Date: Mon, 29 Jul 1996 14:25:20 +0200 (MET DST)
From: Mark Habersack <grendel AT ananke DOT amu DOT edu DOT pl>
To: "Alaric B. Williams" <alaric AT abwillms DOT demon DOT co DOT uk>
cc: djgpp AT delorie DOT com
Subject: Re: __attribute__((section))
In-Reply-To: <838478248.10640.0@abwillms.demon.co.uk>
Message-ID: <Pine.NEB.3.93.960729142119.2833B-100000@ananke.amu.edu.pl>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Sat, 27 Jul 1996, Alaric B. Williams wrote:

>
>
>While we're at it, if we can put things easily into different
>sections, wouldn't that provide a nice easy method of locking memory
>for hardware interrupt handlers? Stick everything in a '.lock'
>section, with sentinel labels in it so that a "lock_everything()" call
>could lock the entire (guaranteed contiguous) section?
That would be great, but alas would require modification to DJGPP startup code
that already deals in a special way with constructors/destructors section.

>
>BTW, is there a better way of getting the extent of a section? Some
>trickery with the linker ADDR and SIZE commands?
Well, maybe it'd be possible to create a variable in, say, startup code (or at
least some piece of soft that is guaranteed to be linked into every app) -
maybe something like vars marking the beginnig and the end of
destructors/constructors section - this variable would get initialized at
runtime (or, better, in the ld script) using ld directives you mentioned. Is
it possible, gurus?

Greetz, Mark

/*******************************/
/** So here I am once more... **************************************/
/** When you grown up and leave your playground, where you kissed **/
/** your Prince and found your Frog - remember the Jester that    **/
/** showed you tears, a Script for Tears... ************************/
/*********************************************/