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  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [day] [month] [year] [list]
Date:   Sun, 3 Jan 2021 23:33:02 +0800
From:   kernel test robot <oliver.sang@...el.com>
To:     Christoph Hellwig <hch@....de>
Cc:     Jens Axboe <axboe@...nel.dk>, Hannes Reinecke <hare@...e.de>,
        LKML <linux-kernel@...r.kernel.org>, lkp@...ts.01.org,
        lkp@...el.com, ying.huang@...el.com, feng.tang@...el.com,
        zhengjun.xing@...el.com
Subject: [loop]  3b4f85d02a:  stress-ng.loop.ops_per_sec -42.1% regression


Greeting,

FYI, we noticed a -42.1% regression of stress-ng.loop.ops_per_sec due to commit:


commit: 3b4f85d02a4bd85cbea999a064235a47694bbb7b ("loop: let set_capacity_revalidate_and_notify update the bdev size")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master


in testcase: stress-ng
on test machine: 48 threads Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz with 112G memory
with following parameters:

	nr_threads: 100%
	disk: 1HDD
	testtime: 30s
	class: device
	cpufreq_governor: performance
	ucode: 0x42e




If you fix the issue, kindly add following tag
Reported-by: kernel test robot <oliver.sang@...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

=========================================================================================
class/compiler/cpufreq_governor/disk/kconfig/nr_threads/rootfs/tbox_group/testcase/testtime/ucode:
  device/gcc-9/performance/1HDD/x86_64-rhel-8.3/100%/debian-10.4-x86_64-20200603.cgz/lkp-ivb-2ep1/stress-ng/30s/0x42e

commit: 
  99473d9db9 ("block: remove the call to __invalidate_device in check_disk_size_change")
  3b4f85d02a ("loop: let set_capacity_revalidate_and_notify update the bdev size")

