Received-SPF: pass (google.com: domain of linux-ext4-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Thu, 5 Aug 2021 14:06:11 +0200
From:   Jan Kara <jack@suse.cz>
To:     Gabriel Krisman Bertazi <krisman@collabora.com>
Cc:     jack@suse.com, amir73il@gmail.com, djwong@kernel.org,
        tytso@mit.edu, david@fromorbit.com, dhowells@redhat.com,
        khazhy@google.com, linux-fsdevel@vger.kernel.org,
        linux-ext4@vger.kernel.org, linux-api@vger.kernel.org,
        kernel@collabora.com
Subject: Re: [PATCH v5 19/23] fanotify: Report fid info for file related file
 system errors
Message-ID: <20210805120611.GK14483@quack2.suse.cz>
References: <20210804160612.3575505-1-krisman@collabora.com>
 <20210804160612.3575505-20-krisman@collabora.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20210804160612.3575505-20-krisman@collabora.com>
User-Agent: Mutt/1.10.1 (2018-07-13)
Precedence: bulk

On Wed 04-08-21 12:06:08, Gabriel Krisman Bertazi wrote:
> Plumb the pieces to add a FID report to error records.  Since all error
> event memory must be pre-allocated, we estimate a file handler size and
> if it is insuficient, we report an invalid FID and increase the
> prediction for the next error slot allocation.
> 
> For errors that don't expose a file handler report it with an invalid
> FID.
> 
> Signed-off-by: Gabriel Krisman Bertazi <krisman@collabora.com>

...

> @@ -715,6 +717,31 @@ static void fanotify_insert_error_event(struct fsnotify_group *group,
>  	fee->fae.type = FANOTIFY_EVENT_TYPE_FS_ERROR;
>  	fee->error = report->error;
>  	fee->err_count = 1;
> +	fee->fsid = fee->sb_mark->fsn_mark.connector->fsid;
> +
> +	/*
> +	 * Error reporting needs to happen in atomic context.  If this
> +	 * inode's file handler length is more than we initially
> +	 * predicted, there is nothing better we can do than report the
> +	 * error with a bad file handler.
> +	 */
> +	fh_len = fanotify_encode_fh_len(inode);
> +	if (fh_len > fee->sb_mark->pred_fh_len) {
> +		pr_warn_ratelimited(
> +			"FH overflows error event. Drop inode information.\n");
> +		/*
> +		 * Update the handler size prediction for the next error
> +		 * event allocation.  This reduces the chance of another
> +		 * overflow.
> +		 */
> +		fee->sb_mark->pred_fh_len = fh_len;
> +
> +		/* For the current error, ignore the inode information. */
> +		inode = NULL;
> +		fh_len = fanotify_encode_fh_len(NULL);
> +	}
> +
> +	fanotify_encode_fh(&fee->object_fh, inode, fh_len, NULL, 0);
>  }

I don't think calling fanotify_encode_fh() from under notification_lock is
a good practice. It calls into a filesystem and god knows which locks it may
need to take to get necessary info... See my reply to patch 18 how we could
handle that better.

								Honza
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR