[PATCH 0/2 v9] reclaim file-backed pages given POSIX_FADV_NOREUSE

[PATCH 0/2 v9] reclaim file-backed pages given POSIX_FADV_NOREUSE

[Jaegeuk Kim]

This patch series does not add new API, but implements POSIX_FADV_NOREUSE where
it keeps the page ranges in the f2fs superblock and add a way for users to
reclaim the pages manually.

Change log from v8:
 - remove new APIs, but use fadvise(POSIX_FADV_NOREUSE)

Jaegeuk Kim (2):
  f2fs: keep POSIX_FADV_NOREUSE ranges
  f2fs: add a sysfs entry to reclaim POSIX_FADV_NOREUSE pages

 Documentation/ABI/testing/sysfs-fs-f2fs |  7 ++
 fs/f2fs/debug.c                         |  3 +
 fs/f2fs/f2fs.h                          | 14 +++-
 fs/f2fs/file.c                          | 60 +++++++++++++++--
 fs/f2fs/inode.c                         | 14 ++++
 fs/f2fs/shrinker.c                      | 90 +++++++++++++++++++++++++
 fs/f2fs/super.c                         |  1 +
 fs/f2fs/sysfs.c                         | 63 +++++++++++++++++
 8 files changed, 246 insertions(+), 6 deletions(-)

-- 
2.48.1.601.g30ceb7b040-goog

[PATCH 1/2] f2fs: keep POSIX_FADV_NOREUSE ranges

[Jaegeuk Kim]

This patch records POSIX_FADV_NOREUSE ranges for users to reclaim the caches
instantly off from LRU.

Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
 fs/f2fs/debug.c |  3 +++
 fs/f2fs/f2fs.h  | 12 +++++++++-
 fs/f2fs/file.c  | 60 ++++++++++++++++++++++++++++++++++++++++++++-----
 fs/f2fs/inode.c | 14 ++++++++++++
 fs/f2fs/super.c |  1 +
 5 files changed, 84 insertions(+), 6 deletions(-)

diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index 468828288a4a..16c2dfb4f595 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -164,6 +164,7 @@ static void update_general_status(struct f2fs_sb_info *sbi)
 	si->ndirty_imeta = get_pages(sbi, F2FS_DIRTY_IMETA);
 	si->ndirty_dirs = sbi->ndirty_inode[DIR_INODE];
 	si->ndirty_files = sbi->ndirty_inode[FILE_INODE];
+	si->ndonate_files = sbi->donate_files;
 	si->nquota_files = sbi->nquota_files;
 	si->ndirty_all = sbi->ndirty_inode[DIRTY_META];
 	si->aw_cnt = atomic_read(&sbi->atomic_files);
@@ -501,6 +502,8 @@ static int stat_show(struct seq_file *s, void *v)
 			   si->compr_inode, si->compr_blocks);
 		seq_printf(s, "  - Swapfile Inode: %u\n",
 			   si->swapfile_inode);
+		seq_printf(s, "  - Donate Inode: %u\n",
+			   si->ndonate_files);
 		seq_printf(s, "  - Orphan/Append/Update Inode: %u, %u, %u\n",
 			   si->orphans, si->append, si->update);
 		seq_printf(s, "\nMain area: %d segs, %d secs %d zones\n",
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 395f9d37449c..3abcb84a0d47 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -850,6 +850,11 @@ struct f2fs_inode_info {
 #endif
 	struct list_head dirty_list;	/* dirty list for dirs and files */
 	struct list_head gdirty_list;	/* linked in global dirty list */
+
+	/* linked in global inode list for cache donation */
+	struct list_head gdonate_list;
+	pgoff_t donate_start, donate_end; /* inclusive */
+
 	struct task_struct *atomic_write_task;	/* store atomic write task */
 	struct extent_tree *extent_tree[NR_EXTENT_CACHES];
 					/* cached extent_tree entry */
@@ -1274,6 +1279,7 @@ enum inode_type {
 	DIR_INODE,			/* for dirty dir inode */
 	FILE_INODE,			/* for dirty regular/symlink inode */
 	DIRTY_META,			/* for all dirtied inode metadata */
+	DONATE_INODE,			/* for all inode to donate pages */
 	NR_INODE_TYPE,
 };
 
