floating point support

Nikhil Sahu

I am using OMAP4430 version A3.i am writing the sensor device driver and i need to use floating point calculation in my driver.

but unfortunately this architecture does not support floating point calculation.

its a way we can do the calculation in user space and connect it to our kernel driver but i do not want this way.

can you people help me how can i use floating point in my kernel driver.is there any flags are present that will enable our architecture to the floating point support.

over 11 years ago

0 Boyko Kazakov over 11 years ago

TI__Genius 10965 points

Hi,

How about floating point emulation ?Did you configure those defines for your kernel build:

CONFIG_FPE_NWFPE
CONFIG_FPE_NWFPE_XP
CONFIG_FPE_FASTFPE
CONFIG_VFP
CONFIG_VFPv3
CONFIG_NEON

Regards,

Boyko

0 Nikhil Sahu over 11 years ago in reply to Boyko Kazakov

Prodigy 230 points

Thx Boyko,

Which one is more suitable for OMAP4430 pandaboard among above all you suggest. and i am using the tool chain arm-eabi- right now.

is there any other tool chain (cross compiler )is available which can add floating point functionality in our kernel?

0 Nikhil Sahu over 11 years ago in reply to Boyko Kazakov

Prodigy 230 points

yes boyko

we configure all those defines for kernel build.

but still floating point calculation is not supported in my kernel-3.2.0.

what is the next step for supporting FPU in kernel

Thanks

Nikhil

0 Boyko Kazakov over 11 years ago in reply to Nikhil Sahu

TI__Genius 10965 points

Hi Nikhil,

Sorry for misleading you, we discussed this topic among our team and came to the conclusion that floating point operations may not be possible to be done within a custom kernel driver.

Is it possible to make those calculation with some integer based algorithm?

Regards,

Boyko

Processors

Processors forum

floating point support