Re: [PATCH v2] scripts: coccinelle: check for !(un)?likely usage

[Denis Efremov]

Hi,

On 06.09.2019 23:19, Julia Lawall wrote:
> 
> 
> On Thu, 29 Aug 2019, Denis Efremov wrote:
> 
>> This patch adds coccinelle script for detecting !likely and
>> !unlikely usage. These notations are confusing. It's better
>> to replace !likely(x) with unlikely(!x) and !unlikely(x) with
>> likely(!x) for readability.
>>
>> The rule transforms !likely(x) to unlikely(!x) based on this logic:
>>   !likely(x) iff
>>   !__builtin_expect(!!(x), 1) iff
>>    __builtin_expect(!!!(x), 0) iff
>>   unlikely(!x)
>>
>> For !unlikely(x) to likely(!x):
>>   !unlikely(x) iff
>>   !__builtin_expect(!!(x), 0) iff
>>   __builtin_expect(!!!(x), 1) iff
>>   likely(!x)
>>
>> Signed-off-by: Denis Efremov <efremov@xxxxxxxxx>
>> Cc: Julia Lawall <Julia.Lawall@xxxxxxx>
>> Cc: Gilles Muller <Gilles.Muller@xxxxxxx>
>> Cc: Nicolas Palix <nicolas.palix@xxxxxxx>
>> Cc: Michal Marek <michal.lkml@xxxxxxxxxxx>
>> Cc: Markus Elfring <Markus.Elfring@xxxxxx>
>> Cc: Joe Perches <joe@xxxxxxxxxxx>
>> Cc: Rasmus Villemoes <linux@xxxxxxxxxxxxxxxxxx>
> 
> Acked-by: Julia Lawall <julia.lawall@xxxxxxx>
> 
> A small improvement though would be to improve the explicit dependency of
> the last four python rules on r1 and r2.  Those rules won't execute unless
> the inherited metavariable has a value, which makes the same dependency.
> 
> julia

I think I will resend this patch as a part of patchset with all warnings fixed
in a couple of days. Hope this will help to create a discussion point with other
developers about readability of "!likely" and "!unlikely".

Thanks,
Denis