* [f2fs-dev] [PATCH 2/6] f2fs: move internal functions into extent_cache.c
2022-12-05 18:54 [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly Jaegeuk Kim
@ 2022-12-05 18:54 ` Jaegeuk Kim
2022-12-08 14:46 ` Chao Yu
2022-12-05 18:54 ` [f2fs-dev] [PATCH 3/6] f2fs: remove unnecessary __init_extent_tree Jaegeuk Kim
` (4 subsequent siblings)
5 siblings, 1 reply; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-05 18:54 UTC (permalink / raw)
To: linux-kernel, linux-f2fs-devel; +Cc: Jaegeuk Kim
No functional change.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
fs/f2fs/extent_cache.c | 88 +++++++++++++++++++++++++++++++++++++-----
fs/f2fs/f2fs.h | 69 +--------------------------------
2 files changed, 81 insertions(+), 76 deletions(-)
diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
index 8cd87aee0292..2a8e31e6d518 100644
--- a/fs/f2fs/extent_cache.c
+++ b/fs/f2fs/extent_cache.c
@@ -15,6 +15,77 @@
#include "node.h"
#include <trace/events/f2fs.h>
+static void __set_extent_info(struct extent_info *ei,
+ unsigned int fofs, unsigned int len,
+ block_t blk, bool keep_clen)
+{
+ ei->fofs = fofs;
+ ei->blk = blk;
+ ei->len = len;
+
+ if (keep_clen)
+ return;
+
+#ifdef CONFIG_F2FS_FS_COMPRESSION
+ ei->c_len = 0;
+#endif
+}
+
+static bool f2fs_may_extent_tree(struct inode *inode)
+{
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+
+ /*
+ * for recovered files during mount do not create extents
+ * if shrinker is not registered.
+ */
+ if (list_empty(&sbi->s_list))
+ return false;
+
+ if (!test_opt(sbi, READ_EXTENT_CACHE) ||
+ is_inode_flag_set(inode, FI_NO_EXTENT) ||
+ (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
+ !f2fs_sb_has_readonly(sbi)))
+ return false;
+
+ return S_ISREG(inode->i_mode);
+}
+
+static void __try_update_largest_extent(struct extent_tree *et,
+ struct extent_node *en)
+{
+ if (en->ei.len <= et->largest.len)
+ return;
+
+ et->largest = en->ei;
+ et->largest_updated = true;
+}
+
+static bool __is_extent_mergeable(struct extent_info *back,
+ struct extent_info *front)
+{
+#ifdef CONFIG_F2FS_FS_COMPRESSION
+ if (back->c_len && back->len != back->c_len)
+ return false;
+ if (front->c_len && front->len != front->c_len)
+ return false;
+#endif
+ return (back->fofs + back->len == front->fofs &&
+ back->blk + back->len == front->blk);
+}
+
+static bool __is_back_mergeable(struct extent_info *cur,
+ struct extent_info *back)
+{
+ return __is_extent_mergeable(back, cur);
+}
+
+static bool __is_front_mergeable(struct extent_info *cur,
+ struct extent_info *front)
+{
+ return __is_extent_mergeable(cur, front);
+}
+
static struct rb_entry *__lookup_rb_tree_fast(struct rb_entry *cached_re,
unsigned int ofs)
{
@@ -591,16 +662,16 @@ static void f2fs_update_extent_tree_range(struct inode *inode,
if (end < org_end && org_end - end >= F2FS_MIN_EXTENT_LEN) {
if (parts) {
- set_extent_info(&ei, end,
- end - dei.fofs + dei.blk,
- org_end - end);
+ __set_extent_info(&ei,
+ end, org_end - end,
+ end - dei.fofs + dei.blk, false);
en1 = __insert_extent_tree(sbi, et, &ei,
NULL, NULL, true);
next_en = en1;
} else {
- en->ei.fofs = end;
- en->ei.blk += end - dei.fofs;
- en->ei.len -= end - dei.fofs;
+ __set_extent_info(&en->ei,
+ end, en->ei.len - (end - dei.fofs),
+ en->ei.blk + (end - dei.fofs), true);
next_en = en;
}
parts++;
@@ -632,8 +703,7 @@ static void f2fs_update_extent_tree_range(struct inode *inode,
/* 3. update extent in extent cache */
if (blkaddr) {
