Re: [PATCH 3/5] mm, page_owner: copy page owner info during migration

From: Vlastimil Babka <vbabka@suse.cz>
To: Hugh Dickins <hughd@google.com>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Joonsoo Kim <iamjoonsoo.kim@lge.com>,
	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 3/5] mm, page_owner: copy page owner info during migration
Date: Thu, 19 Nov 2015 17:44:12 +0100	[thread overview]
Message-ID: <564DFC5C.2090103@suse.cz> (raw)
In-Reply-To: <alpine.LSU.2.11.1511081318110.12914@eggly.anvils>

On 11/08/2015 10:29 PM, Hugh Dickins wrote:
>
> Would it be possible to move that line into migrate_page_copy()?
>
> I don't think it's wrong where you placed it, but that block is really
> about resetting the old page ready for freeing, and I'd prefer to keep
> all the transference of properties from old to new in migrate_page_copy()
> if we can.

OK, makes sense, will do in v2.

> But check how that behaves in the migrate_misplaced_transhuge_page()
> case: I haven't studied long enough, but I think you may have been missing
> to copy_page_owner in that case;

You're right, I missed that path :/

> but beware of its "fail_putback", which
> for some things nastily entails undoing what's already been done.

Yeah, I think I don't need to reset page owner info in the fail_putback 
path, for the same reason I don't reset it from the old page when 
migration is successful. The page is going to be freed anyway, and if it 
somehow hits a bug before that, we will still have something to print 
(after patch 5).

Thanks!