Re: [PATCH 1/7] cleanup: Introduce DEFINE_ACQUIRE() a CLASS() for conditional locking

[Peter Zijlstra <peterz@infradead.org>]

On Tue, May 13, 2025 at 12:46:29PM -0700, Linus Torvalds wrote:
> On Tue, 13 May 2025 at 01:50, Peter Zijlstra <peterz@infradead.org> wrote:
> >
> > +#define __GUARD_IS_ERR(_ptr) \
> > +       ({ unsigned long _var = (__force unsigned long)(_ptr); \
> > +          bool _s; \
> > +          asm_inline volatile ("cmp %[val], %[var]" \
> > +                               : "=@ccns" (_s) \
> > +                               : [val] "i" (-MAX_ERRNO), \
> > +                                 [var] "r" (_var)); \
> > +          unlikely(_s); })
> 
> I think that this might be acceptable if it was some actual common operation.
> 
> But for just the conditional guard test, I think it's cute, but I
> don't think it's really worth it.

Its actually every guard, the destructor is shared between unconditional
and conditional locks.

> Right now IS_ERR_OR_NULL() generates pretty disgusting code, with
> clang doing things like this:
> 
>         testq   %rdi, %rdi
>         sete    %al
>         cmpq    $-4095, %rdi                    # imm = 0xF001
>         setae   %cl
>         orb     %al, %cl
>         je      .LBB3_1

Whee, that is creative indeed :-)