375 ? 3% perf-stat.i.cache-misses
73234568 ? 37% -34.7% 47830132 ? 2% perf-stat.i.cache-references
10261 ? 4% -50.5% 5079 ? 7% perf-stat.i.context-switches
1.29 ? 3% +227.7% 4.23 perf-stat.i.cpi
123.06 -8.1% 113.09 perf-stat.i.cpu-migrations
1474 ? 4% +9.8% 1619 ? 2% perf-stat.i.cycles-between-cache-misses
7.692e+11 ?199% -100.0% 202244 ? 46% perf-stat.i.dTLB-load-misses
5.182e+09 -70.1% 1.551e+09 perf-stat.i.dTLB-loads
107292 ? 95% -66.8% 35623 ? 23% perf-stat.i.dTLB-store-misses
2.564e+09 -76.9% 5.931e+08 ? 2% perf-stat.i.dTLB-stores
2.194e+10 -70.1% 6.557e+09 perf-stat.i.instructions
0.78 ? 2% -68.4% 0.25 perf-stat.i.ipc
130.67 -71.9% 36.71 perf-stat.i.metric.M/sec
93.21 +4.0 97.26 perf-stat.i.node-load-miss-rate%
3910237 ? 5% +52.1% 5948845 ? 3% perf-stat.i.node-load-misses
215305 ? 35% -69.9% 64873 ? 4% perf-stat.i.node-loads
55.52 ? 4% +14.4 69.90 perf-stat.i.node-store-miss-rate%
3175249 ? 9% +151.9% 7999504 ? 3% perf-stat.i.node-store-misses
2444045 ? 4% +32.3% 3234491 ? 2% perf-stat.i.node-stores
3.33 ? 36% +119.4% 7.29 perf-stat.overall.MPKI
30.29 ? 25% +8.1 38.42 ? 4% perf-stat.overall.cache-miss-rate%
1.28 ? 2% +232.8% 4.26 perf-stat.overall.cpi
1399 ? 5% +8.8% 1522 ? 3% perf-stat.overall.cycles-between-cache-misses
0.78 ? 2% -70.0% 0.23 perf-stat.overall.ipc
94.87 +4.0 98.92 perf-stat.overall.node-load-miss-rate%
56.43 ? 4% +14.8 71.21 perf-stat.overall.node-store-miss-rate%
4.699e+09 -71.1% 1.359e+09 perf-stat.ps.branch-instructions
15853475 ? 32% -72.6% 4344260 ? 4% perf-stat.ps.branch-misses
19790250 ? 3% -8.7% 18072625 ? 3% perf-stat.ps.cache-misses
72066501 ? 37% -34.7% 47079258 ? 2% perf-stat.ps.cache-references
10100 ? 4% -50.5% 4999 ? 7% perf-stat.ps.context-switches
120.98 -8.1% 111.24 perf-stat.ps.cpu-migrations
7.486e+11 ?199% -100.0% 198976 ? 46% perf-stat.ps.dTLB-load-misses
5.102e+09 -70.1% 1.526e+09 perf-stat.ps.dTLB-loads
105425 ? 95% -66.8% 34991 ? 23% perf-stat.ps.dTLB-store-misses
2.524e+09 -76.9% 5.837e+08 ? 2% perf-stat.ps.dTLB-stores
2.16e+10 -70.1% 6.454e+09 perf-stat.ps.instructions
3850372 ? 5% +52.1% 5857127 ? 3% perf-stat.ps.node-load-misses
211823 ? 35% -69.9% 63770 ? 4% perf-stat.ps.node-loads
3125462 ? 9% +152.0% 7875415 ? 3% perf-stat.ps.node-store-misses
2404750 ? 4% +32.4% 3183125 ? 2% perf-stat.ps.node-stores
1.364e+12 -70.1% 4.079e+11 perf-stat.total.instructions
18334 ? 40% -45.8% 9942 ? 27% softirqs.CPU0.RCU
25205 ? 59% -74.2% 6513 ? 35% softirqs.CPU1.RCU
11337 ? 24% -64.5% 4025 ? 16% softirqs.CPU10.RCU
3467 ?128% -93.0% 243.40 ? 45% softirqs.CPU11.NET_RX
9325 ? 16% -53.9% 4303 ? 31% softirqs.CPU12.RCU
18711 ? 69% -76.9% 4323 ? 42% softirqs.CPU14.RCU
11424 ? 24% -68.3% 3616 ? 23% softirqs.CPU16.RCU
10296 ? 10% -64.4% 3663 ? 15% softirqs.CPU17.RCU
