GCC port for Intel 8086, 80186 and 80286

GCC port for Intel 8086, 80186 and 80286

[Rask Ingemann Lambertsen]

   Aa week and a half ago, I posted patches to the GCC patches mailing list
with a new back end to generate code for Intel 8086-80286 CPUs. The back end
has not yet been accepted for inclusion into GCC and there is some debate
about how that should be done, but I thought you might want to know anyway.
The messages with the essential parts are here:

http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02108.html
http://gcc.gnu.org/ml/gcc-patches/2007-08/msg00599.html
http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02109.html
http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02110.html
http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02112.html
http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02114.html
http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02117.html
http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02131.html

-- 
Rask Ingemann Lambertsen

Re: GCC port for Intel 8086, 80186 and 80286

[Royce Williams]

Rask Ingemann Lambertsen wrote, on 8/9/2007 12:56 PM:
>    Aa week and a half ago, I posted patches to the GCC patches mailing list
> with a new back end to generate code for Intel 8086-80286 CPUs. The back end

Just the fact that someone is going to the trouble is quite
heartening.  Thanks for letting us know!

Royce

-- 
Royce D. Williams                                  IP Engineering, ACS
http://www.tycho.org/royce/                   - PGP: 3FC087DB/1776A531
   Let him who would move the world first move himself. - Socrates

Re: GCC port for Intel 8086, 80186 and 80286

[Hans]

Hi Rask,

This is quite impressive, what is the reason for this work?

Regards,
Hans.

www.ht-lab.com

----- Original Message ----- 
From: "Rask Ingemann Lambertsen" <rask@sygehus.dk>
To: <linux-8086@vger.kernel.org>
Sent: Thursday, August 09, 2007 9:56 PM
Subject: GCC port for Intel 8086, 80186 and 80286


>   Aa week and a half ago, I posted patches to the GCC patches mailing list
> with a new back end to generate code for Intel 8086-80286 CPUs. The back 
> end
> has not yet been accepted for inclusion into GCC and there is some debate
> about how that should be done, but I thought you might want to know 
> anyway.
> The messages with the essential parts are here:
>
> http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02108.html
> http://gcc.gnu.org/ml/gcc-patches/2007-08/msg00599.html
> http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02109.html
> http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02110.html
> http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02112.html
> http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02114.html
> http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02117.html
> http://gcc.gnu.org/ml/gcc-patches/2007-07/msg02131.html
>
> -- 
> Rask Ingemann Lambertsen
> -
> To unsubscribe from this list: send the line "unsubscribe linux-8086" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: GCC port for Intel 8086, 80186 and 80286

[Rask Ingemann Lambertsen]

On Fri, Aug 10, 2007 at 08:27:59AM +0100, Hans wrote:
> Hi Rask,
> 
> This is quite impressive, what is the reason for this work?

   I have an embedded system with an 80186 and initial experiments with bcc
showed that code quality wasn't good enough.

-- 
Rask Ingemann Lambertsen

Re: GCC port for Intel 8086, 80186 and 80286

[Rask Ingemann Lambertsen]

On Thu, Aug 09, 2007 at 10:56:21PM +0200, Rask Ingemann Lambertsen wrote:
>    Aa week and a half ago, I posted patches to the GCC patches mailing list
> with a new back end to generate code for Intel 8086-80286 CPUs. The back end
> has not yet been accepted for inclusion into GCC and there is some debate
> about how that should be done, but I thought you might want to know anyway.

   I've just posted an update which inludes support for generating ELKS
binaries. Please see
<URL:http://gcc.gnu.org/ml/gcc-patches/2007-08/msg02195.html>.

-- 
Rask Ingemann Lambertsen

Re: GCC port for Intel 8086, 80186 and 80286

[Alan Carvalho de Assis]

Hi Rask,
congratulations!

I hope ELKS become active newly and in future it can be integrated to
main kernel just like happened to uclinux. I known ELKS is based in a
very old kernel source, but we can dream about it.

