Re: hugetlb locking bug.

[Peter Zijlstra <peterz@infradead.org>]

On Fri, 2011-04-15 at 16:57 -0400, Christoph Hellwig wrote:
> On Fri, Apr 15, 2011 at 01:49:04PM -0700, Linus Torvalds wrote:
> > And I really thought we annotated it as such with different
> > "lockdep_set_class()" cases (ie the whole
> > 
> >   lockdep_set_class(&inode->i_mutex,&type->i_mutex_dir_key);
> > 
> > for the S_ISDIR case in unlock_new_inode().
> > 
> > Can somebody more alert than me see why this lockdep issue still
> > triggers with hugetlbfs?
> 
> Because it doesn't use iget or unlock_new_inode, but rather calls
> directly into new_inode().  It and other filesystems not using
> unlock_new_inode will need a local copy of that logic.

Is there a sane reason they do their own magic, and thus need a copy of
the logic, instead of using the generic code that already has it?