Re: [microblaze-uclinux] Double precision support

[John Williams <jwilliams@xxxxxxxxxxxxxx>]

Hi Taewon,

Taeweon Suh wrote:
> One question about double-precision support in uClinux.
> I added mpeg2 decoder user application from mediabench into a user directory.
> While compiling, I ran into a compile error. undefined __gtdf2.
> The function looks like double precision comparion.
> Doesn't the current cross compile library support double precision?


The gtdf2 symbol is a low level floating point suport function - it is 
defined in the toolchain librariy file

/home/bulldog2/suhtw/project/MB_uClinux/microblaze-elf-tools/bin/../lib/gcc-lib/microblaze/2.95.3-4/./../../../../microblaze/lib/libc.a

You can confirm this with objdump

Your problem might be to related to link order of the libraries.  Best 
thing to try is take the mb-gcc command line that is failing (the one 
you posted), and try changing the order of some of the link options 
(-lc, -lm) and explicit libfiles (libc.a, libgcc.a) and so on - see what 
effect that has.  I know this sounds vague, but you'll just have to 
experiment.  If/when you find an ordering that works, let me know and we 
'll see about making it a generic fix.

Regards,

John


___________________________
microblaze-uclinux mailing list
microblaze-uclinux@itee.uq.edu.au
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/