Re: [PATCH] x86: support i386 with Clang

From: Linus Torvalds
Date: Mon May 11 2020 - 14:37:07 EST


On Mon, May 11, 2020 at 11:24 AM Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> I guess I should just test it, since I have that clang tree.

No, clang doesn't seem to handle it even with __builtin_choose_expr(),
and has that

invalid input size for constraint 'qi'

even when it's in a side that is never chosen.

Very annoying. A lot of our magic macros are literally about "pick one
case when the others are not valid for this type".

Linus