All of lore.kernel.org
 help / color / mirror / Atom feed
From: kernel test robot <oliver.sang@intel.com>
To: Zhang Yi <yi.zhang@huaweicloud.com>
Cc: <oe-lkp@lists.linux.dev>, <lkp@intel.com>,
	<linux-ext4@vger.kernel.org>, <linux-fsdevel@vger.kernel.org>,
	<linux-kernel@vger.kernel.org>, <tytso@mit.edu>,
	<adilger.kernel@dilger.ca>, <jack@suse.cz>,
	<ojaswin@linux.ibm.com>, <ritesh.list@gmail.com>,
	<libaokun@linux.alibaba.com>, <yi.zhang@huawei.com>,
	<yi.zhang@huaweicloud.com>, <yizhang089@gmail.com>,
	<yangerkun@huawei.com>, <yukuai@fnnas.com>,
	<oliver.sang@intel.com>
Subject: Re: [PATCH v4 12/13] ext4: move pagecache_isize_extended() out of active handle
Date: Tue, 31 Mar 2026 21:42:11 +0800	[thread overview]
Message-ID: <202603312104.b8cce0af-lkp@intel.com> (raw)
In-Reply-To: <20260327102939.1095257-13-yi.zhang@huaweicloud.com>



Hello,

kernel test robot noticed a 22.1% improvement of fio.write_iops on:


commit: 7cdd1682909c26fb8784755a37688215e8bfbc50 ("[PATCH v4 12/13] ext4: move pagecache_isize_extended() out of active handle")
url: https://github.com/intel-lab-lkp/linux/commits/Zhang-Yi/ext4-add-did_zero-output-parameter-to-ext4_block_zero_page_range/20260328-162522
base: https://git.kernel.org/cgit/linux/kernel/git/tytso/ext4.git dev
patch link: https://lore.kernel.org/all/20260327102939.1095257-13-yi.zhang@huaweicloud.com/
patch subject: [PATCH v4 12/13] ext4: move pagecache_isize_extended() out of active handle

testcase: fio-basic
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:

	runtime: 300s
	disk: 1HDD
	fs: ext4
	nr_task: 1
	test_size: 128G
	rw: write
	bs: 4k
	ioengine: falloc
	cpufreq_governor: performance



Details are as below:
-------------------------------------------------------------------------------------------------->


The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20260331/202603312104.b8cce0af-lkp@intel.com

=========================================================================================
bs/compiler/cpufreq_governor/disk/fs/ioengine/kconfig/nr_task/rootfs/runtime/rw/tbox_group/test_size/testcase:
  4k/gcc-14/performance/1HDD/ext4/falloc/x86_64-rhel-9.4/1/debian-13-x86_64-20250902.cgz/300s/write/lkp-icl-2sp9/128G/fio-basic

commit: 
  ac5088c694 ("ext4: remove ctime/mtime update from ext4_alloc_file_blocks()")
  7cdd168290 ("[PATCH v4 12/13] ext4: move pagecache_isize_extended() out of active handle")

