Re: "obsolete" hardware

Michael Poole (poole+@andrew.cmu.edu)
Wed, 11 Jun 1997 10:47:56 -0400 (EDT)


On Wed, 11 Jun 1997, yuri mironoff wrote:
>
> Forgive me if I didn't make myself clear. But let me say it another
> way. If the choice is between spending the development time working on 386
> backward compatibility or taking advantage of some new feature of a P7 -
> drop the 386 like a hot potato.
>

Since I've seen that argument twice too often, I fell a need to
speak out:
The P7 will not be any reason to drop support for the 386.
Rather, since it will be essentially a RISC chip with an onboard x86
emulator (according to the hype I remember reading from Intel and HP), and
its native instruction set will be different than the x86 series, it will
need a new tree in linux/arch/. Any new features or whatever can be
implemented there; the current x86 code should continue to support the
386.
As many people have said, the 386 code isn't broken, and dropping
support for it won't really gain us anything.