This is a note to let you know that I've just added the patch titled
f2fs: support file pinning for zoned devices
to the 6.6-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
f2fs-support-file-pinning-for-zoned-devices.patch
and it can be found in the queue-6.6 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
commit 01660f99fd3ecf76b21b8c2db069abda71d6f715
Author: Daeho Jeong <daehojeong@xxxxxxxxxx>
Date: Tue Feb 13 09:38:12 2024 -0800
f2fs: support file pinning for zoned devices
[ Upstream commit 9703d69d9d153bb230711d0d577454552aeb13d4 ]
Support file pinning with conventional storage area for zoned devices
Signed-off-by: Daeho Jeong <daehojeong@xxxxxxxxxx>
Signed-off-by: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
Stable-dep-of: aa4074e8fec4 ("f2fs: fix block migration when section is not aligned to pow2")
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 5d6ba12e84482..3558fc3387f54 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -3905,25 +3905,34 @@ static int f2fs_migrate_blocks(struct inode *inode, block_t start_blk,
unsigned int blkofs;
unsigned int blk_per_sec = BLKS_PER_SEC(sbi);
unsigned int secidx = start_blk / blk_per_sec;
- unsigned int end_sec = secidx + blkcnt / blk_per_sec;
+ unsigned int end_sec;
int ret = 0;
+ if (!blkcnt)
+ return 0;
+ end_sec = secidx + (blkcnt - 1) / blk_per_sec;
+
f2fs_down_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]);
filemap_invalidate_lock(inode->i_mapping);
set_inode_flag(inode, FI_ALIGNED_WRITE);
set_inode_flag(inode, FI_OPU_WRITE);
- for (; secidx < end_sec; secidx++) {
+ for (; secidx <= end_sec; secidx++) {
+ unsigned int blkofs_end = secidx == end_sec ?
+ (blkcnt - 1) % blk_per_sec : blk_per_sec - 1;
+
f2fs_down_write(&sbi->pin_sem);
- f2fs_lock_op(sbi);
- f2fs_allocate_new_section(sbi, CURSEG_COLD_DATA_PINNED, false);
- f2fs_unlock_op(sbi);
+ ret = f2fs_allocate_pinning_section(sbi);
+ if (ret) {
+ f2fs_up_write(&sbi->pin_sem);
+ break;
+ }
set_inode_flag(inode, FI_SKIP_WRITES);
- for (blkofs = 0; blkofs < blk_per_sec; blkofs++) {
+ for (blkofs = 0; blkofs <= blkofs_end; blkofs++) {
struct page *page;
unsigned int blkidx = secidx * blk_per_sec + blkofs;
@@ -4012,27 +4021,34 @@ static int check_swap_activate(struct swap_info_struct *sis,
nr_pblocks = map.m_len;
if ((pblock - SM_I(sbi)->main_blkaddr) & sec_blks_mask ||
- nr_pblocks & sec_blks_mask) {
+ nr_pblocks & sec_blks_mask ||
+ !f2fs_valid_pinned_area(sbi, pblock)) {
+ bool last_extent = false;
+
not_aligned++;
nr_pblocks = roundup(nr_pblocks, blks_per_sec);
if (cur_lblock + nr_pblocks > sis->max)
nr_pblocks -= blks_per_sec;
+ /* this extent is last one */
if (!nr_pblocks) {
- /* this extent is last one */
- nr_pblocks = map.m_len;
- f2fs_warn(sbi, "Swapfile: last extent is not aligned to section");
- goto next;
+ nr_pblocks = last_lblock - cur_lblock;
+ last_extent = true;
}
ret = f2fs_migrate_blocks(inode, cur_lblock,
nr_pblocks);
- if (ret)
+ if (ret) {
+ if (ret == -ENOENT)
+ ret = -EINVAL;
goto out;
- goto retry;
+ }
+
+ if (!last_extent)
+ goto retry;
}
-next:
+
if (cur_lblock + nr_pblocks >= sis->max)
nr_pblocks = sis->max - cur_lblock;
@@ -4070,17 +4086,17 @@ static int f2fs_swap_activate(struct swap_info_struct *sis, struct file *file,
sector_t *span)
{
struct inode *inode = file_inode(file);
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
int ret;
if (!S_ISREG(inode->i_mode))
return -EINVAL;
- if (f2fs_readonly(F2FS_I_SB(inode)->sb))
+ if (f2fs_readonly(sbi->sb))
return -EROFS;
- if (f2fs_lfs_mode(F2FS_I_SB(inode))) {
- f2fs_err(F2FS_I_SB(inode),
