# Re: \[PATCH v2 6/7\] cmdline: Gives architectures opportunity to use generically defined boot cmdline manipulation

- From: Christophe Leroy \<hidden\>
- Date: 2021-03-26 06:19:45 
- List: linuxppc-dev
- Page: https://ratatoskr.run/linuxppc-dev/2021/03/6316196
- Thread: https://ratatoskr.run/linuxppc-dev/2021/03/6285815/t.md

```text


Le 25/03/2021 à 20:32, Will Deacon a écrit :
> On Thu, Mar 25, 2021 at 12:18:38PM +0100, Christophe Leroy wrote:
>>
>> - For ARM it means to append the bootloader arguments to the CONFIG_CMDLINE
>> - For Powerpc it means to append the CONFIG_CMDLINE to the bootloader arguments
>> - For SH  it means to append the CONFIG_CMDLINE to the bootloader arguments
>> - For EFI it means to append the bootloader arguments to the CONFIG_CMDLINE
>> - For OF it means to append the CONFIG_CMDLINE to the bootloader arguments
>>
>> So what happens on ARM for instance when it selects CONFIG_OF for instance ?
>
> I think ARM gets different behaviour depending on whether it uses ATAGs or
> FDT.

As far as I can see, ARM uses either ATAGs only or both ATAGs and FDT. ATAGs is forced to 'y' when USE_OF is set. Do I miss something ?

>
>> Or should we consider that EXTEND means APPEND or PREPEND, no matter which ?
>> Because EXTEND is for instance used for:
>>
>> 	config INITRAMFS_FORCE
>> 		bool "Ignore the initramfs passed by the bootloader"
>> 		depends on CMDLINE_EXTEND || CMDLINE_FORCE
>
> Oh man, I didn't spot that one :(
>
> I think I would make the generic options explicit: either APPEND or PREPEND.
> Then architectures which choose to define CMDLINE_EXTEND in their Kconfigs
> can select the generic option that matches their behaviour.
>
> INITRAMFS_FORCE sounds like it should depend on APPEND (assuming that means
> CONFIG_CMDLINE is appended to the bootloader arguments).
>


Christophe
```