software FP support?

[Kumar Gala]

I was wondering what the current state of software floating point support
was in the PowerPC embedded world.

I am trying to build a glibc with soft-fp to get a /bin/sh that does not
setup the fpu state.  I was wondering if someone has already gone through
this trouble or how people have been handling this.

I know that the kernel has support for soft-fp handling which would allow
userland and libc to still use FP code but that seems expensive.

Any information would be helpful.

thanks

 - kumar


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/