99473d9db93a3bf5 3b4f85d02a4bd85cbea999a0642 
---------------- --------------------------- 
       fail:runs  %reproduction    fail:runs
           |             |             |    
          1:4          -25%            :4     kmsg.debugfs:Directory'loop#'with_parent'block'already_present
         %stddev     %change         %stddev
             \          |                \  
    692.75 ± 11%     -41.9%     402.75 ±  6%  stress-ng.loop.ops
     22.52 ± 11%     -42.1%      13.03 ±  5%  stress-ng.loop.ops_per_sec
     13959 ±125%    +173.5%      38175 ± 17%  numa-numastat.node1.other_node
 6.059e+08 ± 11%     -37.4%   3.79e+08 ± 53%  cpuidle.C1E.time
     43932 ± 36%     -40.2%      26251 ± 41%  cpuidle.POLL.usage
    128964 ±  3%     +15.6%     149048        meminfo.AnonHugePages
    245566 ±  3%     +12.4%     275979 ±  7%  meminfo.AnonPages
     10188 ±  4%     -13.4%       8820 ±  6%  softirqs.CPU19.SCHED
     10973 ±  6%     -14.3%       9403 ±  6%  softirqs.CPU2.SCHED
      7422 ±  7%      -9.0%       6753 ±  2%  numa-vmstat.node0.nr_kernel_stack
     31118 ± 12%     +28.5%      39987 ±  5%  numa-vmstat.node1.nr_anon_pages
     33253 ±  8%     +27.3%      42337 ±  6%  numa-vmstat.node1.nr_inactive_anon
     33254 ±  8%     +27.3%      42337 ±  6%  numa-vmstat.node1.nr_zone_inactive_anon
     51597 ±  4%     -16.6%      43024 ±  3%  slabinfo.kmalloc-8.active_objs
     51597 ±  4%     -16.6%      43024 ±  3%  slabinfo.kmalloc-8.num_objs
     11390 ±  4%      -6.5%      10645 ±  4%  slabinfo.proc_inode_cache.active_objs
      9897 ±  3%     +19.0%      11777 ± 11%  slabinfo.vmap_area.active_objs
      7411 ±  7%      -9.0%       6747 ±  2%  numa-meminfo.node0.KernelStack
    133681 ±  4%      -8.2%     122785 ±  2%  numa-meminfo.node0.Slab
    124467 ± 12%     +28.4%     159827 ±  5%  numa-meminfo.node1.AnonPages
    159232 ± 11%     +88.5%     300105 ± 30%  numa-meminfo.node1.AnonPages.max
    133328 ±  8%     +27.2%     169583 ±  6%  numa-meminfo.node1.Inactive
    133043 ±  8%     +27.4%     169517 ±  6%  numa-meminfo.node1.Inactive(anon)
      0.03 ± 30%     -33.0%       0.02 ± 11%  perf-sched.sch_delay.avg.ms.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop.0.do_select
      0.03 ± 43%     -33.6%       0.02 ±  3%  perf-sched.sch_delay.max.ms.__x64_sys_pause.do_syscall_64.entry_SYSCALL_64_after_hwframe.[unknown]
      0.03 ± 46%     -39.1%       0.02 ± 11%  perf-sched.sch_delay.max.ms.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop.0.do_select
     45.50 ± 58%     +56.6%      71.25 ±  2%  perf-sched.wait_and_delay.count.preempt_schedule_common._cond_resched.__anon_vma_prepare.do_anonymous_page.__handle_mm_fault
      0.33 ± 21%     -59.4%       0.13 ± 29%  perf-sched.wait_time.avg.ms.preempt_schedule_common._cond_resched.down_write_killable.vm_mmap_pgoff.ksys_mmap_pgoff
      8.33 ± 34%     -67.9%       2.68 ± 47%  perf-sched.wait_time.max.ms.preempt_schedule_common._cond_resched.down_write_killable.vm_mmap_pgoff.ksys_mmap_pgoff
     61412 ±  3%     +12.1%      68867 ±  7%  proc-vmstat.nr_anon_pages
     66214 ±  2%     +11.2%      73638 ±  6%  proc-vmstat.nr_inactive_anon
     66214 ±  2%     +11.2%      73638 ±  6%  proc-vmstat.nr_zone_inactive_anon
     23919 ± 22%     -21.4%      18797 ±  9%  proc-vmstat.numa_hint_faults
 1.101e+08            +1.3%  1.115e+08        proc-vmstat.pgalloc_normal
 1.101e+08            +1.3%  1.114e+08        proc-vmstat.pgfree
      3444 ±  2%     -26.6%       2527 ±  6%  proc-vmstat.unevictable_pgs_culled
   1395072           +29.1%    1800768 ±  7%  proc-vmstat.unevictable_pgs_scanned
      2.50 ± 22%      +0.8        3.25 ± 16%  perf-profile.calltrace.cycles-pp.drm_atomic_helper_dirtyfb.drm_fb_helper_dirty_work.process_one_work.worker_thread.kthread
      2.56 ± 22%      +0.8        3.33 ± 16%  perf-profile.calltrace.cycles-pp.drm_fb_helper_dirty_work.process_one_work.worker_thread.kthread.ret_from_fork
      2.67 ± 20%      +0.8        3.48 ± 16%  perf-profile.calltrace.cycles-pp.process_one_work.worker_thread.kthread.ret_from_fork
      2.72 ± 19%      +0.8        3.53 ± 16%  perf-profile.calltrace.cycles-pp.worker_thread.kthread.ret_from_fork
      3.35 ± 18%      +0.9        4.22 ± 10%  perf-profile.calltrace.cycles-pp.ret_from_fork
      3.34 ± 18%      +0.9        4.21 ± 10%  perf-profile.calltrace.cycles-pp.kthread.ret_from_fork
      0.12 ± 20%      +0.1        0.19 ± 32%  perf-profile.children.cycles-pp.copy_pte_range
      0.09 ± 40%      +0.1        0.17 ± 46%  perf-profile.children.cycles-pp.drm_atomic_get_plane_state
      0.12 ± 14%      +0.1        0.21 ± 40%  perf-profile.children.cycles-pp.zap_pte_range
      0.16 ± 17%      +0.1        0.25 ± 36%  perf-profile.children.cycles-pp.unmap_page_range
      2.50 ± 22%      +0.8        3.25 ± 16%  perf-profile.children.cycles-pp.drm_atomic_helper_dirtyfb
      2.56 ± 22%      +0.8        3.33 ± 16%  perf-profile.children.cycles-pp.drm_fb_helper_dirty_work
      2.67 ± 20%      +0.8        3.48 ± 16%  perf-profile.children.cycles-pp.process_one_work
      2.72 ± 19%      +0.8        3.53 ± 16%  perf-profile.children.cycles-pp.worker_thread
      3.34 ± 18%      +0.9        4.21 ± 10%  perf-profile.children.cycles-pp.kthread
      3.35 ± 18%      +0.9        4.23 ± 10%  perf-profile.children.cycles-pp.ret_from_fork
      0.07 ± 17%      +0.1        0.12 ± 38%  perf-profile.self.cycles-pp.copy_pte_range
 7.644e+09            +3.0%  7.871e+09        perf-stat.i.branch-instructions
      1.97 ±  4%      -0.4        1.58 ±  7%  perf-stat.i.branch-miss-rate%
  45318734            +6.3%   48178048 ±  3%  perf-stat.i.branch-misses
 1.694e+08            +2.2%   1.73e+08        perf-stat.i.cache-references
 7.529e+09            +2.5%   7.72e+09        perf-stat.i.dTLB-loads
 3.105e+09            +3.3%  3.208e+09        perf-stat.i.dTLB-stores
 3.075e+10            +2.6%  3.156e+10        perf-stat.i.instructions
     19679           -12.4%      17243 ± 17%  perf-stat.i.instructions-per-iTLB-miss
      0.48 ±  8%     +23.3%       0.60 ±  2%  perf-stat.i.metric.K/sec
    385.31            +2.8%     396.25        perf-stat.i.metric.M/sec
     39.32            +1.5       40.79 ±  2%  perf-stat.i.node-store-miss-rate%
      0.30            +3.0%       0.31        perf-stat.overall.ipc
 7.577e+09            +3.1%  7.809e+09        perf-stat.ps.branch-instructions
  45160627            +6.2%   47948462 ±  3%  perf-stat.ps.branch-misses
 3.072e+09            +3.3%  3.173e+09        perf-stat.ps.dTLB-stores
 3.051e+10            +2.7%  3.133e+10        perf-stat.ps.instructions
  5.64e+12            +2.9%  5.803e+12        perf-stat.total.instructions
     10782 ± 20%     -22.0%       8409 ± 18%  interrupts.CPU1.CAL:Function_call_interrupts
    521.75 ±  8%     +54.4%     805.75 ± 23%  interrupts.CPU10.RES:Rescheduling_interrupts
     10321 ±  2%     -14.4%       8833 ± 14%  interrupts.CPU14.TLB:TLB_shootdowns
      3912 ± 18%     -36.5%       2485 ± 14%  interrupts.CPU2.NMI:Non-maskable_interrupts
      3912 ± 18%     -36.5%       2485 ± 14%  interrupts.CPU2.PMI:Performance_monitoring_interrupts
    382.75 ±  8%    +404.0%       1929 ±115%  interrupts.CPU21.RES:Rescheduling_interrupts
      4267 ±  6%     -18.7%       3471 ± 19%  interrupts.CPU23.NMI:Non-maskable_interrupts
      4267 ±  6%     -18.7%       3471 ± 19%  interrupts.CPU23.PMI:Performance_monitoring_interrupts
    462.25 ± 13%     +52.6%     705.25 ± 20%  interrupts.CPU24.RES:Rescheduling_interrupts
    490.50 ±  4%     +18.3%     580.50 ± 10%  interrupts.CPU27.RES:Rescheduling_interrupts
    459.00 ±  5%    +254.8%       1628 ± 86%  interrupts.CPU30.RES:Rescheduling_interrupts
    443.50 ±  2%     +79.3%     795.00 ± 35%  interrupts.CPU35.RES:Rescheduling_interrupts
    531.25 ±  6%     +58.5%     842.00 ± 31%  interrupts.CPU4.RES:Rescheduling_interrupts
      2176 ±  9%     +75.8%       3825 ± 19%  interrupts.CPU40.NMI:Non-maskable_interrupts
      2176 ±  9%     +75.8%       3825 ± 19%  interrupts.CPU40.PMI:Performance_monitoring_interrupts
     16157 ± 72%     -51.5%       7832 ±  9%  interrupts.CPU43.CAL:Function_call_interrupts
     17624 ± 52%     -44.9%       9710 ± 13%  interrupts.CPU44.TLB:TLB_shootdowns
    389.00 ± 15%    +124.7%     874.25 ± 77%  interrupts.CPU47.RES:Rescheduling_interrupts
    533.50 ± 11%     +37.5%     733.75 ± 16%  interrupts.CPU5.RES:Rescheduling_interrupts
    496.75 ±  7%    +161.6%       1299 ± 77%  interrupts.CPU7.RES:Rescheduling_interrupts
      2794 ± 32%     +57.4%       4400 ± 12%  interrupts.CPU9.NMI:Non-maskable_interrupts
      2794 ± 32%     +57.4%       4400 ± 12%  interrupts.CPU9.PMI:Performance_monitoring_interrupts


