Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Feedback-ID: ie3994620:Fastmail
Date:   Sat, 5 Nov 2022 22:51:15 +0300
From:   "Kirill A. Shutemov" <kirill@shutemov.name>
To:     Hugh Dickins <hughd@google.com>
Cc:     Andrew Morton <akpm@linux-foundation.org>,
        Matthew Wilcox <willy@infradead.org>,
        David Hildenbrand <david@redhat.com>,
        Vlastimil Babka <vbabka@suse.cz>, Peter Xu <peterx@redhat.com>,
        Yang Shi <shy828301@gmail.com>,
        John Hubbard <jhubbard@nvidia.com>,
        Mike Kravetz <mike.kravetz@oracle.com>,
        Sidhartha Kumar <sidhartha.kumar@oracle.com>,
        Muchun Song <songmuchun@bytedance.com>,
        Miaohe Lin <linmiaohe@huawei.com>,
        Naoya Horiguchi <naoya.horiguchi@linux.dev>,
        Mina Almasry <almasrymina@google.com>,
        James Houghton <jthoughton@google.com>,
        Zach O'Keefe <zokeefe@google.com>,
        linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: [PATCH 2/3] mm,thp,rmap: simplify compound page mapcount handling
Message-ID: <20221105195115.2d5yvvepdjsqjmmv@box>
References: <5f52de70-975-e94f-f141-543765736181@google.com>
 <47ad693-717-79c8-e1ba-46c3a6602e48@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <47ad693-717-79c8-e1ba-46c3a6602e48@google.com>
Precedence: bulk

On Wed, Nov 02, 2022 at 06:51:38PM -0700, Hugh Dickins wrote:
> Compound page (folio) mapcount calculations have been different for
> anon and file (or shmem) THPs, and involved the obscure PageDoubleMap
> flag.  And each huge mapping and unmapping of a file (or shmem) THP
> involved atomically incrementing and decrementing the mapcount of every
> subpage of that huge page, dirtying many struct page cachelines.
> 
> Add subpages_mapcount field to the struct folio and first tail page,
> so that the total of subpage mapcounts is available in one place near
> the head: then page_mapcount() and total_mapcount() and page_mapped(),
> and their folio equivalents, are so quick that anon and file and hugetlb
> don't need to be optimized differently. Delete the unloved PageDoubleMap.
> 
> page_add and page_remove rmap functions must now maintain the
> subpages_mapcount as well as the subpage _mapcount, when dealing with
> pte mappings of huge pages; and correct maintenance of NR_ANON_MAPPED
> and NR_FILE_MAPPED statistics still needs reading through the subpages,
> using nr_subpages_unmapped() - but only when first or last pmd mapping
> finds subpages_mapcount raised (double-map case, not the common case).
> 
> But are those counts (used to decide when to split an anon THP, and
> in vmscan's pagecache_reclaimable heuristic) correctly maintained?
> Not quite: since page_remove_rmap() (and also split_huge_pmd()) is
> often called without page lock, there can be races when a subpage pte
> mapcount 0<->1 while compound pmd mapcount 0<->1 is scanning - races
> which the previous implementation had prevented. The statistics might
> become inaccurate, and even drift down until they underflow through 0.
> That is not good enough, but is better dealt with in a followup patch.
> 
> Update a few comments on first and second tail page overlaid fields.
> hugepage_add_new_anon_rmap() has to "increment" compound_mapcount, but
> subpages_mapcount and compound_pincount are already correctly at 0,
> so delete its reinitialization of compound_pincount.
> 
> A simple 100 X munmap(mmap(2GB, MAP_SHARED|MAP_POPULATE, tmpfs), 2GB)
> took 18 seconds on small pages, and used to take 1 second on huge pages,
> but now takes 119 milliseconds on huge pages.  Mapping by pmds a second
> time used to take 860ms and now takes 92ms; mapping by pmds after mapping
> by ptes (when the scan is needed) used to take 870ms and now takes 495ms.
> But there might be some benchmarks which would show a slowdown, because
> tail struct pages now fall out of cache until final freeing checks them.
> 
> Signed-off-by: Hugh Dickins <hughd@google.com>

Thanks for doing this!

Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>

And sorry again for PageDoubleMap() :/

Minor nitpick and a question below.

> @@ -829,12 +829,20 @@ static inline int folio_entire_mapcount(struct folio *folio)
>  
>  /*
>   * Mapcount of compound page as a whole, does not include mapped sub-pages.
> - *
> - * Must be called only for compound pages.
> + * Must be called only on head of compound page.
>   */
> -static inline int compound_mapcount(struct page *page)
> +static inline int head_compound_mapcount(struct page *head)
>  {
> -	return folio_entire_mapcount(page_folio(page));
> +	return atomic_read(compound_mapcount_ptr(head)) + 1;
> +}
> +
> +/*
> + * Sum of mapcounts of sub-pages, does not include compound mapcount.
> + * Must be called only on head of compound page.
> + */
> +static inline int head_subpages_mapcount(struct page *head)
> +{
> +	return atomic_read(subpages_mapcount_ptr(head));
>  }
>  
>  /*

Any particular reason these two do not take struct folio as an input?
It would guarantee that it is non-tail page. It will not guarantee
large-folio, but it is something.

> @@ -1265,8 +1288,6 @@ void page_add_new_anon_rmap(struct page *page,
>  		VM_BUG_ON_PAGE(!PageTransHuge(page), page);
>  		/* increment count (starts at -1) */
>  		atomic_set(compound_mapcount_ptr(page), 0);
> -		atomic_set(compound_pincount_ptr(page), 0);
> -

It has to be initialized to 0 on allocation, right?

>  		__mod_lruvec_page_state(page, NR_ANON_THPS, nr);
>  	} else {
>  		/* increment count (starts at -1) */

-- 
  Kiryl Shutsemau / Kirill A. Shutemov