> arm64: support WFET in smp_cond_load_relaxed_timeout()

TL;DR: Including asm/delay-const.h from asm/cmpxchg.h may break the arm64
build of drivers/gpio/gpio-aspeed.c, whose usecs_to_cycles() would conflict
with the new one; it appears unfixed by the series.

> diff --git a/arch/arm64/include/asm/cmpxchg.h b/arch/arm64/include/asm/cmpxchg.h
> index 6cf3cd6873f5..30dca19c9a7b 100644
> --- a/arch/arm64/include/asm/cmpxchg.h
> +++ b/arch/arm64/include/asm/cmpxchg.h
> @@ -12,6 +12,7 @@
>
>  #include <asm/barrier.h>
>  #include <asm/lse.h>
> +#include <asm/delay-const.h>

Can this include break the build of drivers/gpio/gpio-aspeed.c?

asm/cmpxchg.h is pulled into almost every arm64 translation unit through
linux/spinlock.h -> linux/atomic.h -> asm/atomic.h -> asm/cmpxchg.h, so the
static inlines in asm/delay-const.h become visible everywhere.

gpio-aspeed.c already defines a helper with the same name and a different
signature:

drivers/gpio/gpio-aspeed.c:
    static int usecs_to_cycles(struct aspeed_gpio *gpio, unsigned long usecs,
                               u32 *cycles)

while asm/delay-const.h now provides:

arch/arm64/include/asm/delay-const.h:
    static inline u64 usecs_to_cycles(u64 time_usecs)

That looks like it would fail with "conflicting types for 'usecs_to_cycles'".
GPIO_ASPEED depends on ARCH_ASPEED || COMPILE_TEST, and ARCH_ASPEED is an
arm64 platform, so arm64 allmodconfig/allyesconfig and any config enabling
CONFIG_GPIO_ASPEED would stop building at this commit.

The previous patch ("arm64/delay: move, fixup usecs_to_cycles()") renamed the
rpmh-rsc helpers for the same name clash, but gpio-aspeed.c was not touched.
No commit in the rest of the series modifies gpio-aspeed.c or delay-const.h,
so the breakage appears to remain at the end of the series.

Could the gpio-aspeed helper be renamed, could the delay-const.h helpers get
an arch-specific prefix, or could the cycle computation be moved out of line
so that asm/cmpxchg.h doesn't need to include asm/delay-const.h?


---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

CI run summary: https://github.com/kernel-patches/bpf/actions/runs/38867477666