-
- set_extent_info(&ei, fofs, blkaddr, len);
+ __set_extent_info(&ei, fofs, len, blkaddr, false);
if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
__insert_extent_tree(sbi, et, &ei,
insert_p, insert_parent, leftmost);
@@ -692,7 +762,7 @@ void f2fs_update_extent_tree_range_compressed(struct inode *inode,
if (en)
goto unlock_out;
- set_extent_info(&ei, fofs, blkaddr, llen);
+ __set_extent_info(&ei, fofs, llen, blkaddr, true);
ei.c_len = c_len;
if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 1c39f8145b61..04fdf010bb77 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -621,7 +621,7 @@ struct rb_entry {
struct extent_info {
unsigned int fofs; /* start offset in a file */
unsigned int len; /* length of the extent */
- u32 blk; /* start block address of the extent */
+ block_t blk; /* start block address of the extent */
#ifdef CONFIG_F2FS_FS_COMPRESSION
unsigned int c_len; /* physical extent length of compressed blocks */
#endif
@@ -846,17 +846,6 @@ static inline void set_raw_read_extent(struct extent_info *ext,
i_ext->len = cpu_to_le32(ext->len);
}
-static inline void set_extent_info(struct extent_info *ei, unsigned int fofs,
- u32 blk, unsigned int len)
-{
- ei->fofs = fofs;
- ei->blk = blk;
- ei->len = len;
-#ifdef CONFIG_F2FS_FS_COMPRESSION
- ei->c_len = 0;
-#endif
-}
-
static inline bool __is_discard_mergeable(struct discard_info *back,
struct discard_info *front, unsigned int max_len)
{
@@ -876,41 +865,6 @@ static inline bool __is_discard_front_mergeable(struct discard_info *cur,
return __is_discard_mergeable(cur, front, max_len);
}
-static inline bool __is_extent_mergeable(struct extent_info *back,
- struct extent_info *front)
-{
-#ifdef CONFIG_F2FS_FS_COMPRESSION
- if (back->c_len && back->len != back->c_len)
- return false;
- if (front->c_len && front->len != front->c_len)
- return false;
-#endif
- return (back->fofs + back->len == front->fofs &&
- back->blk + back->len == front->blk);
-}
-
-static inline bool __is_back_mergeable(struct extent_info *cur,
- struct extent_info *back)
-{
- return __is_extent_mergeable(back, cur);
-}
-
-static inline bool __is_front_mergeable(struct extent_info *cur,
- struct extent_info *front)
-{
- return __is_extent_mergeable(cur, front);
-}
-
-extern void f2fs_mark_inode_dirty_sync(struct inode *inode, bool sync);
-static inline void __try_update_largest_extent(struct extent_tree *et,
- struct extent_node *en)
-{
- if (en->ei.len > et->largest.len) {
- et->largest = en->ei;
- et->largest_updated = true;
- }
-}
-
/*
* For free nid management
*/
@@ -2581,6 +2535,7 @@ static inline block_t __start_sum_addr(struct f2fs_sb_info *sbi)
return le32_to_cpu(F2FS_CKPT(sbi)->cp_pack_start_sum);
}
+extern void f2fs_mark_inode_dirty_sync(struct inode *inode, bool sync);
static inline int inc_valid_node_count(struct f2fs_sb_info *sbi,
struct inode *inode, bool is_inode)
{
@@ -4403,26 +4358,6 @@ F2FS_FEATURE_FUNCS(casefold, CASEFOLD);
F2FS_FEATURE_FUNCS(compression, COMPRESSION);
F2FS_FEATURE_FUNCS(readonly, RO);
-static inline bool f2fs_may_extent_tree(struct inode *inode)
-{
- struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
-
- if (!test_opt(sbi, READ_EXTENT_CACHE) ||
- is_inode_flag_set(inode, FI_NO_EXTENT) ||
- (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
- !f2fs_sb_has_readonly(sbi)))
- return false;
-
- /*
- * for recovered files during mount do not create extents
- * if shrinker is not registered.
