From mboxrd@z Thu Jan  1 00:00:00 1970
From: marc.zyngier@arm.com (Marc Zyngier)
Date: Thu, 11 Dec 2014 17:23:36 +0000
Subject: [PATCH 4/4] ARM: add 32bit support to GICv3
In-Reply-To: <20141211165711.GA17509@e106794-lin.cambridge.arm.com>
References: <1418307788-17977-1-git-send-email-jean-philippe.brucker@arm.com>
 <1418307788-17977-5-git-send-email-jean-philippe.brucker@arm.com>
 <20141211143205.GP11285@n2100.arm.linux.org.uk> <5489BAEF.5050809@arm.com>
 <20141211165711.GA17509@e106794-lin.cambridge.arm.com>
Message-ID: <5489D318.2060400@arm.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On 11/12/14 16:57, Jean-Philippe Brucker wrote:
> On Thu, Dec 11, 2014 at 03:40:31PM +0000, Marc Zyngier wrote:
>> On 11/12/14 14:32, Russell King - ARM Linux wrote:
>>> On Thu, Dec 11, 2014 at 02:23:08PM +0000, Jean-Philippe Brucker wrote:
>>>> +#ifndef __ASM_ARCH_GICV3_H
>>>> +#define __ASM_ARCH_GICV3_H
>>>> +
>>>> +#ifndef __ASSEMBLY__
>>>> +
>>>> +#include <asm/io.h>
>>>
>>> This should be linux/io.h
>>
>> Also, you should probably move the #ifndef down to the level of the
>> inline functions. The #defines are perfectly usable in assembly context.
>>
> 
> This would require a change in the definition of __ACCESS_CP15, in order
> to distinguish between inlined parameter '%0' and explicit registers.
> I don't see any pleasant ways to do this, apart from adding a 'reg'
> argument to all those #defines.
> Furthermore, all other cp15 accesses from assembly on the arm/ side seem
> to use raw opcodes, instead of macro names.

Ah, fair enough. I guess we can rework that part, should we ever support
KVM on 32bit ARMv8 CPUs...

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...