[linus:master] [memcg] 7e44d00a13: will-it-scale.per_thread_ops 2.6% regression
kernel test robot <[email protected]>
| Newsgroups | dev.linux.lists.oe-lkp,org.kernel.vger.cgroups,org.kernel.vger.linux-kernel,org.kvack.linux-mm |
|---|---|
| Message-ID | <[email protected]> |
Hello,
kernel test robot noticed a 2.6% regression of will-it-scale.per_thread_ops on:
commit: 7e44d00a13ca5691caf4f7c46541ee60bf75b208 ("memcg: use mod_node_page_state to update stats")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
[still regression on linux-next/master 6987d58a9cbc5bd57c983baa514474a86c945d56]
testcase: will-it-scale
config: x86_64-rhel-9.4
compiler: gcc-14
test machine: 64 threads 2 sockets Intel(R) Xeon(R) Gold 6346 CPU @ 3.10GHz (Ice Lake) with 256G memory
parameters:
nr_task: 100%
mode: thread
test: page_fault2
cpufreq_governor: performance
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 <[email protected]>
| Closes: https://lore.kernel.org/oe-lkp/[email protected]
Details are as below:
-------------------------------------------------------------------------------------------------->
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20251210/[email protected]
=========================================================================================
compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase:
gcc-14/performance/x86_64-rhel-9.4/thread/100%/debian-13-x86_64-20250902.cgz/lkp-icl-2sp7/page_fault2/will-it-scale
commit:
3e700b715e ("selftests/mm: gup_test: fix comment regarding origin of FOLL_WRITE")
7e44d00a13 ("memcg: use mod_node_page_state to update stats")
3e700b715e1cef66 7e44d00a13ca5691caf4f7c4654
---------------- ---------------------------
%stddev %change %stddev
\ | \
3453930 -2.6% 3363916 will-it-scale.64.threads
53967 -2.6% 52560 will-it-scale.per_thread_ops
3453930 -2.6% 3363916 will-it-scale.workload
1.053e+09 -2.6% 1.025e+09 proc-vmstat.numa_hit
1.052e+09 -2.6% 1.025e+09 proc-vmstat.numa_local
1.05e+09 -2.6% 1.023e+09 proc-vmstat.pgalloc_normal
1.045e+09 -2.6% 1.018e+09 proc-vmstat.pgfault
1.05e+09 -2.6% 1.023e+09 proc-vmstat.pgfree
3.452e+09 -2.0% 3.383e+09 perf-stat.i.branch-instructions
0.45 +0.0 0.46 perf-stat.i.branch-miss-rate%
4.559e+08 -2.5% 4.446e+08 perf-stat.i.cache-misses
4.696e+08 -2.5% 4.58e+08 perf-stat.i.cache-references
3.88e+10 -2.4% 3.787e+10 perf-stat.i.cpu-cycles
1.741e+10 -1.5% 1.715e+10 perf-stat.i.instructions
107.43 -2.5% 104.76 perf-stat.i.metric.K/sec
3437960 -2.5% 3352362 perf-stat.i.minor-faults
3437961 -2.5% 3352362 perf-stat.i.page-faults
26.18 -34.0% 17.29 ± 70% perf-stat.overall.MPKI
3.441e+09 -34.7% 2.247e+09 ± 70% perf-stat.ps.branch-instructions
4.544e+08 -35.0% 2.953e+08 ± 70% perf-stat.ps.cache-misses
4.68e+08 -35.0% 3.042e+08 ± 70% perf-stat.ps.cache-references
3.867e+10 -34.9% 2.517e+10 ± 70% perf-stat.ps.cpu-cycles
1.736e+10 -34.4% 1.139e+10 ± 70% perf-stat.ps.instructions
3426140 -35.0% 2226448 ± 70% perf-stat.ps.minor-faults
3426140 -35.0% 2226448 ± 70% perf-stat.ps.page-faults
5.293e+12 -34.4% 3.471e+12 ± 70% perf-stat.total.instructions
92.62 -0.2 92.40 perf-profile.calltrace.cycles-pp.asm_exc_page_fault.testcase
1.39 +0.0 1.42 perf-profile.calltrace.cycles-pp.__page_cache_release.folios_put_refs.free_pages_and_swap_cache.__tlb_batch_free_encoded_pages.tlb_flush_mmu
1.89 +0.0 1.92 perf-profile.calltrace.cycles-pp.lru_add.folio_batch_move_lru.__folio_batch_add_and_move.set_pte_range.finish_fault
0.91 +0.0 0.95 perf-profile.calltrace.cycles-pp.lru_gen_del_folio.__page_cache_release.folios_put_refs.free_pages_and_swap_cache.__tlb_batch_free_encoded_pages
1.30 +0.0 1.35 perf-profile.calltrace.cycles-pp.folio_remove_rmap_ptes.zap_present_ptes.zap_pte_range.zap_pmd_range.unmap_page_range
2.38 +0.0 2.43 perf-profile.calltrace.cycles-pp.folios_put_refs.free_pages_and_swap_cache.__tlb_batch_free_encoded_pages.tlb_flush_mmu.zap_pte_range
1.36 +0.1 1.41 perf-profile.calltrace.cycles-pp.lru_gen_add_folio.lru_add.folio_batch_move_lru.__folio_batch_add_and_move.set_pte_range
4.49 +0.2 4.64 perf-profile.calltrace.cycles-pp.zap_present_ptes.zap_pte_range.zap_pmd_range.unmap_page_range.unmap_vmas
7.62 +0.2 7.79 perf-profile.calltrace.cycles-pp.unmap_page_range.unmap_vmas.vms_clear_ptes.vms_complete_munmap_vmas.do_vmi_align_munmap
7.62 +0.2 7.79 perf-profile.calltrace.cycles-pp.unmap_vmas.vms_clear_ptes.vms_complete_munmap_vmas.do_vmi_align_munmap.do_vmi_munmap
7.62 +0.2 7.79 perf-profile.calltrace.cycles-pp.zap_pmd_range.unmap_page_range.unmap_vmas.vms_clear_ptes.vms_complete_munmap_vmas
7.61 +0.2 7.78 perf-profile.calltrace.cycles-pp.zap_pte_range.zap_pmd_range.unmap_page_range.unmap_vmas.vms_clear_ptes
8.17 +0.2 8.34 perf-profile.calltrace.cycles-pp.__vm_munmap.__x64_sys_munmap.do_syscall_64.entry_SYSCALL_64_after_hwframe.__munmap
8.17 +0.2 8.34 perf-profile.calltrace.cycles-pp.__x64_sys_munmap.do_syscall_64.entry_SYSCALL_64_after_hwframe.__munmap
8.14 +0.2 8.31 perf-profile.calltrace.cycles-pp.vms_clear_ptes.vms_complete_munmap_vmas.do_vmi_align_munmap.do_vmi_munmap.__vm_munmap
8.14 +0.2 8.32 perf-profile.calltrace.cycles-pp.vms_complete_munmap_vmas.do_vmi_align_munmap.do_vmi_munmap.__vm_munmap.__x64_sys_munmap
8.17 +0.2 8.34 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__munmap
8.17 +0.2 8.34 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__munmap
8.17 +0.2 8.34 perf-profile.calltrace.cycles-pp.__munmap
8.15 +0.2 8.32 perf-profile.calltrace.cycles-pp.do_vmi_munmap.__vm_munmap.__x64_sys_munmap.do_syscall_64.entry_SYSCALL_64_after_hwframe
8.15 +0.2 8.32 perf-profile.calltrace.cycles-pp.do_vmi_align_munmap.do_vmi_munmap.__vm_munmap.__x64_sys_munmap.do_syscall_64
21.20 +0.3 21.46 perf-profile.calltrace.cycles-pp.finish_fault.do_fault.__handle_mm_fault.handle_mm_fault.do_user_addr_fault
90.18 -0.2 89.94 perf-profile.children.cycles-pp.testcase
86.53 -0.2 86.34 perf-profile.children.cycles-pp.asm_exc_page_fault
1.58 +0.0 1.63 perf-profile.children.cycles-pp.__page_cache_release
1.39 +0.0 1.44 perf-profile.children.cycles-pp.lru_gen_add_folio
1.07 +0.0 1.12 perf-profile.children.cycles-pp.lru_gen_del_folio
1.33 +0.0 1.38 perf-profile.children.cycles-pp.folio_remove_rmap_ptes
1.42 +0.1 1.48 ± 2% perf-profile.children.cycles-pp.__lruvec_stat_mod_folio
0.36 ± 2% +0.1 0.42 perf-profile.children.cycles-pp.__mod_lruvec_state
3.08 +0.1 3.16 perf-profile.children.cycles-pp.folios_put_refs
4.56 +0.1 4.71 perf-profile.children.cycles-pp.zap_present_ptes
7.64 +0.2 7.80 perf-profile.children.cycles-pp.unmap_page_range
7.64 +0.2 7.80 perf-profile.children.cycles-pp.unmap_vmas
7.64 +0.2 7.80 perf-profile.children.cycles-pp.zap_pmd_range
7.64 +0.2 7.80 perf-profile.children.cycles-pp.zap_pte_range
8.17 +0.2 8.34 perf-profile.children.cycles-pp.__x64_sys_munmap
8.14 +0.2 8.31 perf-profile.children.cycles-pp.vms_clear_ptes
8.17 +0.2 8.34 perf-profile.children.cycles-pp.__vm_munmap
8.15 +0.2 8.32 perf-profile.children.cycles-pp.vms_complete_munmap_vmas
8.17 +0.2 8.34 perf-profile.children.cycles-pp.__munmap
8.15 +0.2 8.32 perf-profile.children.cycles-pp.do_vmi_align_munmap
8.15 +0.2 8.33 perf-profile.children.cycles-pp.do_vmi_munmap
8.41 +0.2 8.59 perf-profile.children.cycles-pp.do_syscall_64
8.41 +0.2 8.59 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
21.41 +0.3 21.67 perf-profile.children.cycles-pp.finish_fault
0.00 +0.8 0.78 perf-profile.children.cycles-pp.mod_node_page_state
0.36 ± 3% -0.0 0.34 ± 2% perf-profile.self.cycles-pp.free_pages_and_swap_cache
0.53 ± 2% -0.0 0.50 perf-profile.self.cycles-pp.do_user_addr_fault
0.18 ± 2% -0.0 0.15 ± 2% perf-profile.self.cycles-pp.__page_cache_release
0.49 +0.0 0.53 ± 3% perf-profile.self.cycles-pp.folios_put_refs
3.07 +0.1 3.17 perf-profile.self.cycles-pp.zap_present_ptes
0.00 +0.7 0.73 perf-profile.self.cycles-pp.mod_node_page_state
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