Re: [PATCH v1 4/4] mm: Remove managed_page_count spinlock

[Michal Hocko]

[for some reason I have dropped the rest of the cc list. Restored]

On Tue 06-11-18 15:17:32, Michal Hocko wrote:
> On Fri 26-10-18 16:31:02, Arun KS wrote:
> > Now totalram_pages and managed_pages are atomic varibles. No need
> > of managed_page_count spinlock.
> 
> Yes this is the real improvement. The lock had really a weak consistency
> guarantee. It hasn't been used for anything but the update but no reader
> actually cares about all the values being updated to be in sync. I
> haven't really done an excavation work to see whether it used to be used
> in other contexts in the past but it simply doesn't have any meaning
> anymore.
> 
> > Signed-off-by: Arun KS <arunks@xxxxxxxxxxxxxx>
> 
> Acked-by: Michal Hocko <mhocko@xxxxxxxx>
> 
> Thanks for getting through all the hassles to reach to this end point.
> 
> > ---
> >  include/linux/mmzone.h | 6 ------
> >  mm/page_alloc.c        | 5 -----
> >  2 files changed, 11 deletions(-)
> > 
> > diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
> > index 597b0c7..aa960f6 100644
> > --- a/include/linux/mmzone.h
> > +++ b/include/linux/mmzone.h
> > @@ -428,12 +428,6 @@ struct zone {
> >  	 * Write access to present_pages at runtime should be protected by
> >  	 * mem_hotplug_begin/end(). Any reader who can't tolerant drift of
> >  	 * present_pages should get_online_mems() to get a stable value.
> > -	 *
> > -	 * Read access to managed_pages should be safe because it's unsigned
> > -	 * long. Write access to zone->managed_pages and totalram_pages are
> > -	 * protected by managed_page_count_lock at runtime. Idealy only
> > -	 * adjust_managed_page_count() should be used instead of directly
> > -	 * touching zone->managed_pages and totalram_pages.
> >  	 */
> >  	atomic_long_t		managed_pages;
> >  	unsigned long		spanned_pages;
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index af832de..e29e78f 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -122,9 +122,6 @@
> >  };
> >  EXPORT_SYMBOL(node_states);
> >  
> > -/* Protect totalram_pages and zone->managed_pages */
> > -static DEFINE_SPINLOCK(managed_page_count_lock);
> > -
> >  atomic_long_t _totalram_pages __read_mostly;
> >  unsigned long totalreserve_pages __read_mostly;
> >  unsigned long totalcma_pages __read_mostly;
> > @@ -7062,14 +7059,12 @@ static int __init cmdline_parse_movablecore(char *p)
> >  
> >  void adjust_managed_page_count(struct page *page, long count)
> >  {
> > -	spin_lock(&managed_page_count_lock);
> >  	atomic_long_add(count, &page_zone(page)->managed_pages);
> >  	totalram_pages_add(count);
> >  #ifdef CONFIG_HIGHMEM
> >  	if (PageHighMem(page))
> >  		totalhigh_pages_add(count);
> >  #endif
> > -	spin_unlock(&managed_page_count_lock);
> >  }
> >  EXPORT_SYMBOL(adjust_managed_page_count);
> >  
> > -- 
> > 1.9.1
> 
> -- 
> Michal Hocko
> SUSE Labs

-- 
Michal Hocko
SUSE Labs