This is the mail archive of the crossgcc@sourceware.org mailing list for the crossgcc project.

See the CrossGCC FAQ for lots more information.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: AMCC-440EP floating point error


bora.sahin@ttnet.net.tr said the following on 04/17/2006 05:13 AM:
do you know if --without-fp was used when configuring glibc
your problem looks like that your gcc is not compiled with hard fp
support where as glibc is configured with hard fp. You can try adding
--without-fp to your glibc configure command.

Hmm... I had tried --without-fp, which is the default for 440 in crosstool, before posted a message to the list and crosstool had managed to create a cross toolchain... Below is the default content of 440.dat file:


TARGET=powerpc-440-linux-gnu
TARGET_CFLAGS="-O -mcpu=440"
GCC_EXTRA_CONFIG="--with-cpu=440 --enable-cxx-flags=-mcpu=440"
GLIBC_EXTRA_CONFIG="--without-fp"

So far so everything was OK...

But 440EP has hard floating point support and I wanted to use it so I omitted the GLIBC_EXTRA_CONFIG because glibc uses hard floating point as default. Besides I added GLIBC_EXTRA_CONFIG=â-mhard-floatâ as an extra precaution. Now the contents are

TARGET=powerpc-440-linux-gnu
TARGET_CFLAGS="-O -mcpu=440"
GCC_EXTRA_CONFIG="--with-cpu=440 --enable-cxx-flags=-mcpu=440"
GLIBC_EXTRA_CONFIG=â-mhard-floatâ
It seems you are trying to compile glibc with floating point using a GCC which is not configured to hard floating point.
you need to add --with-fp to config and you can remove -mahrd-float if you want it to use
fpu instructions.
When I use this config I am getting error messages...

After reading your message, then I thought perhaps I should change cpu type because not every 440 cpu has a hard FPU. The new 440.dat file content:

TARGET=powerpc-440-linux-gnu
TARGET_CFLAGS="-O -mcpu=440"
GCC_EXTRA_CONFIG="--with-cpu=440fp --enable-cxx-flags=-mcpu=440fp"
GLIBC_EXTRA_CONFIG=â-mhard-floatâ

Probably you need to configure GCC with floating point too add --with-float=hard to GCC_EXTRA_CONFIG
then you do not need --enable-cxx-flags=-mcpu=440fp


Then again I got the same error message. Perhaps I should specify 440fp in another form like -march?

Thanks...

--
Bora SAHIN

--
For unsubscribe information see http://sourceware.org/lists.html#faq



--
For unsubscribe information see http://sourceware.org/lists.html#faq


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]