On Thu, Feb 26, 2015 at 12:03 PM, Sebastian Huber
wrote:
> On 26/02/15 15:54, Martin Galvan wrote:
>>
>> On Thu, Feb 26, 2015 at 11:47 AM, Sebastian Huber
>> wrote:
>>>
>>> >Hello Martin,
>>> >
>>> >sorry, I should have look at your bsp-start-init-registers.s earlier.
>>> > This
>>> >stuff looks
On 26/02/15 15:54, Martin Galvan wrote:
On Thu, Feb 26, 2015 at 11:47 AM, Sebastian Huber
wrote:
>Hello Martin,
>
>sorry, I should have look at your bsp-start-init-registers.s earlier. This
>stuff looks so generic, that it should go directly into start.S. I guess the
>TMS570 will not remain th
On Thu, Feb 26, 2015 at 11:47 AM, Sebastian Huber
wrote:
> Hello Martin,
>
> sorry, I should have look at your bsp-start-init-registers.s earlier. This
> stuff looks so generic, that it should go directly into start.S. I guess the
> TMS570 will not remain the only ARM core with a lockstep feature.
Hello Martin,
sorry, I should have look at your bsp-start-init-registers.s earlier.
This stuff looks so generic, that it should go directly into start.S. I
guess the TMS570 will not remain the only ARM core with a lockstep feature.
You don't have to save and restore the r0 in a function since
Changes from v1:
* Add bsp_start_* prefixes.
* Add *_banked_fiq suffixes to make it clear it's referring to the FIQ mode
banked regs.
Follow-up from here:
https://lists.rtems.org/pipermail/devel/2015-February/009974.html
This patch adds the macro BSP_START_NEEDS_REGISTER_INITIALIZATION