- */
- if (list_empty(&sbi->s_list))
- return false;
-
- return S_ISREG(inode->i_mode);
-}
-
#ifdef CONFIG_BLK_DEV_ZONED
static inline bool f2fs_blkz_is_seq(struct f2fs_sb_info *sbi, int devi,
block_t blkaddr)
--
2.39.0.rc0.267.gcb52ba06e7-goog
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply related [flat|nested] 19+ messages in thread* [f2fs-dev] [PATCH 3/6] f2fs: remove unnecessary __init_extent_tree
2022-12-05 18:54 [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly Jaegeuk Kim
2022-12-05 18:54 ` [f2fs-dev] [PATCH 2/6] f2fs: move internal functions into extent_cache.c Jaegeuk Kim
@ 2022-12-05 18:54 ` Jaegeuk Kim
2022-12-08 14:48 ` Chao Yu
2022-12-05 18:54 ` [f2fs-dev] [PATCH 4/6] f2fs: refactor extent_cache to support for read and more Jaegeuk Kim
` (3 subsequent siblings)
5 siblings, 1 reply; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-05 18:54 UTC (permalink / raw)
To: linux-kernel, linux-f2fs-devel; +Cc: Jaegeuk Kim
Added into the caller.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
fs/f2fs/extent_cache.c | 21 +++++----------------
1 file changed, 5 insertions(+), 16 deletions(-)
diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
index 2a8e31e6d518..c6810347e205 100644
--- a/fs/f2fs/extent_cache.c
+++ b/fs/f2fs/extent_cache.c
@@ -386,21 +386,6 @@ static struct extent_tree *__grab_extent_tree(struct inode *inode)
return et;
}
-static struct extent_node *__init_extent_tree(struct f2fs_sb_info *sbi,
- struct extent_tree *et, struct extent_info *ei)
-{
- struct rb_node **p = &et->root.rb_root.rb_node;
- struct extent_node *en;
-
- en = __attach_extent_node(sbi, et, ei, NULL, p, true);
- if (!en)
- return NULL;
-
- et->largest = en->ei;
- et->cached_en = en;
- return en;
-}
-
static unsigned int __free_extent_tree(struct f2fs_sb_info *sbi,
struct extent_tree *et)
{
@@ -460,8 +445,12 @@ static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage)
if (atomic_read(&et->node_cnt))
goto out;
- en = __init_extent_tree(sbi, et, &ei);
+ en = __attach_extent_node(sbi, et, &ei, NULL,
+ &et->root.rb_root.rb_node, true);
if (en) {
+ et->largest = en->ei;
+ et->cached_en = en;
+
spin_lock(&sbi->extent_lock);
list_add_tail(&en->list, &sbi->extent_list);
spin_unlock(&sbi->extent_lock);
--
2.39.0.rc0.267.gcb52ba06e7-goog
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply related [flat|nested] 19+ messages in thread* [f2fs-dev] [PATCH 4/6] f2fs: refactor extent_cache to support for read and more
2022-12-05 18:54 [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly Jaegeuk Kim
2022-12-05 18:54 ` [f2fs-dev] [PATCH 2/6] f2fs: move internal functions into extent_cache.c Jaegeuk Kim
2022-12-05 18:54 ` [f2fs-dev] [PATCH 3/6] f2fs: remove unnecessary __init_extent_tree Jaegeuk Kim
@ 2022-12-05 18:54 ` Jaegeuk Kim
2022-12-06 19:05 ` [f2fs-dev] [PATCH 4/6 v2] " Jaegeuk Kim
2022-12-05 18:54 ` [f2fs-dev] [PATCH 5/6] f2fs: allocate the extent_cache by default Jaegeuk Kim
` (2 subsequent siblings)
5 siblings, 1 reply; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-05 18:54 UTC (permalink / raw)
To: linux-kernel, linux-f2fs-devel; +Cc: Jaegeuk Kim
