kpageflags: detect isolated KPF_THP folios

Commit Message

ran xiaokai July 5, 2024, 10:43 a.m. UTC

From: Ran Xiaokai <[e-mail]>

When folio is isolated, the PG_lru bit is cleared. So the PG_lru
check in stable_page_flags() will miss this kind of isolated folios.
Use folio_test_large_rmappable() instead to also include isolated folios.

Since pagecache supports large folios and the introduction of mTHP,
the semantics of KPF_THP have been expanded, now it indicates
not only PMD-sized THP. Update related documentation to clearly state
that KPF_THP indicates multiple order THPs.

v1:
  https://lore.kernel.org/lkml/[e-mail]/
Signed-off-by: Ran Xiaokai <[e-mail]>
---
 Documentation/admin-guide/mm/pagemap.rst |  4 ++--
 fs/proc/page.c                           | 21 +++++++++------------
 2 files changed, 11 insertions(+), 14 deletions(-)

Comments

David Hildenbrand July 5, 2024, 2:03 p.m. UTC | #1

> -	} else if (is_zero_pfn(page_to_pfn(page)))
> +	else if (folio_test_large(folio) &&
> +	         folio_test_large_rmappable(folio)) {
> +		/* Note: we indicate any THPs here, not just PMD-sized ones */
> +		u |= 1 << KPF_THP;
> +	} else if (is_huge_zero_folio(folio)) {
>   		u |= 1 << KPF_ZERO_PAGE;
> +		u |= 1 << KPF_THP;
> +	} else if (is_zero_pfn(page_to_pfn(page))) {

We should also directly switch to "is_zero_folio(folio)" here

> +		u |= 1 << KPF_ZERO_PAGE;
> +	}
>   
>   	/*
>   	 * Caveats on high order pages: PG_buddy and PG_slab will only be set

Especially relevant in context of:

https://lkml.kernel.org/r/[e-mail]

Acked-by: David Hildenbrand <[e-mail]>

Patch

diff --git a/Documentation/admin-guide/mm/pagemap.rst b/Documentation/admin-guide/mm/pagemap.rst
index f5f065c67615..0a8a4decdb72 100644
--- a/Documentation/admin-guide/mm/pagemap.rst
+++ b/Documentation/admin-guide/mm/pagemap.rst
@@ -118,7 +118,7 @@  Short descriptions to the page flags
 21 - KSM
     Identical memory pages dynamically shared between one or more processes.
 22 - THP
-    Contiguous pages which construct transparent hugepages.
+    Contiguous pages which construct THP of any size and mapped by any granularity.
 23 - OFFLINE
     The page is logically offline.
 24 - ZERO_PAGE
@@ -252,7 +252,7 @@  Following flags about pages are currently supported:
 - ``PAGE_IS_PRESENT`` - Page is present in the memory
 - ``PAGE_IS_SWAPPED`` - Page is in swapped
 - ``PAGE_IS_PFNZERO`` - Page has zero PFN
-- ``PAGE_IS_HUGE`` - Page is THP or Hugetlb backed
+- ``PAGE_IS_HUGE`` - Page is PMD-mapped THP or Hugetlb backed
 - ``PAGE_IS_SOFT_DIRTY`` - Page is soft-dirty
 
 The ``struct pm_scan_arg`` is used as the argument of the IOCTL.
diff --git a/fs/proc/page.c b/fs/proc/page.c
index 2fb64bdb64eb..76f2a412aa93 100644
--- a/fs/proc/page.c
+++ b/fs/proc/page.c
@@ -148,19 +148,16 @@  u64 stable_page_flags(const struct page *page)
 		u |= 1 << KPF_COMPOUND_TAIL;
 	if (folio_test_hugetlb(folio))
 		u |= 1 << KPF_HUGE;
-	/*
-	 * We need to check PageLRU/PageAnon
-	 * to make sure a given page is a thp, not a non-huge compound page.
-	 */
-	else if (folio_test_large(folio)) {
-		if ((k & (1 << PG_lru)) || is_anon)
-			u |= 1 << KPF_THP;
-		else if (is_huge_zero_folio(folio)) {
-			u |= 1 << KPF_ZERO_PAGE;
-			u |= 1 << KPF_THP;
-		}
-	} else if (is_zero_pfn(page_to_pfn(page)))
+	else if (folio_test_large(folio) &&
+	         folio_test_large_rmappable(folio)) {
+		/* Note: we indicate any THPs here, not just PMD-sized ones */
+		u |= 1 << KPF_THP;
+	} else if (is_huge_zero_folio(folio)) {
 		u |= 1 << KPF_ZERO_PAGE;
+		u |= 1 << KPF_THP;
+	} else if (is_zero_pfn(page_to_pfn(page))) {
+		u |= 1 << KPF_ZERO_PAGE;
+	}
 
 	/*
 	 * Caveats on high order pages: PG_buddy and PG_slab will only be set