Received-SPF: pass (google.com: domain of linux-kernel+bounces-4230-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Precedence: bulk
MIME-Version: 1.0
References: <20231208061407.2125867-1-yuzhao@google.com> <CAMgjq7BTaV5OvHNjGRVJP2VDxj+PXhfd6957CjS4BJ9J4OY8HA@mail.gmail.com>
 <CAOUHufYwZAUaJh6i8Fazc4gVMSqcsz9JbRNpj0cpx2qR+bZBFw@mail.gmail.com>
 <CAMgjq7AtceR-CXnKFfQHM3qi0y4oGyJ4_sw_uh5EkpXCBzkCXg@mail.gmail.com>
 <CAMgjq7CJ3hYHysyRfHzYU4hOYqhUOttxMYGtg0FxzM_wvvyhFA@mail.gmail.com>
 <CAOUHufa=ybd-EPos9DryLHYyhphN0P7qyV5NY3Pui0dfVSk9tQ@mail.gmail.com>
 <ZXpx0K8Vdo3FJixg@google.com> <CAMgjq7CRf4iEKuW2qWKzbhssMbixBo3UoLPqsSk4b+Tvw8at8A@mail.gmail.com>
 <CAOUHufY6x_Erz02Bzoejfs_g2hcmrMFpiOdjiaWZ97oirz71aQ@mail.gmail.com> <ZXvcgyjTu92Qqk5X@google.com>
In-Reply-To: <ZXvcgyjTu92Qqk5X@google.com>
From: Kairui Song <ryncsn@gmail.com>
Date: Tue, 19 Dec 2023 02:05:38 +0800
Message-ID: <CAMgjq7BsY1tJeOZwSppxUN7Lha-_a7WLfhv1_bxTuU4EuiQyVg@mail.gmail.com>
Subject: Re: [PATCH mm-unstable v1 1/4] mm/mglru: fix underprotected page cache
To: Yu Zhao <yuzhao@google.com>
Cc: Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org, 
	linux-kernel@vger.kernel.org, Charan Teja Kalla <quic_charante@quicinc.com>, 
	Kalesh Singh <kaleshsingh@google.com>, stable@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Yu Zhao <yuzhao@google.com> =E4=BA=8E2023=E5=B9=B412=E6=9C=8815=E6=97=A5=E5=
=91=A8=E4=BA=94 12:56=E5=86=99=E9=81=93=EF=BC=9A
>
> On Thu, Dec 14, 2023 at 04:51:00PM -0700, Yu Zhao wrote:
> > On Thu, Dec 14, 2023 at 11:38=E2=80=AFAM Kairui Song <ryncsn@gmail.com>=
 wrote:
> > >
> > > Yu Zhao <yuzhao@google.com> =E4=BA=8E2023=E5=B9=B412=E6=9C=8814=E6=97=
=A5=E5=91=A8=E5=9B=9B 11:09=E5=86=99=E9=81=93=EF=BC=9A
> > > > On Wed, Dec 13, 2023 at 12:59:14AM -0700, Yu Zhao wrote:
> > > > > On Tue, Dec 12, 2023 at 8:03=E2=80=AFPM Kairui Song <ryncsn@gmail=
.com> wrote:
> > > > > >
> > > > > > Kairui Song <ryncsn@gmail.com> =E4=BA=8E2023=E5=B9=B412=E6=9C=
=8812=E6=97=A5=E5=91=A8=E4=BA=8C 14:52=E5=86=99=E9=81=93=EF=BC=9A
> > > > > > >
> > > > > > > Yu Zhao <yuzhao@google.com> =E4=BA=8E2023=E5=B9=B412=E6=9C=88=
12=E6=97=A5=E5=91=A8=E4=BA=8C 06:07=E5=86=99=E9=81=93=EF=BC=9A
> > > > > > > >
> > > > > > > > On Fri, Dec 8, 2023 at 1:24=E2=80=AFAM Kairui Song <ryncsn@=
gmail.com> wrote:
> > > > > > > > >
> > > > > > > > > Yu Zhao <yuzhao@google.com> =E4=BA=8E2023=E5=B9=B412=E6=
=9C=888=E6=97=A5=E5=91=A8=E4=BA=94 14:14=E5=86=99=E9=81=93=EF=BC=9A
> > > > > > > > > >
> > > > > > > > > > Unmapped folios accessed through file descriptors can b=
e
> > > > > > > > > > underprotected. Those folios are added to the oldest ge=
neration based
> > > > > > > > > > on:
> > > > > > > > > > 1. The fact that they are less costly to reclaim (no ne=
ed to walk the
> > > > > > > > > >    rmap and flush the TLB) and have less impact on perf=
ormance (don't
> > > > > > > > > >    cause major PFs and can be non-blocking if needed ag=
ain).
> > > > > > > > > > 2. The observation that they are likely to be single-us=
e. E.g., for
> > > > > > > > > >    client use cases like Android, its apps parse config=
uration files
> > > > > > > > > >    and store the data in heap (anon); for server use ca=
ses like MySQL,
> > > > > > > > > >    it reads from InnoDB files and holds the cached data=
 for tables in
> > > > > > > > > >    buffer pools (anon).
> > > > > > > > > >
> > > > > > > > > > However, the oldest generation can be very short lived,=
 and if so, it
> > > > > > > > > > doesn't provide the PID controller with enough time to =
respond to a
> > > > > > > > > > surge of refaults. (Note that the PID controller uses w=
eighted
> > > > > > > > > > refaults and those from evicted generations only take a=
 half of the
> > > > > > > > > > whole weight.) In other words, for a short lived genera=
tion, the
> > > > > > > > > > moving average smooths out the spike quickly.
> > > > > > > > > >
> > > > > > > > > > To fix the problem:
> > > > > > > > > > 1. For folios that are already on LRU, if they can be b=
eyond the
> > > > > > > > > >    tracking range of tiers, i.e., five accesses through=
 file
> > > > > > > > > >    descriptors, move them to the second oldest generati=
on to give them
> > > > > > > > > >    more time to age. (Note that tiers are used by the P=
ID controller
> > > > > > > > > >    to statistically determine whether folios accessed m=
ultiple times
> > > > > > > > > >    through file descriptors are worth protecting.)
> > > > > > > > > > 2. When adding unmapped folios to LRU, adjust the place=