16686 ? 80% -73.0% 4497 ? 48% softirqs.CPU18.RCU
22801 ? 52% -75.3% 5620 ? 51% softirqs.CPU19.RCU
19856 ? 66% -76.0% 4775 ? 32% softirqs.CPU2.RCU
17692 ? 63% -67.3% 5787 ? 51% softirqs.CPU23.RCU
10098 ? 11% -24.2% 7656 ? 19% softirqs.CPU24.SCHED
28140 ? 38% -79.2% 5848 ? 77% softirqs.CPU26.RCU
25221 ? 51% -76.7% 5889 ? 64% softirqs.CPU27.RCU
19503 ? 56% -61.0% 7614 ? 53% softirqs.CPU28.RCU
21536 ? 32% -85.6% 3093 ? 13% softirqs.CPU29.RCU
22529 ? 40% -70.4% 6667 ? 64% softirqs.CPU3.RCU
16623 ? 48% -81.0% 3155 ? 14% softirqs.CPU30.RCU
12194 ? 36% -72.9% 3299 ? 3% softirqs.CPU31.RCU
14779 ? 66% -68.1% 4710 ? 41% softirqs.CPU35.RCU
15014 ? 48% -78.2% 3276 ? 13% softirqs.CPU39.RCU
13379 ? 53% -73.9% 3493 ? 28% softirqs.CPU41.RCU
13192 ? 44% -70.0% 3964 ? 45% softirqs.CPU42.RCU
8842 ? 27% -55.8% 3907 ? 44% softirqs.CPU44.RCU
10412 ? 21% -61.1% 4054 ? 24% softirqs.CPU47.RCU
24935 ? 43% -67.6% 8068 ? 75% softirqs.CPU49.RCU
14355 ? 23% -63.9% 5179 ? 77% softirqs.CPU5.RCU
24979 ? 74% -83.7% 4082 ? 28% softirqs.CPU50.RCU
16596 ? 35% -72.4% 4583 ? 37% softirqs.CPU51.RCU
14924 ? 37% -69.3% 4587 ? 45% softirqs.CPU52.RCU
10496 ? 17% -67.7% 3391 ? 23% softirqs.CPU54.RCU
11523 ? 27% -63.5% 4210 ? 37% softirqs.CPU58.RCU
10856 ? 23% -69.6% 3300 ? 14% softirqs.CPU6.RCU
9567 ? 12% -55.3% 4278 ? 34% softirqs.CPU60.RCU
9623 ? 13% -65.3% 3339 ? 20% softirqs.CPU61.RCU
17635 ? 52% -76.6% 4123 ? 38% softirqs.CPU62.RCU
11383 ? 23% -67.1% 3748 ? 16% softirqs.CPU64.RCU
12070 ? 16% -66.9% 3989 ? 9% softirqs.CPU65.RCU
13676 ? 50% -66.3% 4608 ? 54% softirqs.CPU66.RCU
25554 ? 51% -80.7% 4932 ? 38% softirqs.CPU67.RCU
19115 ? 76% -68.3% 6051 ? 34% softirqs.CPU68.RCU
22577 ? 40% -74.8% 5681 ? 50% softirqs.CPU72.RCU
28142 ? 51% -78.7% 6003 ? 56% softirqs.CPU73.RCU
25500 ? 37% -86.6% 3423 ? 28% softirqs.CPU74.RCU
17326 ? 38% -73.1% 4669 ? 36% softirqs.CPU75.RCU
20427 ? 44% -76.4% 4825 ? 34% softirqs.CPU76.RCU
22440 ? 35% -86.0% 3148 ? 14% softirqs.CPU77.RCU
15102 ? 31% -79.1% 3155 ? 12% softirqs.CPU78.RCU
12795 ? 27% -75.0% 3204 ? 11% softirqs.CPU79.RCU
15211 ? 66% -75.0% 3805 ? 23% softirqs.CPU83.RCU
17019 ? 73% -81.4% 3162 ? 30% softirqs.CPU86.RCU
14440 ? 49% -78.2% 3154 ? 21% softirqs.CPU87.RCU
15354 ? 60% -74.2% 3957 ? 61% softirqs.CPU88.RCU
16994 ? 65% -79.9% 3411 ? 32% softirqs.CPU89.RCU
13048 ? 47% -69.2% 4016 ? 59% softirqs.CPU90.RCU
13869 ? 57% -75.8% 3360 ? 29% softirqs.CPU91.RCU
11657 ? 57% -71.4% 3328 ? 32% softirqs.CPU92.RCU
9599 ? 20% -56.5% 4175 ? 43% softirqs.CPU93.RCU
9690 ? 12% -69.3% 2974 ? 15% softirqs.CPU94.RCU
9935 ? 13% -56.1% 4358 ? 48% softirqs.CPU95.RCU
1526333 ? 5% -67.9% 489535 ? 7% softirqs.RCU
59.46 ? 3% -56.6 2.84 ? 4% perf-profile.calltrace.cycles-pp.rename
59.31 ? 3% -56.5 2.80 ? 4% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.rename
