On Tue, 30 Jul 2024 at 11:02, Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On Tue, 30 Jul 2024 at 07:15, Arnd Bergmann <arnd@xxxxxxxxxx> wrote:
> >
> > There is another one that I see with gcc-8 randconfigs (arm64):
>
> So I ended up undoing that part of my patch, so it's a non-issue [..]
I pushed out my current one.
It keeps the old semantics wrt the clamp() static_assert, and it
obviously has the "allow small unsigned types to promote to 'int'"
that I already did earlier.
I still suspect we shouldn't do that relaxed integer promotion rule,
but it's what we used to do, and it's easy to get rid of if we decide
to, and it's a separate issue from the whole "make minmax expansion
more reasonable".
Linus