ment of them so
> > > > > > > > > >    that they are not too close to the tail. The effect =
of this is
> > > > > > > > > >    similar to the above.
> > > > > > > > > >
> > > > > > > > > > On Android, launching 55 apps sequentially:
> > > > > > > > > >                            Before     After      Change
> > > > > > > > > >   workingset_refault_anon  25641024   25598972   0%
> > > > > > > > > >   workingset_refault_file  115016834  106178438  -8%
> > > > > > > > >
> > > > > > > > > Hi Yu,
> > > > > > > > >
> > > > > > > > > Thanks you for your amazing works on MGLRU.
> > > > > > > > >
> > > > > > > > > I believe this is the similar issue I was trying to resol=
ve previously:
> > > > > > > > > https://lwn.net/Articles/945266/
> > > > > > > > > The idea is to use refault distance to decide if the page=
 should be
> > > > > > > > > place in oldest generation or some other gen, which per m=
y test,
> > > > > > > > > worked very well, and we have been using refault distance=
 for MGLRU in
> > > > > > > > > multiple workloads.
> > > > > > > > >
> > > > > > > > > There are a few issues left in my previous RFC series, li=
ke anon pages
> > > > > > > > > in MGLRU shouldn't be considered, I wanted to collect fee=
dback or test
> > > > > > > > > cases, but unfortunately it seems didn't get too much att=
ention
> > > > > > > > > upstream.
> > > > > > > > >
> > > > > > > > > I think both this patch and my previous series are for so=
lving the
> > > > > > > > > file pages underpertected issue, and I did a quick test u=
sing this
> > > > > > > > > series, for mongodb test, refault distance seems still a =
better
> > > > > > > > > solution (I'm not saying these two optimization are mutua=
lly exclusive
> > > > > > > > > though, just they do have some conflicts in implementatio=
n and solving
> > > > > > > > > similar problem):
> > > > > > > > >
> > > > > > > > > Previous result:
> > > > > > > > > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> > > > > > > > > Execution Results after 905 seconds
> > > > > > > > > ---------------------------------------------------------=
---------
> > > > > > > > >                   Executed        Time (=C2=B5s)       Ra=
te
> > > > > > > > >   STOCK_LEVEL     2542            27121571486.2   0.09 tx=
n/s
> > > > > > > > > ---------------------------------------------------------=
---------
> > > > > > > > >   TOTAL           2542            27121571486.2   0.09 tx=
n/s
> > > > > > > > >
> > > > > > > > > This patch:
> > > > > > > > > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> > > > > > > > > Execution Results after 900 seconds
> > > > > > > > > ---------------------------------------------------------=
---------
> > > > > > > > >                   Executed        Time (=C2=B5s)       Ra=
te
> > > > > > > > >   STOCK_LEVEL     1594            27061522574.4   0.06 tx=
n/s
> > > > > > > > > ---------------------------------------------------------=
---------
> > > > > > > > >   TOTAL           1594            27061522574.4   0.06 tx=
n/s
> > > > > > > > >
> > > > > > > > > Unpatched version is always around ~500.
> > > > > > > >
> > > > > > > > Thanks for the test results!
> > > > > > > >
> > > > > > > > > I think there are a few points here:
> > > > > > > > > - Refault distance make use of page shadow so it can bett=
er
> > > > > > > > > distinguish evicted pages of different access pattern (re=
-access
> > > > > > > > > distance).
> > > > > > > > > - Throttled refault distance can help hold part of workin=
gset when
> > > > > > > > > memory is too small to hold the whole workingset.
> > > > > > > > >
> > > > > > > > > So maybe part of this patch and the bits of previous seri=
es can be
> > > > > > > > > combined to work better on this issue, how do you think?
> > > > > > > >
> > > > > > > > I'll try to find some time this week to look at your RFC. I=
t'd be a
> > > > > >
> > > > > > Hi Yu,
> > > > > >
> > > > > > I'm working on V4 of the RFC now, which just update some commen=
ts, and
> > > > > > skip anon page re-activation in refault path for mglru which wa=
s not
> > > > > > very helpful, only some tiny adjustment.
> > > > > > And I found it easier to test with fio, using following test sc=
ript:
> > > > > >
> > > > > > #!/bin/bash
> > > > > > swapoff -a
> > > > > >
> > > > > > modprobe brd rd_nr=3D1 rd_size=3D16777216
> > > > > > mkfs.ext4 /dev/ram0
> > > > > > mount /dev/ram0 /mnt
> > > > > >
> > > > > > mkdir -p /sys/fs/cgroup/benchmark
> > > > > > cd /sys/fs/cgroup/benchmark
