Re: What platform can run m68k COFF binaries?

2018-09-28 Thread Tomasz Rola via cctalk
On Wed, Sep 05, 2018 at 03:39:20PM -0400, Ethan Dicks via cctalk wrote:
> Hi, All,
> 
> I have a backup of some old code that I thought came from a Sun3
> machine, and indeed, there _are_ binaries on there, in a directory
> 'sun' that _do_ run on a Sun3, verified under emulation with "tme".
> 
> The part that puzzles me is the collection of object files and
> binaries in the directory above that.  'file' tells me that they are
> "m68k COFF" files.  From what I've read so far, COFF binaries are from
[...]
> I have a lot of experience with UNIX but my thinnest amount is in the
> m68k arena (mostly some dabbling on Sun3 workstations and a
> Perkin-Elmer 7350).  Thanks for any pointers or tips.
> 
> -ethan

A bit late to the party, but perhaps you could get something from
strings:

  $ strings /bin/tar | less

/lib64/ld-linux-x86-64.so.2
,crS
>gUa
librt.so.1
__gmon_start__
_Jv_RegisterClasses
clock_gettime
libc.so.6

(... and so on, but some of this crap might be useful for your
goal...)

I guess strings is rather easy to write in C, if you do not have
it. Or maybe even could be concocted using a bunch of shell and
pipes.

-- 
Regards,
Tomasz Rola

--
** A C programmer asked whether computer had Buddha's nature.  **
** As the answer, master did "rm -rif" on the programmer's home**
** directory. And then the C programmer became enlightened...  **
** **
** Tomasz Rola  mailto:tomasz_r...@bigfoot.com **


Re: What platform can run m68k COFF binaries?

2018-09-05 Thread Al Kossow via cctalk
The other way to tell would be if they haven't had their symbols
stripped and look at the libaries they are compiled with. This
stuff predates shared libraries. There are probably vendor strings
in there.

On 9/5/18 2:26 PM, Al Kossow via cctalk wrote:
> anything that came from Unisoft Sys V or early Motorola Sys V
> 
> ftp://ftp.oreilly.co.jp/palm/Linux/gcc/gdb-4.16/bfd/coff-aux.c
> 
> for example.
> 
> On 9/5/18 12:39 PM, Ethan Dicks via cctalk wrote:
>> Hi, All,
>>
>> I have a backup of some old code that I thought came from a Sun3
>> machine, and indeed, there _are_ binaries on there, in a directory
>> 'sun' that _do_ run on a Sun3, verified under emulation with "tme".
>>
>> The part that puzzles me is the collection of object files and
>> binaries in the directory above that.  'file' tells me that they are
>> "m68k COFF" files.  From what I've read so far, COFF binaries are from
>> System V Release 2-4.  What I can't reconcile is what they might have
>> been compiled on.  Is it possible these were made for A/UX?  The
>> relevant file dates (late-1988 through mid-1989) do overlap
>> availability of A/UX version 1.
> 



Re: What platform can run m68k COFF binaries?

2018-09-05 Thread Al Kossow via cctalk
anything that came from Unisoft Sys V or early Motorola Sys V

ftp://ftp.oreilly.co.jp/palm/Linux/gcc/gdb-4.16/bfd/coff-aux.c

for example.

On 9/5/18 12:39 PM, Ethan Dicks via cctalk wrote:
> Hi, All,
> 
> I have a backup of some old code that I thought came from a Sun3
> machine, and indeed, there _are_ binaries on there, in a directory
> 'sun' that _do_ run on a Sun3, verified under emulation with "tme".
> 
> The part that puzzles me is the collection of object files and
> binaries in the directory above that.  'file' tells me that they are
> "m68k COFF" files.  From what I've read so far, COFF binaries are from
> System V Release 2-4.  What I can't reconcile is what they might have
> been compiled on.  Is it possible these were made for A/UX?  The
> relevant file dates (late-1988 through mid-1989) do overlap
> availability of A/UX version 1.



Re: What platform can run m68k COFF binaries?

2018-09-05 Thread Brian L. Stuart via cctalk
On Wed, 9/5/18, Ethan Dicks via cctalk  wrote:
> The part that puzzles me is the collection of object files and
> binaries in the directory above that.  'file' tells me that they are
> "m68k COFF" files.  From what I've read so far, COFF binaries are from
> System V Release 2-4.  What I can't reconcile is what they might have
> been compiled on.  Is it possible these were made for A/UX?  The
> relevant file dates (late-1988 through mid-1989) do overlap
> availability of A/UX version 1.

I'm pretty sure the AT&T UNIX PC (7300/3b1) used COFF.
I remember messing around with it when porting an old
version of gcc to the machine.  At that time, GNU was
playing around with something called robotusin (COFF
medicine).

BLS


Re: What platform can run m68k COFF binaries?

2018-09-05 Thread Warner Losh via cctalk
On Wed, Sep 5, 2018 at 1:39 PM Ethan Dicks via cctalk 
wrote:

> Additionally, I'm reading that FreeBSD has a binary compatibility
> layer for COFF but I wonder if that's for Intel binaries only or if it
> extends to m68k.
>

FreeBSD never ran on 68k, so I think that would be no.

NetBSD, however, did and you would likely be able to run them there.
There's even a sys/compat/sunos directory with this tasty comment:
 * the code below is only needed for sun2/sun3 emulation.
in sunos_exec_aout.c which is suggestive. IIRC, I had a friend that had a
special SunOS sun3 binary he ran on NetBSD on one of his 68k machines, but
that was at least 10 years ago or more and I've lost touch with him...

Now, this assumes it's sun3 running SunOS and not some variation of SysV.
That too might work, but might not.

Warner


What platform can run m68k COFF binaries?

2018-09-05 Thread Ethan Dicks via cctalk
Hi, All,

I have a backup of some old code that I thought came from a Sun3
machine, and indeed, there _are_ binaries on there, in a directory
'sun' that _do_ run on a Sun3, verified under emulation with "tme".

The part that puzzles me is the collection of object files and
binaries in the directory above that.  'file' tells me that they are
"m68k COFF" files.  From what I've read so far, COFF binaries are from
System V Release 2-4.  What I can't reconcile is what they might have
been compiled on.  Is it possible these were made for A/UX?  The
relevant file dates (late-1988 through mid-1989) do overlap
availability of A/UX version 1.

If these are for A/UX, it would be handy to verify this, then it would
be more handy to be able to run them (I have source for some of this
but I'm having problems getting modern C compilers to digest
30-year-old crufty code, and for a couple of the utilities I need,
there is no source).

I'm reading of a "shoebill" emulator.  Anyone have any experience with it?

Additionally, I'm reading that FreeBSD has a binary compatibility
layer for COFF but I wonder if that's for Intel binaries only or if it
extends to m68k.

I have a lot of experience with UNIX but my thinnest amount is in the
m68k arena (mostly some dabbling on Sun3 workstations and a
Perkin-Elmer 7350).  Thanks for any pointers or tips.

-ethan