Date: Sat, 12 Nov 1994 17:38:56 -0500
   From: "Chris G. Demetriou" <cgd@alpha.bostic.com>

   [...] the kernel _should_ try to find out whether or not
   an FPU is present, and then do the save/restore of the FP context
   based on that.  the FP context save/restore is the only place in the
   kernel that uses FP.  I take it that that's not currently done?
   [...]

I don't know if it's done in 1.0 or not, but it wasn't in 0.9.

[The followings are to Alice team, rather than to Chris]

I think it'd be easier that Booter determins whether or not an FPU is
available, than that the kernel does it.  Booter can then pass a flag
to the kernel.  If it has to be done in the kernel, it has to try to
execute a floating-point instruction (I used FSAVE in a pre-bootstrap
CPU/FPU identifying routine that I wrote about a three years ago) and
determin whether f-line trap has been occurred, or if not, what type
of FPU the system has from the stack frame FSAVE creates.  I don't
know how much freedom the kernel could have during the above process.
The routine I wrote was free to mangle the vector table and the stack;
it only had to pass the correct CPU/FPU type codes to the ROM
bootstrap loader.

But what I'm concerned a little is whether or not Gestalt (sp?)
returns the correct information when SoftFPU is installed.  Does
anyone know if it does or not?  Maybe we should not be concerned
if SoftFPU is installed or not, since MacBSD cannot boot with
MacOS virtual memory turned on anyway.

Just a thought.

Ken