Skip to contentSkip to conversation

FAILED: patch "[PATCH] f2fs: fix to round down start offset of fallocate for pin" failed to apply to 5.15-stable tree

PATCH New [email protected] 2 participants 3 messages Jul 2 – Jul 3

Subsystems (2)
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y
git checkout FETCH_HEAD
git cherry-pick -x 4275b59673eb60b02eec3997816c83f1f4b909c4
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2026070255-livable-cheese-d39c@gregkh' --subject-prefix 'PATCH 5.15.y' 'HEAD^..'

Possible dependencies:

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 4275b59673eb60b02eec3997816c83f1f4b909c4 Mon Sep 17 00:00:00 2001
From: Sunmin Jeong <[email protected]>
Date: Mon, 22 Jun 2026 14:28:17 +0900
Subject: [PATCH] f2fs: fix to round down start offset of fallocate for pin
file

Currently, the length of fallocate for pin file is section-aligned to keep allocated sections from being selected as victims of GC. However, for the case that the start offset of fallocate is not aligned in section, the allocated sections can't be fully utilized. It's because a new section is allocated by f2fs_allocate_pinning_section() after using blks_per_sec blocks regardless of the start offset. As a result, several unexpected dirty segments may be created, including blocks assigned to the pinned file.

To address this issue, let's round down the start offset of fallocate to the length of section.

The reproducing scenario is as below

chunk=$(((2<<20)+4096)) # 2MB + 4KB
touch test
f2fs_io pinfile set test
f2fs_io fallocate 0 0 $chunk test
f2fs_io fallocate 0 $chunk $chunk test
f2fs_io fallocate 0 $((chunk*2)) $chunk test
f2fs_io fiemap 0 $((chunk*3)) test

Fiemap: offset = 0 len = 12288
logical addr. physical addr. length flags

0   0000000000000000 000000068c600000 0000000000400000 00001088
1   0000000000400000 000000003d400000 0000000000001000 00001088
2   0000000000401000 00000003eb200000 0000000000200000 00001088
3   0000000000601000 00000005e4200000 0000000000001000 00001088
4   0000000000602000 0000000605400000 0000000000200000 00001089
fs/f2fs/file.c +8 -1
index 8acdd94272a0..4b52c56d71f0 100644
@@ -1916,8 +1916,15 @@ static int f2fs_expand_inode_data(struct inode *inode, loff_t offset,
 
 	if (f2fs_is_pinned_file(inode)) {
 		block_t sec_blks = CAP_BLKS_PER_SEC(sbi);
-		block_t sec_len = roundup(map.m_len, sec_blks);
+		block_t sec_len;
 
+		if (map.m_lblk % sec_blks) {
+			map.m_lblk = rounddown(map.m_lblk, sec_blks);
+			map.m_len = pg_end - map.m_lblk;
+			if (off_end)
+				map.m_len++;
+		}
+		sec_len = roundup(map.m_len, sec_blks);
 		map.m_len = sec_blks;
 next_alloc:
 		f2fs_down_write(&sbi->pin_sem);
maintainer
#1 lore raw

PATCH 1/2 f2fs: adjust zone capacity when considering valid block count

From: Jaegeuk Kim <[email protected]>

[ Upstream commit 074b5ea2900ea8e40f8e7a3fd37e0a55ad3d5874 ]

This patch fixes counting unusable blocks set by zone capacity when checking the valid block count in a section.

Stable-dep-of: 4275b59673eb ("f2fs: fix to round down start offset of fallocate for pin file")

---
 fs/f2fs/debug.c   | 2 +-
 fs/f2fs/file.c    | 6 +++---
 fs/f2fs/gc.c      | 4 ++--
 fs/f2fs/segment.c | 7 +++----
 fs/f2fs/segment.h | 8 ++++----
 5 files changed, 13 insertions(+), 14 deletions(-)
fs/f2fs/debug.c +1 -1
index b449c7a372a4b1..cd21b4edeca1cb 100644
@@ -39,7 +39,7 @@ void f2fs_update_sit_info(struct f2fs_sb_info *sbi)
 
 	bimodal = 0;
 	total_vblocks = 0;
-	blks_per_sec = BLKS_PER_SEC(sbi);
+	blks_per_sec = CAP_BLKS_PER_SEC(sbi);
 	hblks_per_sec = blks_per_sec / 2;
 	for (segno = 0; segno < MAIN_SEGS(sbi); segno += sbi->segs_per_sec) {
 		vblocks = get_valid_blocks(sbi, segno, true);
fs/f2fs/file.c +3 -3
index 1c5f2964649f27..d4d485032e6ea6 100644
@@ -1694,7 +1694,7 @@ static int expand_inode_data(struct inode *inode, loff_t offset,
 		return 0;
 
 	if (f2fs_is_pinned_file(inode)) {
-		block_t sec_blks = BLKS_PER_SEC(sbi);
+		block_t sec_blks = CAP_BLKS_PER_SEC(sbi);
 		block_t sec_len = roundup(map.m_len, sec_blks);
 
 		map.m_len = sec_blks;
@@ -2547,7 +2547,7 @@ static int __f2fs_ioc_gc_range(struct file *filp, struct f2fs_gc_range *range)
 			ret = -EAGAIN;
 		goto out;
 	}
-	range->start += BLKS_PER_SEC(sbi);
+	range->start += CAP_BLKS_PER_SEC(sbi);
 	if (range->start <= end)
 		goto do_more;
 out:
@@ -2672,7 +2672,7 @@ static int f2fs_defragment_range(struct f2fs_sb_info *sbi,
 		goto out;
 	}
 
-	sec_num = DIV_ROUND_UP(total, BLKS_PER_SEC(sbi));
+	sec_num = DIV_ROUND_UP(total, CAP_BLKS_PER_SEC(sbi));
 
 	/*
 	 * make sure there are enough free section for LFS allocation, this can
fs/f2fs/gc.c +2 -2
index 9a57754e6e0c1d..69a51fcfbead45 100644
@@ -461,7 +461,7 @@ static void atgc_lookup_victim(struct f2fs_sb_info *sbi,
 	unsigned long long age, u, accu;
 	unsigned long long max_mtime = sit_i->dirty_max_mtime;
 	unsigned long long min_mtime = sit_i->dirty_min_mtime;
-	unsigned int sec_blocks = BLKS_PER_SEC(sbi);
+	unsigned int sec_blocks = CAP_BLKS_PER_SEC(sbi);
 	unsigned int vblocks;
 	unsigned int dirty_threshold = max(am->max_candidate_count,
 					am->candidate_ratio *
@@ -1445,7 +1445,7 @@ static int gc_data_segment(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
 		 */
 		if ((gc_type == BG_GC && has_not_enough_free_secs(sbi, 0, 0)) ||
 			(!force_migrate && get_valid_blocks(sbi, segno, true) ==
-							BLKS_PER_SEC(sbi)))
+							CAP_BLKS_PER_SEC(sbi)))
 			return submitted;
 
 		if (check_valid_map(sbi, segno, off) == 0)
fs/f2fs/segment.c +3 -4
index 52d5915cee9674..95fdad8b19fdaf 100644
@@ -833,7 +833,7 @@ static void __locate_dirty_segment(struct f2fs_sb_info *sbi, unsigned int segno,
 				get_valid_blocks(sbi, segno, true);
 
 			f2fs_bug_on(sbi, unlikely(!valid_blocks ||
-					valid_blocks == BLKS_PER_SEC(sbi)));
+					valid_blocks == CAP_BLKS_PER_SEC(sbi)));
 
 			if (!IS_CURSEC(sbi, secno))
 				set_bit(secno, dirty_i->dirty_secmap);
@@ -869,7 +869,7 @@ static void __remove_dirty_segment(struct f2fs_sb_info *sbi, unsigned int segno,
 			unsigned int secno = GET_SEC_FROM_SEG(sbi, segno);
 
 			if (!valid_blocks ||
-					valid_blocks == BLKS_PER_SEC(sbi)) {
+					valid_blocks == CAP_BLKS_PER_SEC(sbi)) {
 				clear_bit(secno, dirty_i->dirty_secmap);
 				return;
 			}
@@ -4679,7 +4679,6 @@ static void init_dirty_segmap(struct f2fs_sb_info *sbi)
 	struct free_segmap_info *free_i = FREE_I(sbi);
 	unsigned int segno = 0, offset = 0, secno;
 	block_t valid_blocks, usable_blks_in_seg;
-	block_t blks_per_sec = BLKS_PER_SEC(sbi);
 
 	while (1) {
 		/* find dirty segment based on free segmap */
@@ -4708,7 +4707,7 @@ static void init_dirty_segmap(struct f2fs_sb_info *sbi)
 		valid_blocks = get_valid_blocks(sbi, segno, true);
 		secno = GET_SEC_FROM_SEG(sbi, segno);
 
