From: Kairui Song The comment makes it look like copy-paste leftovers from shmem_replace_folio. The first try of the swap doesn't always have a limited zone. So don't drop the restraint, which should make the GFP more accurate. Signed-off-by: Kairui Song --- mm/shmem.c | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-) diff --git a/mm/shmem.c b/mm/shmem.c index eaaeca8f6c39..4f34dfa95a50 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -2043,14 +2043,8 @@ static struct folio *shmem_swap_alloc_folio(struct inode *inode, struct shmem_inode_info *info = SHMEM_I(inode); struct folio *new, *swapcache; int nr_pages = 1 << order; - gfp_t alloc_gfp; + gfp_t alloc_gfp = gfp; - /* - * We have arrived here because our zones are constrained, so don't - * limit chance of success with further cpuset and node constraints. - */ - gfp &= ~GFP_CONSTRAINT_MASK; - alloc_gfp = gfp; if (!IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE)) { if (WARN_ON_ONCE(order)) return ERR_PTR(-EINVAL); --- base-commit: ad92386d5debbde551129dc85f0118f5a521388c change-id: 20260211-shmem-swap-gfp-c68b9d73a90f Best regards, -- Kairui Song