raspberry BSP: Maybe there is a bug in the linker file?

[Christian Mauderer]

On 19/12/2019 15:05, Christian Mauderer wrote:
> On 19/12/2019 14:27, Sebastian Huber wrote:
>> On 19/12/2019 14:24, Christian Mauderer wrote:
>>> Hello,
>>>
>>> triggered by the discussion regarding RTEMS on raspberrypi 3 I did some
>>> tests. I haven't been able to start a RTEMS on my Pi 1 if I tried a
>>> version after commit c5fd79cd4704a4270ba0114a1009ab8556f997c9 from
>>> 29.07.2019. Right before it everything works as expected.
>>>
>>> The commit changes the memory locations. From what I understood, the
>>> bootloader on a raspberry always jumps to address 0x8000 after loading
>>> the application. So I'm not convinced that the change is correct.
>>
>> Is the 0x8000 a fixed address or you configure this in a boot image header?
>>
> 
>>From what I've seen the bootloader just uses a fixed address. All guides
> that I've found just use objcopy to generate a binary from the elf files
> without adding a header. For example a guide for RTEMS:
> 
> 
> http://alanstechnotes.blogspot.com/2013/03/running-your-first-rtems-program-on.html
> 
> A FreeRTOS port mentions this behavior explicitly in a comment in the
> linker file:
> 
> 
> https://github.com/jameswalmsley/RaspberryPi-FreeRTOS/blob/master/raspberrypi.ld#L17
> 
> The "graphics processor" is no typo here. It seems the initialization is
> really done by the graphics processor on the pi.
> _______________________________________________
> devel mailing list
> devel at rtems.org
> http://lists.rtems.org/mailman/listinfo/devel
> 

I just found an option: It is possible to set a "kernel_address" in the
config.txt:


https://www.raspberrypi.org/documentation/configuration/config-txt/boot.md

So the start address can be changed and maybe the adaption to the linker
file isn't necessary.