Re: [PATCH 2/3] mm/page_ref: Ensure page_ref_unfreeze is ordered against prior accesses

[Will Deacon <will.deacon@arm.com>]

On Thu, Jun 08, 2017 at 02:50:59PM +0200, Peter Zijlstra wrote:
> On Thu, Jun 08, 2017 at 12:24:33PM +0100, Will Deacon wrote:
> > [+ PeterZ]
> > 
> > On Thu, Jun 08, 2017 at 01:07:02PM +0200, Vlastimil Babka wrote:
> > > On 06/08/2017 12:40 PM, Kirill A. Shutemov wrote:
> > > > On Thu, Jun 08, 2017 at 11:38:21AM +0200, Vlastimil Babka wrote:
> > > >> On 06/06/2017 07:58 PM, Will Deacon wrote:
> > > >>>  include/linux/page_ref.h | 1 +
> > > >>>  1 file changed, 1 insertion(+)
> > > >>>
> > > >>> diff --git a/include/linux/page_ref.h b/include/linux/page_ref.h
> > > >>> index 610e13271918..74d32d7905cb 100644
> > > >>> --- a/include/linux/page_ref.h
> > > >>> +++ b/include/linux/page_ref.h
> > > >>> @@ -174,6 +174,7 @@ static inline void page_ref_unfreeze(struct page *page, int count)
> > > >>>  	VM_BUG_ON_PAGE(page_count(page) != 0, page);
> > > >>>  	VM_BUG_ON(count == 0);
> > > >>>  
> > > >>> +	smp_mb__before_atomic();
> > > >>>  	atomic_set(&page->_refcount, count);
> 
> So depending on what it actually required, we do have
> atomic_set_release() (atomic_t equivalent to smp_store_release()).

Yeah, I was wondering about that this morning. I think it should do the
trick here, but smp_mb() would be a better fit for the other parts of this
API (page_ref_freeze uses atomic_cmpxchg and page_cache_get_speculative
uses atomic_add_unless).

I'll send a v2 with the full barrier.

Will

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>