On Wed, Jun 22, 2022 at 7:00 PM Jan Kara <jack@xxxxxxx> wrote: > > On Mon 20-06-22 16:45:50, Amir Goldstein wrote: > > Setting flags FAN_ONDIR FAN_EVENT_ON_CHILD in ignore mask has no effect. > > The FAN_EVENT_ON_CHILD flag in mask implicitly applies to ignore mask and > > ignore mask is always implicitly applied to events on directories. > > > > Define a mark flag that replaces this legacy behavior with logic of > > applying the ignore mask according to event flags in ignore mask. > > > > Implement the new logic to prepare for supporting an ignore mask that > > ignores events on children and ignore mask that does not ignore events > > on directories. > > > > To emphasize the change in terminology, also rename ignored_mask mark > > member to ignore_mask and use accessor to get only ignored events or > > events and flags. > > > > This change in terminology finally aligns with the "ignore mask" > > language in man pages and in most of the comments. > > > > Signed-off-by: Amir Goldstein <amir73il@xxxxxxxxx> > > .. > > > @@ -423,7 +425,8 @@ static bool fsnotify_iter_select_report_types( > > * But is *this mark* watching children? > > */ > > if (type == FSNOTIFY_ITER_TYPE_PARENT && > > - !(mark->mask & FS_EVENT_ON_CHILD)) > > + !(mark->mask & FS_EVENT_ON_CHILD) && > > + !(fsnotify_ignore_mask(mark) & FS_EVENT_ON_CHILD)) > > continue; > > So now we have in ->report_mask the FSNOTIFY_ITER_TYPE_PARENT if either > ->mask or ->ignore_mask have FS_EVENT_ON_CHILD set. But I see nothing that > would stop us from applying say ->mask to the set of events we are > interested in if FS_EVENT_ON_CHILD is set only in ->ignore_mask? And I think I spent some time thinking about this and came to a conclusion that 1. It is hard to get all the cases right 2. It is a micro optimization The implication is that the user can set an ignore mask of an object, get no events but still cause performance penalty. Right? So just don't do that... It is easier to maintain the code with the rule that an "interest" in the object is either positive (I want this event) or negative (I don't want this event). If the user has no interest, the user should set nothing in the mark. Do you agree with me that the added complexity would not be worth it? > there's the same problem in the other direction as well. Am I missing Other direction? Thanks, Amir.
