Re: Possibly a stupid question about 21164/21064 instruction diffs.

[Richard Henderson <rth cygnus com>]

On Wed, Dec 30, 1998 at 04:34:07PM -0500, Alvin Starr wrote:
> I have a couple of binaries that have been generateed for a 21164. This is
> a problem since at this point I have a number of 21064/66 systems that I
> do not wish to trash.

There are no 21164 vs 21064 instruction differences.

There are differences from 21164a though.

> Would it not be possible to trap the invalid instruction and execute it in
> the kernel. I know there would be a big speed hit but at least the program
> would run. 

Yes it is possible.  See the opDEC case in do_entIF in
arch/alpha/kernel/traps.c.


r~