> > > > > >
> > > > > > echo 4G > memory.max
> > > > > > echo $$ > cgroup.procs
> > > > > > echo 3 > /proc/sys/vm/drop_caches
> > > > > >
> > > > > > fio -name=3Dmglru --numjobs=3D12 --directory=3D/mnt --size=3D10=
24m \
> > > > > >           --buffered=3D1 --ioengine=3Dio_uring --iodepth=3D128 =
\
> > > > > >           --iodepth_batch_submit=3D32 --iodepth_batch_complete=
=3D32 \
> > > > > >           --rw=3Drandread --random_distribution=3Dzipf:0.5 --no=
randommap \
> > > > > >           --time_based --ramp_time=3D5m --runtime=3D5m --group_=
reporting
> > > > > >
> > > > > > zipf:0.5 is used here to simulate a cached read with slight bia=
s
> > > > > > towards certain pages.
> > > > > > Unpatched 6.7-rc4:
> > > > > > Run status group 0 (all jobs):
> > > > > >    READ: bw=3D6548MiB/s (6866MB/s), 6548MiB/s-6548MiB/s
> > > > > > (6866MB/s-6866MB/s), io=3D1918GiB (2060GB), run=3D300001-300001=
msec
> > > > > >
> > > > > > Patched with RFC v4:
> > > > > > Run status group 0 (all jobs):
> > > > > >    READ: bw=3D7270MiB/s (7623MB/s), 7270MiB/s-7270MiB/s
> > > > > > (7623MB/s-7623MB/s), io=3D2130GiB (2287GB), run=3D300001-300001=
msec
> > > > > >
> > > > > > Patched with this series:
> > > > > > Run status group 0 (all jobs):
> > > > > >    READ: bw=3D7098MiB/s (7442MB/s), 7098MiB/s-7098MiB/s
> > > > > > (7442MB/s-7442MB/s), io=3D2079GiB (2233GB), run=3D300002-300002=
msec
> > > > > >
> > > > > > MGLRU off:
> > > > > > Run status group 0 (all jobs):
> > > > > >    READ: bw=3D6525MiB/s (6842MB/s), 6525MiB/s-6525MiB/s
> > > > > > (6842MB/s-6842MB/s), io=3D1912GiB (2052GB), run=3D300002-300002=
msec
> > > > > >
> > > > > > - If I change zipf:0.5 to random:
> > > > > > Unpatched 6.7-rc4:
> > > > > > Patched with this series:
> > > > > > Run status group 0 (all jobs):
> > > > > >    READ: bw=3D5975MiB/s (6265MB/s), 5975MiB/s-5975MiB/s
> > > > > > (6265MB/s-6265MB/s), io=3D1750GiB (1879GB), run=3D300002-300002=
msec
> > > > > >
> > > > > > Patched with RFC v4:
> > > > > > Run status group 0 (all jobs):
> > > > > >    READ: bw=3D5987MiB/s (6278MB/s), 5987MiB/s-5987MiB/s
> > > > > > (6278MB/s-6278MB/s), io=3D1754GiB (1883GB), run=3D300001-300001=
msec
> > > > > >
> > > > > > Patched with this series:
> > > > > > Run status group 0 (all jobs):
> > > > > >    READ: bw=3D5839MiB/s (6123MB/s), 5839MiB/s-5839MiB/s
> > > > > > (6123MB/s-6123MB/s), io=3D1711GiB (1837GB), run=3D300001-300001=
msec
> > > > > >
> > > > > > MGLRU off:
> > > > > > Run status group 0 (all jobs):
> > > > > >    READ: bw=3D5689MiB/s (5965MB/s), 5689MiB/s-5689MiB/s
> > > > > > (5965MB/s-5965MB/s), io=3D1667GiB (1790GB), run=3D300003-300003=
msec
> > > > > >
> > > > > > fio uses ramdisk so LRU accuracy will have smaller impact. The =
Mongodb
> > > > > > test I provided before uses a SATA SSD so it will have a much h=
igher
> > > > > > impact. I'll provides a script to setup the test case and run i=
t, it's
> > > > > > more complex to setup than fio since involving setting up multi=
ple
> > > > > > replicas and auth and hundreds of GB of test fixtures, I'm curr=
ently
> > > > > > occupied by some other tasks but will try best to send them out=
 as
> > > > > > soon as possible.
> > > > >
> > > > > Thanks! Apparently your RFC did show better IOPS with both access
> > > > > patterns, which was a surprise to me because it had higher refaul=
ts
> > > > > and usually higher refautls result in worse performance.
>
> And thanks for providing the refaults I requested for -- your data
> below confirms what I mentioned above:
>
> For fio:
>                            Your RFC   This series   Change
>   workingset_refault_file  628192729  596790506     -5%
>   IOPS                     1862k      1830k         -2%
>
> For MongoDB:
>                            Your RFC   This series   Change
>   workingset_refault_anon  10512      35277         +30%
>   workingset_refault_file  22751782   20335355      -11%
>   total                    22762294   20370632      -11%
>   TPS                      0.09       0.06          -33%
>
> For MongoDB, this series should be a big win (but apparently it's not),
> especially when using zram, since an anon refault should be a lot
> cheaper than a file refault.
>
> So, I'm baffled...
>
> One important detail I forgot to mention: based on your data from
> lru_gen_full, I think there is another difference between our Kconfigs:
>
>                   Your Kconfig  My Kconfig  Max possible
>   LRU_REFS_WIDTH  1             2           2

