Re: [patch] stop inotify from sending random DELETE_SELF event under load

[John McCutchan <ttb@tentacle.dhs.org>]

On Tue, 2005-09-20 at 05:46 +0100, Al Viro wrote:
> On Tue, Sep 20, 2005 at 12:36:11AM -0400, John McCutchan wrote:
> > On Tue, 2005-09-20 at 05:24 +0100, Al Viro wrote:
> > > On Mon, Sep 19, 2005 at 09:03:36PM -0700, Linus Torvalds wrote:
> > > > One possibility is to mark the dentry deleted in d_flags. That would mean 
> > > > something like this (against the just-pushed-put v2.6.14-rc2, which has 
> > > > my previous hack).
> > > > 
> > > > Untested. Al?
> > >  
> > > Uhh...  I still don't understand which behaviour do you want.
> > 
> > 
> > > 	* removal of this link, at the moment when it stops being accessible
> > > [ none of the above, better done from vfs_...() ]
> > 
> > That is the behaviour we want, how does Linus's second patch not
> > accomplish this? 
> 
> fd = open("foo", 0);
> unlink("foo");
> sleep for ten days
> close(fd);
> 
> 	Linus' patch will send event on close().  Ten days since the moment
> when any lookups on foo would bring you -ENOENT.
> 


Ahh, got it.

> 	Could you please describe the semantics of your events?

DELETE_SELF WD=X

The path you requested a watch on (inotify_add_watch(path,mask) returned
X) has been deleted.

-- 
John McCutchan <ttb@tentacle.dhs.org>