Re: [Xenomai-core] [PATCH 1/3] decouple spinlock stats from XENO_OPT_STATS

[Jan Kiszka <jan.kiszka@domain.hid>]

[-- Attachment #1: Type: text/plain, Size: 1147 bytes --]

Gilles Chanteperdrix wrote:
> Jan Kiszka wrote:
>>  #ifdef CONFIG_XENO_SPINLOCK_DEBUG
>> @@ -343,7 +340,7 @@ static inline spl_t __xnlock_get_irqsave
>>  #else /* !CONFIG_XENO_SPINLOCK_DEBUG */
>>  static inline spl_t __xnlock_get_irqsave (xnlock_t *lock)
>>  {
>> -#endif /* !CONFIG_XENO_SPINLOCK_DEBUG */
>> +#endif /* CONFIG_XENO_SPINLOCK_DEBUG */
> 
> I prefer the notation:
> 
> #if cond
> #endif /* cond */
> 
> and
> 
> #if cond
> #else /* !cond */
> #endif /* !cond */
> 
> to
> 
> #if cond
> #endif /* cond */
> 
> and
> 
> #if cond
> #else /* !cond */
> #endif /* cond */
> 
> Because in the first case, when I see a #endif in the middle of a lot of
> code, I immediately know what predicate is true above the #endif.
> 

Ok, I understand the logic. Nevertheless, I had the impression the first
scheme is used far more often than the second one in Xenomai and I-pipe,
so I took the chance to consolidate the code.

So, what should be the common rule from now on? Is there a kernel-style
rule for this issue we should follow, or are we free and should convert
to your preferred scheme?

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 250 bytes --]