This patch prepares extent_cache to be ready for addition.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
fs/f2fs/data.c | 20 +-
fs/f2fs/debug.c | 65 +++--
fs/f2fs/extent_cache.c | 463 +++++++++++++++++++++---------------
fs/f2fs/f2fs.h | 108 ++++++---
fs/f2fs/file.c | 8 +-
fs/f2fs/gc.c | 4 +-
fs/f2fs/inode.c | 6 +-
fs/f2fs/node.c | 8 +-
fs/f2fs/segment.c | 3 +-
fs/f2fs/shrinker.c | 19 +-
include/trace/events/f2fs.h | 62 +++--
11 files changed, 464 insertions(+), 302 deletions(-)
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 35c19248b1e2..75abd450730b 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -1126,7 +1126,7 @@ void f2fs_update_data_blkaddr(struct dnode_of_data *dn, block_t blkaddr)
{
dn->data_blkaddr = blkaddr;
f2fs_set_data_blkaddr(dn);
- f2fs_update_extent_cache(dn);
+ f2fs_update_read_extent_cache(dn);
}
/* dn->ofs_in_node will be returned with up-to-date last block pointer */
@@ -1195,7 +1195,7 @@ int f2fs_get_block(struct dnode_of_data *dn, pgoff_t index)
struct extent_info ei = {0, };
struct inode *inode = dn->inode;
- if (f2fs_lookup_extent_cache(inode, index, &ei)) {
+ if (f2fs_lookup_read_extent_cache(inode, index, &ei)) {
dn->data_blkaddr = ei.blk + index - ei.fofs;
return 0;
}
@@ -1217,7 +1217,7 @@ struct page *f2fs_get_read_data_page(struct inode *inode, pgoff_t index,
if (!page)
return ERR_PTR(-ENOMEM);
- if (f2fs_lookup_extent_cache(inode, index, &ei)) {
+ if (f2fs_lookup_read_extent_cache(inode, index, &ei)) {
dn.data_blkaddr = ei.blk + index - ei.fofs;
if (!f2fs_is_valid_blkaddr(F2FS_I_SB(inode), dn.data_blkaddr,
DATA_GENERIC_ENHANCE_READ)) {
@@ -1485,7 +1485,7 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
pgofs = (pgoff_t)map->m_lblk;
end = pgofs + maxblocks;
- if (!create && f2fs_lookup_extent_cache(inode, pgofs, &ei)) {
+ if (!create && f2fs_lookup_read_extent_cache(inode, pgofs, &ei)) {
if (f2fs_lfs_mode(sbi) && flag == F2FS_GET_BLOCK_DIO &&
map->m_may_create)
goto next_dnode;
@@ -1695,7 +1695,7 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
if (map->m_flags & F2FS_MAP_MAPPED) {
unsigned int ofs = start_pgofs - map->m_lblk;
- f2fs_update_extent_cache_range(&dn,
+ f2fs_update_read_extent_cache_range(&dn,
start_pgofs, map->m_pblk + ofs,
map->m_len - ofs);
}
@@ -1740,7 +1740,7 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
if (map->m_flags & F2FS_MAP_MAPPED) {
unsigned int ofs = start_pgofs - map->m_lblk;
- f2fs_update_extent_cache_range(&dn,
+ f2fs_update_read_extent_cache_range(&dn,
start_pgofs, map->m_pblk + ofs,
map->m_len - ofs);
}
@@ -2201,7 +2201,7 @@ int f2fs_read_multi_pages(struct compress_ctx *cc, struct bio **bio_ret,
if (f2fs_cluster_is_empty(cc))
goto out;
- if (f2fs_lookup_extent_cache(inode, start_idx, &ei))
+ if (f2fs_lookup_read_extent_cache(inode, start_idx, &ei))
from_dnode = false;
if (!from_dnode)
@@ -2635,7 +2635,7 @@ int f2fs_do_write_data_page(struct f2fs_io_info *fio)
set_new_dnode(&dn, inode, NULL, NULL, 0);
if (need_inplace_update(fio) &&
- f2fs_lookup_extent_cache(inode, page->index, &ei)) {
+ f2fs_lookup_read_extent_cache(inode, page->index, &ei)) {