Hi Yu,

Thanks for the info, my fault, I forgot to update my config as I was
testing some other features.
Buf after I changed LRU_REFS_WIDTH to 2 by disabling IDLE_PAGE, thing
got much worse for MongoDB test:

With LRU_REFS_WIDTH =3D=3D 2:

This patch:
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
Execution Results after 919 seconds
------------------------------------------------------------------
                  Executed        Time (=C2=B5s)       Rate
  STOCK_LEVEL     488             27598136201.9   0.02 txn/s
------------------------------------------------------------------
  TOTAL           488             27598136201.9   0.02 txn/s

memcg    86 /system.slice/docker-1c3a90be9f0a072f5719332419550cd0e1455f2cd5=
863bc2780ca4d3f913ece5.scope
 node     0
          1     948187          0x          0x
                     0          0           0           0           0
         0           0=C2=B7
                     1          0           0           0           0
         0           0=C2=B7
                     2          0           0           0           0
         0           0=C2=B7
                     3          0           0           0           0
         0           0=C2=B7
                                0           0           0           0
         0           0=C2=B7
          2     948187          0     6051788=C2=B7
                     0          0r          0e          0p      11916r
     66442e          0p
                     1          0r          0e          0p        903r
     16888e          0p
                     2          0r          0e          0p        459r
      9764e          0p
                     3          0r          0e          0p          0r
         0e       2874p
                                0           0           0           0
         0           0=C2=B7
          3     948187    1353160        6351=C2=B7
                     0          0           0           0           0
         0           0=C2=B7
                     1          0           0           0           0
         0           0=C2=B7
                     2          0           0           0           0
         0           0=C2=B7
                     3          0           0           0           0
         0           0=C2=B7
                                0           0           0           0
         0           0=C2=B7
          4      73045      23573          12=C2=B7
                     0          0R          0T          0     3498607R
   4868605T          0=C2=B7
                     1          0R          0T          0     3012246R
   3270261T          0=C2=B7
                     2          0R          0T          0     2498608R
   2839104T          0=C2=B7
                     3          0R          0T          0           0R
   1983947T          0=C2=B7
                          1486579L          0O    1380614Y       2945N
      2945F       2734A

