2010-05-14 17:28:41     Possible problem using a float on a 522

Document created by Aaronwu Employee on Sep 26, 2013
Version 1Show Document
  • View in full screen mode

2010-05-14 17:28:41     Possible problem using a float on a 522

Reggy Perrin (UNITED STATES)

Message: 89420   

 

Hi folks,

 

We have working code on a BF526 board that we are porting to a 522 board.  Hardware is almost identical.  However, our program *appears* (I'm not ready to say definitively) to be bailing when we pass a float to a function.

 

In reading the wiki, I've added "-mfast-fp" to the bfin-linux-uclibc-g++ command, but that doesn't seem to change anything.

 

In looking through the kernel and library settings (from make menuconfig), I don't see any options that need to be set when building the kernel and libraries.

 

Am I missing something?

 

Thanks, RP

QuoteReplyEditDelete

 

 

2010-05-15 20:39:29     Re: Possible problem using a float on a 522

Mike Frysinger (UNITED STATES)

Message: 89439   

 

all Blackfin cores are exactly the same when it comes to the ISA and the userspace code generated.  perhaps it's more a matter of using something uninitialized and it just happens to work on the BF526 board.

QuoteReplyEditDelete

Attachments

    Outcomes