Anyone using Ghidra for A/UX executables, how are you doing it? Just extracting the .text/.data sections and importing as raw, or do you have a way to directly load the COFF objects?
Out of the box Ghidra doesn't recognise big-endian COFF, but the patch is an easy fix and at least this gives you the symbols from non-stripped objects. There's no support for A/UX's shared libraries, I guess this would need more work to handle relocations (I don't really understand COFF shlibs, so I have nfi about this). And converting TRAP#0 to decompile to a function call needs knowledge from PAGING/os/sysent.c to correctly handle arguments and then whatever on the Ghidra side to assist the decompiler.
Has anyone tackled this with external support like the mac classic stuff?
Out of the box Ghidra doesn't recognise big-endian COFF, but the patch is an easy fix and at least this gives you the symbols from non-stripped objects. There's no support for A/UX's shared libraries, I guess this would need more work to handle relocations (I don't really understand COFF shlibs, so I have nfi about this). And converting TRAP#0 to decompile to a function call needs knowledge from PAGING/os/sysent.c to correctly handle arguments and then whatever on the Ghidra side to assist the decompiler.
Has anyone tackled this with external support like the mac classic stuff?