59.28 ? 3% -56.5 2.80 ? 4% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.rename
59.20 ? 3% -56.4 2.78 ? 4% perf-profile.calltrace.cycles-pp.__x64_sys_rename.do_syscall_64.entry_SYSCALL_64_after_hwframe.rename
58.75 ? 3% -56.1 2.66 ? 4% perf-profile.calltrace.cycles-pp.do_renameat2.__x64_sys_rename.do_syscall_64.entry_SYSCALL_64_after_hwframe.rename
47.67 ? 3% -47.7 0.00 perf-profile.calltrace.cycles-pp.lock_rename.do_renameat2.__x64_sys_rename.do_syscall_64.entry_SYSCALL_64_after_hwframe
46.96 ? 3% -47.0 0.00 perf-profile.calltrace.cycles-pp.__mutex_lock.lock_rename.do_renameat2.__x64_sys_rename.do_syscall_64
36.08 ? 3% -36.1 0.00 perf-profile.calltrace.cycles-pp.osq_lock.__mutex_lock.lock_rename.do_renameat2.__x64_sys_rename
11.08 ? 9% -11.1 0.00 perf-profile.calltrace.cycles-pp.xfs_inodegc_flush.xfs_fs_statfs.statfs_by_dentry.vfs_statfs.user_statfs
9.71 ? 2% -9.7 0.00 perf-profile.calltrace.cycles-pp.mutex_spin_on_owner.__mutex_lock.lock_rename.do_renameat2.__x64_sys_rename
11.05 ? 4% -8.6 2.43 ? 4% perf-profile.calltrace.cycles-pp.__percpu_counter_sum.xfs_fs_statfs.statfs_by_dentry.vfs_statfs.user_statfs
6.60 ? 15% -6.6 0.00 perf-profile.calltrace.cycles-pp.__flush_work.xfs_inodegc_flush.xfs_fs_statfs.statfs_by_dentry.vfs_statfs
7.71 ? 2% -5.9 1.86 ? 4% perf-profile.calltrace.cycles-pp.vfs_rename.do_renameat2.__x64_sys_rename.do_syscall_64.entry_SYSCALL_64_after_hwframe
6.47 ? 2% -4.9 1.53 ? 4% perf-profile.calltrace.cycles-pp.xfs_vn_rename.vfs_rename.do_renameat2.__x64_sys_rename.do_syscall_64
6.43 ? 2% -4.9 1.53 ? 4% perf-profile.calltrace.cycles-pp.xfs_rename.xfs_vn_rename.vfs_rename.do_renameat2.__x64_sys_rename
3.87 ? 2% -2.9 0.97 ? 23% perf-profile.calltrace.cycles-pp.cpumask_next.__percpu_counter_sum.xfs_fs_statfs.statfs_by_dentry.vfs_statfs
2.49 ? 3% -1.9 0.62 ? 3% perf-profile.calltrace.cycles-pp.__xfs_trans_commit.xfs_rename.xfs_vn_rename.vfs_rename.do_renameat2
2.28 ? 2% -1.7 0.59 ? 4% perf-profile.calltrace.cycles-pp.xlog_cil_commit.__xfs_trans_commit.xfs_rename.xfs_vn_rename.vfs_rename
0.00 +9.1 9.15 ? 2% perf-profile.calltrace.cycles-pp._raw_spin_lock_irq.flush_workqueue_prep_pwqs.flush_workqueue.xfs_fs_statfs.statfs_by_dentry
0.00 +10.1 10.14 ? 2% perf-profile.calltrace.cycles-pp.mutex_spin_on_owner.__mutex_lock.flush_workqueue.xfs_fs_statfs.statfs_by_dentry
0.00 +10.2 10.21 ? 2% perf-profile.calltrace.cycles-pp.flush_workqueue_prep_pwqs.flush_workqueue.xfs_fs_statfs.statfs_by_dentry.vfs_statfs
13.53 ? 2% +27.9 41.41 perf-profile.calltrace.cycles-pp.statvfs64
13.47 ? 2% +27.9 41.39 perf-profile.calltrace.cycles-pp.__statfs.statvfs64
13.44 ? 2% +27.9 41.38 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__statfs.statvfs64
13.44 ? 2% +27.9 41.38 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__statfs.statvfs64
