Re: On setting a lease across a cluster

["J. Bruce Fields" <bfields@fieldses.org>]

On Fri, Jan 04, 2008 at 01:35:50PM -0700, Matthew Wilcox wrote:
> > > > vfs_setlease()
> > > >    if (f_op->setlease())
> > > >       res = f_op->setlease()
> > > >       if (res)
> > > >          return res;
> > > >    lock_kernel()
> > > >    generic_setlease()
> > > >    unlock_kernel()
> > 
> > Why can't the filesystem call into generic_setlease() on its own?
> 
> Because (assuming we're rid of the BKL), fcntl_setlease() needs to
> acquire the spinlock and hold it while generic_setlease() runs, so
> generic_setlease() can't acquire the lock.

So, the problem is that fcntl_setlease() does

	vfs_setlease()
	fasync_helper()

which the bkl held over both, and you want to preserve that?

But what that BKL is doing is a mystery to me--the very first thing that
fasync_helper() does is kmem_cache_alloc(., GFP_KERNEL).  So you won't
be introducing any new problem if you lock those two operations
separately.  Unless I'm totally missing something.

--b.