Re: [PATCHSET v5 0/12] Add support for async buffered reads

From: Sedat Dilek <sedat.dilek@gmail.com>
To: Jens Axboe <axboe@kernel.dk>
Cc: io-uring@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org,
	akpm@linux-foundation.org
Subject: Re: [PATCHSET v5 0/12] Add support for async buffered reads
Date: Mon, 1 Jun 2020 22:18:33 +0200	[thread overview]
Message-ID: <CA+icZUVGJ8jfRqz2iCxLrkvmky79_9N7pGH71eFpg=s+=D4Urw@mail.gmail.com> (raw)
In-Reply-To: <412ec6a7-90c9-5856-bafe-12c8fe2135e7@kernel.dk>

On Mon, Jun 1, 2020 at 4:46 PM Jens Axboe <axboe@kernel.dk> wrote:
>
> On 6/1/20 8:43 AM, Sedat Dilek wrote:
> > On Mon, Jun 1, 2020 at 4:35 PM Jens Axboe <axboe@kernel.dk> wrote:
> >>
> >> On 6/1/20 8:14 AM, Jens Axboe wrote:
> >>> On 6/1/20 8:13 AM, Sedat Dilek wrote:
> >>>> On Mon, Jun 1, 2020 at 4:04 PM Jens Axboe <axboe@kernel.dk> wrote:
> >>>>>
> >>>>> On 6/1/20 7:35 AM, Sedat Dilek wrote:
> >>>>>> Hi Jens,
> >>>>>>
> >>>>>> with Linux v5.7 final I switched to linux-block.git/for-next and reverted...
> >>>>>>
> >>>>>> "block: read-ahead submission should imply no-wait as well"
> >>>>>>
> >>>>>> ...and see no boot-slowdowns.
> >>>>>
> >>>>> Can you try with these patches applied instead? Or pull my async-readahead
> >>>>> branch from the same location.
> >>>>>
> >>>>
> >>>> Yes, I can do that.
> >>>> I pulled from linux-block.git#async-readahead and will report later.
> >>>>
> >>>> Any specific testing desired by you?
> >>>
> >>> Just do your boot timing test and see if it works, thanks.
> >>
> >> Actually, can you just re-test with the current async-buffered.6 branch?
> >> I think the major surgery should wait for 5.9, we can do this a bit
> >> easier without having to touch everything around us.
> >>
> >
> > With linux-block.git#async-readahead:
> >
> >   mycompiler -Wp,-MD,kernel/.sys.o.d -nostdinc -isystem
> > /home/dileks/src/llvm-toolchain/install/lib/clang/10.0.1rc1/include
> > -I./arch/x86/include -I./arch/x86/include/generated  -I./include
> > -I./arch/x86/include/uapi -I./arch/x86/include/generated/uapi
> > -I./include/uapi -I./include/generated/uapi -include
> > ./include/linux/kconfig.h -include ./include/linux/compiler_types.h
> > -D__KERNEL__ -Qunused-arguments -Wall -Wundef
> > -Werror=strict-prototypes -Wno-trigraphs -fno-strict-aliasing
> > -fno-common -fshort-wchar -fno-PIE
> > -Werror=implicit-function-declaration -Werror=implicit-int
> > -Wno-format-security -std=gnu89 -no-integrated-as
> > -Werror=unknown-warning-option -mno-sse -mno-mmx -mno-sse2 -mno-3dnow
> > -mno-avx -m64 -mno-80387 -mstack-alignment=8 -mtune=generic
> > -mno-red-zone -mcmodel=kernel -Wno-sign-compare
> > -fno-asynchronous-unwind-tables -mretpoline-external-thunk
> > -fno-delete-null-pointer-checks -Wno-address-of-packed-member -O2
> > -Wframe-larger-than=2048 -fstack-protector-strong
> > -Wno-format-invalid-specifier -Wno-gnu -mno-global-merge
> > -Wno-unused-const-variable -g -gz=zlib -pg -mfentry -DCC_USING_FENTRY
> > -Wdeclaration-after-statement -Wvla -Wno-pointer-sign
> > -Wno-array-bounds -fno-strict-overflow -fno-merge-all-constants
> > -fno-stack-check -Werror=date-time -Werror=incompatible-pointer-types
> > -fmacro-prefix-map=./= -fcf-protection=none -Wno-initializer-overrides
> > -Wno-format -Wno-sign-compare -Wno-format-zero-length
> > -Wno-tautological-constant-out-of-range-compare
> > -DKBUILD_MODFILE='"kernel/sys"' -DKBUILD_BASENAME='"sys"'
> > -DKBUILD_MODNAME='"sys"' -c -o kernel/sys.o kernel/sys.c
> > fs/9p/vfs_addr.c:112:4: error: use of undeclared identifier 'filp'
> >                         filp->private_data);
> >                         ^
> > 1 error generated.
> > make[5]: *** [scripts/Makefile.build:267: fs/9p/vfs_addr.o] Error 1
> > make[4]: *** [scripts/Makefile.build:488: fs/9p] Error 2
> > make[3]: *** [Makefile:1735: fs] Error 2
> > make[3]: *** Waiting for unfinished jobs....
> >
> > I guess block.git#async-buffered.6 needs the same revert of "block:
> > read-ahead submission should imply no-wait as well".
>
> Sorry, forgot to push out the updated version. But as per previous
> email, I think that major work should wait. Just try the updated
> async-buffered.6 branch instead.
>
> --
> Jens Axboe
>

Looks good.

$ cat systemd-analyze_time_5.7.0-2-amd64-clang.txt
Startup finished in 6.054s (kernel) + 42.626s (userspace) = 48.680s
graphical.target reached after 42.605s in userspace

$ cat systemd-analyze_blame_5.7.0-2-amd64-clang.txt
25.777s udisks2.service
24.527s accounts-daemon.service
18.228s polkit.service
17.593s avahi-daemon.service
17.554s NetworkManager.service
17.526s rtkit-daemon.service
17.417s dev-sdc2.device
17.200s switcheroo-control.service
17.186s wpa_supplicant.service
17.174s systemd-logind.service
15.714s smartmontools.service
14.255s zramswap.service
13.495s NetworkManager-wait-online.service
11.092s rsyslog.service
 8.900s networking.service
 7.817s gpm.service
 6.571s ModemManager.service
 6.403s packagekit.service
 5.701s e2scrub_reap.service
 3.714s alsa-restore.service
 3.509s systemd-udevd.service
 3.286s apparmor.service
 3.028s atd.service
 2.361s exim4.service
 1.532s systemd-tmpfiles-setup.service
 1.213s systemd-journal-flush.service
 1.163s keyboard-setup.service
 1.011s bluetooth.service
 1.008s binfmt-support.service
  995ms systemd-modules-load.service
  987ms pppd-dns.service
  919ms systemd-journald.service
  875ms upower.service
  856ms systemd-sysusers.service
  750ms systemd-udev-trigger.service
  728ms dev-hugepages.mount
  725ms dev-mqueue.mount
  724ms sys-kernel-debug.mount
  723ms sys-kernel-tracing.mount
  648ms modprobe@drm.service
  640ms ifupdown-wait-online.service
  609ms systemd-timesyncd.service
  579ms systemd-tmpfiles-setup-dev.service
  521ms systemd-backlight@backlight:intel_backlight.service
  477ms systemd-remount-fs.service
  454ms systemd-sysctl.service
  441ms systemd-random-seed.service
  389ms console-setup.service
  306ms systemd-rfkill.service
  266ms kmod-static-nodes.service
  244ms user@1000.service
  150ms proc-sys-fs-binfmt_misc.mount
  131ms systemd-update-utmp.service
   98ms systemd-user-sessions.service
   26ms user-runtime-dir@1000.service
   15ms systemd-update-utmp-runlevel.service
    6ms sys-fs-fuse-connections.mount
    3ms ifupdown-pre.service

$ cat systemd-analyze_critical-chain_5.7.0-2-amd64-clang.txt
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.

graphical.target @42.605s
└─multi-user.target @42.604s
  └─exim4.service @40.240s +2.361s
    └─network-online.target @40.237s
      └─NetworkManager-wait-online.service @26.740s +13.495s
        └─NetworkManager.service @9.177s +17.554s
          └─dbus.service @9.170s
            └─basic.target @9.081s
              └─sockets.target @9.080s
                └─dbus.socket @9.080s
                  └─sysinit.target @8.814s

└─systemd-backlight@backlight:intel_backlight.service @23.030s +521ms
                      └─system-systemd\x2dbacklight.slice @22.251s
                        └─system.slice @2.429s
                          └─-.slice @2.429s

Hope this helps you.

- Sedat -