13.41 ? 2% +28.0 41.37 perf-profile.calltrace.cycles-pp.__do_sys_statfs.do_syscall_64.entry_SYSCALL_64_after_hwframe.__statfs.statvfs64
13.41 +30.4 43.83 perf-profile.calltrace.cycles-pp.__statfs
13.21 +30.6 43.78 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__statfs
13.18 +30.6 43.77 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__statfs
13.09 +30.6 43.74 perf-profile.calltrace.cycles-pp.__do_sys_statfs.do_syscall_64.entry_SYSCALL_64_after_hwframe.__statfs
26.34 +58.7 85.07 perf-profile.calltrace.cycles-pp.user_statfs.__do_sys_statfs.do_syscall_64.entry_SYSCALL_64_after_hwframe.__statfs
0.00 +59.8 59.78 perf-profile.calltrace.cycles-pp.osq_lock.__mutex_lock.flush_workqueue.xfs_fs_statfs.statfs_by_dentry
23.44 ? 2% +61.0 84.49 perf-profile.calltrace.cycles-pp.vfs_statfs.user_statfs.__do_sys_statfs.do_syscall_64.entry_SYSCALL_64_after_hwframe
23.40 ? 2% +61.1 84.47 perf-profile.calltrace.cycles-pp.statfs_by_dentry.vfs_statfs.user_statfs.__do_sys_statfs.do_syscall_64
23.25 ? 2% +61.2 84.44 perf-profile.calltrace.cycles-pp.xfs_fs_statfs.statfs_by_dentry.vfs_statfs.user_statfs.__do_sys_statfs
0.00 +70.5 70.49 perf-profile.calltrace.cycles-pp.__mutex_lock.flush_workqueue.xfs_fs_statfs.statfs_by_dentry.vfs_statfs
0.00 +81.1 81.11 perf-profile.calltrace.cycles-pp.flush_workqueue.xfs_fs_statfs.statfs_by_dentry.vfs_statfs.user_statfs
59.50 ? 3% -56.6 2.85 ? 4% perf-profile.children.cycles-pp.rename
59.21 ? 3% -56.4 2.78 ? 4% perf-profile.children.cycles-pp.__x64_sys_rename
58.78 ? 3% -56.1 2.67 ? 4% perf-profile.children.cycles-pp.do_renameat2
47.67 ? 3% -47.6 0.10 ? 21% perf-profile.children.cycles-pp.lock_rename
11.20 ? 9% -10.5 0.75 ? 4% perf-profile.children.cycles-pp.xfs_inodegc_flush
11.50 ? 4% -9.0 2.55 ? 4% perf-profile.children.cycles-pp.__percpu_counter_sum
7.15 ? 14% -7.2 0.00 perf-profile.children.cycles-pp.__flush_work
7.73 ? 2% -5.9 1.87 ? 4% perf-profile.children.cycles-pp.vfs_rename
7.34 -5.6 1.75 ? 4% perf-profile.children.cycles-pp.cpumask_next
6.47 ? 2% -4.9 1.54 ? 4% perf-profile.children.cycles-pp.xfs_vn_rename
6.45 ? 2% -4.9 1.54 ? 4% perf-profile.children.cycles-pp.xfs_rename
4.65 -3.5 1.12 ? 3% perf-profile.children.cycles-pp._find_next_bit
2.71 ? 2% -2.0 0.75 ? 4% perf-profile.children.cycles-pp.xfs_inodegc_queue_all
2.51 ? 2% -1.9 0.63 ? 4% perf-profile.children.cycles-pp.__xfs_trans_commit
2.12 ? 7% -1.7 0.39 ? 4% perf-profile.children.cycles-pp._raw_spin_lock_irqsave
2.31 ? 2% -1.7 0.61 ? 4% perf-profile.children.cycles-pp.xlog_cil_commit
2.08 ? 3% -1.7 0.41 perf-profile.children.cycles-pp._raw_spin_lock
1.78 ? 6% -1.5 0.30 ? 3% perf-profile.children.cycles-pp.user_path_at_empty
1.40 ? 3% -1.3 0.10 ? 9% perf-profile.children.cycles-pp.__might_resched
