Re: STM32F4 and Dynamic loading.

Hi Valentin

On 28/02/14 10:15, Valentin BOUSSON wrote:

> Actually, yes it could be sufficient for my project, and digging around
> my problem, I tried to compile the CYGPKG_OBJLOADER package too.
> And this is with this one I have a conflict problems :
> 
> This package requires CYGINT_SERVICES_OBJLOAD_RELOCATOR == 1,
> which is implemented by :
> - CYGBLD_OBJLOADER_ARCHITECTURE_{POWERPC,I386,ARM}
> 
> The problem now is the template for stm32f4discovery seems to "inherit"
> from CYGPKG_HAL_CORTEXM while CYGBLD_OBJLOADER_ARCHITECTURE_ARM is
> calculated from CYGPKG_HAL_ARM only.
> 
> Is there a structural incompatibility between HAL_ARM and HAL_CORTEXM ?
> Or that's a thing I can modify/hack in the ecos.
> 
> Or (I don't hope so, I'm still a beginner) do I have to implement
> something to create the CYGBLD_OBJLOADER_ARCHITECTURE_CORTEXM ?

Good question. It depends whether relocate_arm.c supports the various
relocation types used with Cortex-M targets. The code calls CYG_ASSERT
if it encounters relocation types it does not support, so you might try
experimenting with cdl_option CYGBLD_OBJLOADER_ARCHITECTURE_ARM modified
as follows and see if it works:

    calculated CYGPKG_HAL_ARM || CYGPKG_HAL_CORTEXM

Of course, you should enable CYGDBG_USE_ASSERTS.

If you decide to experiment, please report back to the ecos-discuss list.

I hope this helps...

John Dallaway
eCos maintainer
http://www.dallaway.org.uk/john

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss