Linux-Fsdevel Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Amir Goldstein <amir73il@gmail.com>
To: Mel Gorman <mgorman@techsingularity.net>
Cc: Jan Kara <jack@suse.cz>, linux-fsdevel <linux-fsdevel@vger.kernel.org>
Subject: Re: fsnotify: minimise overhead when there are no marks related to sb
Date: Sun, 26 Jul 2020 18:20:26 +0300	[thread overview]
Message-ID: <CAOQ4uxgYpufPyhivOQyEhUQ0g+atKLwAAuefkSwaWXYAyMgw5Q@mail.gmail.com> (raw)
In-Reply-To: <CAOQ4uxi5Zpp7rCKdOkdw9Nkd=uGC-K2AuLqOFc0WQc_CgJQP2Q@mail.gmail.com>

On Thu, Jul 9, 2020 at 8:56 PM Amir Goldstein <amir73il@gmail.com> wrote:
>
> > > > Otherwise the patch looks good. One observation though: The (mask &
> > > > FS_MODIFY) check means that all vfs_write() calls end up going through the
> > > > "slower" path iterating all mark types and checking whether there are marks
> > > > anyway. That could be relatively simply optimized using a hidden mask flag
> > > > like FS_ALWAYS_RECEIVE_MODIFY which would be set when there's some mark
> > > > needing special handling of FS_MODIFY... Not sure if we care enough at this
> > > > point...
> > >
> > > Yeh that sounds low hanging.
> > > Actually, I Don't think we need to define a flag for that.
> > > __fsnotify_recalc_mask() can add FS_MODIFY to the object's mask if needed.
> >
> > Yes, that would be even more elegant.
> >
> > > I will take a look at that as part of FS_PRE_MODIFY work.
> > > But in general, we should fight the urge to optimize theoretic
> > > performance issues...
> >
> > Agreed. I just suspect this may bring measurable benefit for hackbench pipe
> > or tiny tmpfs writes after seeing Mel's results. But as I wrote this is a
> > separate idea and without some numbers confirming my suspicion I don't
> > think the complication is worth it so I don't want you to burn time on this
> > unless you're really interested :).
> >
>
> You know me ;-)
> FS_MODIFY optimization pushed to fsnotify_pre_modify branch.
> Only tested that LTP tests pass.
>
> Note that this is only expected to improve performance in case there *are*
> marks, but not marks with ignore mask, because there is an earlier
> optimization in fsnotify() for the no marks case.
>

Hi Mel,

After following up on Jan's suggestion above, I realized there is another
low hanging optimization we can make.

As you may remember, one of the solutions we considered was to exclude
special or internal sb's from notifications based on some SB flag, but making
assumptions about which sb are expected to provide notifications turned out
to be a risky game.

We can however, keep a counter on sb to *know* there are no watches
on any object in this sb, so the test:

        if (!sb->s_fsnotify_marks &&
            (!mnt || !mnt->mnt_fsnotify_marks) &&
            (!inode || !inode->i_fsnotify_marks))
                return 0;

Which is not so nice for inlining, can be summarized as:

        if (atomic_long_read(&inode->i_sb->s_fsnotify_obj_refs) == 0)
                return 0;

Which is nicer for inlining.

I am not sure if you had a concrete reason for:
"fs: Do not check if there is a fsnotify watcher on pseudo inodes"
or if you did it for the sport.

I have made the change above for the sport and for now I do not
intend to post it for review unless a good reason comes up.

If you are interested or curious to queue this code to Suse perf testing,
I pushed it to branch fsnotify-perf [1]. It may be interesting to see if it
won back the cpu cycles lost in the revert of your patch.

This branch is based on some changes already in Jan's tree and some
changes in my development tree (fsnotify_pre_modify), but you already
fed this development branch to perf test machine once and reported back
that there was no significant degradation.

I can also provide the optimization patches based on Linus' tree if needed.

Thanks,
Amir.

