[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <alpine.DEB.2.21.2003241215580.34058@chino.kir.corp.google.com>
Date: Tue, 24 Mar 2020 12:17:11 -0700 (PDT)
From: David Rientjes <rientjes@...gle.com>
To: "Kirill A. Shutemov" <kirill@...temov.name>
cc: Hui Zhu <teawater@...il.com>, hannes@...xchg.org,
mhocko@...nel.org, vdavydov.dev@...il.com,
akpm@...ux-foundation.org, hughd@...gle.com,
yang.shi@...ux.alibaba.com, dan.j.williams@...el.com,
aneesh.kumar@...ux.ibm.com, sean.j.christopherson@...el.com,
thellstrom@...are.com, guro@...com, shakeelb@...gle.com,
chris@...isdown.name, tj@...nel.org, tglx@...utronix.de,
linux-kernel@...r.kernel.org, cgroups@...r.kernel.org,
linux-mm@...ck.org, Hui Zhu <teawaterz@...ux.alibaba.com>
Subject: Re: [PATCH] mm, memcg: Add memory.transparent_hugepage_disabled
On Tue, 24 Mar 2020, Kirill A. Shutemov wrote:
> On Tue, Mar 24, 2020 at 06:31:56PM +0800, Hui Zhu wrote:
> > /sys/kernel/mm/transparent_hugepage/enabled is the only interface to
> > control if the application can use THP in system level.
>
> No. We have prctl(PR_SET_THP_DISABLE) too.
>
Yeah, I think this is the key point since this can be inherited across
fork and thus you already have the ability to disable thp for a process
where you cannot even modify the binary to do MADV_NOHUGEPAGE.
Powered by blists - more mailing lists