@@ -1629,6 +1635,9 @@ struct f2fs_sb_info {
 	unsigned int warm_data_age_threshold;
 	unsigned int last_age_weight;
 
+	/* control donate caches */
+	unsigned int donate_files;
+
 	/* basic filesystem units */
 	unsigned int log_sectors_per_block;	/* log2 sectors per block */
 	unsigned int log_blocksize;		/* log2 block size */
@@ -3968,7 +3977,8 @@ struct f2fs_stat_info {
 	unsigned long long allocated_data_blocks;
 	int ndirty_node, ndirty_dent, ndirty_meta, ndirty_imeta;
 	int ndirty_data, ndirty_qdata;
-	unsigned int ndirty_dirs, ndirty_files, nquota_files, ndirty_all;
+	unsigned int ndirty_dirs, ndirty_files, ndirty_all;
+	unsigned int nquota_files, ndonate_files;
 	int nats, dirty_nats, sits, dirty_sits;
 	int free_nids, avail_nids, alloc_nids;
 	int total_count, utilization;
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index f2f298c75921..014cb7660a9a 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -2450,6 +2450,52 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg)
 	return ret;
 }
 
+static void f2fs_keep_noreuse_range(struct inode *inode,
+				loff_t offset, loff_t len)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	u64 max_bytes = F2FS_BLK_TO_BYTES(max_file_blocks(inode));
+	u64 start, end;
+
+	if (!S_ISREG(inode->i_mode))
+		return;
+
+	if (offset >= max_bytes || len > max_bytes ||
+	    (offset + len) > max_bytes)
+		return;
+
+	start = offset >> PAGE_SHIFT;
+	end = DIV_ROUND_UP(offset + len, PAGE_SIZE);
+
+	inode_lock(inode);
+	if (f2fs_is_atomic_file(inode)) {
+		inode_unlock(inode);
+		return;
+	}
+
+	spin_lock(&sbi->inode_lock[DONATE_INODE]);
+	/* let's remove the range, if len = 0 */
+	if (!len) {
+		if (!list_empty(&F2FS_I(inode)->gdonate_list)) {
+			list_del_init(&F2FS_I(inode)->gdonate_list);
+			sbi->donate_files--;
+		}
+	} else {
+		if (list_empty(&F2FS_I(inode)->gdonate_list)) {
+			list_add_tail(&F2FS_I(inode)->gdonate_list,
+					&sbi->inode_list[DONATE_INODE]);
+			sbi->donate_files++;
+		} else {
+			list_move_tail(&F2FS_I(inode)->gdonate_list,
+					&sbi->inode_list[DONATE_INODE]);
+		}
+		F2FS_I(inode)->donate_start = start;
+		F2FS_I(inode)->donate_end = end - 1;
+	}
+	spin_unlock(&sbi->inode_lock[DONATE_INODE]);
+	inode_unlock(inode);
+}
+
 static int f2fs_ioc_fitrim(struct file *filp, unsigned long arg)
 {
 	struct inode *inode = file_inode(filp);
@@ -5168,12 +5214,16 @@ static int f2fs_file_fadvise(struct file *filp, loff_t offset, loff_t len,
 	}
 
 	err = generic_fadvise(filp, offset, len, advice);
-	if (!err && advice == POSIX_FADV_DONTNEED &&
-		test_opt(F2FS_I_SB(inode), COMPRESS_CACHE) &&
-		f2fs_compressed_file(inode))
-		f2fs_invalidate_compress_pages(F2FS_I_SB(inode), inode->i_ino);
+	if (err)
+		return err;
 
