Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
IronPort-SDR: /z9VP3Z9XWd9spGq7ieG/ZovY6IWirwWTutrfPzcmkTMvKo6WIxIanBfWYHtP/FpgNu8x0wPKq
 otft++t6h8RQ==
IronPort-SDR: kmEMWjItn+e/O1kjKTQTYR6gbQ84vIFG7T5GjuUqndEu/Lu2iK+OcbxwSWEhHjUyTAqaceVcKf
 LVXTgpkw9rPQ==
Subject: Re: [PATCH 4/6] mm/page_alloc: Scale the number of pages that are
 batch freed
To:     Mel Gorman <mgorman@techsingularity.net>,
        Linux-MM <linux-mm@kvack.org>
Cc:     Dave Hansen <dave.hansen@linux.intel.com>,
        Matthew Wilcox <willy@infradead.org>,
        Vlastimil Babka <vbabka@suse.cz>,
        Michal Hocko <mhocko@kernel.org>,
        Nicholas Piggin <npiggin@gmail.com>,
        LKML <linux-kernel@vger.kernel.org>
References: <20210521102826.28552-1-mgorman@techsingularity.net>
 <20210521102826.28552-5-mgorman@techsingularity.net>
From:   Dave Hansen <dave.hansen@intel.com>
Autocrypt: addr=dave.hansen@intel.com; keydata=
 xsFNBE6HMP0BEADIMA3XYkQfF3dwHlj58Yjsc4E5y5G67cfbt8dvaUq2fx1lR0K9h1bOI6fC
 oAiUXvGAOxPDsB/P6UEOISPpLl5IuYsSwAeZGkdQ5g6m1xq7AlDJQZddhr/1DC/nMVa/2BoY
 2UnKuZuSBu7lgOE193+7Uks3416N2hTkyKUSNkduyoZ9F5twiBhxPJwPtn/wnch6n5RsoXsb
 ygOEDxLEsSk/7eyFycjE+btUtAWZtx+HseyaGfqkZK0Z9bT1lsaHecmB203xShwCPT49Blxz
 VOab8668QpaEOdLGhtvrVYVK7x4skyT3nGWcgDCl5/Vp3TWA4K+IofwvXzX2ON/Mj7aQwf5W
 iC+3nWC7q0uxKwwsddJ0Nu+dpA/UORQWa1NiAftEoSpk5+nUUi0WE+5DRm0H+TXKBWMGNCFn
 c6+EKg5zQaa8KqymHcOrSXNPmzJuXvDQ8uj2J8XuzCZfK4uy1+YdIr0yyEMI7mdh4KX50LO1
 pmowEqDh7dLShTOif/7UtQYrzYq9cPnjU2ZW4qd5Qz2joSGTG9eCXLz5PRe5SqHxv6ljk8mb
 ApNuY7bOXO/A7T2j5RwXIlcmssqIjBcxsRRoIbpCwWWGjkYjzYCjgsNFL6rt4OL11OUF37wL
 QcTl7fbCGv53KfKPdYD5hcbguLKi/aCccJK18ZwNjFhqr4MliQARAQABzShEYXZpZCBDaHJp
 c3RvcGhlciBIYW5zZW4gPGRhdmVAc3I3MS5uZXQ+wsF7BBMBAgAlAhsDBgsJCAcDAgYVCAIJ
 CgsEFgIDAQIeAQIXgAUCTo3k0QIZAQAKCRBoNZUwcMmSsMO2D/421Xg8pimb9mPzM5N7khT0
 2MCnaGssU1T59YPE25kYdx2HntwdO0JA27Wn9xx5zYijOe6B21ufrvsyv42auCO85+oFJWfE
 K2R/IpLle09GDx5tcEmMAHX6KSxpHmGuJmUPibHVbfep2aCh9lKaDqQR07gXXWK5/yU1Dx0r
 VVFRaHTasp9fZ9AmY4K9/BSA3VkQ8v3OrxNty3OdsrmTTzO91YszpdbjjEFZK53zXy6tUD2d
 e1i0kBBS6NLAAsqEtneplz88T/v7MpLmpY30N9gQU3QyRC50jJ7LU9RazMjUQY1WohVsR56d
