Re: [PATCH RFC] kernel/locking, fs/direct-io: Introduce and use down_write_nolockdep()

[Bart Van Assche <bvanassche@acm.org>]

On Fri, 2018-10-26 at 10:43 -0700, Matthew Wilcox wrote:
> On Fri, Oct 26, 2018 at 09:49:05AM -0700, Bart Van Assche wrote:
> > +++ b/include/linux/rwsem.h
> > @@ -41,6 +41,10 @@ struct rw_semaphore {
> >  #endif
> >  #ifdef CONFIG_DEBUG_LOCK_ALLOC
> >  	struct lockdep_map	dep_map;
> > +	/*
> > +	 * Number of up_write() calls that must skip rwsem_release().
> > +	 */
> > +	unsigned		nolockdep;
> 
> This reads a bit weird.  By definition, only one writer is allowed
> at a time.  And you can't call up_write() before down_write().
> So functionally, this is a bool, and the comment should at least
> ackowledge that, even if it remains implemented as an unsigned int.
> 
> I'd suggest the implementation uses '= 1' and '= 0' rather than ++ and --.

Hi Matthew,

That sounds like a good idea to me.

> > diff --git a/mm/rmap.c b/mm/rmap.c
> > index 1e79fac3186b..2a953d3b7431 100644
> > --- a/mm/rmap.c
> > +++ b/mm/rmap.c
> > @@ -81,6 +81,7 @@ static inline struct anon_vma *anon_vma_alloc(void)
> >  
> >  	anon_vma = kmem_cache_alloc(anon_vma_cachep, GFP_KERNEL);
> >  	if (anon_vma) {
> > +		init_rwsem(&anon_vma->rwsem);
> >  		atomic_set(&anon_vma->refcount, 1);
> >  		anon_vma->degree = 1;	/* Reference for first vma */
> >  		anon_vma->parent = anon_vma;
> 
> Why is this needed?  The anon_vma_ctor() already calls init_rwsem().
> 
> (I suspect this is one of those ctors that isn't actually useful and
> should be inlined into anon_vma_alloc())

Without that call I noticed that the "nolockdep" variable was sometimes set
when down_write() got called. Does that mean that it can happen that an
anon_vma structure is freed without releasing anon_vma->rwsem?

Thanks,

Bart.