                                                                                
                           stress-ng.loop.ops_per_sec                           
                                                                                
  30 +----------------------------------------------------------------------+   
     |                                                                      |   
  25 |-+                          +..            +..+                       |   
     |                +.. .+     :               :   +               +    +.|   
     |.+..+.+..+     +   +  :    :   +.+..  +.. :     +.+.. .+..+   + :  +  |   
  20 |-+       :    +       :  .+          +    :          +     + +  : +   |   
     |         :    :        +.           +    +                  +    +    |   
  15 |-+        :  :       O                                                |   
     | O  O O  O:  :O O  O   O  O O  O O  O O    O  O O                     |   
  10 |-+        :O :                           O                            |   
     |          : :                                                         |   
     |          : :                                                         |   
   5 |-+         ::                                                         |   
     |           :                                                          |   
   0 +----------------------------------------------------------------------+   
                                                                                
                                                                                
[*] bisect-good sample
[O] bisect-bad  sample



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,
Oliver Sang


View attachment "config-5.10.0-rc4-00031-g3b4f85d02a4b" of type "text/plain" (171021 bytes)

View attachment "job-script" of type "text/plain" (7994 bytes)

View attachment "job.yaml" of type "text/plain" (5584 bytes)

View attachment "reproduce" of type "text/plain" (392 bytes)

Powered by blists - more mailing lists