Re: [PATCH 3/4] fs: add fput_queue

[Jeff Layton <jlayton@poochiereds.net>]

On Mon, 14 Sep 2015 17:39:54 +0100
Al Viro <viro@ZenIV.linux.org.uk> wrote:

> On Mon, Sep 14, 2015 at 10:19:18AM -0400, Jeff Layton wrote:
> > > > + * borrow the infrastructure used by kthreads, and the task can then just
> > > > + * called flush_delayed_fput to ensure that the final fput has completed.
> > > 
> > > Are you sure that it's not a typo?
> > 
> > I don't think so, but it could be clearer. Something like this maybe?
> > 
> >      "then we can't queue it via task_work_add."
> 
> Huh?
> 
> task_work_add() callbacks *will* run before we return to userland

Right, but only just before. We need it to run before we try to set the
lease in the context of a fcntl() call. How about this text instead
then? I'll fix up the patch if this sounds reasonable:

"When fput is called in the context of a userland process, it'll queue
the actual work (__fput()) to be done just before returning to userland. In
some cases however, we need to ensure that the __fput runs before that
point. There is no safe way to flush work that has been queued via
task_work_add however, so to do this we borrow the delayed_fput
infrastructure that kthreads use. The userland process can use
fput_queue() on one or more struct files and then call
flush_delayed_fput() to ensure that they are completely closed."

-- 
Jeff Layton <jlayton@poochiereds.net>