Re: [PATCH 1/5] mm, page_owner: print migratetype of a page, not pageblock

From: Joonsoo Kim <iamjoonsoo.kim@lge.com>
To: Vlastimil Babka <vbabka@suse.cz>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Minchan Kim <minchan@kernel.org>,
	Sasha Levin <sasha.levin@oracle.com>,
	"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>,
	Mel Gorman <mgorman@suse.de>
Subject: Re: [PATCH 1/5] mm, page_owner: print migratetype of a page, not pageblock
Date: Thu, 5 Nov 2015 17:19:55 +0900	[thread overview]
Message-ID: <20151105081955.GA26034@js1304-P5Q-DELUXE> (raw)
In-Reply-To: <563B1005.3070203@suse.cz>

On Thu, Nov 05, 2015 at 09:15:01AM +0100, Vlastimil Babka wrote:
> On 11/05/2015 09:09 AM, Joonsoo Kim wrote:
> > On Wed, Nov 04, 2015 at 04:00:57PM +0100, Vlastimil Babka wrote:
> >> The information in /sys/kernel/debug/page_owner includes the migratetype
> >> declared during the page allocation via gfp_flags. This is also checked against
> >> the pageblock's migratetype, and reported as Fallback allocation if these two
> >> differ (although in fact fallback allocation is not the only reason why they
> >> can differ).
> >> 
> >> However, the migratetype actually printed is the one of the pageblock, not of
> >> the page itself, so it's the same for all pages in the pageblock. This is
> >> apparently a bug, noticed when working on other page_owner improvements. Fixed.
> > 
> > We can guess page migratetype through gfp_mask output although it isn't
> > easy task for now. But, there is no way to know pageblock migratetype.
> > I used this to know how memory is fragmented.
> 
> Ah, I see. How bout just we print both migratetypes then and remove the
> "Fallback" part, which can be trivially deduced from them (and as I noted it's
> somewhat misleading anyway)?

I'm okay with your new suggestion.

Thanks.