Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Tue, 28 Mar 2023 09:54:40 +0200
From:   Jan Kara <jack@suse.cz>
To:     Christian Brauner <brauner@kernel.org>
Cc:     linux-fsdevel@vger.kernel.org, Jan Kara <jack@suse.cz>,
        Amir Goldstein <amir73il@gmail.com>,
        Matthew Bobrowski <repnop@google.com>,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH 3/3] fanotify: use pidfd_prepare()
Message-ID: <20230328075440.ho55dt2xhvbz7yog@quack3>
References: <20230327-pidfd-file-api-v1-0-5c0e9a3158e4@kernel.org>
 <20230327-pidfd-file-api-v1-3-5c0e9a3158e4@kernel.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230327-pidfd-file-api-v1-3-5c0e9a3158e4@kernel.org>
Precedence: bulk

On Mon 27-03-23 20:22:53, Christian Brauner wrote:
> We generally try to avoid installing a file descriptor into the caller's
> file descriptor table just to close it again via close_fd() in case an
> error occurs. Instead we reserve a file descriptor but don't install it
> into the caller's file descriptor table yet. If we fail for other,
> unrelated reasons we can just close the reserved file descriptor and if
> we make it past all meaningful error paths we just install it. Fanotify
> gets this right already for one fd type but not for pidfds.
> 
> Use the new pidfd_prepare() helper to reserve a pidfd and a pidfd file
> and switch to the more common fd allocation and installation pattern.
> 
> Signed-off-by: Christian Brauner <brauner@kernel.org>

Thanks for the improvement! It looks good to me. Feel free to add:

Acked-by: Jan Kara <jack@suse.cz>

								Honza

> ---
>  fs/notify/fanotify/fanotify_user.c | 13 +++++++++----
>  1 file changed, 9 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/notify/fanotify/fanotify_user.c b/fs/notify/fanotify/fanotify_user.c
> index 8f430bfad487..22fb1cf7e1fc 100644
> --- a/fs/notify/fanotify/fanotify_user.c
> +++ b/fs/notify/fanotify/fanotify_user.c
> @@ -663,7 +663,7 @@ static ssize_t copy_event_to_user(struct fsnotify_group *group,
>  	struct fanotify_info *info = fanotify_event_info(event);
>  	unsigned int info_mode = FAN_GROUP_FLAG(group, FANOTIFY_INFO_MODES);
>  	unsigned int pidfd_mode = info_mode & FAN_REPORT_PIDFD;
> -	struct file *f = NULL;
> +	struct file *f = NULL, *pidfd_file = NULL;
>  	int ret, pidfd = FAN_NOPIDFD, fd = FAN_NOFD;
>  
>  	pr_debug("%s: group=%p event=%p\n", __func__, group, event);
> @@ -718,7 +718,7 @@ static ssize_t copy_event_to_user(struct fsnotify_group *group,
>  		    !pid_has_task(event->pid, PIDTYPE_TGID)) {
>  			pidfd = FAN_NOPIDFD;
>  		} else {
> -			pidfd = pidfd_create(event->pid, 0);
> +			pidfd = pidfd_prepare(event->pid, 0, &pidfd_file);
>  			if (pidfd < 0)
>  				pidfd = FAN_EPIDFD;
>  		}
> @@ -751,6 +751,9 @@ static ssize_t copy_event_to_user(struct fsnotify_group *group,
>  	if (f)
>  		fd_install(fd, f);
>  
> +	if (pidfd_file)
> +		fd_install(pidfd, pidfd_file);
> +
>  	return metadata.event_len;
>  
>  out_close_fd:
> @@ -759,8 +762,10 @@ static ssize_t copy_event_to_user(struct fsnotify_group *group,
>  		fput(f);
>  	}
>  
> -	if (pidfd >= 0)
> -		close_fd(pidfd);
> +	if (pidfd >= 0) {
> +		put_unused_fd(pidfd);
> +		fput(pidfd_file);
> +	}
>  
>  	return ret;
>  }
> 
> -- 
> 2.34.1
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR