[TUHS] [pups] gcc-3.4.6 and old unix

[TUHS] [pups] gcc-3.4.6 and old unix

[patv]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 2106 bytes --]

If this helps at all, I've been working (very, very slowly) on a port of
v32 to Intel platforms.  At first I used gcc for some kernel work, but
quickly realized that it would be overwhelming to the final v7 system. 
Since I don't want to do the work twice, I looked for a different compiler
suite.  I switched to the ACK compiler suite and just finished the WinXP
cross compiler work.  It has a pdp11 back end, which I have yet to try,
that may be useful.

It isn't gcc, but ir does do ANSI C and the i386 assembler seems to be
pretty complete.  Let me know if there's any interest and I'll put it up
on my site for download.

Pat


> Toby Thain napisał(a):
> 
> >On 24-Apr-06, at 9:05 AM, Wesley Parish wrote:
> >
> >  
> >
> >>Quoting Andrzej Popielewicz <vasco at icpnet.pl>:
> >>
> >>    
> >>
> >>>Wesley Parish napisał(a):
> >>>
> >>>      
> >>>
> >
> >It can't be done.
> >
> >As others point out, the program is many times (100x or more?) too  
> >big -- likely even gcc 1.x is far too big, but gcc {2,3,4}.x are all  
> >meant for large 32-bit systems.
> >
> >However, cross-compilation can certainly be easily done. I have made  
> >a PDP-11 back-end for lcc[1] (not quite complete but shows that it  
> >can be done), which is an ANSI (c89) compiler[2]. lcc is a much  
> >smaller and simpler compiler than gcc, but its executables are still  
> >massively outsize for PDP-11 systems.
> >  
> >
> Yes, even running vi or csh in Ultrix (in simh pdp11) produced message : 
> too big. After setting cpu to 3072K it worked(setting to 4096 K hanged 
> the system BTW).
> Cross compilation has also this advantage , that You have better editors 
> to Your disposal and You can work faster.
> Well native cc seems to be good enough, using pdp11 in emulator we have 
> anyway only hobbyst license .
> 
> Andrzej
> _______________________________________________
> TUHS mailing list
> TUHS at minnie.tuhs.org
> https://minnie.tuhs.org/mailman/listinfo/tuhs
> 
> 


---------------------------------------------
This message was sent using Monmouth Internet MI-Webmail.
http://www.monmouth.com/

[TUHS] [pups] gcc-3.4.6 and old unix

[Charlie ROOT]

On Tue, 25 Apr 2006 patv at monmouth.com wrote:

> If this helps at all, I've been working (very, very slowly) on a port of
> v32 to Intel platforms.  At first I used gcc for some kernel work, but
> quickly realized that it would be overwhelming to the final v7 system. 

If you're interested in running V7 on x86, you should check out
the 286-port on the TUHS FTP site, as it was fully operational 
until the author, according to his report, messed up the file 
system code.

For a 32-bit Unix, the Quasijarus project would be better
starting point, as it is more seasoned as a 32-bit operating
system.  The project project was started by Michael Sokolov, with
the primary goal of extending 4.3BSD-Tahoe to run on newer VAX
hardware.  You can find the source, as well as the mailing list,
from the web page at:  http://ifctfvax.harhan.org/Quasijarus/

Because NetBSD and especially GCC have long since outgrown all
but the most powerful VAX hardware, including my VAXstation
4000-60, I've been looking into the possibility of getting
Quasijarus to run on the machine - very slowly, of course.
I've only managed to hack the NetBSD kernel into running the
binaries properly - it should support 32V-ones as well, for
that matter.

I'm also interested in 386-ports of the classical Unix utilities,
but my kernel-side focus is on a brand new, non-portable kernel
written in assembly language for compactness and flexibility of
running, examining and debugging code that excessively picky
operating systems choke at - e.g. real- and kernel-mode code.

> Since I don't want to do the work twice, I looked for a different compiler
> suite.  I switched to the ACK compiler suite and just finished the WinXP
> cross compiler work.  It has a pdp11 back end, which I have yet to try,
> that may be useful.
> 
> It isn't gcc, but ir does do ANSI C and the i386 assembler seems to be
> pretty complete.  Let me know if there's any interest and I'll put it up
> on my site for download.

Did you make other improvements than XP cross-compilation?

-aw

[TUHS] [pups] gcc-3.4.6 and old unix

[Pat Villani]

It's actually a GNU port, with small changes to the actual sources.  
Mainly configure and make file changes so that it properly builds on 
cygwin and linux.  Creates minix binaries.

Pat


Charlie ROOT wrote:
>
> Did you make other improvements than XP cross-compilation?
>
> -aw
>   

[TUHS] [pups] gcc-3.4.6 and old unix

[Andrzej Popielewicz]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1116 bytes --]

Toby Thain napisał(a):

>On 24-Apr-06, at 9:05 AM, Wesley Parish wrote:
>
>  
>
>>Quoting Andrzej Popielewicz <vasco at icpnet.pl>:
>>
>>    
>>
>>>Wesley Parish napisał(a):
>>>
>>>      
>>>
>
>It can't be done.
>
>As others point out, the program is many times (100x or more?) too  
>big -- likely even gcc 1.x is far too big, but gcc {2,3,4}.x are all  
>meant for large 32-bit systems.
>
>However, cross-compilation can certainly be easily done. I have made  
>a PDP-11 back-end for lcc[1] (not quite complete but shows that it  
>can be done), which is an ANSI (c89) compiler[2]. lcc is a much  
>smaller and simpler compiler than gcc, but its executables are still  
>massively outsize for PDP-11 systems.
>  
>
Yes, even running vi or csh in Ultrix (in simh pdp11) produced message : 
too big. After setting cpu to 3072K it worked(setting to 4096 K hanged 
the system BTW).
Cross compilation has also this advantage , that You have better editors 
to Your disposal and You can work faster.
Well native cc seems to be good enough, using pdp11 in emulator we have 
anyway only hobbyst license .

