X-Authentication-Warning: delorie.com: mailnull set sender to djgpp-workers-bounces using -f
From: sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Message-Id: <10202082038.AA22206@clio.rice.edu>
Subject: Re: Alignment problem
To: djgpp-workers AT delorie DOT com
Date: Fri, 8 Feb 2002 14:38:02 -0600 (CST)
In-Reply-To: <200202081853.g18IrgO08699@envy.delorie.com> from "DJ Delorie" at Feb 08, 2002 01:53:42 PM
X-Mailer: ELM [version 2.5 PL2]
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> It doesn't align to 8 bytes.  It rounds the *size* up to 8 bytes.
> 
> Malloc doesn't need to align to "optimum" alignment.  It only needs
> align to "required" alignment.  If we need to increase the alignment,
> then we will.

I think we should increase the alignment to 8 bytes when we get around
to it since it does have a huge impact on some operations (such as
floating loads/stores).

I haven't looked at our malloc() at all, but if it uses before and 
after longword boundary tags then 8-byte alignment doesn't waste
any memory either (just potentially 4 bytes after a new sbrk()).