- "Swapfile not supported in LFS mode");
+ if (f2fs_lfs_mode(sbi) && !f2fs_sb_has_blkzoned(sbi)) {
+ f2fs_err(sbi, "Swapfile not supported in LFS mode");
return -EINVAL;
}
@@ -4093,13 +4109,17 @@ static int f2fs_swap_activate(struct swap_info_struct *sis, struct file *file,
f2fs_precache_extents(inode);
+ ret = filemap_fdatawrite(inode->i_mapping);
+ if (ret < 0)
+ return ret;
+
ret = check_swap_activate(sis, file, span);
if (ret < 0)
return ret;
stat_inc_swapfile_inode(inode);
set_inode_flag(inode, FI_PIN_FILE);
- f2fs_update_time(F2FS_I_SB(inode), REQ_TIME);
+ f2fs_update_time(sbi, REQ_TIME);
return ret;
}
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 9b80ceaf39502..82f54ffde4cd4 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -3704,7 +3704,8 @@ void f2fs_get_new_segment(struct f2fs_sb_info *sbi,
unsigned int *newseg, bool new_sec, int dir);
void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
unsigned int start, unsigned int end);
-void f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force);
+int f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force);
+int f2fs_allocate_pinning_section(struct f2fs_sb_info *sbi);
void f2fs_allocate_new_segments(struct f2fs_sb_info *sbi);
int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range);
bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi,
@@ -3880,6 +3881,9 @@ void f2fs_stop_gc_thread(struct f2fs_sb_info *sbi);
block_t f2fs_start_bidx_of_node(unsigned int node_ofs, struct inode *inode);
int f2fs_gc(struct f2fs_sb_info *sbi, struct f2fs_gc_control *gc_control);
void f2fs_build_gc_manager(struct f2fs_sb_info *sbi);
+int f2fs_gc_range(struct f2fs_sb_info *sbi,
+ unsigned int start_seg, unsigned int end_seg,
+ bool dry_run, unsigned int dry_run_sections);
int f2fs_resize_fs(struct file *filp, __u64 block_count);
int __init f2fs_create_garbage_collection_cache(void);
void f2fs_destroy_garbage_collection_cache(void);
@@ -4536,6 +4540,17 @@ static inline bool f2fs_lfs_mode(struct f2fs_sb_info *sbi)
return F2FS_OPTION(sbi).fs_mode == FS_MODE_LFS;
}
+static inline bool f2fs_valid_pinned_area(struct f2fs_sb_info *sbi,
+ block_t blkaddr)
+{
+ if (f2fs_sb_has_blkzoned(sbi)) {
+ int devi = f2fs_target_device_index(sbi, blkaddr);
+
+ return !bdev_is_zoned(FDEV(devi).bdev);
+ }
+ return true;
+}
+
static inline bool f2fs_low_mem_mode(struct f2fs_sb_info *sbi)
{
return F2FS_OPTION(sbi).memory_mode == MEMORY_MODE_LOW;
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index 2671a61556951..20dc11b58bb69 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -1729,9 +1729,11 @@ static int f2fs_expand_inode_data(struct inode *inode, loff_t offset,
f2fs_down_write(&sbi->pin_sem);
- f2fs_lock_op(sbi);
- f2fs_allocate_new_section(sbi, CURSEG_COLD_DATA_PINNED, false);
- f2fs_unlock_op(sbi);
+ err = f2fs_allocate_pinning_section(sbi);
+ if (err) {
+ f2fs_up_write(&sbi->pin_sem);
+ goto out_err;
+ }
map.m_seg_type = CURSEG_COLD_DATA_PINNED;
err = f2fs_map_blocks(inode, &map, F2FS_GET_BLOCK_PRE_DIO);
@@ -3184,6 +3186,7 @@ int f2fs_pin_file_control(struct inode *inode, bool inc)
static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg)
{
struct inode *inode = file_inode(filp);
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
__u32 pin;
int ret = 0;
@@ -3193,7 +3196,7 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg)
if (!S_ISREG(inode->i_mode))
return -EINVAL;
- if (f2fs_readonly(F2FS_I_SB(inode)->sb))
+ if (f2fs_readonly(sbi->sb))
return -EROFS;
ret = mnt_want_write_file(filp);
