linux-kernel - Re: [PATCH] mm/min_free_kbytes: modify min_free

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: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 19 Jun 2023 11:25:38 +0800
From:   "Huang, Ying" <ying.huang@...el.com>
To:     liuq <liuq131@...natelecom.cn>
Cc:     akpm@...ux-foundation.org, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org, Vlastimil Babka <vbabka@...e.cz>,
        Mel Gorman <mgorman@...hsingularity.net>
Subject: Re: [PATCH] mm/min_free_kbytes: modify min_free_kbytes calculation
 rules


Added Mel and Vlastimil.

liuq <liuq131@...natelecom.cn> writes:

> The current calculation of min_free_kbytes only uses ZONE_DMA and
> ZONE_NORMAL pages,but the ZONE_MOVABLE zone->_watermark[WMARK_MIN]
> will also divide part of min_free_kbytes.This will cause the min
> watermark of ZONE_NORMAL to be too small in the presence of ZONE_MOVEABLE.

This seems like a real problem per my understanding.  Can you show the
contents of /proc/zoneinfo on a problem system?

But, per my understanding, min_free_kbytes are used for __GFP_HIGH and
PF_MEMALLOC allocations.  While ZONE_MOVABLE will not be used for them
usually.  So I think we should treat ZONE_MOVABLE as ZONE_HIGHMEM in
__setup_per_zone_wmarks().

Best Regards,
Huang, Ying

> Signed-off-by: liuq <liuq131@...natelecom.cn>
> ---
>  include/linux/mm.h |  1 +
>  mm/khugepaged.c    |  2 +-
>  mm/page_alloc.c    | 15 ++++++++++++++-
>  3 files changed, 16 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index cf3d0d673f6b..1f91d035bcaf 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -863,6 +863,7 @@ void split_page(struct page *page, unsigned int order);
>  void folio_copy(struct folio *dst, struct folio *src);
>  
>  unsigned long nr_free_buffer_pages(void);
> +unsigned long nr_free_pagecache_pages(void);
>  
>  /*
>   * Compound pages have a destructor function.  Provide a
> diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> index 16be62d493cd..6632264b951c 100644
> --- a/mm/khugepaged.c
> +++ b/mm/khugepaged.c
> @@ -2342,7 +2342,7 @@ static void set_recommended_min_free_kbytes(void)
>  
>  	/* don't ever allow to reserve more than 5% of the lowmem */
>  	recommended_min = min(recommended_min,
> -			      (unsigned long) nr_free_buffer_pages() / 20);
> +			      (unsigned long) nr_free_pagecache_pages() / 20);
>  	recommended_min <<= (PAGE_SHIFT-10);
>  
>  	if (recommended_min > min_free_kbytes) {
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index e008a3df0485..489b564526dd 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -5775,6 +5775,19 @@ unsigned long nr_free_buffer_pages(void)
>  }
>  EXPORT_SYMBOL_GPL(nr_free_buffer_pages);
>  
> +/**
> + * nr_free_pagecache_pages - count number of pages beyond high watermark
> + *
> + * nr_free_pagecache_pages() counts the number of pages which are beyond the
> + * high watermark within all zones.
> + *
> + * Return: number of pages beyond high watermark within all zones.
> + */
> +unsigned long nr_free_pagecache_pages(void)
> +{
> +	return nr_free_zone_pages(gfp_zone(GFP_HIGHUSER_MOVABLE));
> +}
> +
>  static inline void show_node(struct zone *zone)
>  {
>  	if (IS_ENABLED(CONFIG_NUMA))
> @@ -8651,7 +8664,7 @@ void calculate_min_free_kbytes(void)
>  	unsigned long lowmem_kbytes;
>  	int new_min_free_kbytes;
>  
> -	lowmem_kbytes = nr_free_buffer_pages() * (PAGE_SIZE >> 10);
> +	lowmem_kbytes = nr_free_pagecache_pages() * (PAGE_SIZE >> 10);
>  	new_min_free_kbytes = int_sqrt(lowmem_kbytes * 16);
>  
>  	if (new_min_free_kbytes > user_min_free_kbytes)