gcc-3.3 VS gcc-2.95.3 soft floating point on mipsel?

[David Wuertele]

For some time now, I have been using a cross compiler for my mipsel
embedded system.  The toolchain is based on gcc-2.95.3 (plus redhat
patches), and I have been pretty happy with it.

Trying to modernize my setup, I am now trying crosstool-0.28-rc31 with
gcc-3.3.3 and glibc-2.3.2.  But I'm finding that my display libraries,
which rely heavily on floating point operations, go haywire (garbage
on the display) when I compile them with gcc-3.3.3 -msoft-float.

By just removing the -msoft-float from my compilation of those
libraries, everything works fine.  The only difference is that now I'm
relying on kernel FP emulation, which is much slower.  The performance
of my display routines is unacceptable.

What has changed in terms of mipsel -msoft-float between 2.95.3 and
3.3?  Is there a patch that I should be applying to fix -msoft-float?

Thanks,
Dave


------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sources.redhat.com