Subject: Re: /usr/include/machine
To: None <>
From: Matt Thomas <>
List: tech-kern
Date: 07/13/2001 09:28:31
At 07:41 PM 7/13/2001 +0900, Izumi Tsutsui wrote:
>In article <>
> wrote:
> > > But new arm ports are broken for one month already, and the consensus
> > > to fix the problem is still not made.
> >
> > Yep it's been a month or more, but I've not seen any complaints about it.
>I'm being annoyed for a month, but I cannot see where is the goal,
>so I don't know what is the right fix.

The goal is to have no cats-specific, arm26-specific, etal includes or defines
in /usr/include/arm and /usr/include/machine.  Therefore all the ARM platforms
look identical to userland and all userland and be built on any ARM machine.

> > Anyway my currently list of problems is:
> > gnu/lib/libstc++/config - needs makefile tweaking to install 
> _G_config.h in
> > /usr/include/arm (I'm not to sure if that's the best place as currently 
> arm26
> > is built with 2.95.x and arm32 with the intree egcs, so they may differ?)
>It could be fixed by changing INCSDIR to /usr/include/machine, but
>arm32_drain_writebuf.c and arm32_icache.c under lib/libarch/arm32
>include machine/sysarch.h so they cannot be compiled.

Since arm26 doesn't have sysarch.h, the right solution is to move the
arm32 sysarch.h to arm and install it.  arm26 will need to have a
dummy sysarch(2) syscall but since there are part of the arm ABI it
should have them anyway.

>Anyway, how many people know the current status of these changes?
>Is there any discussion _before_ these changes were done?

On this list.

> >  biggest pain with machine/param.h is that everything seems to have a
> > dependancy on it!
>IMHO, if machine/param.h can be shared,
>these ports did not have to be split.

Only those things that user programs needs to have defined are
being in defined in the common param.h
Matt Thomas               Internet:
3am Software Foundry      WWW URL:
Cupertino, CA             Disclaimer: I avow all knowledge of this message