Re: [buildd] Etch?
On Fri, 4 Aug 2006, Wouter Verhelst wrote:
> I don't think that will be much of an issue. There are a few cases in
> which CF is different from classic 68k (try comparing the address
> register indirect with postincrement or predecrement addressing modes on
> CF and classic 68k for A7, if you want a very nice example); but such
> issues can be worked around by either not using them at all, or just by
> using them only in certain circumstances (though I don't think that will
> be possible for the above example). It will just boil down to making
> sure the compiler does that, and to making sure the assembler refuses to
> create opcodes that behave too differently on both architectures; this
> is much easier to accomplish than trying to avoid that the compiler
> makes mistakes in the optimizer, like it currently does.
While it's possible to avoid these instructions, it would mean possibly
very larger code and thus even slower code. Currently we at least assume
an 68020 cpu, but for CF support we had to go back to the 68000 and then
there are still a few instructions missing (mostly byte/word operations).
To be honest I'm not looking forward to this prospect.