On Wed 22-07-20 15:58:41, Amir Goldstein wrote:
> When reporting event with parent/name info, we should not skip sb/mount
> marks mask if event has FAN_EVENT_ON_CHILD in the mask.
>
> This check is a leftover from the time when the event on child was
> reported in a separate callback than the event on parent and we did
> not want to get duplicate events for sb/mount mark.
>
> Fixes: eca4784cbb18 ("fsnotify: send event to parent and child with single callback")
> Signed-off-by: Amir Goldstein <amir73il@xxxxxxxxx>
OK, I've decided to just drop "fanotify: report both events on parent and
child with single callback" because it didn't improve anything and amend
eca4784cbb18 with this change...
Honza
> ---
> fs/notify/fanotify/fanotify.c | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/fs/notify/fanotify/fanotify.c b/fs/notify/fanotify/fanotify.c
> index a24f08a9c50f..36ea0cd6387e 100644
> --- a/fs/notify/fanotify/fanotify.c
> +++ b/fs/notify/fanotify/fanotify.c
> @@ -265,13 +265,11 @@ static u32 fanotify_group_event_mask(struct fsnotify_group *group,
> continue;
>
> /*
> - * If the event is for a child and this mark doesn't care about
> - * events on a child, don't send it!
> - * The special object type "child" always cares about events on
> - * a child, because it refers to the child inode itself.
> + * If the event is for a child and this mark is on a parent not
> + * watching children, don't send it!
> */
> if (event_mask & FS_EVENT_ON_CHILD &&
> - type != FSNOTIFY_OBJ_TYPE_CHILD &&
> + type == FSNOTIFY_OBJ_TYPE_INODE &&
> !(mark->mask & FS_EVENT_ON_CHILD))
> continue;
>
> --
> 2.17.1
>
--
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR
