hi, Al Viro, besides below performance report, we noticed stderr "A_dependency_job_for_nfs-idmapd.service_failed.See'journalctl-xe'for_details" while testing which we didn't observe on parent. we also noticed there are something like below in attached kmsg.xz: kern :warn : [ 61.938032] NFSD: Unable to initialize client recovery tracking! (-110) kern :info : [ 61.946579] NFSD: starting 90-second grace period (net f0000000) kern :warn : [ 63.146060] NFSD: Unable to initialize client recovery tracking! (-110) kern :info : [ 63.154531] NFSD: starting 90-second grace period (net f0000000) which we didn't observed in parent tests, either. we are not sure whether these impact the performance results, all these FYI. Hello, kernel test robot noticed a 17.5% regression of fsmark.app_overhead on: commit: 1d8f830b7f27ba92a700a86988695e965f73b15a ("rpc_pipe: get rid of __rpc_lookup_create_exclusive()") https://git.kernel.org/cgit/linux/kernel/git/viro/vfs.git work.simple_recursive_removal testcase: fsmark test machine: 48 threads 2 sockets Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz (Ivy Bridge-EP) with 112G memory parameters: iterations: 1x nr_threads: 32t disk: 1SSD fs: xfs fs2: nfsv4 filesize: 8K test_size: 400M sync_method: fsyncBeforeClose nr_directories: 16d nr_files_per_directory: 256fpd cpufreq_governor: performance In addition to that, the commit also has significant impact on the following tests: +------------------+-------------------------------------------------------------------------------------------------+ | testcase: change | fsmark: fsmark.app_overhead 20.3% regression | | test machine | 48 threads 2 sockets Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz (Ivy Bridge-EP) with 112G memory | | test parameters | cpufreq_governor=performance | | | disk=1SSD | | | filesize=9B | | | fs2=nfsv4 | | | fs=xfs | | | iterations=1x | | | nr_directories=16d | | | nr_files_per_directory=256fpd | | | nr_threads=32t | | | sync_method=fsyncBeforeClose | | | test_size=400M | +------------------+-------------------------------------------------------------------------------------------------+ 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@xxxxxxxxx> | Closes: https://lore.kernel.org/oe-lkp/202403011849.c1afb89c-oliver.sang@xxxxxxxxx Details are as below: --------------------------------------------------------------------------------------------------> The kernel config and materials to reproduce are available at: https://download.01.org/0day-ci/archive/20240301/202403011849.c1afb89c-oliver.sang@xxxxxxxxx ========================================================================================= compiler/cpufreq_governor/disk/filesize/fs2/fs/iterations/kconfig/nr_directories/nr_files_per_directory/nr_threads/rootfs/sync_method/tbox_group/test_size/testcase: gcc-12/performance/1SSD/8K/nfsv4/xfs/1x/x86_64-rhel-8.3/16d/256fpd/32t/debian-12-x86_64-20240206.cgz/fsyncBeforeClose/lkp-ivb-2ep1/400M/fsmark commit: 82a439e321 ("rpc_pipe: fold __rpc_mkpipe_dentry() into its sole caller") 1d8f830b7f ("rpc_pipe: get rid of __rpc_lookup_create_exclusive()") 82a439e3213e7095 1d8f830b7f27ba92a700a869886 ---------------- --------------------------- fail:runs %reproduction fail:runs | | | :6 100% 6:6 stderr.A_dependency_job_for_nfs-idmapd.service_failed.See'journalctl-xe'for_details :6 100% 6:6 stderr.has_stderr %stddev %change %stddev \ | \ 2398327 +12.3% 2692538 cpuidle..usage 107204 +140.2% 257509 ± 16% meminfo.AnonHugePages 304743 +9.2% 332637 ± 2% vmstat.system.cs 413300 ± 6% +14.4% 472827 ± 3% numa-numastat.node0.local_node 445362 ± 5% +12.9% 502812 ± 2% numa-numastat.node0.numa_hit 445128 ± 5% +13.1% 503240 ± 2% numa-vmstat.node0.numa_hit 413066 ± 6% +14.6% 473254 ± 3% numa-vmstat.node0.numa_local 66.15 ± 2% +164.6% 175.04 uptime.boot 2769 ± 2% +188.3% 7983 uptime.idle 65542525 +17.5% 77003387 fsmark.app_overhead 55.33 +4.2% 57.67 fsmark.time.percent_of_cpu_this_job_got 288191 +5.1% 302810 fsmark.time.voluntary_context_switches 74.50 +2.9% 76.68 iostat.cpu.idle 17.01 -10.9% 15.15 ± 2% iostat.cpu.iowait 2.59 -14.6% 2.21 iostat.cpu.user 19.75 -2.3 17.47 ± 2% mpstat.cpu.all.iowait% 2.63 -0.4 2.23 mpstat.cpu.all.usr% 12.34 ± 2% -17.5% 10.18 ± 2% mpstat.max_utilization_pct 53999 ± 2% +201.1% 162576 sched_debug.cpu_clk 53400 ± 2% +203.3% 161979 ± 2% sched_debug.ktime 54525 ± 2% +199.2% 163131 ± 2% sched_debug.sched_clk 52.21 +140.6% 125.62 ± 16% proc-vmstat.nr_anon_transparent_hugepages 8420 -3.2% 8154 proc-vmstat.nr_mapped 41883 +2.1% 42754 proc-vmstat.nr_slab_reclaimable 31420 +8.4% 34074 proc-vmstat.nr_slab_unreclaimable 704537 +7.9% 760346 proc-vmstat.numa_hit 654821 +8.5% 710630 proc-vmstat.numa_local 1039627 +6.1% 1103455 proc-vmstat.pgalloc_normal 729116 ± 6% +16.2% 846939 proc-vmstat.pgfree 1.13 ± 3% +9.8% 1.24 ± 2% perf-stat.i.MPKI 2.134e+09 -5.1% 2.026e+09 perf-stat.i.branch-instructions 4.84 -0.2 4.63 perf-stat.i.branch-miss-rate% 1.03e+08 -8.7% 94072597 perf-stat.i.branch-misses 6.16 ± 2% -0.3 5.86 ± 3% perf-stat.i.cache-miss-rate% 11302040 ± 2% +6.8% 12067382 ± 3% perf-stat.i.cache-misses 2.199e+08 +9.6% 2.409e+08 perf-stat.i.cache-references 360033 +14.8% 413385 perf-stat.i.context-switches 2239 ± 4% +13.0% 2530 ± 4% perf-stat.i.cpu-migrations 1472 ± 2% -6.0% 1384 ± 2% perf-stat.i.cycles-between-cache-misses 1.039e+10 -5.7% 9.795e+09 perf-stat.i.instructions 0.64 -5.0% 0.60 perf-stat.i.ipc 7.68 +14.3% 8.78 perf-stat.i.metric.K/sec 9527 ± 3% -13.3% 8258 ± 7% perf-stat.i.minor-faults 9528 ± 3% -13.3% 8258 ± 7% perf-stat.i.page-faults 1.09 ± 3% +13.1% 1.23 ± 2% perf-stat.overall.MPKI 4.83 -0.2 4.65 perf-stat.overall.branch-miss-rate% 1.59 +6.5% 1.70 perf-stat.overall.cpi 1464 ± 2% -5.9% 1378 ± 2% perf-stat.overall.cycles-between-cache-misses 0.63 -6.1% 0.59 perf-stat.overall.ipc 1.955e+09 -5.0% 1.857e+09 perf-stat.ps.branch-instructions 94404569 -8.6% 86246777 perf-stat.ps.branch-misses 10360850 ± 2% +6.8% 11061496 ± 3% perf-stat.ps.cache-misses 2.017e+08 +9.5% 2.209e+08 perf-stat.ps.cache-references 330250 +14.8% 379036 perf-stat.ps.context-switches 2054 ± 4% +12.9% 2320 ± 4% perf-stat.ps.cpu-migrations 9.517e+09 -5.7% 8.979e+09 perf-stat.ps.instructions 8697 ± 3% -13.1% 7557 ± 7% perf-stat.ps.minor-faults 8697 ± 3% -13.1% 7558 ± 7% perf-stat.ps.page-faults 1.151e+11 -6.0% 1.082e+11 perf-stat.total.instructions *************************************************************************************************** lkp-ivb-2ep1: 48 threads 2 sockets Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz (Ivy Bridge-EP) with 112G memory ========================================================================================= compiler/cpufreq_governor/disk/filesize/fs2/fs/iterations/kconfig/nr_directories/nr_files_per_directory/nr_threads/rootfs/sync_method/tbox_group/test_size/testcase: gcc-12/performance/1SSD/9B/nfsv4/xfs/1x/x86_64-rhel-8.3/16d/256fpd/32t/debian-12-x86_64-20240206.cgz/fsyncBeforeClose/lkp-ivb-2ep1/400M/fsmark commit: 82a439e321 ("rpc_pipe: fold __rpc_mkpipe_dentry() into its sole caller") 1d8f830b7f ("rpc_pipe: get rid of __rpc_lookup_create_exclusive()") 82a439e3213e7095 1d8f830b7f27ba92a700a869886 ---------------- --------------------------- fail:runs %reproduction fail:runs | | | :6 100% 6:6 stderr.A_dependency_job_for_nfs-idmapd.service_failed.See'journalctl-xe'for_details :6 100% 6:6 stderr.has_stderr %stddev %change %stddev \ | \ 4652582 +12.3% 5226642 cpuidle..usage 1.26 ± 74% +2.3 3.56 ± 31% perf-profile.calltrace.cycles-pp.mmap_region.do_mmap.vm_mmap_pgoff.ksys_mmap_pgoff.do_syscall_64 188802 ± 2% +63.3% 308290 ± 19% meminfo.AnonHugePages 140056 +10.9% 155322 meminfo.SUnreclaim 5.50 +0.6 6.05 ± 2% mpstat.cpu.all.sys% 12.67 -16.1% 10.63 mpstat.max_utilization_pct 655519 ± 2% +14.2% 748457 ± 2% numa-numastat.node0.local_node 680860 +14.6% 779999 ± 2% numa-numastat.node0.numa_hit 681047 +14.6% 780819 ± 2% numa-vmstat.node0.numa_hit 655706 ± 2% +14.3% 749281 ± 2% numa-vmstat.node0.numa_local 77.80 ± 3% +133.7% 181.80 uptime.boot 3182 ± 4% +156.6% 8168 uptime.idle 349730 +15.0% 402139 vmstat.system.cs 71199 +5.9% 75404 vmstat.system.in 1.219e+08 +20.3% 1.467e+08 fsmark.app_overhead 55.83 +3.3% 57.67 fsmark.time.percent_of_cpu_this_job_got 574683 +3.9% 597195 fsmark.time.voluntary_context_switches 18.20 -5.0% 17.30 iostat.cpu.iowait 6.17 +8.8% 6.72 iostat.cpu.system 2.48 -11.1% 2.21 iostat.cpu.user 56144 ± 4% +184.2% 159591 sched_debug.cpu_clk 55548 ± 4% +186.2% 158995 sched_debug.ktime 56677 ± 4% +182.5% 160122 sched_debug.sched_clk 91.96 ± 2% +64.1% 150.93 ± 18% proc-vmstat.nr_anon_transparent_hugepages 8527 -4.7% 8128 proc-vmstat.nr_mapped 35029 +10.9% 38832 proc-vmstat.nr_slab_unreclaimable 980307 +11.0% 1087887 proc-vmstat.numa_hit 930596 +11.4% 1036273 proc-vmstat.numa_local 1446464 +7.9% 1560287 proc-vmstat.pgalloc_normal 1.06 ± 2% +8.4% 1.15 ± 3% perf-stat.i.MPKI 5.00 -0.2 4.83 perf-stat.i.branch-miss-rate% 1.016e+08 -2.9% 98670000 perf-stat.i.branch-misses 10378740 ± 2% +9.1% 11323437 ± 3% perf-stat.i.cache-misses 2.348e+08 +9.3% 2.566e+08 perf-stat.i.cache-references 396967 +13.9% 452135 perf-stat.i.context-switches 1.70 +2.9% 1.74 perf-stat.i.cpi 1.647e+10 +4.9% 1.728e+10 perf-stat.i.cpu-cycles 2382 +13.4% 2702 ± 2% perf-stat.i.cpu-migrations 0.61 -4.0% 0.58 perf-stat.i.ipc 8.36 +14.0% 9.52 perf-stat.i.metric.K/sec 1.05 ± 3% +9.2% 1.14 ± 3% perf-stat.overall.MPKI 4.98 -0.2 4.83 perf-stat.overall.branch-miss-rate% 1.66 +5.1% 1.75 perf-stat.overall.cpi 0.60 -4.8% 0.57 perf-stat.overall.ipc 96867729 -2.8% 94186227 perf-stat.ps.branch-misses 9895781 ± 2% +9.2% 10807588 ± 3% perf-stat.ps.cache-misses 2.239e+08 +9.4% 2.449e+08 perf-stat.ps.cache-references 378576 +14.0% 431601 perf-stat.ps.context-switches 1.571e+10 +5.0% 1.65e+10 perf-stat.ps.cpu-cycles 2271 +13.5% 2579 ± 2% perf-stat.ps.cpu-migrations 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
Attachment:
kmsg.xz
Description: application/xz