Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
In-Reply-To: <154149744356.18189.13444618369955494886.stgit@localhost.localdomain>
References: <154149673368.18189.14752124997476502777.stgit@localhost.localdomain>
 <154149744356.18189.13444618369955494886.stgit@localhost.localdomain>
From:   Miklos Szeredi <miklos@szeredi.hu>
Date:   Wed, 7 Nov 2018 15:20:42 +0100
Message-ID: <CAJfpegtnCTY_eWsyNPRPaE0=wPeyMPAXScWQ4scV7vXXcWyamA@mail.gmail.com>
Subject: Re: [PATCH 5/6] fuse: Protect fuse_inode::nlookup with fuse_inode::lock
To:     Kirill Tkhai <ktkhai@virtuozzo.com>
Cc:     linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Nov 6, 2018 at 10:44 AM, Kirill Tkhai <ktkhai@virtuozzo.com> wrote:
> This continues previous patch and introduces the same
> protection for nlookup field. It goes as separate patch
> since it's separate logic change (sadly, but it looks
> impossible to split previous patch more then in this way).

Well, the way you can split things up better is to just add the fine
grained lock first, which you can do in multiple patches, since all of
it is going to be a no-op.  And when that is done, remove the coarse
grained lock in yet another patch, which hopefully will result in
better performance and no other change.

>
> Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
> ---
>  fs/fuse/dir.c     |    4 ++--
>  fs/fuse/inode.c   |    4 ++--
>  fs/fuse/readdir.c |    4 ++--
>  3 files changed, 6 insertions(+), 6 deletions(-)
>
> diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
> index 35f3b3d1e044..ac8519285327 100644
> --- a/fs/fuse/dir.c
> +++ b/fs/fuse/dir.c
> @@ -222,9 +222,9 @@ static int fuse_dentry_revalidate(struct dentry *entry, unsigned int flags)
>                                 fuse_queue_forget(fc, forget, outarg.nodeid, 1);
>                                 goto invalid;
>                         }
> -                       spin_lock(&fc->lock);
> +                       spin_lock(&fi->lock);
>                         fi->nlookup++;
> -                       spin_unlock(&fc->lock);
> +                       spin_unlock(&fi->lock);
>                 }
>                 kfree(forget);
>                 if (ret == -ENOMEM)
> diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
> index 5f488b019cd9..b8092d49a4b2 100644
> --- a/fs/fuse/inode.c
> +++ b/fs/fuse/inode.c
> @@ -327,9 +327,9 @@ struct inode *fuse_iget(struct super_block *sb, u64 nodeid,
>         }
>
>         fi = get_fuse_inode(inode);
> -       spin_lock(&fc->lock);
> +       spin_lock(&fi->lock);
>         fi->nlookup++;
> -       spin_unlock(&fc->lock);
> +       spin_unlock(&fi->lock);
>         fuse_change_attributes(inode, attr, attr_valid, attr_version);
>
>         return inode;
> diff --git a/fs/fuse/readdir.c b/fs/fuse/readdir.c
> index ab18b78f4755..574d03f8a573 100644
> --- a/fs/fuse/readdir.c
> +++ b/fs/fuse/readdir.c
> @@ -213,9 +213,9 @@ static int fuse_direntplus_link(struct file *file,
>                 }
>
>                 fi = get_fuse_inode(inode);
> -               spin_lock(&fc->lock);
> +               spin_lock(&fi->lock);
>                 fi->nlookup++;
> -               spin_unlock(&fc->lock);
> +               spin_unlock(&fi->lock);
>
>                 forget_all_cached_acls(inode);
>                 fuse_change_attributes(inode, &o->attr,
>