ROM image trouble (gensh2)

Fernando RUIZ CASAS (E-mail) fernando.ruiz at ctv.es
Fri Jun 8 07:32:42 UTC 2001


Hi,

 I am resolved this concern in three ways.

 LInkcmds for ram,
 linkcmds for rom
 and linkcmds for ramrom (Rom(8bits)shadow in ram(16bits))

To make this is necesary touch the start/start.S and make a litle
transformation in
start.c
Here the first step is set stack pointer.
make a cli() and after
call to init_pfc() and init_bsc().
A posible move romcode to ramspace
and initvars() moving the initial values to ram area
and let that start zeroes the bss area.

Once this is done the RTEMS is ready to enter.

To place the romvects section is easy placing it also in start.c like a
special section.

In linkcmds you can place it at 0x00000000.


I send you the bsp.cfg and the BSP.

Fernando RUIZ CASAS
home:correo at fernando-ruiz.com
work:fernando.ruiz at ctv.es




> -----Mensaje original-----
> De: Radzislaw Galler [mailto:rgaller at ET.PUT.Poznan.PL]
> Enviado el: viernes, 08 de junio de 2001 8:42
> Para: 'Rtems-Users (E-mail)
> Asunto: ROM image trouble (gensh2)
>
>
> Hi,
>
> Some time ago I managed to put RTMES on evaluation board - it
> was not so big
> trouble because gensh2 BSP was developed for it. Now I have
> built my own
> board with SH2 and some extra SRAM, and there is the probelm.
>
> For the correct start of the processor it is necessary to put
> interrupt
> vector table at the very begining. I copied gensh2 BSP and modified
> necessary files (also linkcmds) and it compiled successfuly.
> All necessary
> initialization of I/O ports was done. All I look for is the
> vector table to
> put in right place. Putting just my own created structure
> doesn't help (it
> has only pointer to start routine and its stack located in
> the internal
> processor memory). There is no possibilty to check what's
> going on in the
> processor (no monitor).
>
> Any clues are welcome.
>
> Radek
>
>




More information about the users mailing list