Received-SPF: pass (google.com: domain of linux-kernel+bounces-120577-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Message-ID: <b5b09acc-8bca-4e8c-82d4-6542fc7e9aec@suse.cz>
Date: Wed, 27 Mar 2024 09:54:01 +0100
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH 10/10] mm: page_alloc: consolidate free page accounting
Content-Language: en-US
To: Johannes Weiner <hannes@cmpxchg.org>,
 Andrew Morton <akpm@linux-foundation.org>
Cc: Mel Gorman <mgorman@techsingularity.net>, Zi Yan <ziy@nvidia.com>,
 "Huang, Ying" <ying.huang@intel.com>, David Hildenbrand <david@redhat.com>,
 linux-mm@kvack.org, linux-kernel@vger.kernel.org
References: <20240320180429.678181-1-hannes@cmpxchg.org>
 <20240320180429.678181-11-hannes@cmpxchg.org>
From: Vlastimil Babka <vbabka@suse.cz>
In-Reply-To: <20240320180429.678181-11-hannes@cmpxchg.org>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On 3/20/24 7:02 PM, Johannes Weiner wrote:
> Free page accounting currently happens a bit too high up the call
> stack, where it has to deal with guard pages, compaction capturing,
> block stealing and even page isolation. This is subtle and fragile,
> and makes it difficult to hack on the code.
> 
> Now that type violations on the freelists have been fixed, push the
> accounting down to where pages enter and leave the freelist.

Awesome!

> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>

Reviewed-by: Vlastimil Babka <vbabka@suse.cz>

Just some nits:

> @@ -1314,10 +1349,10 @@ static inline void expand(struct zone *zone, struct page *page,
>  		 * Corresponding page table entries will not be touched,
>  		 * pages will stay not present in virtual address space
>  		 */
> -		if (set_page_guard(zone, &page[size], high, migratetype))
> +		if (set_page_guard(zone, &page[size], high))
>  			continue;
>  
> -		add_to_free_list(&page[size], zone, high, migratetype);
> +		add_to_free_list(&page[size], zone, high, migratetype, false);

This is account_freepages() in the hot loop, what if we instead used
__add_to_free_list(), sum up nr_pages and called account_freepages() once
outside of the loop?

>  		set_buddy_order(&page[size], high);
>  	}
>  }

<snip>

> diff --git a/mm/page_isolation.c b/mm/page_isolation.c
> index 042937d5abe4..914a71c580d8 100644
> --- a/mm/page_isolation.c
> +++ b/mm/page_isolation.c
> @@ -252,7 +252,8 @@ static void unset_migratetype_isolate(struct page *page, int migratetype)
>  		 * Isolating this block already succeeded, so this
>  		 * should not fail on zone boundaries.
>  		 */
> -		WARN_ON_ONCE(!move_freepages_block_isolate(zone, page, migratetype));
> +		WARN_ON_ONCE(!move_freepages_block_isolate(zone, page,
> +							   migratetype));
>  	} else {
>  		set_pageblock_migratetype(page, migratetype);
>  		__putback_isolated_page(page, order, migratetype);

Looks like a drive-by edit of an extra file just to adjust identation.