[1] https://github.com/amir73il/linux/commits/fsnotify-perf

  reply	other threads:[~2020-07-26 15:20 UTC|newest]

Thread overview: 35+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-12  9:33 [PATCH 00/20] Prep work for fanotify named events Amir Goldstein
2020-06-12  9:33 ` [PATCH 01/20] fsnotify: Rearrange fast path to minimise overhead when there is no watcher Amir Goldstein
2020-07-03 14:03   ` Jan Kara
2020-07-04  9:30     ` Amir Goldstein
2020-07-06 11:05       ` Jan Kara
2020-07-09 17:56         ` fsnotify: minimise overhead when there are no marks with ignore mask Amir Goldstein
2020-07-26 15:20           ` Amir Goldstein [this message]
2020-07-27  7:44             ` fsnotify: minimise overhead when there are no marks related to sb Jan Kara
2020-07-27 10:02               ` Amir Goldstein
2020-06-12  9:33 ` [PATCH 02/20] fsnotify: fold fsnotify() call into fsnotify_parent() Amir Goldstein
2020-06-12  9:33 ` [PATCH 03/20] fsnotify: return non const from fsnotify_data_inode() Amir Goldstein
2020-06-12  9:33 ` [PATCH 04/20] nfsd: use fsnotify_data_inode() to get the unlinked inode Amir Goldstein
2020-06-12 10:25   ` Amir Goldstein
2020-06-12  9:33 ` [PATCH 05/20] kernfs: do not call fsnotify() with name without a parent Amir Goldstein
2020-06-29 13:27   ` Tejun Heo
2020-06-29 16:11   ` Greg Kroah-Hartman
2020-06-12  9:33 ` [PATCH 06/20] inotify: do not use objectid when comparing events Amir Goldstein
2020-06-12  9:33 ` [PATCH 07/20] fanotify: create overflow event type Amir Goldstein
2020-06-12  9:33 ` [PATCH 08/20] fanotify: break up fanotify_alloc_event() Amir Goldstein
2020-06-12  9:33 ` [PATCH 09/20] fsnotify: pass dir argument to handle_event() callback Amir Goldstein
2020-07-03 14:49   ` Jan Kara
2020-06-12  9:33 ` [PATCH 10/20] fanotify: generalize the handling of extra event flags Amir Goldstein
2020-06-12  9:33 ` [PATCH 11/20] fanotify: generalize merge logic of events on dir Amir Goldstein
2020-06-12  9:33 ` [PATCH 12/20] fanotify: distinguish between fid encode error and null fid Amir Goldstein
2020-06-12  9:33 ` [PATCH 13/20] fanotify: generalize test for FAN_REPORT_FID Amir Goldstein
2020-06-12  9:33 ` [PATCH 14/20] fanotify: mask out special event flags from ignored mask Amir Goldstein
2020-06-12  9:33 ` [PATCH 15/20] fanotify: prepare for implicit event flags in mark mask Amir Goldstein
2020-06-12  9:33 ` [PATCH 16/20] fanotify: use FAN_EVENT_ON_CHILD as implicit flag on sb/mount/non-dir marks Amir Goldstein
2020-06-12  9:33 ` [PATCH 17/20] fanotify: remove event FAN_DIR_MODIFY Amir Goldstein
2020-06-12  9:33 ` [PATCH 18/20] fsnotify: add object type "child" to object type iterator Amir Goldstein
2020-06-12  9:33 ` [PATCH 19/20] fanotify: move event name into fanotify_fh Amir Goldstein
2020-07-03 16:02   ` Jan Kara
2020-07-06  8:21     ` Amir Goldstein
2020-07-06 15:24       ` Jan Kara
2020-06-12  9:33 ` [PATCH 20/20] fanotify: no external fh buffer in fanotify_name_event Amir Goldstein

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=CAOQ4uxgYpufPyhivOQyEhUQ0g+atKLwAAuefkSwaWXYAyMgw5Q@mail.gmail.com \
    --to=amir73il@gmail.com \
    --cc=jack@suse.cz \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=mgorman@techsingularity.net \
    --subject='Re: fsnotify: minimise overhead when there are no marks related to sb' \
    /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

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).