[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20160921192201.GL24210@dhcp22.suse.cz>
Date: Wed, 21 Sep 2016 21:22:01 +0200
From: Michal Hocko <mhocko@...nel.org>
To: Dave Hansen <dave.hansen@...ux.intel.com>
Cc: Mike Kravetz <mike.kravetz@...cle.com>,
Gerald Schaefer <gerald.schaefer@...ibm.com>,
Andrew Morton <akpm@...ux-foundation.org>,
Naoya Horiguchi <n-horiguchi@...jp.nec.com>,
linux-mm@...ck.org, linux-kernel@...r.kernel.org,
"Kirill A . Shutemov" <kirill.shutemov@...ux.intel.com>,
Vlastimil Babka <vbabka@...e.cz>,
"Aneesh Kumar K . V" <aneesh.kumar@...ux.vnet.ibm.com>,
Martin Schwidefsky <schwidefsky@...ibm.com>,
Heiko Carstens <heiko.carstens@...ibm.com>,
Rui Teng <rui.teng@...ux.vnet.ibm.com>
Subject: Re: [PATCH 0/1] memory offline issues with hugepage size > memory
block size
On Wed 21-09-16 11:27:48, Dave Hansen wrote:
> On 09/21/2016 11:20 AM, Michal Hocko wrote:
> > I would even question the per page block offlining itself. Why would
> > anybody want to offline few blocks rather than the whole node? What is
> > the usecase here?
>
> The original reason was so that you could remove a DIMM or a riser card
> full of DIMMs, which are certainly a subset of a node.
OK, I see, thanks for the clarification! I was always thinking more in
node rather than physical memory range hot-remove. I do agree that it
makes sense to free the whole gigantic huge page if we encounter a tail
page for the above use case because losing the gigantic page is
justified when the whole dim goes away.
Thanks!
--
Michal Hocko
SUSE Labs
Powered by blists - more mailing lists