Andrzej

[TUHS] [pups] gcc-3.4.6 and old unix

[Toby Thain]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 2746 bytes --]


On 24-Apr-06, at 9:05 AM, Wesley Parish wrote:

> Quoting Andrzej Popielewicz <vasco at icpnet.pl>:
>
>> Wesley Parish napisał(a):
>>
>>> djggp relies on a DOS extender to run its 32-bit programs in a 16- 
>>> bit
>>> operating environment on a 32-bit machine. go32 is Delorie's own DOS
>> extender
>>> - OpenWatcom offers at least two other DOS extenders to pull the  
>>> same
>> kind of
>>> trick.
>>>
>>>
>> I know that, because I have used djgpp for a quite lot of time.
>>
>> I thought , that Bill wanted to "scale down" gcc compiler to 16 bit
>> environment, and probably djgpp could be a good starting point.
>> I understand , that gcc is not "directly" portable.But because gcc  
>> can
>> create pdp code in cross compile environment, perhaps it is not
>> impossible .
>
> In that case, all I can suggest is that the gcc source files are
> cross-compiled to pdp11, and error messages noted.  Then the files get
> rewritten for the pdp11 ... I'm sorry I can't help - I'm neither a  
> gcc guru
> nor a pdp11 guru.


It can't be done.

As others point out, the program is many times (100x or more?) too  
big -- likely even gcc 1.x is far too big, but gcc {2,3,4}.x are all  
meant for large 32-bit systems.

However, cross-compilation can certainly be easily done. I have made  
a PDP-11 back-end for lcc[1] (not quite complete but shows that it  
can be done), which is an ANSI (c89) compiler[2]. lcc is a much  
smaller and simpler compiler than gcc, but its executables are still  
massively outsize for PDP-11 systems.

--Toby

[1] http://www.telegraphics.com.au/sw/info/lcc-pdp11.html
[2] http://www.cs.princeton.edu/software/lcc/


>>
>> BTW I have just ported Bob Supnik pdp11 emulator to Coherent  
>> 4.2.10 and
>> it seems to work(compile works etc). Coherent being 32 bit for 386
>> processors has nonetheless 286 support builtin in kernel, so it  
>> seems to
>>
>> be good environment for this simulator (?).
>
> That's good news!
>> I have the same problem as Bill, I have to kill pdp11 process on  
>> another
>>
>> console to exit.
>
> You're not the only one, either.
>
> Wesley Parish
>>
>>
>> Andrzej
>> _______________________________________________
>> PUPS mailing list
>> PUPS at minnie.tuhs.org
>> https://minnie.tuhs.org/mailman/listinfo/pups
>>
>
>
>
> "Sharpened hands are happy hands.
> "Brim the tinfall with mirthful bands"
> - A Deepness in the Sky, Vernor Vinge
>
> "I me.  Shape middled me.  I would come out into hot!"
> I from the spicy that day was overcasked mockingly - it's a symbol  
> of the
> other horizon. - emacs : meta x dissociated-press
> _______________________________________________
> PUPS mailing list
> PUPS at minnie.tuhs.org
> https://minnie.tuhs.org/mailman/listinfo/pups

[TUHS] [pups] gcc-3.4.6 and old unix

[Wesley Parish]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1896 bytes --]

Quoting Andrzej Popielewicz <vasco at icpnet.pl>: 
 
> Wesley Parish napisał(a): 
>  
> >djggp relies on a DOS extender to run its 32-bit programs in a 16-bit  
> >operating environment on a 32-bit machine. go32 is Delorie's own DOS 
> extender  
> >- OpenWatcom offers at least two other DOS extenders to pull the same 
> kind of  
> >trick.  
> >  
> > 
> I know that, because I have used djgpp for a quite lot of time. 
>  
> I thought , that Bill wanted to "scale down" gcc compiler to 16 bit  
> environment, and probably djgpp could be a good starting point. 
> I understand , that gcc is not "directly" portable.But because gcc can  
> create pdp code in cross compile environment, perhaps it is not 
> impossible . 
 
In that case, all I can suggest is that the gcc source files are 
cross-compiled to pdp11, and error messages noted.  Then the files get 
rewritten for the pdp11 ... I'm sorry I can't help - I'm neither a gcc guru 
nor a pdp11 guru. 
>  
> BTW I have just ported Bob Supnik pdp11 emulator to Coherent 4.2.10 and  
> it seems to work(compile works etc). Coherent being 32 bit for 386  
> processors has nonetheless 286 support builtin in kernel, so it seems to 
>  
> be good environment for this simulator (?). 
 
That's good news! 
> I have the same problem as Bill, I have to kill pdp11 process on another 
>  
> console to exit. 
 
You're not the only one, either. 
 
Wesley Parish 
>  
>  
> Andrzej 
> _______________________________________________ 
> PUPS mailing list 
> PUPS at minnie.tuhs.org 
> https://minnie.tuhs.org/mailman/listinfo/pups 
>   
 
 
 
"Sharpened hands are happy hands. 
"Brim the tinfall with mirthful bands"  
- A Deepness in the Sky, Vernor Vinge 
 
"I me.  Shape middled me.  I would come out into hot!"  
I from the spicy that day was overcasked mockingly - it's a symbol of the  
other horizon. - emacs : meta x dissociated-press