hardfloat - support?

[thomask77]

Hi!

Is there any support for -mfloat-abi=hard yet?

This would be a real killer feature for SAT, because the free version of the commercial toolchains always omit this calling convention. There are also cheap STM32F4 discovery boards which would really benefit from that.

[Jeremy Herbert]

I have been trying to get this working (unsuccessfully) over the last few days also. +1

[Piotr Esden-Tempski]

I know that this is a broadly requested feature. I sadly did not have time till now to take a look into that, and I am not sure when I will. :(

[Jeremy Herbert]

I have had success using the hard floating point unit with MikeSmith's fork of this project.

https://github.com/MikeSmith/summon-arm-toolchain

for the record, the flags you should use are:

-mcpu=cortex-m4 -mthumb-interwork -mfloat-abi=hard -mfpu=fpv4-sp-d16 -mlittle-endian

[soonhooi]

I tried to compile summon-arm-toolchain in windows with Git Bash; but encounter the "Found unknown OS. Aborting!" error after running ./summon-arm-toolchain.

Is there a workaround for this? Thanks!

[Michael Pratt]

I managed to get a working hardfloat toolchain, though I have heard from some that they are unable to get it to build.

https://github.com/prattmic/arm-cortex-m4-hardfloat-toolchain

[thomask77]

Since some time, there's also an official(?) toolchain from ARM(?) with hardfloat-multilibs:

https://launchpad.net/gcc-arm-embedded

I'm using it for all my new projects, replacing the CodeSourcery toolchain.

[Piotr Esden-Tempski]

Hardfloat should be working now. Please open a new issue if this still persists. :)