-	return err;
+	if (advice == POSIX_FADV_DONTNEED &&
+	    (test_opt(F2FS_I_SB(inode), COMPRESS_CACHE) &&
+	     f2fs_compressed_file(inode)))
+		f2fs_invalidate_compress_pages(F2FS_I_SB(inode), inode->i_ino);
+	else if (advice == POSIX_FADV_NOREUSE)
+		f2fs_keep_noreuse_range(inode, offset, len);
+	return 0;
 }
 
 #ifdef CONFIG_COMPAT
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index 3f03b1c9b7f1..543fb942bec1 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -814,6 +814,19 @@ int f2fs_write_inode(struct inode *inode, struct writeback_control *wbc)
 	return 0;
 }
 
+static void f2fs_remove_donate_inode(struct inode *inode)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+
+	if (list_empty(&F2FS_I(inode)->gdonate_list))
+		return;
+
+	spin_lock(&sbi->inode_lock[DONATE_INODE]);
+	list_del_init(&F2FS_I(inode)->gdonate_list);
+	sbi->donate_files--;
+	spin_unlock(&sbi->inode_lock[DONATE_INODE]);
+}
+
 /*
  * Called at the last iput() if i_nlink is zero
  */
@@ -848,6 +861,7 @@ void f2fs_evict_inode(struct inode *inode)
 
 	f2fs_bug_on(sbi, get_dirty_pages(inode));
 	f2fs_remove_dirty_inode(inode);
+	f2fs_remove_donate_inode(inode);
 
 	if (!IS_DEVICE_ALIASING(inode))
 		f2fs_destroy_extent_tree(inode);
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 759b642991d4..f5c69cc2de72 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -1441,6 +1441,7 @@ static struct inode *f2fs_alloc_inode(struct super_block *sb)
 	spin_lock_init(&fi->i_size_lock);
 	INIT_LIST_HEAD(&fi->dirty_list);
 	INIT_LIST_HEAD(&fi->gdirty_list);
+	INIT_LIST_HEAD(&fi->gdonate_list);
 	init_f2fs_rwsem(&fi->i_gc_rwsem[READ]);
 	init_f2fs_rwsem(&fi->i_gc_rwsem[WRITE]);
 	init_f2fs_rwsem(&fi->i_xattr_sem);
-- 
2.48.1.601.g30ceb7b040-goog

[PATCH 2/2] f2fs: add a sysfs entry to reclaim POSIX_FADV_NOREUSE pages

[Jaegeuk Kim]

1. fadvise(fd1, POSIX_FADV_NOREUSE, {0,3});
2. fadvise(fd2, POSIX_FADV_NOREUSE, {1,2});
3. fadvise(fd3, POSIX_FADV_NOREUSE, {3,1});
4. echo 1024 > /sys/fs/f2fs/tuning/reclaim_caches_kb

This gives a way to reclaim file-backed pages by iterating all f2fs mounts until
reclaiming 1MB page cache ranges, registered by #1, #2, and #3.

5. cat /sys/fs/f2fs/tuning/reclaim_caches_kb
-> gives total number of registered file ranges.

Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
 Documentation/ABI/testing/sysfs-fs-f2fs |  7 ++
 fs/f2fs/f2fs.h                          |  2 +
 fs/f2fs/shrinker.c                      | 90 +++++++++++++++++++++++++
 fs/f2fs/sysfs.c                         | 63 +++++++++++++++++
 4 files changed, 162 insertions(+)

diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs
index 3e1630c70d8a..81deae2af84d 100644
--- a/Documentation/ABI/testing/sysfs-fs-f2fs
+++ b/Documentation/ABI/testing/sysfs-fs-f2fs
@@ -828,3 +828,10 @@ Date:		November 2024
 Contact:	"Chao Yu" <chao@kernel.org>
 Description:	It controls max read extent count for per-inode, the value of threshold
 		is 10240 by default.
+
+What:		/sys/fs/f2fs/tuning/reclaim_caches_kb
+Date:		February 2025
+Contact:	"Jaegeuk Kim" <jaegeuk@kernel.org>
+Description:	It reclaims the given KBs of file-backed pages registered by
+		ioctl(F2FS_IOC_DONATE_RANGE).
+		For example, writing N tries to drop N KBs spaces in LRU.
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 3abcb84a0d47..05879c6dc4d6 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -4243,6 +4243,8 @@ unsigned long f2fs_shrink_count(struct shrinker *shrink,
 			struct shrink_control *sc);
 unsigned long f2fs_shrink_scan(struct shrinker *shrink,
 			struct shrink_control *sc);
+unsigned int f2fs_donate_files(void);
+void f2fs_reclaim_caches(unsigned int reclaim_caches_kb);
 void f2fs_join_shrinker(struct f2fs_sb_info *sbi);
 void f2fs_leave_shrinker(struct f2fs_sb_info *sbi);
 
diff --git a/fs/f2fs/shrinker.c b/fs/f2fs/shrinker.c
index fc1bbef418ce..9c8d3aee89af 100644
--- a/fs/f2fs/shrinker.c
+++ b/fs/f2fs/shrinker.c
@@ -130,6 +130,96 @@ unsigned long f2fs_shrink_scan(struct shrinker *shrink,
 	return freed;
 }
 
+unsigned int f2fs_donate_files(void)
+{
+	struct f2fs_sb_info *sbi;
+	struct list_head *p;
+	unsigned int donate_files = 0;
+
+	spin_lock(&f2fs_list_lock);
+	p = f2fs_list.next;
+	while (p != &f2fs_list) {
+		sbi = list_entry(p, struct f2fs_sb_info, s_list);
+
+		/* stop f2fs_put_super */
+		if (!mutex_trylock(&sbi->umount_mutex)) {
+			p = p->next;
+			continue;
+		}
+		spin_unlock(&f2fs_list_lock);
+
+		donate_files += sbi->donate_files;
+
+		spin_lock(&f2fs_list_lock);
+		p = p->next;
+		mutex_unlock(&sbi->umount_mutex);
+	}
+	spin_unlock(&f2fs_list_lock);
+
+	return donate_files;
+}
+
+static unsigned int do_reclaim_caches(struct f2fs_sb_info *sbi,
+				unsigned int reclaim_caches_kb)
+{
+	struct inode *inode;
+	struct f2fs_inode_info *fi;
+	unsigned int nfiles = sbi->donate_files;
+	pgoff_t npages = reclaim_caches_kb >> (PAGE_SHIFT - 10);
+
+	while (npages && nfiles--) {
+		pgoff_t len;
+
+		spin_lock(&sbi->inode_lock[DONATE_INODE]);
+		if (list_empty(&sbi->inode_list[DONATE_INODE])) {
+			spin_unlock(&sbi->inode_lock[DONATE_INODE]);
+			break;
+		}
+		fi = list_first_entry(&sbi->inode_list[DONATE_INODE],
+					struct f2fs_inode_info, gdonate_list);
+		list_move_tail(&fi->gdonate_list, &sbi->inode_list[DONATE_INODE]);
+		inode = igrab(&fi->vfs_inode);
+		spin_unlock(&sbi->inode_lock[DONATE_INODE]);
+
+		if (!inode)
+			continue;
+
+		len = fi->donate_end - fi->donate_start + 1;
+		npages = npages < len ? 0 : npages - len;
+		invalidate_inode_pages2_range(inode->i_mapping,
+					fi->donate_start, fi->donate_end);
+		iput(inode);
+		cond_resched();
+	}
+	return npages << (PAGE_SHIFT - 10);
+}
+
+void f2fs_reclaim_caches(unsigned int reclaim_caches_kb)
+{
+	struct f2fs_sb_info *sbi;
+	struct list_head *p;
+
+	spin_lock(&f2fs_list_lock);
+	p = f2fs_list.next;
+	while (p != &f2fs_list && reclaim_caches_kb) {
+		sbi = list_entry(p, struct f2fs_sb_info, s_list);
+
+		/* stop f2fs_put_super */
+		if (!mutex_trylock(&sbi->umount_mutex)) {
+			p = p->next;
+			continue;
+		}
+		spin_unlock(&f2fs_list_lock);
+
+		reclaim_caches_kb = do_reclaim_caches(sbi, reclaim_caches_kb);
+
+		spin_lock(&f2fs_list_lock);
+		p = p->next;
+		mutex_unlock(&sbi->umount_mutex);
+	}
+	spin_unlock(&f2fs_list_lock);
+}
+
 void f2fs_join_shrinker(struct f2fs_sb_info *sbi)
 {
 	spin_lock(&f2fs_list_lock);
diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
index b419555e1ea7..b27336acf519 100644
--- a/fs/f2fs/sysfs.c
+++ b/fs/f2fs/sysfs.c
@@ -916,6 +916,39 @@ static struct f2fs_base_attr f2fs_base_attr_##_name = {		\
 	.show	= f2fs_feature_show,				\
 }
 
