From: Kevin Brown (kevin@sccsi.com)
Date: 03/14/93


From: kevin@sccsi.com (Kevin Brown)
Subject: Re: a.out format compatiblity
Date: Mon, 15 Mar 1993 03:58:06 GMT

In article <1993Mar11.192558.9745@rtfm.mlb.fl.us> luckey@rtfm.mlb.fl.us (Jon Luckey) writes:
>I recall a while back a postin in comp.os.coherent that some
>one was using Coherent to develop programs to run on xenix.
>
>It was much more cost effective for him to buy coherent for
>$100, than the N-thousand dollar SCO C development suite, and
>that the a.out execuatble binary formats were compatable.
>
>Does Linux (and/or BSD386 for that matter) maintain that
>compatiblity in a.out format? Even if just in small model
>programs? Can binarys migrate in one direction and not the
>reverse (such as linux using 386 instructions that the 286
>based coherent can't cope with)?

Linux does use a.out format, but I don't know that it is in any way
compatible with Xenix's executable format.

If you're interested in doing cross development, you might be interested
in getting the gcc sources and building a Linux to Xenix cross compiler.
I'm sure there are ways to configure the tools to generate the right kind
of executable output.

However, you'll need to insure that you can generate xenix system calls
from the cross compiler. Since I expect the system call code is
proprietary, it's not clear that you can do anything other than compile
to object files and perform linking on the target machine. Since you
have to have the linker in order to rebuild the kernel (which is usually
how kernel configuration is done), there's some chance that this would
work even if the target machine didn't have the software development
system (though it's not clear that the libraries would be supplied in
that event).

-- 
Kevin Brown                                     kevin@nuchat.sccsi.com
This is your .signature virus: < begin 644 .signature (9V]T8VAA(0K0z end >
            This is your .signature virus on drugs: <>
                        Any questions?