fio->old_blkaddr = ei.blk + page->index - ei.fofs;
if (!f2fs_is_valid_blkaddr(fio->sbi, fio->old_blkaddr,
@@ -3359,7 +3359,7 @@ static int prepare_write_begin(struct f2fs_sb_info *sbi,
} else if (locked) {
err = f2fs_get_block(&dn, index);
} else {
- if (f2fs_lookup_extent_cache(inode, index, &ei)) {
+ if (f2fs_lookup_read_extent_cache(inode, index, &ei)) {
dn.data_blkaddr = ei.blk + index - ei.fofs;
} else {
/* hole case */
@@ -3400,7 +3400,7 @@ static int __find_data_block(struct inode *inode, pgoff_t index,
set_new_dnode(&dn, inode, ipage, ipage, 0);
- if (f2fs_lookup_extent_cache(inode, index, &ei)) {
+ if (f2fs_lookup_read_extent_cache(inode, index, &ei)) {
dn.data_blkaddr = ei.blk + index - ei.fofs;
} else {
/* hole case */
diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index a216dcdf6941..a9baa121d829 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -72,15 +72,23 @@ static void update_general_status(struct f2fs_sb_info *sbi)
si->main_area_zones = si->main_area_sections /
le32_to_cpu(raw_super->secs_per_zone);
- /* validation check of the segment numbers */
+ /* general extent cache stats */
+ for (i = 0; i < NR_EXTENT_CACHES; i++) {
+ struct extent_tree_info *eti = &sbi->extent_tree[i];
+
+ si->hit_cached[i] = atomic64_read(&sbi->read_hit_cached[i]);
+ si->hit_rbtree[i] = atomic64_read(&sbi->read_hit_rbtree[i]);
+ si->total_ext[i] = atomic64_read(&sbi->total_hit_ext[i]);
+ si->hit_total[i] = si->hit_cached[i] + si->hit_rbtree[i];
+ si->ext_tree[i] = atomic_read(&eti->total_ext_tree);
+ si->zombie_tree[i] = atomic_read(&eti->total_zombie_tree);
+ si->ext_node[i] = atomic_read(&eti->total_ext_node);
+ }
+ /* read extent_cache only */
si->hit_largest = atomic64_read(&sbi->read_hit_largest);
- si->hit_cached = atomic64_read(&sbi->read_hit_cached);
- si->hit_rbtree = atomic64_read(&sbi->read_hit_rbtree);
- si->hit_total = si->hit_largest + si->hit_cached + si->hit_rbtree;
- si->total_ext = atomic64_read(&sbi->total_hit_ext);
- si->ext_tree = atomic_read(&sbi->total_ext_tree);
- si->zombie_tree = atomic_read(&sbi->total_zombie_tree);
- si->ext_node = atomic_read(&sbi->total_ext_node);
+ si->hit_total[EX_READ] += si->hit_largest;
+
+ /* validation check of the segment numbers */
si->ndirty_node = get_pages(sbi, F2FS_DIRTY_NODES);
si->ndirty_dent = get_pages(sbi, F2FS_DIRTY_DENTS);
si->ndirty_meta = get_pages(sbi, F2FS_DIRTY_META);
@@ -294,10 +302,16 @@ static void update_mem_info(struct f2fs_sb_info *sbi)
sizeof(struct nat_entry_set);
for (i = 0; i < MAX_INO_ENTRY; i++)
si->cache_mem += sbi->im[i].ino_num * sizeof(struct ino_entry);
- si->cache_mem += atomic_read(&sbi->total_ext_tree) *
+
+ for (i = 0; i < NR_EXTENT_CACHES; i++) {
+ struct extent_tree_info *eti = &sbi->extent_tree[i];
+
+ si->ext_mem[i] = atomic_read(&eti->total_ext_tree) *
sizeof(struct extent_tree);
- si->cache_mem += atomic_read(&sbi->total_ext_node) *
+ si->ext_mem[i] += atomic_read(&eti->total_ext_node) *
sizeof(struct extent_node);
+ si->cache_mem += si->ext_mem[i];
+ }
si->page_mem = 0;
if (sbi->node_inode) {
@@ -490,16 +504,18 @@ static int stat_show(struct seq_file *s, void *v)
