From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1753239AbdKVWp7 (ORCPT <rfc822;w@1wt.eu>);
        Wed, 22 Nov 2017 17:45:59 -0500
Received: from mga06.intel.com ([134.134.136.31]:48841 "EHLO mga06.intel.com"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1752016AbdKVWp5 (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
        Wed, 22 Nov 2017 17:45:57 -0500
X-ExtLoop1: 1
X-IronPort-AV: E=Sophos;i="5.44,438,1505804400"; 
   d="scan'208";a="152260703"
Subject: Re: [PATCH 08/30] x86, kaiser: unmap kernel from userspace page
 tables (core patch)
To: Thomas Gleixner <tglx@linutronix.de>
References: <20171110193058.BECA7D88@viggo.jf.intel.com>
 <20171110193112.6A962D6A@viggo.jf.intel.com>
 <alpine.DEB.2.20.1711201518490.1734@nanos>
Cc: linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        richard.fellner@student.tugraz.at, moritz.lipp@iaik.tugraz.at,
        daniel.gruss@iaik.tugraz.at, michael.schwarz@iaik.tugraz.at,
        luto@kernel.org, torvalds@linux-foundation.org, keescook@google.com,
        hughd@google.com, x86@kernel.org
From: Dave Hansen <dave.hansen@linux.intel.com>
Message-ID: <82d2d82f-32a0-64e6-03b2-3ca0a9f97de6@linux.intel.com>
Date: Wed, 22 Nov 2017 14:45:54 -0800
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.4.0
MIME-Version: 1.0
In-Reply-To: <alpine.DEB.2.20.1711201518490.1734@nanos>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 11/20/2017 09:21 AM, Thomas Gleixner wrote:
>> +	pgd = native_get_shadow_pgd(pgd_offset_k(0UL));
>> +	for (i = PTRS_PER_PGD / 2; i < PTRS_PER_PGD; i++) {
>> +		unsigned long addr = PAGE_OFFSET + i * PGDIR_SIZE;
> This looks wrong. The kernel address space gets incremented by PGDIR_SIZE
> and does not make a jump from PAGE_OFFSET to PAGE_OFFSET + 256 * PGDIR_SIZE
> 
> 	int i, j;
> 
> 	for (i = PTRS_PER_PGD / 2, j = 0; i < PTRS_PER_PGD; i++, j++) {
> 		unsigned long addr = PAGE_OFFSET + j * PGDIR_SIZE;
> 
> Not that is has any effect right now. Neither p4d_alloc_one() nor
> pud_alloc_one() are using the 'addr' argument.

Ahh, you're saying that 'i' is effectively starting *at* PAGE_OFFSET
since it's halfway up the address space already doing PTRS_PER_PGD/2.
Adding PAGE_OFFSET to PAGE_OFFSET is nonsense.

Would it just be simpler to do:

>         for (i = PTRS_PER_PGD / 2; i < PTRS_PER_PGD; i++) {
>                 unsigned long addr = i * PGDIR_SIZE;

?