[meta-ti] Is it intended to build TI bootloaders with march=armv5?

[Andreas Müller]

On Friday, December 16, 2011 02:44:07 PM you wrote:
> Do you have the disassembly around the access?
> ________________________________________
> From: meta-ti-bounces at yoctoproject.org [meta-ti-bounces at yoctoproject.org]
> on behalf of Andreas Müller [schnitzeltony at gmx.de] Sent: Friday, December
> 16, 2011 4:21 AM
> To: meta-ti at yoctoproject.org
> Subject: [meta-ti] Is it intended to build TI bootloaders with march=armv5?
> 
> Hi
> 
> as some of you might know, I am desperately working on u-boot SPL support
> for gumstix overo and having very strange effects: as soon as a stacic
> integer variable is accessed, the bootloader hangs [1].
> 
> In the compile logs I found ( checked also for beagleboard ) that
> 
> march=armv5
> 
> is used ( x-load is build with march=armv7-a ). Is this intended? Could
> this explain my trouble?
> 
> Andreas
> 
> [1] http://lists.denx.de/pipermail/u-boot/2011-December/113491.html
> 
> 
Please give me a hint how/where to get this.

Andreas