Re: [PATCH -mm -v2] mm, swap: Use page-cluster as max window of VMA based swap readahead

[Minchan Kim]

On Wed, Oct 11, 2017 at 03:08:47PM +0800, Huang, Ying wrote:
> From: Huang Ying <ying.huang@xxxxxxxxx>
> 
> When the VMA based swap readahead was introduced, a new knob
> 
>   /sys/kernel/mm/swap/vma_ra_max_order
> 
> was added as the max window of VMA swap readahead.  This is to make it
> possible to use different max window for VMA based readahead and
> original physical readahead.  But Minchan Kim pointed out that this
> will cause a regression because setting page-cluster sysctl to zero
> cannot disable swap readahead with the change.
> 
> To fix the regression, the page-cluster sysctl is used as the max
> window of both the VMA based swap readahead and original physical swap
> readahead.  If more fine grained control is needed in the future, more
> knobs can be added as the subordinate knobs of the page-cluster
> sysctl.
> 
> The vma_ra_max_order knob is deleted.  Because the knob was
> introduced in v4.14-rc1, and this patch is targeting being merged
> before v4.14 releasing, there should be no existing users of this
> newly added ABI.
> 
> Cc: Johannes Weiner <hannes@xxxxxxxxxxx>
> Cc: Rik van Riel <riel@xxxxxxxxxx>
> Cc: Shaohua Li <shli@xxxxxxxxxx>
> Cc: Hugh Dickins <hughd@xxxxxxxxxx>
> Cc: Fengguang Wu <fengguang.wu@xxxxxxxxx>
> Cc: Tim Chen <tim.c.chen@xxxxxxxxx>
> Cc: Dave Hansen <dave.hansen@xxxxxxxxx>
> Reported-by: Minchan Kim <minchan@xxxxxxxxxx>
> Signed-off-by: "Huang, Ying" <ying.huang@xxxxxxxxx>

Acked-by: Minchan Kim <minchan@xxxxxxxxxx>

Below a question:

> ---
>  Documentation/ABI/testing/sysfs-kernel-mm-swap | 10 -------
>  mm/swap_state.c                                | 41 +++++---------------------
>  2 files changed, 7 insertions(+), 44 deletions(-)
> 
> diff --git a/Documentation/ABI/testing/sysfs-kernel-mm-swap b/Documentation/ABI/testing/sysfs-kernel-mm-swap
> index 587db52084c7..94672016c268 100644
> --- a/Documentation/ABI/testing/sysfs-kernel-mm-swap
> +++ b/Documentation/ABI/testing/sysfs-kernel-mm-swap
> @@ -14,13 +14,3 @@ Description:	Enable/disable VMA based swap readahead.
>  		still used for tmpfs etc. other users.  If set to
>  		false, the global swap readahead algorithm will be
>  		used for all swappable pages.
> -
> -What:		/sys/kernel/mm/swap/vma_ra_max_order
> -Date:		August 2017
> -Contact:	Linux memory management mailing list <linux-mm@xxxxxxxxx>
> -Description:	The max readahead size in order for VMA based swap readahead
> -
> -		VMA based swap readahead algorithm will readahead at
> -		most 1 << max_order pages for each readahead.  The
> -		real readahead size for each readahead will be scaled
> -		according to the estimation algorithm.
> diff --git a/mm/swap_state.c b/mm/swap_state.c
> index ed91091d1e68..05b6803f0cce 100644
> --- a/mm/swap_state.c
> +++ b/mm/swap_state.c
> @@ -39,10 +39,6 @@ struct address_space *swapper_spaces[MAX_SWAPFILES];
>  static unsigned int nr_swapper_spaces[MAX_SWAPFILES];
>  bool swap_vma_readahead = true;
>  
> -#define SWAP_RA_MAX_ORDER_DEFAULT	3
> -
> -static int swap_ra_max_order = SWAP_RA_MAX_ORDER_DEFAULT;
> -
>  #define SWAP_RA_WIN_SHIFT	(PAGE_SHIFT / 2)
>  #define SWAP_RA_HITS_MASK	((1UL << SWAP_RA_WIN_SHIFT) - 1)
>  #define SWAP_RA_HITS_MAX	SWAP_RA_HITS_MASK
> @@ -664,6 +660,13 @@ struct page *swap_readahead_detect(struct vm_fault *vmf,
>  	pte_t *tpte;
>  #endif
>  
> +	max_win = 1 << min_t(unsigned int, READ_ONCE(page_cluster),
> +			     SWAP_RA_ORDER_CEILING);

Why do we need READ_ONCE in here? IOW, without it, what happens?

> +	if (max_win == 1) {
> +		swap_ra->win = 1;
> +		return NULL;
> +	}