@@ -3206,9 +3209,18 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg)
clear_inode_flag(inode, FI_PIN_FILE);
f2fs_i_gc_failures_write(inode, 0);
goto done;
+ } else if (f2fs_is_pinned_file(inode)) {
+ goto done;
}
- if (f2fs_should_update_outplace(inode, NULL)) {
+ if (f2fs_sb_has_blkzoned(sbi) && F2FS_HAS_BLOCKS(inode)) {
+ ret = -EFBIG;
+ goto out;
+ }
+
+ /* Let's allow file pinning on zoned device. */
+ if (!f2fs_sb_has_blkzoned(sbi) &&
+ f2fs_should_update_outplace(inode, NULL)) {
ret = -EINVAL;
goto out;
}
@@ -3230,7 +3242,7 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg)
set_inode_flag(inode, FI_PIN_FILE);
ret = F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN];
done:
- f2fs_update_time(F2FS_I_SB(inode), REQ_TIME);
+ f2fs_update_time(sbi, REQ_TIME);
out:
inode_unlock(inode);
mnt_drop_write_file(filp);
diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
index 2a3d64f4253ee..9a4ec639e58a7 100644
--- a/fs/f2fs/gc.c
+++ b/fs/f2fs/gc.c
@@ -1979,10 +1979,12 @@ void f2fs_build_gc_manager(struct f2fs_sb_info *sbi)
init_atgc_management(sbi);
}
-static int f2fs_gc_range(struct f2fs_sb_info *sbi,
- unsigned int start_seg, unsigned int end_seg, bool dry_run)
+int f2fs_gc_range(struct f2fs_sb_info *sbi,
+ unsigned int start_seg, unsigned int end_seg,
+ bool dry_run, unsigned int dry_run_sections)
{
unsigned int segno;
+ unsigned int gc_secs = dry_run_sections;
for (segno = start_seg; segno <= end_seg; segno += SEGS_PER_SEC(sbi)) {
struct gc_inode_list gc_list = {
@@ -1990,11 +1992,15 @@ static int f2fs_gc_range(struct f2fs_sb_info *sbi,
.iroot = RADIX_TREE_INIT(gc_list.iroot, GFP_NOFS),
};
- do_garbage_collect(sbi, segno, &gc_list, FG_GC, true);
+ do_garbage_collect(sbi, segno, &gc_list, FG_GC,
+ dry_run_sections == 0);
put_gc_inode(&gc_list);
if (!dry_run && get_valid_blocks(sbi, segno, true))
return -EAGAIN;
+ if (dry_run && dry_run_sections &&
+ !get_valid_blocks(sbi, segno, true) && --gc_secs == 0)
+ break;
if (fatal_signal_pending(current))
return -ERESTARTSYS;
@@ -2032,7 +2038,7 @@ static int free_segment_range(struct f2fs_sb_info *sbi,
f2fs_allocate_segment_for_resize(sbi, type, start, end);
/* do GC to move out valid blocks in the range */
- err = f2fs_gc_range(sbi, start, end, dry_run);
+ err = f2fs_gc_range(sbi, start, end, dry_run, 0);
if (err || dry_run)
goto out;
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 523c3a91bdf21..52b281e9f472f 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -2638,7 +2638,7 @@ static int is_next_segment_free(struct f2fs_sb_info *sbi,
* This function should be returned with success, otherwise BUG
*/
static void get_new_segment(struct f2fs_sb_info *sbi,
- unsigned int *newseg, bool new_sec)
+ unsigned int *newseg, bool new_sec, bool pinning)
{
struct free_segmap_info *free_i = FREE_I(sbi);
unsigned int segno, secno, zoneno;
@@ -2656,6 +2656,16 @@ static void get_new_segment(struct f2fs_sb_info *sbi,
if (segno < GET_SEG_FROM_SEC(sbi, hint + 1))
goto got_it;
}
+
+ /*
+ * If we format f2fs on zoned storage, let's try to get pinned sections
+ * from beginning of the storage, which should be a conventional one.
+ */
+ if (f2fs_sb_has_blkzoned(sbi)) {
+ segno = pinning ? 0 : max(first_zoned_segno(sbi), *newseg);
+ hint = GET_SEC_FROM_SEG(sbi, segno);
+ }
+
find_other_zone:
secno = find_next_zero_bit(free_i->free_secmap, MAIN_SECS(sbi), hint);
if (secno >= MAIN_SECS(sbi)) {
@@ -2754,21 +2764,30 @@ static unsigned int __get_next_segno(struct f2fs_sb_info *sbi, int type)
* Allocate a current working segment.
* This function always allocates a free segment in LFS manner.