si->bg_node_blks);
seq_printf(s, "BG skip : IO: %u, Other: %u\n",
si->io_skip_bggc, si->other_skip_bggc);
- seq_puts(s, "\nExtent Cache:\n");
+ seq_puts(s, "\nExtent Cache (Read):\n");
seq_printf(s, " - Hit Count: L1-1:%llu L1-2:%llu L2:%llu\n",
- si->hit_largest, si->hit_cached,
- si->hit_rbtree);
+ si->hit_largest, si->hit_cached[EX_READ],
+ si->hit_rbtree[EX_READ]);
seq_printf(s, " - Hit Ratio: %llu%% (%llu / %llu)\n",
- !si->total_ext ? 0 :
- div64_u64(si->hit_total * 100, si->total_ext),
- si->hit_total, si->total_ext);
+ !si->total_ext[EX_READ] ? 0 :
+ div64_u64(si->hit_total[EX_READ] * 100,
+ si->total_ext[EX_READ]),
+ si->hit_total[EX_READ], si->total_ext[EX_READ]);
seq_printf(s, " - Inner Struct Count: tree: %d(%d), node: %d\n",
- si->ext_tree, si->zombie_tree, si->ext_node);
+ si->ext_tree[EX_READ], si->zombie_tree[EX_READ],
+ si->ext_node[EX_READ]);
seq_puts(s, "\nBalancing F2FS Async:\n");
seq_printf(s, " - DIO (R: %4d, W: %4d)\n",
si->nr_dio_read, si->nr_dio_write);
@@ -566,8 +582,10 @@ static int stat_show(struct seq_file *s, void *v)
(si->base_mem + si->cache_mem + si->page_mem) >> 10);
seq_printf(s, " - static: %llu KB\n",
si->base_mem >> 10);
- seq_printf(s, " - cached: %llu KB\n",
+ seq_printf(s, " - cached all: %llu KB\n",
si->cache_mem >> 10);
+ seq_printf(s, " - read extent cache: %llu KB\n",
+ si->ext_mem[EX_READ] >> 10);
seq_printf(s, " - paged : %llu KB\n",
si->page_mem >> 10);
}
@@ -600,10 +618,15 @@ int f2fs_build_stats(struct f2fs_sb_info *sbi)
si->sbi = sbi;
sbi->stat_info = si;
- atomic64_set(&sbi->total_hit_ext, 0);
- atomic64_set(&sbi->read_hit_rbtree, 0);
+ /* general extent cache stats */
+ for (i = 0; i < NR_EXTENT_CACHES; i++) {
+ atomic64_set(&sbi->total_hit_ext[i], 0);
+ atomic64_set(&sbi->read_hit_rbtree[i], 0);
+ atomic64_set(&sbi->read_hit_cached[i], 0);
+ }
+
+ /* read extent_cache only */
atomic64_set(&sbi->read_hit_largest, 0);
- atomic64_set(&sbi->read_hit_cached, 0);
atomic_set(&sbi->inline_xattr, 0);
atomic_set(&sbi->inline_inode, 0);
diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
index c6810347e205..654a14ab8977 100644
--- a/fs/f2fs/extent_cache.c
+++ b/fs/f2fs/extent_cache.c
@@ -17,21 +17,37 @@
static void __set_extent_info(struct extent_info *ei,
unsigned int fofs, unsigned int len,
- block_t blk, bool keep_clen)
+ block_t blk, bool keep_clen,
+ enum extent_type type)
{
ei->fofs = fofs;
- ei->blk = blk;
ei->len = len;
- if (keep_clen)
- return;
-
+ if (type == EX_READ) {
+ ei->blk = blk;
+ if (keep_clen)
+ return;
#ifdef CONFIG_F2FS_FS_COMPRESSION
- ei->c_len = 0;
+ ei->c_len = 0;
#endif
+ }
+}
+
+static bool __may_read_extent_tree(struct inode *inode)
+{
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+
+ if (!test_opt(sbi, READ_EXTENT_CACHE))
+ return false;
+ if (is_inode_flag_set(inode, FI_NO_EXTENT))
+ return false;
+ if (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
+ !f2fs_sb_has_readonly(sbi))
+ return false;
+ return S_ISREG(inode->i_mode);
}
-static bool f2fs_may_extent_tree(struct inode *inode)
+static bool __may_extent_tree(struct inode *inode, enum extent_type type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
