Re: [PATCH v3] locking/local_lock, mm: Replace localtry_ helpers with local_trylock_t type

[Alexei Starovoitov]

On Wed, Apr 2, 2025 at 8:12 PM Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> On Wed,  2 Apr 2025 19:55:14 -0700 Alexei Starovoitov <alexei.starovoitov@xxxxxxxxx> wrote:
>
> > Partially revert commit 0aaddfb06882 ("locking/local_lock: Introduce localtry_lock_t").
> > Remove localtry_*() helpers, since localtry_lock() name might
> > be misinterpreted as "try lock".
> >
>
> So many macros grumble.
>
> +#define local_trylock_init(lock)       __local_trylock_init(lock)
> +#define local_trylock(lock)            __local_trylock(lock)
> +#define local_trylock_irqsave(lock, flags)                     \
> +#define __local_trylock_init(lock) __local_lock_init(lock.llock)
> +#define __local_lock_acquire(lock)                                     \
> +#define __local_trylock(lock)                                  \
> +#define __local_trylock_irqsave(lock, flags)                   \
> +#define __local_lock_release(lock)                                     \
> +#define __local_unlock(lock)                                   \
> +#define __local_unlock_irq(lock)                               \
> +#define __local_unlock_irqrestore(lock, flags)                 \
> +#define __local_lock_nested_bh(lock)                           \
> +#define __local_unlock_nested_bh(lock)                         \
> +#define __local_trylock_init(l)                        __local_lock_init(l)
> +#define __local_trylock(lock)                                  \
> +#define __local_trylock_irqsave(lock, flags)                   \
>
> I expect many of these could have been implemented as static inlines.
>
> Oh well, that's a separate project for someone sometime.

They need to be macroses otherwise _Generic() trick won't work.

Thanks for applying v3.

Do you want to take "mm/page_alloc: Avoid second trylock of zone->lock"
fix as well ?

https://lore.kernel.org/all/20250331002809.94758-1-alexei.starovoitov@xxxxxxxxx/

or should I ?