From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <owner-linux-mm@kvack.org>
Received: from mail-pl1-f199.google.com (mail-pl1-f199.google.com [209.85.214.199])
	by kanga.kvack.org (Postfix) with ESMTP id 8C19D6B577C
	for <linux-mm@kvack.org>; Fri, 31 Aug 2018 10:49:54 -0400 (EDT)
Received: by mail-pl1-f199.google.com with SMTP id b93-v6so6171579plb.10
        for <linux-mm@kvack.org>; Fri, 31 Aug 2018 07:49:54 -0700 (PDT)
Received: from mga01.intel.com (mga01.intel.com. [192.55.52.88])
        by mx.google.com with ESMTPS id g3-v6si9696711pll.395.2018.08.31.07.49.53
        for <linux-mm@kvack.org>
        (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128);
        Fri, 31 Aug 2018 07:49:53 -0700 (PDT)
Subject: Re: [RFC PATCH v3 12/24] x86/mm: Modify ptep_set_wrprotect and
 pmdp_set_wrprotect for _PAGE_DIRTY_SW
References: <ce051b5b-feef-376f-e085-11f65a5f2215@linux.intel.com>
 <1535649960.26689.15.camel@intel.com>
 <33d45a12-513c-eba2-a2de-3d6b630e928e@linux.intel.com>
 <1535651666.27823.6.camel@intel.com>
 <CAG48ez3ixWROuQc6WZze6qPL6q0e_gCnMU4XF11JUWziePsBJg@mail.gmail.com>
 <1535660494.28258.36.camel@intel.com>
 <CAG48ez0yOuDhqxB779aO3Kss3gQ3cZTJL1VphDXQm+_M9jFPvQ@mail.gmail.com>
 <1535662366.28781.6.camel@intel.com>
 <CAG48ez0mkr95_TbLQnDGuGUd6G+eJVLZ-fEjDkwA6dSrm+9tLw@mail.gmail.com>
 <CAG48ez3S3+DzAyo_SnoUW1GO0Cpd_x0A83MOx2p_MkogoAatLQ@mail.gmail.com>
 <20180831095300.GF24124@hirez.programming.kicks-ass.net>
 <1535726032.32537.0.camel@intel.com>
From: Dave Hansen <dave.hansen@linux.intel.com>
Message-ID: <f5a36e32-7c5f-91fe-9e98-fb44867fda11@linux.intel.com>
Date: Fri, 31 Aug 2018 07:47:29 -0700
MIME-Version: 1.0
In-Reply-To: <1535726032.32537.0.camel@intel.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Sender: owner-linux-mm@kvack.org
List-ID: <linux-mm.kvack.org>
To: Yu-cheng Yu <yu-cheng.yu@intel.com>, Peter Zijlstra <peterz@infradead.org>, Jann Horn <jannh@google.com>
Cc: the arch/x86 maintainers <x86@kernel.org>, "H . Peter Anvin" <hpa@zytor.com>, Thomas Gleixner <tglx@linutronix.de>, Ingo Molnar <mingo@redhat.com>, kernel list <linux-kernel@vger.kernel.org>, linux-doc@vger.kernel.org, Linux-MM <linux-mm@kvack.org>, linux-arch <linux-arch@vger.kernel.org>, Linux API <linux-api@vger.kernel.org>, Arnd Bergmann <arnd@arndb.de>, Andy Lutomirski <luto@amacapital.net>, Balbir Singh <bsingharora@gmail.com>, Cyrill Gorcunov <gorcunov@gmail.com>, Florian Weimer <fweimer@redhat.com>, hjl.tools@gmail.com, Jonathan Corbet <corbet@lwn.net>, keescook@chromiun.org, Mike Kravetz <mike.kravetz@oracle.com>, Nadav Amit <nadav.amit@gmail.com>, Oleg Nesterov <oleg@redhat.com>, Pavel Machek <pavel@ucw.cz>, ravi.v.shankar@intel.com, vedvyas.shanbhogue@intel.com

On 08/31/2018 07:33 AM, Yu-cheng Yu wrote:
> Please use the form:
> 
> 	pte_t new_pte, pte = READ_ONCE(*ptep);
> 	do {
> 		new_pte = /* ... */;
> 	} while (!try_cmpxchg(ptep, &pte, new_pte);

It's probably also worth doing some testing to see if you can detect the
cost of the cmpxchg.  It's definitely more than the old code.

A loop that does mprotect(PROT_READ) followed by
mprotect(PROT_READ|PROT_WRITE) should do it.