@@ -42,18 +58,16 @@ static bool f2fs_may_extent_tree(struct inode *inode)
if (list_empty(&sbi->s_list))
return false;
- if (!test_opt(sbi, READ_EXTENT_CACHE) ||
- is_inode_flag_set(inode, FI_NO_EXTENT) ||
- (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
- !f2fs_sb_has_readonly(sbi)))
- return false;
-
- return S_ISREG(inode->i_mode);
+ if (type == EX_READ)
+ return __may_read_extent_tree(inode);
+ return false;
}
static void __try_update_largest_extent(struct extent_tree *et,
struct extent_node *en)
{
+ if (et->type != EX_READ)
+ return;
if (en->ei.len <= et->largest.len)
return;
@@ -62,28 +76,31 @@ static void __try_update_largest_extent(struct extent_tree *et,
}
static bool __is_extent_mergeable(struct extent_info *back,
- struct extent_info *front)
+ struct extent_info *front, enum extent_type type)
{
+ if (type == EX_READ) {
#ifdef CONFIG_F2FS_FS_COMPRESSION
- if (back->c_len && back->len != back->c_len)
- return false;
- if (front->c_len && front->len != front->c_len)
- return false;
+ if (back->c_len && back->len != back->c_len)
+ return false;
+ if (front->c_len && front->len != front->c_len)
+ return false;
#endif
- return (back->fofs + back->len == front->fofs &&
- back->blk + back->len == front->blk);
+ return (back->fofs + back->len == front->fofs &&
+ back->blk + back->len == front->blk);
+ }
+ return false;
}
static bool __is_back_mergeable(struct extent_info *cur,
- struct extent_info *back)
+ struct extent_info *back, enum extent_type type)
{
- return __is_extent_mergeable(back, cur);
+ return __is_extent_mergeable(back, cur, type);
}
static bool __is_front_mergeable(struct extent_info *cur,
- struct extent_info *front)
+ struct extent_info *front, enum extent_type type)
{
- return __is_extent_mergeable(cur, front);
+ return __is_extent_mergeable(cur, front, type);
}
static struct rb_entry *__lookup_rb_tree_fast(struct rb_entry *cached_re,
@@ -308,6 +325,7 @@ static struct extent_node *__attach_extent_node(struct f2fs_sb_info *sbi,
struct rb_node *parent, struct rb_node **p,
bool leftmost)
{
+ struct extent_tree_info *eti = &sbi->extent_tree[et->type];
struct extent_node *en;
en = f2fs_kmem_cache_alloc(extent_node_slab, GFP_ATOMIC, false, sbi);
@@ -321,16 +339,18 @@ static struct extent_node *__attach_extent_node(struct f2fs_sb_info *sbi,
rb_link_node(&en->rb_node, parent, p);
rb_insert_color_cached(&en->rb_node, &et->root, leftmost);
atomic_inc(&et->node_cnt);
- atomic_inc(&sbi->total_ext_node);
+ atomic_inc(&eti->total_ext_node);
return en;
}
static void __detach_extent_node(struct f2fs_sb_info *sbi,
struct extent_tree *et, struct extent_node *en)
{
+ struct extent_tree_info *eti = &sbi->extent_tree[et->type];
+
rb_erase_cached(&en->rb_node, &et->root);
atomic_dec(&et->node_cnt);
- atomic_dec(&sbi->total_ext_node);
+ atomic_dec(&eti->total_ext_node);
if (et->cached_en == en)
et->cached_en = NULL;
@@ -346,42 +366,47 @@ static void __detach_extent_node(struct f2fs_sb_info *sbi,
static void __release_extent_node(struct f2fs_sb_info *sbi,
struct extent_tree *et, struct extent_node *en)