-		if (!valid_blocks || valid_blocks == blks_per_sec)
+		if (!valid_blocks || valid_blocks == CAP_BLKS_PER_SEC(sbi))
 			continue;
 		if (IS_CURSEC(sbi, secno))
 			continue;
fs/f2fs/segment.h +4 -4
index e59b330ed83351..5fdb393524c888 100644
@@ -605,10 +605,10 @@ static inline bool has_not_enough_free_secs(struct f2fs_sb_info *sbi,
 					get_pages(sbi, F2FS_DIRTY_DENTS) +
 					get_pages(sbi, F2FS_DIRTY_IMETA);
 	unsigned int total_dent_blocks = get_pages(sbi, F2FS_DIRTY_DENTS);
-	unsigned int node_secs = total_node_blocks / BLKS_PER_SEC(sbi);
-	unsigned int dent_secs = total_dent_blocks / BLKS_PER_SEC(sbi);
-	unsigned int node_blocks = total_node_blocks % BLKS_PER_SEC(sbi);
-	unsigned int dent_blocks = total_dent_blocks % BLKS_PER_SEC(sbi);
+	unsigned int node_secs = total_node_blocks / CAP_BLKS_PER_SEC(sbi);
+	unsigned int dent_secs = total_dent_blocks / CAP_BLKS_PER_SEC(sbi);
+	unsigned int node_blocks = total_node_blocks % CAP_BLKS_PER_SEC(sbi);
+	unsigned int dent_blocks = total_dent_blocks % CAP_BLKS_PER_SEC(sbi);
 	unsigned int free, need_lower, need_upper;
 
 	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
Signature · 1 line
2.53.0
maintainer
#2 lore raw

PATCH 2/2 f2fs: fix to round down start offset of fallocate for pin file

From: Sunmin Jeong <[email protected]>

[ Upstream commit 4275b59673eb60b02eec3997816c83f1f4b909c4 ]

Currently, the length of fallocate for pin file is section-aligned to keep allocated sections from being selected as victims of GC. However, for the case that the start offset of fallocate is not aligned in section, the allocated sections can't be fully utilized. It's because a new section is allocated by f2fs_allocate_pinning_section() after using blks_per_sec blocks regardless of the start offset. As a result, several unexpected dirty segments may be created, including blocks assigned to the pinned file.

To address this issue, let's round down the start offset of fallocate to the length of section.

The reproducing scenario is as below

chunk=$(((2<<20)+4096)) # 2MB + 4KB
touch test
f2fs_io pinfile set test
f2fs_io fallocate 0 0 $chunk test
f2fs_io fallocate 0 $chunk $chunk test
f2fs_io fallocate 0 $((chunk*2)) $chunk test
f2fs_io fiemap 0 $((chunk*3)) test

Fiemap: offset = 0 len = 12288
logical addr. physical addr. length flags

0   0000000000000000 000000068c600000 0000000000400000 00001088
1   0000000000400000 000000003d400000 0000000000001000 00001088
2   0000000000401000 00000003eb200000 0000000000200000 00001088
3   0000000000601000 00000005e4200000 0000000000001000 00001088
4   0000000000602000 0000000605400000 0000000000200000 00001089
---
 fs/f2fs/file.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
fs/f2fs/file.c +8 -1
index d4d485032e6ea6..3894c5889df8fe 100644
@@ -1695,8 +1695,15 @@ static int expand_inode_data(struct inode *inode, loff_t offset,
 
 	if (f2fs_is_pinned_file(inode)) {
 		block_t sec_blks = CAP_BLKS_PER_SEC(sbi);
-		block_t sec_len = roundup(map.m_len, sec_blks);
+		block_t sec_len;
 
+		if (map.m_lblk % sec_blks) {
+			map.m_lblk = rounddown(map.m_lblk, sec_blks);
+			map.m_len = pg_end - map.m_lblk;
+			if (off_end)
+				map.m_len++;
+		}
+		sec_len = roundup(map.m_len, sec_blks);
 		map.m_len = sec_blks;
 next_alloc:
 		if (has_not_enough_free_secs(sbi, 0,
Signature · 1 line
2.53.0