ac5088c694659c49 7cdd1682909c26fb8784755a376 
---------------- --------------------------- 
         %stddev     %change         %stddev
             \          |                \  
      0.01            -0.0        0.00        fio.latency_250us%
      0.11 ±  3%      -0.0        0.08 ±  7%  fio.latency_2us%
      0.07 ±  2%      -0.0        0.05 ±  3%  fio.latency_4us%
     34.85 ±  2%     -18.0%      28.58        fio.time.elapsed_time
     34.85 ±  2%     -18.0%      28.58        fio.time.elapsed_time.max
     24.00           -24.1%      18.22        fio.time.system_time
      3744 ±  2%     -17.8%       3078        fio.time.voluntary_context_switches
      3807 ±  2%     +22.1%       4649        fio.write_bw_MBps
    846.67           -20.9%     669.33        fio.write_clat_90%_ns
    854.67           -21.1%     674.67        fio.write_clat_95%_ns
    870.67           -21.0%     688.00        fio.write_clat_99%_ns
    834.60           -21.3%     656.75        fio.write_clat_mean_ns
    974606 ±  2%     +22.1%    1190280        fio.write_iops
 2.331e+09 ±  2%     -17.4%  1.926e+09        cpuidle..time
      1.10            -5.8%       1.03        iostat.cpu.system
    112890 ±  2%     -14.9%      96120        turbostat.IRQ
      1472 ±  4%     +15.6%       1702        vmstat.io.bo
      1.10            -0.1        1.04        mpstat.cpu.all.sys%
      0.54 ±  2%      +0.1        0.64 ±  3%  mpstat.cpu.all.usr%
     49.52 ± 21%     -31.3%      34.03 ± 33%  perf-sched.total_wait_time.max.ms
     49.52 ± 21%     -31.3%      34.03 ± 33%  perf-sched.wait_time.max.ms.[unknown].[unknown].[unknown].[unknown].[unknown]
    123.26            -5.1%     116.96        uptime.boot
      7526            -5.3%       7128        uptime.idle
      0.03 ±  7%     +17.9%       0.03 ±  5%  perf-stat.i.MPKI
      0.29 ±  8%      +0.1        0.38        perf-stat.i.branch-miss-rate%
   4759878 ±  7%     +33.3%    6346860        perf-stat.i.branch-misses
    201390 ±  4%     +19.8%     241185 ±  6%  perf-stat.i.cache-misses
    891976           +14.7%    1023336        perf-stat.i.cache-references
      0.48 ±  2%      +3.1%       0.49        perf-stat.i.cpi
 3.807e+09            +1.5%  3.866e+09        perf-stat.i.cpu-cycles
     70.29            +2.0%      71.71        perf-stat.i.cpu-migrations
    382.22 ±  3%     +19.4%     456.21 ±  4%  perf-stat.i.minor-faults
    382.22 ±  3%     +19.4%     456.20 ±  4%  perf-stat.i.page-faults
      0.03 ±  6%     +21.4%       0.03 ±  6%  perf-stat.overall.MPKI
      0.28 ±  9%      +0.1        0.38        perf-stat.overall.branch-miss-rate%
     18830 ±  5%     -15.0%      16001 ±  5%  perf-stat.overall.cycles-between-cache-misses
      8436           -19.3%       6810        perf-stat.overall.path-length
   4633589 ±  7%     +32.5%    6139073        perf-stat.ps.branch-misses
    197159 ±  5%     +18.9%     234444 ±  6%  perf-stat.ps.cache-misses
    871990           +13.9%     993519        perf-stat.ps.cache-references
     68.47            +1.4%      69.43        perf-stat.ps.cpu-migrations
    379.37 ±  3%     +18.3%     448.62 ±  3%  perf-stat.ps.minor-faults
    379.37 ±  3%     +18.3%     448.62 ±  3%  perf-stat.ps.page-faults
 2.831e+11           -19.3%  2.285e+11        perf-stat.total.instructions
      8.95 ± 91%      -6.7        2.23 ±141%  perf-profile.calltrace.cycles-pp.do_file_open.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe
      8.95 ± 91%      -6.7        2.23 ±141%  perf-profile.calltrace.cycles-pp.path_openat.do_file_open.do_sys_openat2.__x64_sys_openat.do_syscall_64
      6.48 ± 86%      -5.4        1.04 ±223%  perf-profile.calltrace.cycles-pp.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe
      6.48 ± 86%      -5.4        1.04 ±223%  perf-profile.calltrace.cycles-pp.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe
      5.42 ±109%      -4.6        0.79 ±223%  perf-profile.calltrace.cycles-pp.lruvec_stat_mod_folio.folio_remove_rmap_ptes.zap_present_ptes.zap_pte_range.zap_pmd_range
      5.46 ± 47%      -3.6        1.84 ±143%  perf-profile.calltrace.cycles-pp.tlb_finish_mmu.exit_mmap.__mmput.exit_mm.do_exit
      4.28 ± 73%      -3.2        1.04 ±223%  perf-profile.calltrace.cycles-pp.open_last_lookups.path_openat.do_file_open.do_sys_openat2.__x64_sys_openat
      4.28 ± 73%      -2.4        1.84 ±143%  perf-profile.calltrace.cycles-pp.__tlb_batch_free_encoded_pages.tlb_finish_mmu.exit_mmap.__mmput.exit_mm
      4.28 ± 73%      -2.4        1.84 ±143%  perf-profile.calltrace.cycles-pp.free_pages_and_swap_cache.__tlb_batch_free_encoded_pages.tlb_finish_mmu.exit_mmap.__mmput
      3.40 ±101%      -2.4        1.04 ±223%  perf-profile.calltrace.cycles-pp.lookup_open.open_last_lookups.path_openat.do_file_open.do_sys_openat2
      2.99 ±101%      -1.2        1.84 ±143%  perf-profile.calltrace.cycles-pp.folios_put_refs.free_pages_and_swap_cache.__tlb_batch_free_encoded_pages.tlb_finish_mmu.exit_mmap
      2.99 ±101%      -0.6        2.43 ±143%  perf-profile.calltrace.cycles-pp.vfs_statx.vfs_fstatat.__do_sys_newfstatat.do_syscall_64.entry_SYSCALL_64_after_hwframe
      8.96 ± 91%      -6.7        2.23 ±141%  perf-profile.children.cycles-pp.do_file_open
      8.96 ± 91%      -6.7        2.23 ±141%  perf-profile.children.cycles-pp.path_openat
      5.46 ± 47%      -3.6        1.84 ±143%  perf-profile.children.cycles-pp.tlb_finish_mmu
      4.28 ± 73%      -3.2        1.04 ±223%  perf-profile.children.cycles-pp.open_last_lookups
      3.40 ±101%      -2.4        1.04 ±223%  perf-profile.children.cycles-pp.lookup_open
      3.40 ±101%      -2.2        1.19 ±223%  perf-profile.children.cycles-pp.link_path_walk
      4.14 ±102%      -2.0        2.18 ±149%  perf-profile.children.cycles-pp.lruvec_stat_mod_folio
      3.88 ±105%      -1.7        2.18 ±149%  perf-profile.children.cycles-pp.kernfs_fop_readdir
      3.92 ±109%      -1.5        2.43 ±143%  perf-profile.children.cycles-pp.__do_sys_newfstatat
      3.92 ±109%      -1.5        2.43 ±143%  perf-profile.children.cycles-pp.vfs_fstatat
      2.99 ±101%      -0.6        2.43 ±143%  perf-profile.children.cycles-pp.vfs_statx
      5.80 ± 74%      +6.5       12.30 ± 22%  perf-profile.children.cycles-pp.exc_page_fault
      3.88 ±105%      -3.9        0.00        perf-profile.self.cycles-pp.kernfs_fop_readdir




