linux-kernel - Re: [linus:master] [mm] 8d59d2214c: vm-scalability.throughput -36.6% regression

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 linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <ZbDJsfsZt2ITyo61@xsang-OptiPlex-9020>
Date: Wed, 24 Jan 2024 16:26:25 +0800
From: Oliver Sang <oliver.sang@...el.com>
To: Yosry Ahmed <yosryahmed@...gle.com>
CC: <oe-lkp@...ts.linux.dev>, <lkp@...el.com>, <linux-kernel@...r.kernel.org>,
	Andrew Morton <akpm@...ux-foundation.org>, Johannes Weiner
	<hannes@...xchg.org>, Domenico Cerasuolo <cerasuolodomenico@...il.com>,
	Shakeel Butt <shakeelb@...gle.com>, Chris Li <chrisl@...nel.org>, Greg Thelen
	<gthelen@...gle.com>, Ivan Babrou <ivan@...udflare.com>, Michal Hocko
	<mhocko@...nel.org>, Michal Koutny <mkoutny@...e.com>, Muchun Song
	<muchun.song@...ux.dev>, Roman Gushchin <roman.gushchin@...ux.dev>, Tejun Heo
	<tj@...nel.org>, Waiman Long <longman@...hat.com>, Wei Xu
	<weixugc@...gle.com>, <cgroups@...r.kernel.org>, <linux-mm@...ck.org>,
	<ying.huang@...el.com>, <feng.tang@...el.com>, <fengwei.yin@...el.com>,
	<oliver.sang@...el.com>
Subject: Re: [linus:master] [mm] 8d59d2214c: vm-scalability.throughput -36.6%
 regression

hi, Yosry Ahmed,

On Mon, Jan 22, 2024 at 11:42:04PM -0800, Yosry Ahmed wrote:
> > > Oliver, would you be able to test if the attached patch helps? It's
> > > based on 8d59d2214c236.
> >
> > the patch failed to compile:
> >
> > build_errors:
> >   - "mm/memcontrol.c:731:38: error: 'x' undeclared (first use in this function)"
> 
> Apologizes, apparently I sent the patch with some pending diff in my
> tree that I hadn't committed. Please find a fixed patch attached.

the regression disappears after applying the patch.

Tested-by: kernel test robot <oliver.sang@...el.com>

for 1st regression we reported (details is attached as vm-scalability):

=========================================================================================
compiler/cpufreq_governor/kconfig/rootfs/runtime/size/tbox_group/test/testcase:
  gcc-12/performance/x86_64-rhel-8.3/debian-11.1-x86_64-20220510.cgz/300s/1T/lkp-cpl-4sp2/lru-shm/vm-scalability

commit:
  e0bf1dc859fdd mm: memcg: move vmstats structs definition above flushing code
  8d59d2214c236 mm: memcg: make stats flushing threshold per-memcg
  0cba55e237ba6 mm: memcg: optimize parent iteration in memcg_rstat_updated()

e0bf1dc859fdd08e 8d59d2214c2362e7a9d185d80b6 0cba55e237ba61489c0a29f7d27
---------------- --------------------------- ---------------------------
         %stddev     %change         %stddev     %change         %stddev
             \          |                \          |                \
    946447           -37.8%     588327            -1.1%     936279        vm-scalability.median
 2.131e+08           -36.6%  1.351e+08            -1.4%  2.102e+08        vm-scalability.throughput



for 2nd regression (details is attached as will-it-scale-tlb_flush2):

=========================================================================================
compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase:
  gcc-12/performance/x86_64-rhel-8.3/process/50%/debian-11.1-x86_64-20220510.cgz/lkp-skl-fpga01/tlb_flush2/will-it-scale

commit:
  e0bf1dc859fdd mm: memcg: move vmstats structs definition above flushing code
  8d59d2214c236 mm: memcg: make stats flushing threshold per-memcg
  0cba55e237ba6 mm: memcg: optimize parent iteration in memcg_rstat_updated()

e0bf1dc859fdd08e 8d59d2214c2362e7a9d185d80b6 0cba55e237ba61489c0a29f7d27
---------------- --------------------------- ---------------------------
         %stddev     %change         %stddev     %change         %stddev
             \          |                \          |                \
   2132437           -32.3%    1444430            +0.9%    2151460        will-it-scale.52.processes
     41008           -32.3%      27776            +0.9%      41373        will-it-scale.per_process_ops



interesting thing is, it also helps on will-it-scale:fallocate tests which we
ignored in original report (details is attached as will-it-scale-fallocate1):

=========================================================================================
compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase:
  gcc-12/performance/x86_64-rhel-8.3/process/100%/debian-11.1-x86_64-20220510.cgz/lkp-cpl-4sp2/fallocate1/will-it-scale

commit:
  e0bf1dc859fdd mm: memcg: move vmstats structs definition above flushing code
  8d59d2214c236 mm: memcg: make stats flushing threshold per-memcg
  0cba55e237ba6 mm: memcg: optimize parent iteration in memcg_rstat_updated()

e0bf1dc859fdd08e 8d59d2214c2362e7a9d185d80b6 0cba55e237ba61489c0a29f7d27
---------------- --------------------------- ---------------------------
         %stddev     %change         %stddev     %change         %stddev
             \          |                \          |                \
   5426049           -33.8%    3590953            +3.3%    5605429        will-it-scale.224.processes
     24222           -33.8%      16030            +3.3%      25023        will-it-scale.per_process_ops

> 
> Thanks.



View attachment "vm-scalability" of type "text/plain" (82648 bytes)

View attachment "will-it-scale-tlb_flush2" of type "text/plain" (38804 bytes)

View attachment "will-it-scale-fallocate1" of type "text/plain" (33790 bytes)