+static ssize_t f2fs_tune_show(struct f2fs_base_attr *a, char *buf)
+{
+	unsigned int res = 0;
+
+	if (!strcmp(a->attr.name, "reclaim_caches_kb"))
+		res = f2fs_donate_files();
+
+	return sysfs_emit(buf, "%u\n", res);
+}
+
+static ssize_t f2fs_tune_store(struct f2fs_base_attr *a,
+			const char *buf, size_t count)
+{
+	unsigned long t;
+	int ret;
+
+	ret = kstrtoul(skip_spaces(buf), 0, &t);
+	if (ret)
+		return ret;
+
+	if (!strcmp(a->attr.name, "reclaim_caches_kb"))
+		f2fs_reclaim_caches(t);
+
+	return count;
+}
+
+#define F2FS_TUNE_RW_ATTR(_name)				\
+static struct f2fs_base_attr f2fs_base_attr_##_name = {		\
+	.attr = {.name = __stringify(_name), .mode = 0644 },	\
+	.show	= f2fs_tune_show,				\
+	.store	= f2fs_tune_store,				\
+}
+
 static ssize_t f2fs_sb_feature_show(struct f2fs_attr *a,
 		struct f2fs_sb_info *sbi, char *buf)
 {
@@ -1368,6 +1401,14 @@ static struct attribute *f2fs_sb_feat_attrs[] = {
 };
 ATTRIBUTE_GROUPS(f2fs_sb_feat);
 
+F2FS_TUNE_RW_ATTR(reclaim_caches_kb);
+
+static struct attribute *f2fs_tune_attrs[] = {
+	BASE_ATTR_LIST(reclaim_caches_kb),
+	NULL,
+};
+ATTRIBUTE_GROUPS(f2fs_tune);
+
 static const struct sysfs_ops f2fs_attr_ops = {
 	.show	= f2fs_attr_show,
 	.store	= f2fs_attr_store,
@@ -1401,6 +1442,20 @@ static struct kobject f2fs_feat = {
 	.kset	= &f2fs_kset,
 };
 
+static const struct sysfs_ops f2fs_tune_attr_ops = {
+	.show	= f2fs_base_attr_show,
+	.store	= f2fs_base_attr_store,
+};
+
+static const struct kobj_type f2fs_tune_ktype = {
+	.default_groups = f2fs_tune_groups,
+	.sysfs_ops	= &f2fs_tune_attr_ops,
+};
+
+static struct kobject f2fs_tune = {
+	.kset	= &f2fs_kset,
+};
+
 static ssize_t f2fs_stat_attr_show(struct kobject *kobj,
 				struct attribute *attr, char *buf)
 {
@@ -1637,6 +1692,11 @@ int __init f2fs_init_sysfs(void)
 	if (ret)
 		goto put_kobject;
 