1.66 ? 5% -1.3 0.39 ? 10% perf-profile.children.cycles-pp.dput
1.33 ? 7% -1.1 0.18 ? 4% perf-profile.children.cycles-pp.filename_lookup
1.44 ? 4% -1.1 0.32 ? 7% perf-profile.children.cycles-pp.xfs_trans_alloc
1.29 ? 7% -1.1 0.17 ? 4% perf-profile.children.cycles-pp.path_lookupat
1.35 ? 4% -1.1 0.29 ? 3% perf-profile.children.cycles-pp.__lookup_hash
1.28 ? 3% -1.0 0.29 ? 7% perf-profile.children.cycles-pp.xfs_trans_reserve
0.99 ? 4% -1.0 0.04 ? 50% perf-profile.children.cycles-pp.__cond_resched
1.01 ? 9% -0.9 0.10 ? 7% perf-profile.children.cycles-pp.complete_walk
1.16 ? 6% -0.9 0.27 ? 12% perf-profile.children.cycles-pp.lockref_put_return
0.97 ? 10% -0.9 0.10 ? 10% perf-profile.children.cycles-pp.try_to_unlazy
0.92 ? 10% -0.8 0.08 ? 9% perf-profile.children.cycles-pp.__legitimize_path
0.90 ? 4% -0.8 0.07 ? 9% perf-profile.children.cycles-pp.__might_sleep
1.09 ? 5% -0.8 0.27 ? 13% perf-profile.children.cycles-pp.path_put
0.97 ? 6% -0.8 0.19 ? 9% perf-profile.children.cycles-pp.xfs_log_reserve
0.83 ? 11% -0.8 0.05 ? 9% perf-profile.children.cycles-pp.lockref_get_not_dead
0.98 ? 5% -0.7 0.28 ? 4% perf-profile.children.cycles-pp.xlog_cil_insert_items
0.88 ? 3% -0.6 0.23 ? 5% perf-profile.children.cycles-pp.filename_parentat
0.83 ? 2% -0.6 0.22 ? 6% perf-profile.children.cycles-pp.path_parentat
0.84 -0.6 0.23 ? 4% perf-profile.children.cycles-pp.xfs_trans_log_inode
0.80 -0.6 0.22 ? 6% perf-profile.children.cycles-pp.getname_flags
0.66 ? 2% -0.5 0.15 ? 7% perf-profile.children.cycles-pp.kmem_cache_alloc
0.70 ? 2% -0.5 0.19 ? 8% perf-profile.children.cycles-pp.xfs_dir_createname
0.65 ? 3% -0.5 0.17 ? 13% perf-profile.children.cycles-pp.__vsnprintf_chk
0.59 ? 8% -0.5 0.11 ? 7% perf-profile.children.cycles-pp.xfs_vn_lookup
0.60 ? 4% -0.4 0.17 ? 4% perf-profile.children.cycles-pp.link_path_walk
0.54 ? 2% -0.4 0.15 ? 7% perf-profile.children.cycles-pp.strncpy_from_user
0.54 ? 4% -0.4 0.15 ? 10% perf-profile.children.cycles-pp.xfs_dir2_sf_addname
0.53 ? 6% -0.4 0.14 ? 6% perf-profile.children.cycles-pp.d_move
0.53 ? 2% -0.4 0.16 ? 4% perf-profile.children.cycles-pp.xfs_log_ticket_ungrant
0.49 ? 5% -0.4 0.12 ? 11% perf-profile.children.cycles-pp.vfprintf
1.22 ? 38% -0.3 0.87 ? 3% perf-profile.children.cycles-pp.__hrtimer_run_queues
0.47 ? 5% -0.3 0.14 ? 4% perf-profile.children.cycles-pp.xfs_inode_item_format
0.42 ? 4% -0.3 0.10 ? 11% perf-profile.children.cycles-pp.path_init
0.76 ? 11% -0.3 0.46 ? 5% perf-profile.children.cycles-pp.__softirqentry_text_start
0.40 ? 4% -0.3 0.10 ? 11% perf-profile.children.cycles-pp.lookup_dcache
0.39 ? 4% -0.3 0.10 ? 9% perf-profile.children.cycles-pp.d_lookup
0.39 ? 12% -0.3 0.10 ? 8% perf-profile.children.cycles-pp.xfs_mod_fdblocks
0.36 ? 13% -0.3 0.09 ? 13% perf-profile.children.cycles-pp.xfs_lookup
