Subject: Re: Floating point in the kernel
To: Jason Thorpe <>
From: Brian C. Grayson <>
List: tech-kern
Date: 09/18/1998 15:13:20
On Fri, Sep 18, 1998 at 12:37:52PM -0700, Jason Thorpe wrote:
> On Fri, 18 Sep 1998 14:49:29 -0500 
>  "Brian C. Grayson" <> wrote:
>  >   Or is a scheme like this what you meant by ``very careful about
>  > saves and restores''?  :)
> ...but it still doens't solve the problem of not even having FP in hardware.

  Right.  I was only trying to address ``FP can't be done in the
kernel on _any_ hardware for interrupt-performance reasons, so
Andreas is out of luck with his proposed LKM for a software

  I assume people without hardware FP wouldn't be able to use
his software modem, as emulated FP is probably way too slow to
satisfy the real-time demands.  Just like people without SMP
motherboards won't be able to use SMP -- it's a hardware
requirement.  But that won't (or shouldn't) stop work on either
project!  :)

"We do good work at a profit if we can, a loss if we must, but always good
work."  --  a car repair shop's slogan