+	ret = kobject_init_and_add(&f2fs_tune, &f2fs_tune_ktype,
+				   NULL, "tuning");
+	if (ret)
+		goto put_kobject;
+
 	f2fs_proc_root = proc_mkdir("fs/f2fs", NULL);
 	if (!f2fs_proc_root) {
 		ret = -ENOMEM;
@@ -1644,7 +1704,9 @@ int __init f2fs_init_sysfs(void)
 	}
 
 	return 0;
+
 put_kobject:
+	kobject_put(&f2fs_tune);
 	kobject_put(&f2fs_feat);
 	kset_unregister(&f2fs_kset);
 	return ret;
@@ -1652,6 +1714,7 @@ int __init f2fs_init_sysfs(void)
 
 void f2fs_exit_sysfs(void)
 {
+	kobject_put(&f2fs_tune);
 	kobject_put(&f2fs_feat);
 	kset_unregister(&f2fs_kset);
 	remove_proc_entry("fs/f2fs", NULL);
-- 
2.48.1.601.g30ceb7b040-goog

Re: [f2fs-dev] [PATCH 1/2] f2fs: keep POSIX_FADV_NOREUSE ranges

[Chao Yu]

On 2/12/25 10:31, Jaegeuk Kim via Linux-f2fs-devel wrote:
> This patch records POSIX_FADV_NOREUSE ranges for users to reclaim the caches
> instantly off from LRU.
> 
> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>

Reviewed-by: Chao Yu <chao@kernel.org>

Thanks,

Re: [f2fs-dev] [PATCH 2/2] f2fs: add a sysfs entry to reclaim POSIX_FADV_NOREUSE pages

[Chao Yu]

On 2/12/25 10:31, Jaegeuk Kim via Linux-f2fs-devel wrote:
> 1. fadvise(fd1, POSIX_FADV_NOREUSE, {0,3});
> 2. fadvise(fd2, POSIX_FADV_NOREUSE, {1,2});
> 3. fadvise(fd3, POSIX_FADV_NOREUSE, {3,1});
> 4. echo 1024 > /sys/fs/f2fs/tuning/reclaim_caches_kb
> 
> This gives a way to reclaim file-backed pages by iterating all f2fs mounts until
> reclaiming 1MB page cache ranges, registered by #1, #2, and #3.
> 
> 5. cat /sys/fs/f2fs/tuning/reclaim_caches_kb
> -> gives total number of registered file ranges.
> 
> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>

Reviewed-by: Chao Yu <chao@kernel.org>

Thanks,

Re: [PATCH 0/2 v9] reclaim file-backed pages given POSIX_FADV_NOREUSE

[Christoph Hellwig]

This still has a file system sysfs HACK, you're still not Ccing the
right list, etc.

Can you pleae at least try to get it right?

On Wed, Feb 12, 2025 at 02:31:55AM +0000, Jaegeuk Kim wrote:
> This patch series does not add new API, but implements POSIX_FADV_NOREUSE where
> it keeps the page ranges in the f2fs superblock and add a way for users to
> reclaim the pages manually.
> 
> Change log from v8:
>  - remove new APIs, but use fadvise(POSIX_FADV_NOREUSE)
> 
> Jaegeuk Kim (2):
>   f2fs: keep POSIX_FADV_NOREUSE ranges
>   f2fs: add a sysfs entry to reclaim POSIX_FADV_NOREUSE pages
> 
>  Documentation/ABI/testing/sysfs-fs-f2fs |  7 ++
>  fs/f2fs/debug.c                         |  3 +
>  fs/f2fs/f2fs.h                          | 14 +++-
>  fs/f2fs/file.c                          | 60 +++++++++++++++--
>  fs/f2fs/inode.c                         | 14 ++++
>  fs/f2fs/shrinker.c                      | 90 +++++++++++++++++++++++++
>  fs/f2fs/super.c                         |  1 +
>  fs/f2fs/sysfs.c                         | 63 +++++++++++++++++
>  8 files changed, 246 insertions(+), 6 deletions(-)
> 
> -- 
> 2.48.1.601.g30ceb7b040-goog
> 
> 
---end quoted text---

Re: [PATCH 0/2 v9] reclaim file-backed pages given POSIX_FADV_NOREUSE

[Jaegeuk Kim]

On 02/18, Christoph Hellwig wrote:
> This still has a file system sysfs HACK, you're still not Ccing the
> right list, etc.
> 
> Can you pleae at least try to get it right?

