Re: [PATCH 09/25] Calculate the migratetype for allocation only once

[KOSAKI Motohiro]

> On Tue, Apr 21, 2009 at 04:37:28PM +0900, KOSAKI Motohiro wrote:
> > > GFP mask is converted into a migratetype when deciding which pagelist to
> > > take a page from. However, it is happening multiple times per
> > > allocation, at least once per zone traversed. Calculate it once.
> > > 
> > > Signed-off-by: Mel Gorman <mel@xxxxxxxxx>
> > > ---
> > >  mm/page_alloc.c |   43 ++++++++++++++++++++++++++-----------------
> > >  1 files changed, 26 insertions(+), 17 deletions(-)
> > > 
> > > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > > index b27bcde..f960cf5 100644
> > > --- a/mm/page_alloc.c
> > > +++ b/mm/page_alloc.c
> > > @@ -1065,13 +1065,13 @@ void split_page(struct page *page, unsigned int order)
> > >   * or two.
> > >   */
> > >  static struct page *buffered_rmqueue(struct zone *preferred_zone,
> > > -			struct zone *zone, int order, gfp_t gfp_flags)
> > > +			struct zone *zone, int order, gfp_t gfp_flags,
> > > +			int migratetype)
> > >  {
> > >  	unsigned long flags;
> > >  	struct page *page;
> > >  	int cold = !!(gfp_flags & __GFP_COLD);
> > >  	int cpu;
> > > -	int migratetype = allocflags_to_migratetype(gfp_flags);
> > 
> > hmmm....
> > 
> > allocflags_to_migratetype() is very cheap function and buffered_rmqueue()
> > and other non-inline static function isn't guranteed inlined.
> > 
> 
> A later patch makes them inlined due to the fact there is only one call
> site.

Oh, I see.
I drop my claim. thanks.



> > -------------------------------------------------------------------
> > /* Convert GFP flags to their corresponding migrate type */
> > static inline int allocflags_to_migratetype(gfp_t gfp_flags)
> > {
> >         WARN_ON((gfp_flags & GFP_MOVABLE_MASK) == GFP_MOVABLE_MASK);
> > 
> >         if (unlikely(page_group_by_mobility_disabled))
> >                 return MIGRATE_UNMOVABLE;
> > 
> >         /* Group based on mobility */
> >         return (((gfp_flags & __GFP_MOVABLE) != 0) << 1) |
> >                 ((gfp_flags & __GFP_RECLAIMABLE) != 0);
> > }
> > -------------------------------------------------------------------
> > 
> > s/WARN_ON/VM_BUG_ON/ is better?
> > 
> 
> I wanted to catch out-of-tree drivers but it's been a while so maybe VM_BUG_ON
> wouldn't hurt. I can add a patch that does that a pass 2 of improving the
> allocator or would you prefer to see it now?

no. another patch is better :)


> > GFP_MOVABLE_MASK makes 3. 3 mean MIGRATE_RESERVE. it seems obviously bug.
> > 
> 
> Short answer;
> No, GFP flags that result in MIGRATE_RESERVE is a bug. The caller should
> never want to be allocating from there.
> 
> Longer answer;
> The size of the MIGRATE_RESERVE depends on the number of free pages that
> must be kept in the zone. Because GFP flags never result in here, the
> area is only used when the alternative is to fail the allocation and the
> watermarks are still met. The intention is that high-order atomic
> allocations that were short lived may be allocated from here. This was
> to preserve a behaviour in the allocator before MIGRATE_RESERVE was
> introduced. It makes no sense for a caller to allocate directly out of
> here and in fact the fallback list for MIGRATE_RESERVE are useless

Yeah.
My past mail is too poor. I agree it is caller's bug.
I mean VM_BUG_ON is better because
  - obviously caller bug 
  - VM_BUG_ON is no runtime impact when VM_DEBUG is off.



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/