0.35 ? 3% -0.3 0.08 ? 5% perf-profile.children.cycles-pp.d_alloc
0.35 ? 13% -0.3 0.09 ? 13% perf-profile.children.cycles-pp.xfs_dir_lookup
0.76 ? 19% -0.3 0.50 ? 7% perf-profile.children.cycles-pp.irq_exit_rcu
0.34 ? 6% -0.3 0.09 perf-profile.children.cycles-pp.xfs_dir_removename
0.72 ? 43% -0.2 0.48 ? 4% perf-profile.children.cycles-pp.tick_sched_handle
0.32 ? 5% -0.2 0.08 ? 12% perf-profile.children.cycles-pp._IO_default_xsputn
0.31 ? 10% -0.2 0.08 ? 7% perf-profile.children.cycles-pp.rcu_core
0.28 ? 3% -0.2 0.05 perf-profile.children.cycles-pp.__d_alloc
0.31 ? 6% -0.2 0.08 ? 5% perf-profile.children.cycles-pp.inode_permission
0.69 ? 40% -0.2 0.47 ? 3% perf-profile.children.cycles-pp.update_process_times
0.31 ? 6% -0.2 0.10 ? 9% perf-profile.children.cycles-pp.xlog_grant_add_space
0.28 ? 13% -0.2 0.07 ? 10% perf-profile.children.cycles-pp.down_read
0.24 ? 7% -0.2 0.07 ? 11% perf-profile.children.cycles-pp.__d_move
0.29 ? 10% -0.2 0.12 ? 8% perf-profile.children.cycles-pp.kthread
0.21 ? 7% -0.2 0.05 ? 50% perf-profile.children.cycles-pp.rcu_do_batch
0.29 ? 10% -0.2 0.13 ? 6% perf-profile.children.cycles-pp.ret_from_fork
0.19 ? 6% -0.2 0.03 ? 81% perf-profile.children.cycles-pp.xlog_space_left
0.19 ? 7% -0.2 0.03 ? 82% perf-profile.children.cycles-pp.xlog_grant_push_ail
0.23 ? 6% -0.2 0.07 ? 7% perf-profile.children.cycles-pp.kmem_cache_free
0.21 ? 4% -0.2 0.05 ? 9% perf-profile.children.cycles-pp.__kmalloc
0.19 ? 6% -0.2 0.03 ? 82% perf-profile.children.cycles-pp.xlog_grant_push_threshold
0.22 ? 4% -0.2 0.06 ? 12% perf-profile.children.cycles-pp.__check_object_size
0.21 ? 4% -0.2 0.06 ? 6% perf-profile.children.cycles-pp.kmem_alloc
0.21 ? 9% -0.2 0.06 ? 6% perf-profile.children.cycles-pp.xfs_dir2_sf_removename
0.19 ? 6% -0.1 0.05 perf-profile.children.cycles-pp.make_kuid
0.20 ? 7% -0.1 0.05 ? 9% perf-profile.children.cycles-pp.map_id_range_down
0.19 ? 6% -0.1 0.05 ? 9% perf-profile.children.cycles-pp.xfs_lock_inodes
0.29 ? 7% -0.1 0.15 ? 12% perf-profile.children.cycles-pp.mutex_unlock
0.24 ? 8% -0.1 0.10 ? 9% perf-profile.children.cycles-pp.__d_lookup
0.17 ? 7% -0.1 0.03 ? 81% perf-profile.children.cycles-pp.__entry_text_start
0.20 ? 5% -0.1 0.07 ? 12% perf-profile.children.cycles-pp.__schedule
0.19 ? 7% -0.1 0.05 ? 7% perf-profile.children.cycles-pp.generic_permission
0.18 ? 9% -0.1 0.04 ? 51% perf-profile.children.cycles-pp.schedule
0.18 ? 6% -0.1 0.05 ? 52% perf-profile.children.cycles-pp.fsnotify_get_cookie
0.35 ? 4% -0.1 0.23 ? 6% perf-profile.children.cycles-pp.mutex_lock
0.17 ? 6% -0.1 0.05 perf-profile.children.cycles-pp.syscall_return_via_sysret
0.40 ? 24% -0.1 0.29 ? 3% perf-profile.children.cycles-pp.scheduler_tick
0.23 ? 9% -0.1 0.12 ? 8% perf-profile.children.cycles-pp.load_balance
