Re: [PATCH v7 03/10] mm/lru: replace pgdat lru_lock with lruvec lock

[Daniel Jordan]

Hi Alex,

On Wed, Dec 25, 2019 at 05:04:19PM +0800, Alex Shi wrote:
> @@ -900,6 +904,29 @@ static inline struct lruvec *mem_cgroup_page_lruvec(struct page *page,
>  {
>  	return &pgdat->__lruvec;
>  }
> +#define lock_page_lruvec_irq(page)			\
> +({							\
> +	struct pglist_data *pgdat = page_pgdat(page);	\
> +	spin_lock_irq(&pgdat->__lruvec.lru_lock);	\
> +	&pgdat->__lruvec;				\
> +})
> +
> +#define lock_page_lruvec_irqsave(page, flagsp)			\
> +({								\
> +	struct pglist_data *pgdat = page_pgdat(page);		\
> +	spin_lock_irqsave(&pgdat->__lruvec.lru_lock, *flagsp);	\
> +	&pgdat->__lruvec;					\
> +})
> +
> +#define unlock_page_lruvec_irq(lruvec)			\
> +({							\
> +	spin_unlock_irq(&lruvec->lru_lock);		\
> +})
> +
> +#define unlock_page_lruvec_irqrestore(lruvec, flags)		\
> +({								\
> +	spin_unlock_irqrestore(&lruvec->lru_lock, flags);	\
> +})

Noticed this while testing your series.  These are safe as inline functions, so
I think you may have gotten the wrong impression when Johannes made this point:

https://lore.kernel.org/linux-mm/20191119164448.GA396644@xxxxxxxxxxx/