Re: [PATCH v3 20/20] mm: stop maintaining the per-page mapcount of large folios (CONFIG_NO_PAGE_MAPCOUNT)

On Tue, Oct 14, 2025 at 02:59:30PM +0200, David Hildenbrand wrote:

On 14.10.25 14:23, Wei Yang wrote:

quoted

On Mon, Mar 03, 2025 at 05:30:13PM +0100, David Hildenbrand wrote:
[...]

quoted

@@ -1678,6 +1726,22 @@ static __always_inline void __folio_remove_rmap(struct folio *folio,

		break;
	case RMAP_LEVEL_PMD:
	case RMAP_LEVEL_PUD:
+		if (IS_ENABLED(CONFIG_NO_PAGE_MAPCOUNT)) {
+			last = atomic_add_negative(-1, &folio->_entire_mapcount);
+			if (level == RMAP_LEVEL_PMD && last)
+				nr_pmdmapped = folio_large_nr_pages(folio);
+			nr = folio_dec_return_large_mapcount(folio, vma);
+			if (!nr) {
+				/* Now completely unmapped. */
+				nr = folio_large_nr_pages(folio);
+			} else {
+				partially_mapped = last &&
+						   nr < folio_large_nr_pages(folio);

Hi, David

Hi!

quoted

Do you think this is better to be?

	partially_mapped = last && nr < nr_pmdmapped;

I see what you mean, it would be similar to the CONFIG_PAGE_MAPCOUNT case
below.

But probably it could then be

partially_mapped = nr < nr_pmdmapped;

because nr_pmdmapped is only set when "last = true".

I'm not sure if there is a good reason to change it at this point though.
Smells like a micro-optimization for PUD, which we probably shouldn't worry
about.

quoted

As commit 349994cf61e6 mentioned, we don't support partially mapped PUD-sized
folio yet.

We do support partially mapped PUD-sized folios I think, but not anonymous
PUD-sized folios.

So consequently the partially_mapped variable will never really be used later
on, because the folio_test_anon() will never hit in the PUD case.

Ok, folio_test_anon() takes care of it. We won't add it to defer list by
accident.

-- 
Cheers

David / dhildenb

-- 
Wei Yang
Help you, Help me