linux-kernel - Re: [PATCH V3] MM: make vmpressure

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:	Mon, 12 May 2014 09:55:37 +0200
From:	Michal Hocko <mhocko@...e.cz>
To:	Fabian Frederick <fabf@...net.be>
Cc:	linux-kernel <linux-kernel@...r.kernel.org>,
	Tejun Heo <tj@...nel.org>, akpm <akpm@...ux-foundation.org>,
	Davidlohr Bueso <davidlohr@...com>,
	Li Zefan <lizefan@...wei.com>,
	cgroups <cgroups@...r.kernel.org>
Subject: Re: [PATCH V3] MM: make vmpressure_win dynamic

On Sat 10-05-14 08:52:17, Fabian Frederick wrote:
> This patch addresses TODO featuring in original version :

This is the 3rd version of the patch and I still fail to see the
motivation other than addressing a TODO which doesn't sound like a
sufficient justification to me. Why is the new scheme needed in the
first place? Why is calculate_normal_threshold a proper scheme?

> "Make the window size depend on machine size, as we do for vmstat"
> 
> It initializes vmpressure_win in vmstat using calculate_normal_threshold()
> based on each zone/cpu * SWAP_CLUSTER_SIZE instead of
> static SWAP_CLUSTER_MAX * 16 = 512
> 
> Some values :
> 
> CPU	MB	VMPRESSURE
> 8	512	1024
> 1	1024 	320
> 2	1024 	640
> 2	2048 	768
> 4	2048	1152
> 1	4096 	640
> 4	4096 	1920
> ...
> 
> (No maximum vmpressure value currently defined)
> 
> cf mm/vmstat.c: calculate_normal_threshold for further values
> 
> vmpressure_win is refreshed through cpu notifier and exposed read-only in
> /proc/sys/vm/vmpressure_win
> 
> Cc: Tejun Heo <tj@...nel.org>
> Cc: Andrew Morton <akpm@...ux-foundation.org>
> Cc: Michal Hocko <mhocko@...e.cz>
> Cc: Davidlohr Bueso <davidlohr@...com>
> Cc: Li Zefan <lizefan@...wei.com>
> Cc: cgroups@...r.kernel.org
> Signed-off-by: Fabian Frederick <fabf@...net.be>
> ---
> V3:
>   -Further description
>   -sysctl handler / ulong
>   -Cc list
> 
> V2: Suggestions by Andrew Morton
>   -Expose vmpressure_win
>   -Remove spinlock
> 
>  include/linux/swap.h | 4 ++++
>  kernel/sysctl.c      | 9 +++++++++
>  mm/vmpressure.c      | 4 +---
>  mm/vmstat.c          | 6 ++++++
>  4 files changed, 20 insertions(+), 3 deletions(-)
> 
> diff --git a/include/linux/swap.h b/include/linux/swap.h
> index 3507115..886897d 100644
> --- a/include/linux/swap.h
> +++ b/include/linux/swap.h
> @@ -354,6 +354,10 @@ extern int vm_swappiness;
>  extern int remove_mapping(struct address_space *mapping, struct page *page);
>  extern unsigned long vm_total_pages;
>  
> +#ifdef CONFIG_MEMCG
> +extern unsigned long vmpressure_win;
> +#endif
> +
>  #ifdef CONFIG_NUMA
>  extern int zone_reclaim_mode;
>  extern int sysctl_min_unmapped_ratio;
> diff --git a/kernel/sysctl.c b/kernel/sysctl.c
> index 74f5b58..c0c5ea6 100644
> --- a/kernel/sysctl.c
> +++ b/kernel/sysctl.c
> @@ -1202,6 +1202,15 @@ static struct ctl_table vm_table[] = {
>  		.mode           = 0444 /* read-only */,
>  		.proc_handler   = pdflush_proc_obsolete,
>  	},
> +#ifdef CONFIG_MEMCG
> +	{
> +		.procname	= "vmpressure_window",
> +		.data		= &vmpressure_win,
> +		.maxlen		= sizeof(vmpressure_win),
> +		.mode		= 0444,
> +		.proc_handler	= proc_doulongvec_minmax,
> +	},
> +#endif
>  	{
>  		.procname	= "swappiness",
>  		.data		= &vm_swappiness,
> diff --git a/mm/vmpressure.c b/mm/vmpressure.c
> index d4042e7..d738670 100644
> --- a/mm/vmpressure.c
> +++ b/mm/vmpressure.c
> @@ -35,10 +35,8 @@
>   * As the vmscan reclaimer logic works with chunks which are multiple of
>   * SWAP_CLUSTER_MAX, it makes sense to use it for the window size as well.
>   *
> - * TODO: Make the window size depend on machine size, as we do for vmstat
> - * thresholds. Currently we set it to 512 pages (2MB for 4KB pages).
>   */
> -static const unsigned long vmpressure_win = SWAP_CLUSTER_MAX * 16;
> +unsigned long vmpressure_win = SWAP_CLUSTER_MAX * 16;
>  
>  /*
>   * These thresholds are used when we account memory pressure through
> diff --git a/mm/vmstat.c b/mm/vmstat.c
> index 302dd07..ede68fd 100644
> --- a/mm/vmstat.c
> +++ b/mm/vmstat.c
> @@ -20,6 +20,7 @@
>  #include <linux/writeback.h>
>  #include <linux/compaction.h>
>  #include <linux/mm_inline.h>
> +#include <linux/swap.h>
>  
>  #include "internal.h"
>  
> @@ -163,11 +164,13 @@ void refresh_zone_stat_thresholds(void)
>  	struct zone *zone;
>  	int cpu;
>  	int threshold;
> +	unsigned long new_vmpressure_win = 0;
>  
>  	for_each_populated_zone(zone) {
>  		unsigned long max_drift, tolerate_drift;
>  
>  		threshold = calculate_normal_threshold(zone);
> +		new_vmpressure_win += threshold;
>  
>  		for_each_online_cpu(cpu)
>  			per_cpu_ptr(zone->pageset, cpu)->stat_threshold
> @@ -184,6 +187,9 @@ void refresh_zone_stat_thresholds(void)
>  			zone->percpu_drift_mark = high_wmark_pages(zone) +
>  					max_drift;
>  	}
> +#ifdef CONFIG_MEMCG
> +	vmpressure_win = new_vmpressure_win * SWAP_CLUSTER_MAX;
> +#endif
>  }
>  
>  void set_pgdat_percpu_threshold(pg_data_t *pgdat,
> -- 
> 1.8.4.5

-- 
Michal Hocko
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/