building a toolchian with hard float support
adam at os.inf.tu-dresden.de
Wed Dec 10 22:44:48 CET 2008
On Wed Dec 10, 2008 at 04:54:25 -0800, Gabi Voiculescu wrote:
> I have to create a toolchain with hard fpu support for arm1176's VFPv2.
> That's because, by looking at the ARM documentation I don't see how I can perform
> mathematic operations on the FPU (like FABSx, FADDx without using these mnemonics), (as it is the case for fpu load/store operations), so as far as see, I have to build a toolchain that recognized these mnemonics.
> I dont' know where to start.
> Has anybody used a script like crosstools for this? I hope by creating a new dat file with proper GCC and GLIBC argumets I could do the job myself, but I have no idea what would those arguments be or where to find them:
> GCC += ?? --with-float=hard --fpu=vfpv2 --cpu=arm1176jzf ??
> GLIBC += ??
> Where can I find what gcc version might include support for my hard fpu? I tried the man pages in gcc-4.1.2 and gcc-3.4.4 without success and I can't figure out where to look for this on gnu.org.
> Can you point me in the right direction to use.
Codesourcery is known to have recent pre-built ARM compilers including
Adam adam at os.inf.tu-dresden.de
More information about the l4-hackers