Re: double (longlong) problems on ARM

[Dan Kegel <dank at kegel dot com>]

Craig A. Vanderborgh wrote:
> I have a gcc-3.3.2 based toolchain that I constructed with the aid of
> Dan Kegel's crosstool-0.27.  It all works very well but I have one huge
> problem I cannot figure out.
>
> My crossgcc is generating code so that the longlongs (and doubles) have
> the constituent 4-byte words BACKWARDS from what is expected on the
> target (a netwinder).
>
> How do I configure the compiler/binutils build so that the 4-byte words
> for longs/doubles are REVERSED from what they are now.  I have
> debugged/prognosticated the situation enough to know that if I could do
> this, I'd have what the netwinder is expecting to see.

This sounds like what was discussed in http://www.spinics.net/lists/arm/msg08396.html
Does figure 9.3.3 of
http://netwinder.osuosl.org/pub/netwinder/docs/arm/ARM7500FEvB_5.pdf
explain things at all (at least for floating point)?

I wonder if building for softfloat or vfp would change this.

BTW you should be using crosstool-0.28-rc37, not crosstool-0.27.
It has better support for vfp on arm; see the comments in
http://kegel.com/crosstool/crosstool-0.28-rc37/patches/glibc-2.3.2/glibc-fp-byteorder.patch
- Dan

--
Trying to get a job as a c++ developer?  See http://kegel.com/academy/getting-hired.html

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