Does anyone know why we have m68k-specific implementations of the
insque and remque functions?  

Until today, I thought that they used the cas or cas2 instructions to
make the link/unlink atomic to match the original vax implementation.
But much to my surpise, I discovered that the assembly code is almost
identical to what gcc emits (the only difference is the assmbly code
does away with the stack frame).

If the assembly code is identical to the C version, I don't think we
should have it.

	--jtc