Re: [PATCH v5 15/17] fanotify: support events with data type FSNOTIFY_EVENT_INODE

[Jan Kara <jack@suse.cz>]

On Thu 10-01-19 19:04:42, Amir Goldstein wrote:
> When event data type is FSNOTIFY_EVENT_INODE, we don't have a refernece
> to the mount, so we will not be able to open a file descriptor when user
> reads the event. However, if the listener has enabled reporting file
> identifier with the FAN_REPORT_FID init flag, we allow reporting those
> events and we use an identifier inode to encode fid.
> 
> The inode to use as identifier when reporting fid depends on the event.
> For dirent modification events, we report the modified directory inode
> and we report the "victim" inode otherwise.
> For example:
> FS_ATTRIB reports the child inode even if reported on a watched parent.
> FS_CREATE reports the modified dir inode and not the created inode.
> 
> Signed-off-by: Amir Goldstein <amir73il@gmail.com>
> ---
>  fs/notify/fanotify/fanotify.c      | 67 ++++++++++++++++++++----------
>  fs/notify/fanotify/fanotify.h      |  2 +-
>  fs/notify/fanotify/fanotify_user.c |  3 +-
>  3 files changed, 48 insertions(+), 24 deletions(-)
> 
> diff --git a/fs/notify/fanotify/fanotify.c b/fs/notify/fanotify/fanotify.c
> index fcb98ea99508..e3ca1632feb8 100644
> --- a/fs/notify/fanotify/fanotify.c
> +++ b/fs/notify/fanotify/fanotify.c
> @@ -96,7 +96,7 @@ static int fanotify_get_response(struct fsnotify_group *group,
>  
>  	pr_debug("%s: group=%p event=%p about to return ret=%d\n", __func__,
>  		 group, event, ret);
> -	
> +
>  	return ret;
>  }
>  
> @@ -106,9 +106,10 @@ static int fanotify_get_response(struct fsnotify_group *group,
>   * been included within the event mask, but have not been explicitly
>   * requested by the user, will not be present in the returned mask.
>   */
> -static u32 fanotify_group_event_mask(struct fsnotify_iter_info *iter_info,
> -				       u32 event_mask, const void *data,
> -				       int data_type)
> +static u32 fanotify_group_event_mask(struct fsnotify_group *group,
> +				     struct fsnotify_iter_info *iter_info,
> +				     u32 event_mask, const void *data,
> +				     int data_type)
>  {
>  	__u32 marks_mask = 0, marks_ignored_mask = 0;
>  	const struct path *path = data;
> @@ -118,14 +119,14 @@ static u32 fanotify_group_event_mask(struct fsnotify_iter_info *iter_info,
>  	pr_debug("%s: report_mask=%x mask=%x data=%p data_type=%d\n",
>  		 __func__, iter_info->report_mask, event_mask, data, data_type);
>  
> -	/* If we don't have enough info to send an event to userspace say no */
> -	if (data_type != FSNOTIFY_EVENT_PATH)
> -		return 0;
> -
> -	/* Sorry, fanotify only gives a damn about files and dirs */
> -	if (!d_is_reg(path->dentry) &&
> -	    !d_can_lookup(path->dentry))
> +	if (data_type == FSNOTIFY_EVENT_PATH) {
> +		/* Path type events are only relevant for files and dirs */
> +		if (!d_is_reg(path->dentry) && !d_can_lookup(path->dentry))
> +			return 0;

Hum, does this mean that fanotify won't report O_PATH open on symlink
unlike inotify? So shouldn't the condition rather be:

	if (!FAN_GROUP_FLAG(group, FAN_REPORT_FID)) {
		if (data_type != FSNOTIFY_EVENT_PATH)
			return 0;
		if (!d_is_reg(path->dentry) && !d_can_lookup(path->dentry))
			return 0;
	}

?
								Honza

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