0.20 ? 5% -0.1 0.10 ? 7% perf-profile.children.cycles-pp.find_busiest_group
0.19 ? 5% -0.1 0.10 ? 7% perf-profile.children.cycles-pp.update_sd_lb_stats
0.16 ? 34% -0.1 0.09 ? 11% perf-profile.children.cycles-pp.percpu_counter_add_batch
0.12 ? 61% -0.1 0.05 ? 52% perf-profile.children.cycles-pp.rcu_sched_clock_irq
0.17 ? 11% -0.1 0.10 ? 11% perf-profile.children.cycles-pp.osq_unlock
0.13 ? 18% -0.1 0.06 ? 56% perf-profile.children.cycles-pp.start_kernel
0.10 ? 28% -0.1 0.04 ? 50% perf-profile.children.cycles-pp.process_one_work
0.12 ? 26% -0.1 0.07 ? 12% perf-profile.children.cycles-pp.worker_thread
0.12 ? 17% -0.0 0.07 ? 6% perf-profile.children.cycles-pp.update_rq_clock
0.07 ? 31% -0.0 0.03 ? 81% perf-profile.children.cycles-pp.update_blocked_averages
0.07 ? 37% -0.0 0.04 ? 50% perf-profile.children.cycles-pp.run_rebalance_domains
0.00 +0.1 0.08 ? 15% perf-profile.children.cycles-pp._IO_setb
9.72 ? 2% +0.4 10.16 perf-profile.children.cycles-pp.mutex_spin_on_owner
86.17 ? 2% +2.0 88.15 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
86.13 ? 2% +2.0 88.13 perf-profile.children.cycles-pp.do_syscall_64
2.35 ? 40% +6.9 9.23 ? 2% perf-profile.children.cycles-pp._raw_spin_lock_irq
0.00 +10.3 10.26 ? 2% perf-profile.children.cycles-pp.flush_workqueue_prep_pwqs
46.97 ? 3% +23.5 70.52 perf-profile.children.cycles-pp.__mutex_lock
36.10 ? 3% +23.7 59.81 perf-profile.children.cycles-pp.osq_lock
13.54 ? 2% +27.9 41.41 perf-profile.children.cycles-pp.statvfs64
26.95 +58.3 85.24 perf-profile.children.cycles-pp.__statfs
26.51 +58.6 85.12 perf-profile.children.cycles-pp.__do_sys_statfs
26.34 +58.7 85.07 perf-profile.children.cycles-pp.user_statfs
23.44 ? 2% +61.0 84.49 perf-profile.children.cycles-pp.vfs_statfs
23.40 ? 2% +61.1 84.47 perf-profile.children.cycles-pp.statfs_by_dentry
23.26 ? 2% +61.2 84.44 perf-profile.children.cycles-pp.xfs_fs_statfs
0.00 +81.1 81.12 perf-profile.children.cycles-pp.flush_workqueue
5.07 ? 6% -4.0 1.04 ? 4% perf-profile.self.cycles-pp.__percpu_counter_sum
3.96 -3.0 0.96 ? 3% perf-profile.self.cycles-pp._find_next_bit
2.75 ? 2% -2.1 0.65 ? 6% perf-profile.self.cycles-pp.cpumask_next
2.04 ? 3% -1.6 0.40 perf-profile.self.cycles-pp._raw_spin_lock
1.28 ? 3% -1.2 0.10 ? 14% perf-profile.self.cycles-pp.__might_resched
1.44 ? 6% -1.1 0.38 ? 4% perf-profile.self.cycles-pp._raw_spin_lock_irqsave
1.15 ? 6% -0.9 0.27 ? 12% perf-profile.self.cycles-pp.lockref_put_return
1.21 ? 2% -0.9 0.34 ? 5% perf-profile.self.cycles-pp.xfs_inodegc_queue_all
0.82 ? 11% -0.8 0.05 ? 9% perf-profile.self.cycles-pp.lockref_get_not_dead
0.73 ? 5% -0.7 0.06 ? 6% perf-profile.self.cycles-pp.__might_sleep
0.62 ? 4% -0.4 0.17 ? 8% perf-profile.self.cycles-pp.xfs_trans_log_inode
0.86 ? 2% -0.4 0.43 ? 3% perf-profile.self.cycles-pp.__mutex_lock
