From mboxrd@z Thu Jan  1 00:00:00 1970
From: "H. Peter Anvin" <hpa@zytor.com>
Subject: Re: [PATCH 01/35] lmb: prepare x86 to use lmb to replace early_res
Date: Mon, 17 May 2010 15:12:49 -0700
Message-ID: <4BF1BF61.4030605@zytor.com>
References: <1273796396-29649-1-git-send-email-yinghai@kernel.org>	 <1273796396-29649-2-git-send-email-yinghai@kernel.org>	 <1273803143.21352.353.camel@pasglop>  <4BECEB8D.8070600@oracle.com>	 <1273824578.21352.579.camel@pasglop>  <4BF18500.60105@zytor.com> <1274133749.21352.756.camel@pasglop>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Return-path: <linux-arch-owner@vger.kernel.org>
Received: from terminus.zytor.com ([198.137.202.10]:38037 "EHLO mail.zytor.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1751996Ab0EQWUQ (ORCPT <rfc822;linux-arch@vger.kernel.org>);
	Mon, 17 May 2010 18:20:16 -0400
In-Reply-To: <1274133749.21352.756.camel@pasglop>
Sender: linux-arch-owner@vger.kernel.org
List-ID: <linux-arch.vger.kernel.org>
To: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Yinghai <yinghai.lu@oracle.com>, Ingo Molnar <mingo@elte.hu>, Thomas Gleixner <tglx@linutronix.de>, Andrew Morton <akpm@linux-foundation.org>, David Miller <davem@davemloft.net>, Linus Torvalds <torvalds@linux-foundation.org>, Johannes Weiner <hannes@cmpxchg.org>, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org

On 05/17/2010 03:02 PM, Benjamin Herrenschmidt wrote:
> On Mon, 2010-05-17 at 11:03 -0700, H. Peter Anvin wrote:
>> On 05/14/2010 01:09 AM, Benjamin Herrenschmidt wrote:
>>>
>>> No. That is not the point. Read the rest of my email !
>>>
>>> We need to -sanitize- those errors. _Maybe_ exposing LMB_ERROR is the
>>> right way to do so, but in that case, we need to make -all- function use
>>> the same error code. Right now, some fail with 0 and some with
>>> LMB_ERROR.
>>>
>>
>> Using errnos like the rest of the kernel seems like the right thing to
>> do, IMO.
> 
> Maybe. The allocator/find functions return a physical address. If we all
> agree that a physical address between -PAGE_SIZE and 0 is never valid,
> then we can overlay the negative errno codes like we do for pointers.
> 
> I'll have a look at that, it shouldn't be very hard.
> 

For x86 with 64-bit resource_t, this is always true (physical addresses
are never negative.)  For x86 with 32-bit resource_t negative physical
addresses can be negative, but the top of the address space will always
be occupied by the boot ROM (it's a hardware constraint.)

	-hpa