From: "Yoshinori K. Okuji" <okuji@enbug.org>
Organization: enbug.org
To: The development of GRUB 2 <grub-devel@gnu.org>
Date: Thu, 14 Jul 2005 10:27:00 +0200
User-Agent: KMail/1.7.2
References: <42D4129A.6010802@yahoo.fr> <42D59268.4090005@yahoo.fr>
	<23f63ec505913b950c9734eee2477a7b@penguinppc.org>
In-Reply-To: <23f63ec505913b950c9734eee2477a7b@penguinppc.org>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200507141027.00171.okuji@enbug.org>
Subject: Re: [PATCH] sparc64 (common & specific files)
Precedence: list
Reply-To: The development of GRUB 2 <grub-devel@gnu.org>

On Thursday 14 July 2005 02:43, Hollis Blanchard wrote:
> On Jul 13, 2005, at 5:15 PM, Vincent Pelletier wrote:
> > Marco Gerards wrote:
> >> For which reason?
> >>
> >> But if it should be more precise, it should be a grub_uint64_t.
> >
> > If the architecture give us a 32 bit number, we won't invent the
> > remaining 32 bits...
>
> We don't invent them; they are 0.

Please do not extend it to 64-bit on i386. 64-bit variables consume more 
registers and enlarge the binary code. GRUB does not have to be very precise 
about the time, so I prefer 32-bit here for all architectures. If you have a 
good reason to use 64-bit except that the native representation is 64-bit, 
let me know.

Okuji