Trouble building RTEMS from master (TMS570 and LPC1768)

[Pavel Pisa]

Hello Martin and Gedare,

On Wednesday 04 of November 2015 22:58:57 Martin Galvan wrote:
> On Wed, Nov 4, 2015 at 6:57 PM, Gedare Bloom <gedare at rtems.org> wrote:
> > Did you bootstrap -c and then bootstrap? I think I saw this error
> > earlier today too, with a new sparc toolchain, but then it went away
> > after i rm-rf'd my build tree and tried over.
>
> Yeah, I did this several times and I'm still seeing it.

I take the complete bootstrap sequence as

  ./bootstrap -c
  ./bootstrap -p
  ./bootstrap

But I have not checked latest mater in past weeks.

By the way, we have already working LwIP driver
for RTESM TMS570 BSP. I hope we do final clean up
for initial RC publication. We have some corrections
for header files as well, already available in Premek's
github repos, but these should go to both master and 4.11.
We have one suitable solution for vectors redirection
between bootloader and application. We have implemented
and tested at least five or may it be even more options
but many of them have problems with stability because
use HW a way for which it has not been initially intended.
We need to decide what should be done with master and 4.11.
The actual included variant of solution with POM is highly
problematic.

So it is important what are intended use cases.
The one where RTEMS image starts from address 0
is simple and should no use POM nor VIM interrupt bypassing.
But it requires integration with HAL startup for now.
Solution with separate setup/loader requires vectors
remapping. I am prepared to discuss possible and tested
options. I have got and expect some more feedback from Ti
as well.

Best wishes,

              Pavel