mm-huge_memory-initialise-workingset-state-before-folio-split.patch removed from -mm tree
The quilt patch titled
Subject: mm/huge_memory: initialise workingset state before folio split
has been removed from the -mm tree. Its filename was
mm-huge_memory-initialise-workingset-state-before-folio-split.patch
This patch was dropped because it was merged into the mm-hotfixes-stable branch of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
------------------------------------------------------
From: Matt Fleming <[email protected]>
Subject: mm/huge_memory: initialise workingset state before folio split
Date: Sat, 25 Jul 2026 11:14:19 +0100
xas_try_split() adds __GFP_ACCOUNT for page-cache xa_nodes, but __folio_split() leaves the xa_state's xa_lru unset. That lets a live, memcg-charged xa_node exist without being linked into the mapping's shadow_nodes list_lru; when reclaim later walks the list_lru it trips VM_WARN_ON(!css_is_dying()).
Use mapping_set_update() to install both the workingset update callback and the shadow_nodes list_lru on the xa_state.
---
mm/huge_memory.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
mm/huge_memory.c renamed from mm/huge_memory.c~mm-huge_memory-initialise-workingset-state-before-folio-split +3 -1
@@ -4076,7 +4076,7 @@ static int __folio_split(struct folio *fgfp_t gfp; mapping = folio->mapping; - min_order =mapping_min_folio_order (folio->mapping); + min_order =mapping_min_folio_order (mapping);if (new_order < min_order) { ret = -EINVAL ; goto out;
@@ -4090,6 +4090,8 @@ static int __folio_split(struct folio *f goto out; } +mapping_set_update (&xas, mapping); +if (split_type ==SPLIT_TYPE_UNIFORM ) {xas_set_order (&xas, folio->index, new_order);xas_split_alloc (&xas, folio, old_order, gfp);
_
Patches currently in -mm which might be from [email protected] are