Re: [PATCH v2 09/11] mm/memcg, swap: store cgroup id in cluster table directly

[YoungJun Park]

> +		if (IS_ENABLED(CONFIG_MEMCG) && !memcg_table) {
> +			swap_table_free(table);
> +			return -ENOMEM;
> +		}

Hi Kairui,

Nit: 
(Just a readability nit. purely my preference, feel free to ignore.)
the checks around swap_memcg_table_alloc() reduce to two
equivalent forms of the same memcg success/failure question:

(!IS_ENABLED(CONFIG_MEMCG) || memcg_table)   /* success */
(IS_ENABLED(CONFIG_MEMCG) && !memcg_table)   /* failure */

A macro for the failure side would let the call sites read as plain
positive/negative:

#define SWAP_MEMCG_TABLE_ALLOC_FAILED(t) \
	(IS_ENABLED(CONFIG_MEMCG) && !(t))

SWAP_MEMCG_TABLE_ALLOC_FAILED(memcg_table)     /* failure */
!SWAP_MEMCG_TABLE_ALLOC_FAILED(memcg_table)    /* success */

Equivalently, the same macro can be expressed by splitting on
CONFIG_MEMCG.

#ifdef CONFIG_MEMCG
#define SWAP_MEMCG_TABLE_ALLOC_FAILED(t)  (!(t))
#else
#define SWAP_MEMCG_TABLE_ALLOC_FAILED(t)  (0)
#endif

What do you think?

Thanks,
Youngjun Park