Re: [PATCH 2/2] mm/page_alloc: DEBUG_VM checks for free_list placement of CMA and RESERVE pages

[Mel Gorman]

On Thu, May 08, 2014 at 02:54:21PM +0900, Joonsoo Kim wrote:
> > >> Furthermore, I think there's a problem that
> > >> setup_zone_migrate_reserve() operates on pageblocks, but as MAX_ODER
> > >> is higher than pageblock_order, RESERVE pages might be merged with
> > >> buddies of different migratetype and end up on their free_list. That
> > >> seems to me like a flaw in the design of reserves, but perhaps
> > >> others won't think it's serious enough to fix?
> 
> I wanna know who want MIGRATE_RESERVE. On my previous testing, one
> pageblock for MIGRATE_RESERVE is merged with buddies of different
> migratetype during boot-up and never come back again. But my system works
> well. :)
> 

It's important for short-lived high-order atomic allocations.
MIGRATE_RESERVE preserves a property of the buddy allocator prior to the
merging of fragmentation avoidance. Most users will not notice as not
many drivers depend on these allocations working. If they are getting
destroyed at boot-up, it's a bug.

-- 
Mel Gorman
SUSE Labs
--
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/