Hi Renaud,
On Thu, Jan 22, 2015 at 11:53:56AM +0000, Renaud Barbier wrote:
>
> I am doing my first port to an ARM CPU (Broadcom chip). For testing I
> built the friendlyarm-tiny210 and created an objdump to see what the
> code look like:
>
> barebox: file format elf32-littlearm
>
>
> Disassembly of section .text:
>
> 23e00000 <start>:
> 23e00000: ea000012 b 23e00050 <start+0x50>
> ...
> 23e00050: ea000013 b 23e000a4 <barebox_arm_reset_vector>
> ...
>
> 23e000a4 <barebox_arm_reset_vector>:
> 23e000a4: e52de004 push {lr} ; (str lr, [sp, #-4]!)
>
> I am a bit surprise to see a push to stack as the first instruction of
> barebox_arm_reset_vector as the stack pointer has not been defined yet.
> Or is there a pre boot loader that is not built as part of the
> friendlyarm-tiny210 build.
>
> Other BSP I have built do not produce this push instruction.
> Where is the difference coming from in the build?
>
> Are all BSP having a lowlevel.c file assumed to have a pre-bootloader?
barebox_arm_reset_vector should have a __naked like the other boards
have, then the stack won't be used. Does the board work? If yes then the
stack pointer seems to be initialized with some value where it doesn't
hurt to push to, but it doesn't look intentional.
Sascha
--
Pengutronix e.K. | |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
_______________________________________________
barebox mailing list
barebox@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/barebox
