Re: "dummy-flash" device tree possible bug?


Diego Sueiro
 

Kumar,

When you are using one RAM region for both memory & code, are you splitting the region up some how from Zephyr’s perspective?
Rather than ‘dummy-flash’ we need to handle the use case correctly.

From the source code itself, there is no difference. For now, I'm only testing the shell subsys sample application. I'm not storing anything.
The linker does all the necessary job to generate a working/bootable binary.

I tried to use the same memory region for both flash and sram in the dts but the zephyr binary doesn't boot. I compared the elf header between the two approaches and they are different, but since I'm not an expert in linker and program memory layout I haven't looked into it in detail. 


Regards,

--
*dS
Diego Sueiro

CEO do Embarcados
www.embarcados.com.br

/*long live rock 'n roll*/

On 9 March 2018 at 14:16, Kumar Gala <kumar.gala@...> wrote:

> On Mar 9, 2018, at 12:00 AM, Diego Sueiro <diego.sueiro@...> wrote:
>
> Just to give you an update on this.
>
> It works when using 2 different RAMs regions for flash and. But I still want to understand the usage of the "dummy-flash" and if I'm setting it correctly because I'm having configuration issues #6365  (for now I submitted the PR #6366 to fix this problem).
>
> There will be use cases where just one RAM will be available in the system for Zephyr and this RAM will have to accommodate both code and data. And to get this working I had to apply the mentioned PR and omit the "zephyr,flash" property from dts.
>

When you are using one RAM region for both memory & code, are you splitting the region up some how from Zephyr’s perspective?

Rather than ‘dummy-flash’ we need to handle the use case correctly.

- k

Join devel@lists.zephyrproject.org to automatically receive all group messages.