Sender: bill AT taniwha DOT tssc DOT co DOT nz
Message-ID: <3752415C.3645C239@taniwha.tssc.co.nz>
Date: Mon, 31 May 1999 19:59:24 +1200
From: Bill Currie <bill AT taniwha DOT tssc DOT co DOT nz>
X-Mailer: Mozilla 4.05 [en] (X11; I; Linux 2.2.6 i486)
MIME-Version: 1.0
To: djgpp-workers AT delorie DOT com
CC: "Mark E." <snowball3 AT bigfoot DOT com>
Subject: Re: gcc 3.0
References: <Pine DOT SUN DOT 3 DOT 91 DOT 990530115215 DOT 12571A-100000 AT is>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: djgpp-workers AT delorie DOT com

Eli Zaretskii wrote:
> 
> On Thu, 27 May 1999, Mark E. wrote:
> 
> > > But if GCC supports something like __attribute__((interrupt)), then I
> > > agree we should go for that.  I'm not privy enough to GCC/EGCS
> > > developments to know whether in fact such an attribute exists.
> > >
> >
> > It does for several targets.
> 
> That's good news.
> 
> > Would this be something that would work the same for all x86 targets?
> 
> Probably not.  At least the wrapper code is DPMI- and DJGPP-specific
> in some of its aspects.  I think.

I'm currently working on __attribute__((naked)) for *ALL* egcs ports. 
This will produce functions that have *NO* entry or exit code (main code
only), not even a `ret', I believe.  Of course, the programmer will be
required to supply the appropriate code via asm statements.  Ideal for
interrupt functions (especially in cross platform code), and probably
library startup code (eg .init in elf).

I've actually got most of the work done, I just have to figure out where
exactly to put the tests for nakedness and not emit the prolog/epilog
code.

Bill
-- 
Leave others their otherness.