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


  reply	other threads:[~2026-03-31 13:42 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-27 10:29 [PATCH v4 00/13] ext4: refactor partial block zero-out for iomap conversion Zhang Yi
2026-03-27 10:29 ` [PATCH v4 01/13] ext4: add did_zero output parameter to ext4_block_zero_page_range() Zhang Yi
2026-03-27 10:29 ` [PATCH v4 02/13] ext4: rename and extend ext4_block_truncate_page() Zhang Yi
2026-03-27 10:29 ` [PATCH v4 03/13] ext4: factor out journalled block zeroing range Zhang Yi
2026-03-27 10:29 ` [PATCH v4 04/13] ext4: rename ext4_block_zero_page_range() to ext4_block_zero_range() Zhang Yi
2026-03-27 10:29 ` [PATCH v4 05/13] ext4: move ordered data handling out of ext4_block_do_zero_range() Zhang Yi
2026-03-27 10:29 ` [PATCH v4 06/13] ext4: remove handle parameters from zero partial block functions Zhang Yi
2026-03-27 10:29 ` [PATCH v4 07/13] ext4: pass allocate range as loff_t to ext4_alloc_file_blocks() Zhang Yi
2026-03-27 10:29 ` [PATCH v4 08/13] ext4: move zero partial block range functions out of active handle Zhang Yi
2026-03-27 10:29 ` [PATCH v4 09/13] ext4: ensure zeroed partial blocks are persisted in SYNC mode Zhang Yi
2026-04-01 17:06   ` Jan Kara
2026-04-02  1:21     ` Zhang Yi
2026-03-27 10:29 ` [PATCH v4 10/13] ext4: unify SYNC mode checks in fallocate paths Zhang Yi
2026-04-01 17:06   ` Jan Kara
2026-03-27 10:29 ` [PATCH v4 11/13] ext4: remove ctime/mtime update from ext4_alloc_file_blocks() Zhang Yi
2026-04-01 17:09   ` Jan Kara
2026-03-27 10:29 ` [PATCH v4 12/13] ext4: move pagecache_isize_extended() out of active handle Zhang Yi
2026-03-31 13:42   ` kernel test robot [this message]
2026-04-01 17:21   ` Jan Kara
2026-03-27 10:29 ` [PATCH v4 13/13] ext4: zero post-EOF partial block before appending write Zhang Yi
2026-04-10 15:18 ` [PATCH v4 00/13] ext4: refactor partial block zero-out for iomap conversion Theodore Ts'o

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=202603312104.b8cce0af-lkp@intel.com \
    --to=oliver.sang@intel.com \
    --cc=adilger.kernel@dilger.ca \
    --cc=jack@suse.cz \
    --cc=libaokun@linux.alibaba.com \
    --cc=linux-ext4@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lkp@intel.com \
    --cc=oe-lkp@lists.linux.dev \
    --cc=ojaswin@linux.ibm.com \
    --cc=ritesh.list@gmail.com \
    --cc=tytso@mit.edu \
    --cc=yangerkun@huawei.com \
    --cc=yi.zhang@huawei.com \
    --cc=yi.zhang@huaweicloud.com \
    --cc=yizhang089@gmail.com \
    --cc=yukuai@fnnas.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.