Re: [PATCH v1] cleanup: adjust scoped_guard() to avoid potential warning

[Dmitry Torokhov]

Hi Przemek,

On Thu, Oct 03, 2024 at 01:39:06PM +0200, Przemek Kitszel wrote:
> @@ -167,14 +172,25 @@ static inline class_##_name##_t class_##_name##ext##_constructor(_init_args) \
>  	CLASS(_name, __UNIQUE_ID(guard))
>  
>  #define __guard_ptr(_name) class_##_name##_lock_ptr
> +#define __is_cond_ptr(_name) class_##_name##_is_conditional
> +
> +#define __scoped_guard_labeled(_label, _name, args...)			\
> +	for (CLASS(_name, scope)(args);					\
> +	     __guard_ptr(_name)(&scope) || !__is_cond_ptr(_name);	\

It would be great if you added the comment that "!__is_cond_ptr(_name)"
condition ensures that the compiler does not believe that it is possible
to skip the loop body because it does not realize that
"__guard_ptr(_name)(&scope)" will never return 0 for unconditional
locks. You have the explanation in the patch description, but I think it
is worth to reiterate here as well.

> +		     ({ goto _label; }))				\
> +		if (0)							\
> +		_label:							\
> +			break;						\
> +		else
> +

Reviewed-by: Dmitry Torokhov <dmitry.torokhov@xxxxxxxxx>

Thanks.

-- 
Dmitry