workingset_refault_anon 0
workingset_refault_file 18130598

              total        used        free      shared  buff/cache   avail=
able
Mem:          31978        6705         312          20       24960       2=
4786
Swap:         31977           4       31973

RFC:
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
Execution Results after 908 seconds
------------------------------------------------------------------
                  Executed        Time (=C2=B5s)       Rate
  STOCK_LEVEL     2252            27159962888.2   0.08 txn/s
------------------------------------------------------------------
  TOTAL           2252            27159962888.2   0.08 txn/s

workingset_refault_anon 22585
workingset_refault_file 22715256

memcg    66 /system.slice/docker-0989446ff78106e32d3f400a0cf371c9a703281bde=
d86d6d6bb1af706ebb25da.scope
 node     0
         22     563007       2274     1198225=C2=B7
                     0          0r          1e          0p          0r
    697076e          0p
                     1          0r          0e          0p          0r
         0e     325661p
                     2          0r          0e          0p          0r
         0e     888728p
                     3          0r          0e          0p          0r
         0e    3602238p
                                0           0           0           0
         0           0=C2=B7
         23     532222       7525     4948747=C2=B7
                     0          0           0           0           0
         0           0=C2=B7
                     1          0           0           0           0
         0           0=C2=B7
                     2          0           0           0           0
         0           0=C2=B7
                     3          0           0           0           0
         0           0=C2=B7
                                0           0           0           0
         0           0=C2=B7
         24     500367    1214667        3292=C2=B7
                     0          0           0           0           0
         0           0=C2=B7
                     1          0           0           0           0
         0           0=C2=B7
                     2          0           0           0           0
         0           0=C2=B7
                     3          0           0           0           0
         0           0=C2=B7
                                0           0           0           0
         0           0=C2=B7
         25     469692      40797         466=C2=B7
                     0          0R        271T          0           0R
   1162165T          0=C2=B7
                     1          0R          0T          0      774028R
   1205332T          0=C2=B7
                     2          0R          0T          0           0R
    932484T          0=C2=B7
                     3          0R          1T          0           0R
   4252158T          0=C2=B7
                         25178380L     156515O   23953602Y      59234N
     49391F      48664A

              total        used        free      shared  buff/cache   avail=
able
Mem:          31978        6968         338           5       24671       2=
4555
Swap:         31977        1533       30444

Using same mongodb config (a 3 replica cluster using the same config):
{
    "net": {
        "bindIpAll": true,
        "ipv6": false,
        "maxIncomingConnections": 10000,
    },
    "setParameter": {
        "disabledSecureAllocatorDomains": "*"
    },
    "replication": {
        "oplogSizeMB": 10480,
        "replSetName": "issa-tpcc_0"
    },
    "security": {
        "keyFile": "/data/db/keyfile"
    },
    "storage": {
        "dbPath": "/data/db/",
        "syncPeriodSecs": 60,
        "directoryPerDB": true,
        "wiredTiger": {
            "engineConfig": {
                "cacheSizeGB": 5
            }
        }
    },
    "systemLog": {
        "destination": "file",
        "logAppend": true,
        "logRotate": "rename",
        "path": "/data/db/mongod.log",
        "verbosity": 0
    }
}

The test environment have 32g memory and 16 core.

Per my analyze, the access pattern for the mongodb test is that page
will be re-access long after it's evicted so PID controller won't
protect higher tier. That RFC will make use of the long existing
shadow to do feedback to PID/Gen so the result will be much better.
Still need more adjusting though, will try to do a rebase on top of
mm-unstable which includes your patch.

I've no idea why the workingset_refault_* is higher in the better
case, this a clearly an IO bound workload, Memory and IO is busy while
CPU is not full...

I've uploaded my local reproducer here:
https://github.com/ryncsn/emm-test-project/tree/master/mongo-cluster
https://github.com/ryncsn/py-tpcc