*/
-static void new_curseg(struct f2fs_sb_info *sbi, int type, bool new_sec)
+static int new_curseg(struct f2fs_sb_info *sbi, int type, bool new_sec)
{
struct curseg_info *curseg = CURSEG_I(sbi, type);
unsigned int segno = curseg->segno;
+ bool pinning = type == CURSEG_COLD_DATA_PINNED;
if (curseg->inited)
write_sum_page(sbi, curseg->sum_blk, GET_SUM_BLOCK(sbi, segno));
+
segno = __get_next_segno(sbi, type);
- get_new_segment(sbi, &segno, new_sec);
+ get_new_segment(sbi, &segno, new_sec, pinning);
+ if (new_sec && pinning &&
+ !f2fs_valid_pinned_area(sbi, START_BLOCK(sbi, segno))) {
+ __set_free(sbi, segno);
+ return -EAGAIN;
+ }
+
curseg->next_segno = segno;
reset_curseg(sbi, type, 1);
curseg->alloc_type = LFS;
if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK)
curseg->fragment_remained_chunk =
get_random_u32_inclusive(1, sbi->max_fragment_chunk);
+ return 0;
}
static int __next_free_blkoff(struct f2fs_sb_info *sbi,
@@ -3041,7 +3060,7 @@ void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
f2fs_up_read(&SM_I(sbi)->curseg_lock);
}
-static void __allocate_new_segment(struct f2fs_sb_info *sbi, int type,
+static int __allocate_new_segment(struct f2fs_sb_info *sbi, int type,
bool new_sec, bool force)
{
struct curseg_info *curseg = CURSEG_I(sbi, type);
@@ -3051,21 +3070,49 @@ static void __allocate_new_segment(struct f2fs_sb_info *sbi, int type,
!curseg->next_blkoff &&
!get_valid_blocks(sbi, curseg->segno, new_sec) &&
!get_ckpt_valid_blocks(sbi, curseg->segno, new_sec))
- return;
+ return 0;
old_segno = curseg->segno;
- new_curseg(sbi, type, true);
+ if (new_curseg(sbi, type, true))
+ return -EAGAIN;
stat_inc_seg_type(sbi, curseg);
locate_dirty_segment(sbi, old_segno);
+ return 0;
}
-void f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force)
+int f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force)
{
+ int ret;
+
f2fs_down_read(&SM_I(sbi)->curseg_lock);
down_write(&SIT_I(sbi)->sentry_lock);
- __allocate_new_segment(sbi, type, true, force);
+ ret = __allocate_new_segment(sbi, type, true, force);
up_write(&SIT_I(sbi)->sentry_lock);
f2fs_up_read(&SM_I(sbi)->curseg_lock);
+
+ return ret;
+}
+
+int f2fs_allocate_pinning_section(struct f2fs_sb_info *sbi)
+{
+ int err;
+ bool gc_required = true;
+
+retry:
+ f2fs_lock_op(sbi);
+ err = f2fs_allocate_new_section(sbi, CURSEG_COLD_DATA_PINNED, false);
+ f2fs_unlock_op(sbi);
+
+ if (f2fs_sb_has_blkzoned(sbi) && err && gc_required) {
+ f2fs_down_write(&sbi->gc_lock);
+ f2fs_gc_range(sbi, 0, GET_SEGNO(sbi, FDEV(0).end_blk), true, 1);
+ f2fs_up_write(&sbi->gc_lock);
+
+ gc_required = false;
+ goto retry;
+ }
+
+ return err;
}
void f2fs_allocate_new_segments(struct f2fs_sb_info *sbi)
@@ -3431,6 +3478,10 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
* new segment.
*/
if (segment_full) {
+ if (type == CURSEG_COLD_DATA_PINNED &&
+ !((curseg->segno + 1) % sbi->segs_per_sec))
+ goto skip_new_segment;
+
if (from_gc) {
get_atssr_segment(sbi, type, se->type,
AT_SSR, se->mtime);
@@ -3442,6 +3493,8 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
stat_inc_seg_type(sbi, curseg);
}
}
+
+skip_new_segment:
/*
* segment dirty status should be updated after segment allocation,
* so we just need to update status only one time after previous
diff --git a/fs/f2fs/segment.h b/fs/f2fs/segment.h
index 93ffb62e45f4d..4595f1cc03828 100644
--- a/fs/f2fs/segment.h
+++ b/fs/f2fs/segment.h
@@ -956,3 +956,13 @@ static inline void wake_up_discard_thread(struct f2fs_sb_info *sbi, bool force)
dcc->discard_wake = true;
wake_up_interruptible_all(&dcc->discard_wait_queue);
}
+
+static inline unsigned int first_zoned_segno(struct f2fs_sb_info *sbi)
+{
+ int devi;
+
+ for (devi = 0; devi < sbi->s_ndevs; devi++)
+ if (bdev_is_zoned(FDEV(devi).bdev))
+ return GET_SEGNO(sbi, FDEV(devi).start_blk);
+ return 0;
+}
