[PATCH 1/4] locking/atomic/x86: Silence intentional wrapping addition

Kees Cook keescook at chromium.org
Thu Apr 25 10:19:24 PDT 2024


On Thu, Apr 25, 2024 at 11:15:17AM +0100, Mark Rutland wrote:
> To be clear, I dislike the function annotation because then it applies to
> *everything* within the function, which is overly broad and the intent becomes
> unclear. That makes it painful to refactor the code (since e.g. if we want to
> add another operation to the function which *should not* wrap, that gets
> silenced too).

Yeah, I find that a convincing argument for larger functions, but it
seemed to me that for these 1-line implementations it was okay. But
regardless, yup, no function-level annotation here.

> I'm happy with something that applies to specific types/variables or specific
> operations (which is what these patches do).

Thanks!

-- 
Kees Cook



More information about the linux-arm-kernel mailing list