Just dreaming :-)

Alan

2007/8/30, Rask Ingemann Lambertsen <rask@sygehus.dk>:
> On Thu, Aug 09, 2007 at 10:56:21PM +0200, Rask Ingemann Lambertsen wrote:
> >    Aa week and a half ago, I posted patches to the GCC patches mailing list
> > with a new back end to generate code for Intel 8086-80286 CPUs. The back end
> > has not yet been accepted for inclusion into GCC and there is some debate
> > about how that should be done, but I thought you might want to know anyway.
>
>    I've just posted an update which inludes support for generating ELKS
> binaries. Please see
> <URL:http://gcc.gnu.org/ml/gcc-patches/2007-08/msg02195.html>.
>
> --
> Rask Ingemann Lambertsen
> -
> To unsubscribe from this list: send the line "unsubscribe linux-8086" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>


-- 
------------------------------------------------------------
|                     Alan Carvalho de Assis                |
------------------------------------------------------------
--
Não importa o que os outros irão pensar,
A cura para a infelicidade é a felicidade
-
To unsubscribe from this list: send the line "unsubscribe linux-8086" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: GCC port for Intel 8086, 80186 and 80286

[Chad]

2007/8/30, Rask Ingemann Lambertsen <rask@sygehus.dk>:

>    I've just posted an update which inludes support for generating ELKS
> binaries. Please see
> <URL:http://gcc.gnu.org/ml/gcc-patches/2007-08/msg02195.html>.

Awesome - I would so have loved to have this way back when.  Now to
find time to try it out...

On 8/30/07, Alan Carvalho de Assis <acassis@gmail.com> wrote:
> Hi Rask,
> congratulations!
>
> I hope ELKS become active newly and in future it can be integrated to
> main kernel just like happened to uclinux. I known ELKS is based in a
> very old kernel source, but we can dream about it.
>
> Just dreaming :-)

If you're interested in working on mainline tree stuff look into a
ucLinux port to i386, if there isn't already one*, and in addition a
cool ucLinux-centric distribution/packaging.  OpenEmbedded would be a
good base since you can build for many different archs, but it can be
a bit interesting to wrangle ;)

(* - yes this sounds nuts... but it might be faster than regular Linux
and probably smaller.  At least Samsung's people measured that while
working on mmu-less ARM ucLinux.)

- Chad

Re: GCC port for Intel 8086, 80186 and 80286

[Alan Cox]

On Thu, 30 Aug 2007 13:43:52 -0300
"Alan Carvalho de Assis" <acassis@gmail.com> wrote:

> Hi Rask,
> congratulations!
> 
> I hope ELKS become active newly and in future it can be integrated to
> main kernel just like happened to uclinux. I known ELKS is based in a
> very old kernel source, but we can dream about it.

Well the first step to sanity would be moving the kernel code to the 8086
gcc, getting ansi prototypes and sane C. That means sorting the asm out
with gcc asm which frankly scares me 8)

Re: GCC port for Intel 8086, 80186 and 80286

[Eddy]

Alan Cox wrote:
> On Thu, 30 Aug 2007 13:43:52 -0300
> "Alan Carvalho de Assis" <acassis@gmail.com> wrote:
>
>   
>
> Well the first step to sanity would be moving the kernel code to the 8086
> gcc, getting ansi prototypes and sane C. That means sorting the asm out
> with gcc asm which frankly scares me 8)
> -
>   


I don't know very well what are the differences between the 2 codes, but
I think that is possible to use a recursive bash script to automatically
replace lines that have to be replaced and go.
I thought at something like this:
#!/bin/bash

for i in `grep -R "text that is in files to be replaced" elks/* | cut
-f1-1 -d ":"`
do
sed -i -e 's,old code,new code,g' $i
sed -i -e 's,old code,new code,g' $i
...
done

It is a good idea?
I hope yes.

Re: GCC port for Intel 8086, 80186 and 80286

[Vikas N Kumar]

You forget that assembler directives are different for Gnu Assembler
(gas) and the bcc assembler. Not to forget the fact that gas accepts
instructions like movl, movd etc and many other assemblers do not.
they only accept mov, and figure out the size of the data based on the
registers used as operands. It will have to be done manually and
checked thoroughly.

On 8/31/07, Eddy <eddyx89@gmail.com> wrote:
> Alan Cox wrote:
> > On Thu, 30 Aug 2007 13:43:52 -0300
> > "Alan Carvalho de Assis" <acassis@gmail.com> wrote:
> >
> >
> >
> > Well the first step to sanity would be moving the kernel code to the 8086
> > gcc, getting ansi prototypes and sane C. That means sorting the asm out
> > with gcc asm which frankly scares me 8)
> > -
> >
>
>
> I don't know very well what are the differences between the 2 codes, but
> I think that is possible to use a recursive bash script to automatically
> replace lines that have to be replaced and go.
> I thought at something like this:
> #!/bin/bash
>
> for i in `grep -R "text that is in files to be replaced" elks/* | cut
> -f1-1 -d ":"`
> do
> sed -i -e 's,old code,new code,g' $i
> sed -i -e 's,old code,new code,g' $i
> ...
> done
>
> It is a good idea?
> I hope yes.
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-8086" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>


-- 
http://www.vikaskumar.org/