use hard float in Genode

Stefan Kalkowski stefan.kalkowski at
Tue Oct 30 10:40:22 CET 2018


On Tue, Oct 30, 2018 at 02:38:57PM +0800, lzSun wrote:
> Hello,
> I wonder whether I can use hard float in Genode. I am now using Genode 18.05 and was working on tz_vmm demo. I add "-CC_OPT += -mfloat-abi=hard -mfpu=vfp" in "/repos/base/mk/" file. It seems Genode tool chain can recognize this arguments,but print error message like this:
> "/usr/local/genode-gcc/bin/genode-arm-ld: error: supc++.o.tmp uses VFP register arguments, /usr/local/genode-gcc/bin/../lib/gcc/arm-none-eabi/6.3.0/../../../../arm-none-eabi/lib/libsupc++.a(class_type_info.o) does not"
> Form the error message, I guess it can't use hard float in genode because some static library in genode tool chain are build by soft float.Am I right? Any help will be OK, thanks. 

yes, you're right. If you want to compile with hard-float and link
against the C++ runtime library delivered by Genode's toolchain, it
will fail. You need to compile the Genode toolchain yourself, if this
is a hard requirement for you. Have a look at the tool/tool_chain
script in Genode's source code repository therefore.

Please, be aware that the kernel needs to switch the FPU then too.
This is not the case on all ARM platforms within our own base-hw
kernel, and also not on i.MX53. There is a topic branch to generally
enable the FPU on all of our ARM platforms of base-hw here:

but it needs more work to be done. Actually, I'm at it, and hope to
close that gap quite soon.

Best regards

> ------------------
> Best wishes

> _______________________________________________
> Genode users mailing list
> users at

Stefan Kalkowski
Genode labs |

More information about the users mailing list