 ORqFxS8ChhyJs7BI34vQusYHDTp6PnZHUppb9WIzjeWlC7Jc8lSBDlEWodmqQQgp5+6AfhTD
 kDv1a+W5+ncq+Uo63WHRiCPuyt4di4/0zo28RVcjtzlGBZtmz2EIC3vUfmoZbO/Gn6EKbYAn
 rzz3iU/JWV8DwQ+sZSGu0HmvYMt6t5SmqWQo/hyHtA7uF5Wxtu1lCgolSQw4t49ZuOyOnQi5
 f8R3nE7lpVCSF1TT+h8kMvFPv3VG7KunyjHr3sEptYxQs4VRxqeirSuyBv1TyxT+LdTm6j4a
 mulOWf+YtFRAgIYyyN5YOepDEBv4LUM8Tz98lZiNMlFyRMNrsLV6Pv6SxhrMxbT6TNVS5D+6
 UorTLotDZKp5+M7BTQRUY85qARAAsgMW71BIXRgxjYNCYQ3Xs8k3TfAvQRbHccky50h99TUY
 sqdULbsb3KhmY29raw1bgmyM0a4DGS1YKN7qazCDsdQlxIJp9t2YYdBKXVRzPCCsfWe1dK/q
 66UVhRPP8EGZ4CmFYuPTxqGY+dGRInxCeap/xzbKdvmPm01Iw3YFjAE4PQ4hTMr/H76KoDbD
 cq62U50oKC83ca/PRRh2QqEqACvIH4BR7jueAZSPEDnzwxvVgzyeuhwqHY05QRK/wsKuhq7s
 UuYtmN92Fasbxbw2tbVLZfoidklikvZAmotg0dwcFTjSRGEg0Gr3p/xBzJWNavFZZ95Rj7Et
 db0lCt0HDSY5q4GMR+SrFbH+jzUY/ZqfGdZCBqo0cdPPp58krVgtIGR+ja2Mkva6ah94/oQN
 lnCOw3udS+Eb/aRcM6detZr7XOngvxsWolBrhwTQFT9D2NH6ryAuvKd6yyAFt3/e7r+HHtkU
 kOy27D7IpjngqP+b4EumELI/NxPgIqT69PQmo9IZaI/oRaKorYnDaZrMXViqDrFdD37XELwQ
 gmLoSm2VfbOYY7fap/AhPOgOYOSqg3/Nxcapv71yoBzRRxOc4FxmZ65mn+q3rEM27yRztBW9
 AnCKIc66T2i92HqXCw6AgoBJRjBkI3QnEkPgohQkZdAb8o9WGVKpfmZKbYBo4pEAEQEAAcLB
 XwQYAQIACQUCVGPOagIbDAAKCRBoNZUwcMmSsJeCEACCh7P/aaOLKWQxcnw47p4phIVR6pVL
 e4IEdR7Jf7ZL00s3vKSNT+nRqdl1ugJx9Ymsp8kXKMk9GSfmZpuMQB9c6io1qZc6nW/3TtvK
 pNGz7KPPtaDzvKA4S5tfrWPnDr7n15AU5vsIZvgMjU42gkbemkjJwP0B1RkifIK60yQqAAlT
 YZ14P0dIPdIPIlfEPiAWcg5BtLQU4Wg3cNQdpWrCJ1E3m/RIlXy/2Y3YOVVohfSy+4kvvYU3
 lXUdPb04UPw4VWwjcVZPg7cgR7Izion61bGHqVqURgSALt2yvHl7cr68NYoFkzbNsGsye9ft
 M9ozM23JSgMkRylPSXTeh5JIK9pz2+etco3AfLCKtaRVysjvpysukmWMTrx8QnI5Nn5MOlJj
 1Ov4/50JY9pXzgIDVSrgy6LYSMc4vKZ3QfCY7ipLRORyalFDF3j5AGCMRENJjHPD6O7bl3Xo
 4DzMID+8eucbXxKiNEbs21IqBZbbKdY1GkcEGTE7AnkA3Y6YB7I/j9mQ3hCgm5muJuhM/2Fr
 OPsw5tV/LmQ5GXH0JQ/TZXWygyRFyyI2FqNTx4WHqUn3yFj8rwTAU1tluRUYyeLy0ayUlKBH
 ybj0N71vWO936MqP6haFERzuPAIpxj2ezwu0xb1GjTk4ynna6h5GjnKgdfOWoRtoWndMZxbA
 z5cecg==
Message-ID: <8646d3ad-345f-7ec7-fe4a-ada2680487a3@intel.com>
Date:   Fri, 21 May 2021 15:36:05 -0700
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.10.0
MIME-Version: 1.0
In-Reply-To: <20210521102826.28552-5-mgorman@techsingularity.net>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk

...
> +static int nr_pcp_free(struct per_cpu_pages *pcp, int high, int batch)
> +{
> +	int min_nr_free, max_nr_free;
> +
> +	/* Check for PCP disabled or boot pageset */
> +	if (unlikely(high < batch))
> +		return 1;
> +
> +	min_nr_free = batch;
> +	max_nr_free = high - batch;

I puzzled over this for a minute.  I *think* it means to say: "Leave at
least one batch worth of pages in the pcp at all times so that the next
allocation can still be satisfied from this pcp."

> +	batch <<= pcp->free_factor;
> +	if (batch < max_nr_free)
> +		pcp->free_factor++;
> +	batch = clamp(batch, min_nr_free, max_nr_free);
> +
> +	return batch;
> +}
> +
>  static void free_unref_page_commit(struct page *page, unsigned long pfn,
>  				   int migratetype)
>  {
>  	struct zone *zone = page_zone(page);
>  	struct per_cpu_pages *pcp;
> +	int high;
>  
>  	__count_vm_event(PGFREE);
>  	pcp = this_cpu_ptr(zone->per_cpu_pageset);
>  	list_add(&page->lru, &pcp->lists[migratetype]);
>  	pcp->count++;
> -	if (pcp->count >= READ_ONCE(pcp->high))
> -		free_pcppages_bulk(zone, READ_ONCE(pcp->batch), pcp);
> +	high = READ_ONCE(pcp->high);
> +	if (pcp->count >= high) {
> +		int batch = READ_ONCE(pcp->batch);
> +
> +		free_pcppages_bulk(zone, nr_pcp_free(pcp, high, batch), pcp);
> +	}
>  }
>  
>  /*
> @@ -3531,6 +3555,7 @@ static struct page *rmqueue_pcplist(struct zone *preferred_zone,
>  
>  	local_lock_irqsave(&pagesets.lock, flags);
>  	pcp = this_cpu_ptr(zone->per_cpu_pageset);
> +	pcp->free_factor >>= 1;
>  	list = &pcp->lists[migratetype];
>  	page = __rmqueue_pcplist(zone,  migratetype, alloc_flags, pcp, list);
>  	local_unlock_irqrestore(&pagesets.lock, flags);

A high-level description of the algorithm in the changelog would also be
nice.  I *think* it's basically:

After hitting the high pcp mark, free one pcp->batch at a time.  But, as
subsequent pcp free operations occur, keep doubling the size of the
freed batches.  Cap them so that they always leave at least one
pcp->batch worth of pages.  Scale the size back down by half whenever an
allocation that consumes a page from the pcp occurs.

While I'd appreciate another comment or two, I do think this is worth
doing, and the approach seems sound:

Acked-by: Dave Hansen <dave.hansen@linux.intel.com>