From mboxrd@z Thu Jan  1 00:00:00 1970
From: compudj@krystal.dyndns.org (Mathieu Desnoyers)
Date: Mon, 15 Feb 2010 09:46:15 -0500
Subject: [ltt-dev] [PATCH 03/11] add urcu/arch_defaults.h
In-Reply-To: <4B78FD7E.8050402@redhat.com>
References: <1266081392-8732-1-git-send-email-pbonzini@redhat.com>
	<1266081392-8732-4-git-send-email-pbonzini@redhat.com>
	<20100214143459.GD5871@Krystal> <4B78FD7E.8050402@redhat.com>
Message-ID: <20100215144615.GB13436@Krystal>

* Paolo Bonzini (pbonzini at redhat.com) wrote:
> On 02/14/2010 03:34 PM, Mathieu Desnoyers wrote:
>> +#if !defined(mc)&&  !defined(rmc)&&  !defined(wmc)
>>
>> instead ?
>
> Ok.
>
>>> >  +/*
>>> >  + * Architectures without cache coherency need something like the following:
>>> >  + *
>>> >  + * #define mb()	mc()
>>> >  + * #define rmb()	rmc()
>>> >  + * #define wmb()	wmc()
>>> >  + * #define mc()	arch_cache_flush()        // mandatory
>>
>> The // comment style should be avoided. (Following Linux kernel
>> guide-lines)
>
> Note that this is within a comment; I used it reluctantly because C  
> comments do not nest.  I could replace it with "<-- mandatory" but I  
> think // is better in this case.

I'm not quite sure what this "mandatory" comment implies. I think
expressing it more verbosely in a sentence above the defines would be
better.

Thanks,

Mathieu

>
> Paolo
>

-- 
Mathieu Desnoyers
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68