[kaffe] XScale jthreads code

Helmer Krämer hkraemer at freenet.de
Sat Sep 18 02:03:13 PDT 2004

On Fri, 17 Sep 2004 21:46:48 +0100
Peter Naulls <peter at chocky.org> wrote:


> In fact, most ARMs don't have FP support (it's emulated) including
> StrongARM variants on which Kaffe for ARM would mostly have been run in
> the past. Even when they do have genuine FP support, it doesn't change
> the generated code, and really there's almost nothing between ARMs that
> changes how application code operates.
> There are a number of reasons why a change might be required or seen:
>   One is the change in soft versus hard FP support in GCC's output (the
>   former means GCC tries to avoid actual FP instructions since they are
>   slow because of emulation), and might have sparked the confusion.
>   This depends upon the toolchain setup, and is certainly not XScale
>   specific.
>   Changes between GCC 2.95/3.3/3.4 meant the code came out differently
>   and different numbers of items needed to be saved on the stack.
>   Something else.
> In any case, I recommend you find the true cause, and take away any code
> that checks for an XScale.

to avoid the confusion, we should probably remove the --enable-xscale
configure option and have a --with-native-fpu (or whatever) option
instead that would default to true on i386 but to false on arm. The
different SP_OFFSETs in config/arm/threads.h would then depend on
whether an fpu is available and not on whether kaffe is built for
an xscale. Does that make more sense?


More information about the kaffe mailing list