How to use floating point operation on AM437x

Nomo

Hello,

Please let me know how to use floating point operation such as the "float" type or "double" type on AM437x.
I want to build a floating point operation in to the kernel space software (driver) with Linux SDK v8.
However, when the kernel is building, the following error is output.

-------------------------------------
undefined reference to `__aeabi_dadd'
undefined reference to `__aeabi_dsub'
...
-------------------------------------

Best Regards,

Nomoto

over 10 years ago

0 Biser Gatchev-XID over 10 years ago

TI__Guru**** 393215 points

Hi Nomoto-san,

As far as I know the Linux kernel does not support floating point. I will ask the SW team to check this.

0 Yordan Kovachev over 10 years ago in reply to Biser Gatchev-XID

TI__Guru**** 161600 points

Hi,

I confirm Biser's answer.

Using floating points operations in kernel is strongly discouraged.

There are a couple of hacks you can try, see the following community discussions:
stackoverflow.com/.../use-of-floating-point-in-the-linux-kernel
www.linuxsmiths.com/blog
yarchive.net/.../kernel_fp.html

But generally I would advise against using floating point operations inside kernel space.

Best Regards,
Yordan

0 Matthijs van Duin over 10 years ago in reply to Yordan Kovachev

Mastermind 8040 points

Most of those links are x86-specific though.

There's also a specific doc on Neon in the kernel.

Processors

Processors forum

How to use floating point operation on AM437x