Re: [PATCH 01/43] sched: Clean up and standardize #if/#else/#endif markers in sched/autogroup.[ch]

[Sohil Mehta <sohil.mehta@intel.com>]

On 5/28/2025 1:08 AM, Ingo Molnar wrote:
>  - Use the standard #ifdef marker format for larger blocks,
>    where appropriate:
> 
>         #if CONFIG_FOO
>         ...
>         #else /* !CONFIG_FOO: */

I am trying to understand if this #else comment style is a standard in
the kernel? Any documentation link?

The comment next to the #else could relate to ending of the previous
section or related to beginning of the next section. I see 3 formats
commonly used:

1) #else /* CONFIG_FOO */

  $ grep -Pr '^(?=.*#else)(?=.*\/\*)(?=.*CONFIG_)(?!.*!)(?!.*:).*$' .

  Count: 1122

2) #else /* !CONFIG_FOO */

  $ grep -Pr '^(?=.*#else)(?=.*\/\*)(?=.*CONFIG_)(?=.*!)(?!.*:).*$' .

  Count: 642

3) #else /* !CONFIG_FOO: */

  $ grep -Pr '^(?=.*#else)(?=.*\/\*)(?=.*CONFIG_)(?=.*!)(?=.*:).*$'

  Count: 43

Even though the counts do not account for the #ifndef cases, they
provide a rough approximation of the kernel usage. #1 and #2 already
conflict with each other. I feel adding more to the 3rd type would
probably make it more confusing for readers and developers.

I am assuming that the ':' at the end is to signify that the comment
pertains to the beginning of the following section. If we want to
standardize #3 wouldn't it be better to first converge between #1 and #2?


>         ...
>         #endif /* !CONFIG_FOO */
>