[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180722035156.GA12125@bombadil.infradead.org>
Date: Sat, 21 Jul 2018 20:51:56 -0700
From: Matthew Wilcox <willy@...radead.org>
To: David Rientjes <rientjes@...gle.com>
Cc: Yang Shi <yang.shi@...ux.alibaba.com>,
Andrew Morton <akpm@...ux-foundation.org>,
kirill@...temov.name, hughd@...gle.com, aaron.lu@...el.com,
linux-mm@...ck.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm: thp: remove use_zero_page sysfs knob
On Fri, Jul 20, 2018 at 02:05:52PM -0700, David Rientjes wrote:
> The huge zero page can be reclaimed under memory pressure and, if it is,
> it is attempted to be allocted again with gfp flags that attempt memory
> compaction that can become expensive. If we are constantly under memory
> pressure, it gets freed and reallocated millions of times always trying to
> compact memory both directly and by kicking kcompactd in the background.
>
> It likely should also be per node.
Have you benchmarked making the non-huge zero page per-node?
Powered by blists - more mailing lists