Linux/DM3730: floating point exception

ranchu

Guru 20755 points

Part Number: DM3730

Tool/software: Linux

Hello,

I am trying to invoke floating point exception in bare metal but without success yet.

In the TRM (sprugn4) I could not find any details about the floating point exception.

I do have success in catching other exception (undefined command for example), but not floating point exception.

This is what I did this far:

1. I verified that floating point is supported in compiler

2. I also see that the numbers are printed as float, and when trying to divide 2 floats, 0/0 it gives inf in the result (but no exception).

Can anyone please give some hints how we can generate a floating point exception ?

Is it masked somehow by default ?

Thanks,

Ran

over 8 years ago

0 Steve Kipisz over 8 years ago

TI__Genius 15510 points

You probably need to look in an ARM reference manual for more information. I have not looked at this in a long time, but I recall there is cp15 register where you have to enable cp10 and cp11 for floating point to work.

Steve K.

0 ranchu over 8 years ago in reply to Steve Kipisz

Guru 20755 points

Hi Steve,

I see in arm web site the following comment:

infocenter.arm.com/.../index.jsp

VFPv3 is backwards compatible with VFPv2, except that it cannot trap floating point exceptions.

Does it mean it is not supported in cortex A8 ?

Thanks,
Ran

Processors

Processors forum

Linux/DM3730: floating point exception