[EMAIL PROTECTED] wrote:
> On Wed, Nov 29, 2000 at 05:24:40PM -0700, Gordon McNutt wrote:
> > __start_rtlinux_funcs is, apparently, some kind of implicit macro which
>
> It's a ELF section symbol not a macro.
>
Ok.
>
> > references the beginning of the "rtlinux_funcs" section. This section is
> > established by the RTLINUX_EXPORT macro used in rtlinux.c. By doing an
> > EXPORT_SYMBOL on it, the patch makes it available for the rtl module to
> > grab ahold of the rtlinux_funcs section (and all the variable pointers
> > stashed there by the RTLINUX_EXPORT macro calls).
> >
> > Obvious.
>
> Not so obvious to me anywahy, I really had to study what Linus did for the mm
> fixup to figure it out. Obvious once you get it. Good analysis.
>
Just a (friendly) joke.
>
> > Not sure why they don't just EXPORT_SYMBOL the symbols placed in
> > rtlinux_funcs. Perhaps if I stare at it some more the reason will appear.
>
> No deep reason other than modularity: we want to keep the rtl symbols
> out of the soup of Linux symbols and we sure don't want random modules
> latching on to those.
>
Hmm... I just convinced myself that you did it so that you could patch the 'jmp'
instructions laid down by the hw_irq.h macros with relative (instead of
absolute) addresses.
--Gordon
-- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
---
For more information on Real-Time Linux see:
http://www.rtlinux.org/rtlinux/