0.52 ? 2% -0.4 0.16 ? 4% perf-profile.self.cycles-pp.xfs_log_ticket_ungrant
0.47 ? 5% -0.4 0.11 ? 10% perf-profile.self.cycles-pp.vfprintf
0.41 ? 4% -0.3 0.10 ? 10% perf-profile.self.cycles-pp.path_init
0.30 ? 6% -0.2 0.09 ? 8% perf-profile.self.cycles-pp.xfs_inode_item_format
0.31 ? 6% -0.2 0.10 ? 9% perf-profile.self.cycles-pp.xlog_grant_add_space
0.28 ? 6% -0.2 0.08 ? 12% perf-profile.self.cycles-pp._IO_default_xsputn
0.26 ? 2% -0.2 0.07 ? 15% perf-profile.self.cycles-pp.strncpy_from_user
0.24 ? 14% -0.2 0.06 ? 12% perf-profile.self.cycles-pp.down_read
0.23 ? 3% -0.2 0.06 ? 12% perf-profile.self.cycles-pp.kmem_cache_alloc
0.22 ? 3% -0.2 0.06 ? 15% perf-profile.self.cycles-pp.link_path_walk
0.19 ? 5% -0.2 0.03 ? 81% perf-profile.self.cycles-pp.xlog_space_left
0.29 ? 6% -0.1 0.15 ? 12% perf-profile.self.cycles-pp.mutex_unlock
0.18 ? 5% -0.1 0.05 ? 52% perf-profile.self.cycles-pp.fsnotify_get_cookie
0.18 ? 6% -0.1 0.05 perf-profile.self.cycles-pp.map_id_range_down
0.22 ? 8% -0.1 0.09 ? 12% perf-profile.self.cycles-pp.__d_lookup
0.17 ? 6% -0.1 0.05 perf-profile.self.cycles-pp.syscall_return_via_sysret
0.33 ? 4% -0.1 0.21 ? 4% perf-profile.self.cycles-pp.mutex_lock
0.17 ? 11% -0.1 0.10 ? 11% perf-profile.self.cycles-pp.osq_unlock
0.14 ? 7% -0.1 0.07 ? 9% perf-profile.self.cycles-pp.update_sd_lb_stats
0.15 ? 34% -0.1 0.09 ? 11% perf-profile.self.cycles-pp.percpu_counter_add_batch
0.13 ? 8% -0.0 0.11 ? 8% perf-profile.self.cycles-pp.native_sched_clock
0.00 +0.1 0.08 ? 12% perf-profile.self.cycles-pp._IO_setb
9.67 ? 2% +0.5 10.13 ? 2% perf-profile.self.cycles-pp.mutex_spin_on_owner
0.00 +1.1 1.07 ? 4% perf-profile.self.cycles-pp.flush_workqueue_prep_pwqs
2.33 ? 40% +6.9 9.18 ? 2% perf-profile.self.cycles-pp._raw_spin_lock_irq
35.94 ? 3% +23.7 59.65 perf-profile.self.cycles-pp.osq_lock
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.
---
0DAY/LKP+ Test Infrastructure Open Source Technology Center
https://lists.01.org/hyperkitty/list/[email protected] Intel Corporation
Thanks,
Oliver Sang
Hi,
On 2022/1/31 23:49, kernel test robot wrote:
> 0.00 +59.8 59.78 perf-profile.calltrace.cycles-pp.osq_lock.__mutex_lock.flush_workqueue.xfs_fs_statfs.statfs_by_dentry
> 23.44 ± 2% +61.0 84.49 perf-profile.calltrace.cycles-pp.vfs_statfs.user_statfs.__do_sys_statfs.do_syscall_64.entry_SYSCALL_64_after_hwframe
> 23.40 ± 2% +61.1 84.47 perf-profile.calltrace.cycles-pp.statfs_by_dentry.vfs_statfs.user_statfs.__do_sys_statfs.do_syscall_64
We did further check of the perf data. From above three lines,
the patch brought higher lock content (+59.8) in this extreme cases.
Regarding the patch could fix race issue, we will not tracking this
report. Let us know if you have any comments. Thanks.
Regards
Yin, Fengwei