> OK, nice trick but for this series, I'd like to keep the original ignore > mask behavior (bug to bug compatibility) or possibly let parent's ignore > mask be applied only for events being sent to the parent due to its > FS_EVENT_ON_CHILD. That should be easy if we set the FS_EVENT_ON_CHILD flag only for the case of a watching parent. And I believe that would make the flag meaningful again and not redundant as you now see it. Please note that the series is already not "bug compatible", because the patch " send event to parent and child with single callback" already fixes the combination parent watch + child ignore, which did not work before and this is declared in commit message. > Can you please fix that up? Will do. I will also take care of LTP test coverage for the ignore mask cases that have been fixed and for dnotify/inotify parent+child watching case. > I won't get to it before I > leave for vacation but once I return, I'd like to just pick the fixed up > commit and push everything to linux-next... Thanks! > Thanks a lot for everything! Amir.
