All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
To: Liu Bo <bo.liu@linux.alibaba.com>, mszeredi@redhat.com
Cc: virtio-fs@redhat.com
Subject: Re: [Virtio-fs] [PATCH] virtiofsd: enable PARALLEL_DIROPS during INIT
Date: Fri, 7 Jun 2019 19:33:10 +0100	[thread overview]
Message-ID: <20190607183310.GY2631@work-vm> (raw)
In-Reply-To: <20190607181533.r7xkszkdi377o5ha@US-160370MP2.local>

* Liu Bo (bo.liu@linux.alibaba.com) wrote:
> On Fri, Jun 07, 2019 at 04:21:47PM +0100, Dr. David Alan Gilbert wrote:
> > * Liu Bo (bo.liu@linux.alibaba.com) wrote:
> > > lookup is a RO operations, PARALLEL_DIROPS can be enabled.
> > > 
> > > Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
> > 
> > I'm not too sure what this is doing; and why doesn't upstream libfuse
> > have this?
> 
> I'm not sure why libfuse doesn't include the capability, but AFAICT this
> is needed for practical use.

Miklos: Do you know what the rules are around PARALLEL_DIROPS and why
the upstream libfuse doesn't have it eanbled?

Dave


> Say that we have 100 threads doing 'ls /mnt/virtiofs/', then they are
> serialized by a mutex in
> ---
> fuse_lookup
>   fuse_lock_inode
>     if (!get_fuse_conn(inode)->parallel_dirops)
>       mutex_lock(&get_fuse_inode(inode)->mutex);
> ---
> 
> What fuse_lookup does is a plain lookup with loading inode if necessary,
> so I don't see anything wrong making operations parallel.
> 
> This is found by a fsmark test,
> ----
> NFILES=100000
> 
> time $FSMARK  -D  10000  -S0  -n  $NFILES  -s  0  -L  5 -l /tmp/fs_log.txt \
>         -d  $MNT/0  -d  $MNT/1 \
>         -d  $MNT/2  -d  $MNT/3 \
>         -d  $MNT/4  -d  $MNT/5 \
>         -d  $MNT/6  -d  $MNT/7 \
>         -d  $MNT/8  -d  $MNT/9 \
>         -d  $MNT/10  -d  $MNT/11 \
>         -d  $MNT/12  -d  $MNT/13 \
>         -d  $MNT/14  -d  $MNT/15
> ----
> 
> the vanilla virtiofs was super slow due to the mutex lock.
> 
> thanks,
> -liubo
> 
> 
> > 
> > Dave
> > 
> > > ---
> > >  contrib/virtiofsd/fuse_lowlevel.c | 2 ++
> > >  1 file changed, 2 insertions(+)
> > > 
> > > diff --git a/contrib/virtiofsd/fuse_lowlevel.c b/contrib/virtiofsd/fuse_lowlevel.c
> > > index c460c4c..93ce788 100644
> > > --- a/contrib/virtiofsd/fuse_lowlevel.c
> > > +++ b/contrib/virtiofsd/fuse_lowlevel.c
> > > @@ -2089,6 +2089,8 @@ static void do_init(fuse_req_t req, fuse_ino_t nodeid,
> > >  
> > >  	if (se->conn.want & FUSE_CAP_ASYNC_READ)
> > >  		outarg.flags |= FUSE_ASYNC_READ;
> > > +        if (se->conn.want & FUSE_CAP_PARALLEL_DIROPS)
> > > +                outarg.flags |= FUSE_PARALLEL_DIROPS;
> > >  	if (se->conn.want & FUSE_CAP_POSIX_LOCKS)
> > >  		outarg.flags |= FUSE_POSIX_LOCKS;
> > >  	if (se->conn.want & FUSE_CAP_ATOMIC_O_TRUNC)
> > > -- 
> > > 1.8.3.1
> > > 
> > --
> > Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK


  reply	other threads:[~2019-06-07 18:33 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-06-07  2:38 [Virtio-fs] [PATCH] virtiofsd: enable PARALLEL_DIROPS during INIT Liu Bo
2019-06-07 15:21 ` Dr. David Alan Gilbert
2019-06-07 18:15   ` Liu Bo
2019-06-07 18:33     ` Dr. David Alan Gilbert [this message]
2019-06-07 19:17       ` Miklos Szeredi
2019-06-07 19:23         ` Dr. David Alan Gilbert
2019-06-07 19:29           ` Liu Bo
2019-06-07 19:35           ` Miklos Szeredi
2019-06-10  9:04             ` Dr. David Alan Gilbert

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=20190607183310.GY2631@work-vm \
    --to=dgilbert@redhat.com \
    --cc=bo.liu@linux.alibaba.com \
    --cc=mszeredi@redhat.com \
    --cc=virtio-fs@redhat.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.