Re: [PATCH] treewide: drop CONFIG_EMBEDDED

[Christophe Leroy]

Hi,

Le 19/08/2023 à 05:33, Jesse T a écrit :
>>>
>>>> Should there be a warning here to update change it instead of removal?
>>>
>>> kconfig doesn't have a warning mechanism AFAIK.
>>> Do you have an idea of how this would work?
> 
> No, unfortunately. As you said without a warning it would be overlooked so
> a change would not be necessary.
> 
> A possible solution is to check in a header file with:
> 
> #ifdef CONFIG_EMBEDDED
> #warning "CONFIG_EMBEDDED has changed to CONFIG_EXPERT"
> #endif
> 
> Does anyone else have an opinion on this?

My opinion is that has happen several times in the past and will happen 
again. It is not a big deal, whoever updates to a new kernel will make a 
savedefconfig and compare with previous defconfig and see what has 
changed. Once you see that CONFIG_EMBEDDED is disappearing you look at 
kernel history to find out why CONFIG_EMBEDDED disappears, and you 
understand from the commit message that you have to select CONFIG_EXPERT 
instead.

A couple examples I have in mind from the past:
- CONFIG_FORCE_MAX_ZONEORDER became CONFIG_ARCH_FORCE_MAX_ORDER
- CONFIG_MTD_NAND became CONFIG_MTD_RAW_NAND

> Since kconfig doesn't have a warning mechanism the patch seems fine as is.

So yes the patch is fine as is IMHO.

Christophe