For some reason, Reko was not able to decompile this code into a C representation

That's likely because it's one of those (of which many existed) which attempt to dumbly pattern-match against what a typical C compiler of the time (with equally dumb and extremely inefficient code generation) would do, but that routine clearly looks like handwritten Asm. I've never seen a C compiler from that era generate a LOOP instruction, for example, and of course "cli" nor the I/O instructions are not expressable except perhaps as intrinsics. Ghidra might be a bit better at this, as it's a generalised decompiler.

In fact, when the compiler (RPGC.EXE) compiles some RPG source code, it seems to copy the parallel port routine from itself into the compiled program.

This reminds me of the classic Ken Thompson attack.