I was modifying the patch having 1) declaring a static global list, 2) adding
some fields to superblock and inode structures to keep the given range in the
inode through fadvise, 3) adding hooks in evict_inode to handle the list, 4)
exploring which sysfs entry in MM to reclaim them explicitly.

But, I stopped at some point, as it looks not good at all. Moreover, I started
to be questioning why not just doing in F2FS back, given sementically I didn't
change anything  on general behavior of fadvise(POSIX_FADV_NOREUSE), IIUC, which
moves pages back to LRU. In addiiton to that, I'd like to keep the range hint in
a filesystem and provide a sysfs entry to manage the hints additionally.
In addition, I don't think there's rule that filesystem cannot reclaim file-back
pages, as it just uses the exported symbol that all filesystems are using in
various different purpose. Hence, I don't get the point which is wrong.

Thanks,

> 
> On Wed, Feb 12, 2025 at 02:31:55AM +0000, Jaegeuk Kim wrote:
> > This patch series does not add new API, but implements POSIX_FADV_NOREUSE where
> > it keeps the page ranges in the f2fs superblock and add a way for users to
> > reclaim the pages manually.
> > 
> > Change log from v8:
> >  - remove new APIs, but use fadvise(POSIX_FADV_NOREUSE)
> > 
> > Jaegeuk Kim (2):
> >   f2fs: keep POSIX_FADV_NOREUSE ranges
> >   f2fs: add a sysfs entry to reclaim POSIX_FADV_NOREUSE pages
> > 
> >  Documentation/ABI/testing/sysfs-fs-f2fs |  7 ++
> >  fs/f2fs/debug.c                         |  3 +
> >  fs/f2fs/f2fs.h                          | 14 +++-
> >  fs/f2fs/file.c                          | 60 +++++++++++++++--
> >  fs/f2fs/inode.c                         | 14 ++++
> >  fs/f2fs/shrinker.c                      | 90 +++++++++++++++++++++++++
> >  fs/f2fs/super.c                         |  1 +
> >  fs/f2fs/sysfs.c                         | 63 +++++++++++++++++
> >  8 files changed, 246 insertions(+), 6 deletions(-)
> > 
> > -- 
> > 2.48.1.601.g30ceb7b040-goog
> > 
> > 
> ---end quoted text---

Re: [f2fs-dev] [PATCH 0/2 v9] reclaim file-backed pages given POSIX_FADV_NOREUSE

[patchwork-bot+f2fs]

Hello:

This series was applied to jaegeuk/f2fs.git (dev)
by Jaegeuk Kim <jaegeuk@kernel.org>:

On Wed, 12 Feb 2025 02:31:55 +0000 you wrote:
> This patch series does not add new API, but implements POSIX_FADV_NOREUSE where
> it keeps the page ranges in the f2fs superblock and add a way for users to
> reclaim the pages manually.
> 
> Change log from v8:
>  - remove new APIs, but use fadvise(POSIX_FADV_NOREUSE)
> 
> [...]

Here is the summary with links:
  - [f2fs-dev,1/2] f2fs: keep POSIX_FADV_NOREUSE ranges
    https://git.kernel.org/jaegeuk/f2fs/c/ef0c333cad8d
  - [f2fs-dev,2/2] f2fs: add a sysfs entry to reclaim POSIX_FADV_NOREUSE pages
    https://git.kernel.org/jaegeuk/f2fs/c/a907f3a68ee2

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html