The quilt patch titled
Subject: mm: shmem: shmem_writepage() split folio at EOF before swapout
has been removed from the -mm tree. Its filename was
mm-shmem-support-large-folio-swap-out-fix.patch
This patch was dropped because it was folded into mm-shmem-support-large-folio-swap-out.patch
------------------------------------------------------
From: Hugh Dickins <hughd@xxxxxxxxxx>
Subject: mm: shmem: shmem_writepage() split folio at EOF before swapout
Date: Sun, 25 Aug 2024 16:14:17 -0700 (PDT)
Working in a constrained (size= or nr_blocks=) huge=always tmpfs relies
on swapout to split a large folio at EOF, to trim off its excess before
hitting premature ENOSPC: shmem_unused_huge_shrink() contains no code to
handle splitting huge swap blocks, and nobody would want that to be added.
Link: https://lkml.kernel.org/r/aef55f8d-6040-692d-65e3-16150cce4440@xxxxxxxxxx
Signed-off-by: Hugh Dickins <hughd@xxxxxxxxxx>
Cc: Baolin Wang <baolin.wang@xxxxxxxxxxxxxxxxx>
Cc: Barry Song <baohua@xxxxxxxxxx>
Cc: Chris Li <chrisl@xxxxxxxxxx>
Cc: David Hildenbrand <david@xxxxxxxxxx>
Cc: "Huang, Ying" <ying.huang@xxxxxxxxx>
Cc: Kefeng Wang <wangkefeng.wang@xxxxxxxxxx>
Cc: Lance Yang <ioworker0@xxxxxxxxx>
Cc: Matthew Wilcox <willy@xxxxxxxxxxxxx>
Cc: Pankaj Raghav <p.raghav@xxxxxxxxxxx>
Cc: Ryan Roberts <ryan.roberts@xxxxxxx>
Cc: Yang Shi <shy828301@xxxxxxxxx>
Cc: Zi Yan <ziy@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/shmem.c | 15 ++++++++++++++-
1 file changed, 14 insertions(+), 1 deletion(-)
--- a/mm/shmem.c~mm-shmem-support-large-folio-swap-out-fix
+++ a/mm/shmem.c
@@ -1459,6 +1459,7 @@ static int shmem_writepage(struct page *
swp_entry_t swap;
pgoff_t index;
int nr_pages;
+ bool split = false;
/*
* Our capabilities prevent regular writeback or sync from ever calling
@@ -1480,8 +1481,20 @@ static int shmem_writepage(struct page *
* If /sys/kernel/mm/transparent_hugepage/shmem_enabled is "always" or
* "force", drivers/gpu/drm/i915/gem/i915_gem_shmem.c gets huge pages,
* and its shmem_writeback() needs them to be split when swapping.
+ *
+ * And shrinkage of pages beyond i_size does not split swap, so
+ * swapout of a large folio crossing i_size needs to split too
+ * (unless fallocate has been used to preallocate beyond EOF).
*/
- if (wbc->split_large_folio && folio_test_large(folio)) {
+ if (folio_test_large(folio)) {
+ split = wbc->split_large_folio;
+ index = shmem_fallocend(inode,
+ DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE));
+ if (index > folio->index && index < folio_next_index(folio))
+ split = true;
+ }
+
+ if (split) {
try_split:
/* Ensure the subpages are still dirty */
folio_test_set_dirty(folio);
_
Patches currently in -mm which might be from hughd@xxxxxxxxxx are
mm-shmem-support-large-folio-swap-out.patch
mm-restart-if-multiple-traversals-raced-fix.patch
mm-shmem-fix-minor-off-by-one-in-shrinkable-calculation.patch
mm-shmem-extend-shmem_unused_huge_shrink-to-all-sizes.patch