{
- spin_lock(&sbi->extent_lock);
+ struct extent_tree_info *eti = &sbi->extent_tree[et->type];
+
+ spin_lock(&eti->extent_lock);
f2fs_bug_on(sbi, list_empty(&en->list));
list_del_init(&en->list);
- spin_unlock(&sbi->extent_lock);
+ spin_unlock(&eti->extent_lock);
__detach_extent_node(sbi, et, en);
}
-static struct extent_tree *__grab_extent_tree(struct inode *inode)
+static struct extent_tree *__grab_extent_tree(struct inode *inode,
+ enum extent_type type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+ struct extent_tree_info *eti = &sbi->extent_tree[type];
struct extent_tree *et;
nid_t ino = inode->i_ino;
- mutex_lock(&sbi->extent_tree_lock);
- et = radix_tree_lookup(&sbi->extent_tree_root, ino);
+ mutex_lock(&eti->extent_tree_lock);
+ et = radix_tree_lookup(&eti->extent_tree_root, ino);
if (!et) {
et = f2fs_kmem_cache_alloc(extent_tree_slab,
GFP_NOFS, true, NULL);
- f2fs_radix_tree_insert(&sbi->extent_tree_root, ino, et);
+ f2fs_radix_tree_insert(&eti->extent_tree_root, ino, et);
memset(et, 0, sizeof(struct extent_tree));
et->ino = ino;
+ et->type = type;
et->root = RB_ROOT_CACHED;
et->cached_en = NULL;
rwlock_init(&et->lock);
INIT_LIST_HEAD(&et->list);
atomic_set(&et->node_cnt, 0);
- atomic_inc(&sbi->total_ext_tree);
+ atomic_inc(&eti->total_ext_tree);
} else {
- atomic_dec(&sbi->total_zombie_tree);
+ atomic_dec(&eti->total_zombie_tree);
list_del_init(&et->list);
}
- mutex_unlock(&sbi->extent_tree_lock);
+ mutex_unlock(&eti->extent_tree_lock);
/* never died until evict_inode */
- F2FS_I(inode)->extent_tree = et;
+ F2FS_I(inode)->extent_tree[type] = et;
return et;
}
@@ -415,35 +440,38 @@ static void __drop_largest_extent(struct extent_tree *et,
}
/* return true, if inode page is changed */
-static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage)
+static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage,
+ enum extent_type type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+ struct extent_tree_info *eti = &sbi->extent_tree[type];
struct f2fs_extent *i_ext = ipage ? &F2FS_INODE(ipage)->i_ext : NULL;
struct extent_tree *et;
struct extent_node *en;
struct extent_info ei;
- if (!f2fs_may_extent_tree(inode)) {
- /* drop largest extent */
- if (i_ext && i_ext->len) {
+ if (!__may_extent_tree(inode, type)) {
+ /* drop largest read extent */
+ if (type == EX_READ && i_ext && i_ext->len) {
f2fs_wait_on_page_writeback(ipage, NODE, true, true);
i_ext->len = 0;
set_page_dirty(ipage);
- return;
}
- return;
+ goto out;
}
- et = __grab_extent_tree(inode);
+ et = __grab_extent_tree(inode, type);
if (!i_ext || !i_ext->len)
- return;
+ goto out;
+
+ BUG_ON(type != EX_READ);
get_read_extent_info(&ei, i_ext);
write_lock(&et->lock);
if (atomic_read(&et->node_cnt))
- goto out;
+ goto unlock_out;
en = __attach_extent_node(sbi, et, &ei, NULL,
&et->root.rb_root.rb_node, true);
@@ -451,37 +479,40 @@ static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage)
et->largest = en->ei;
et->cached_en = en;
- spin_lock(&sbi->extent_lock);
- list_add_tail(&en->list, &sbi->extent_list);
- spin_unlock(&sbi->extent_lock);
+ spin_lock(&eti->extent_lock);
+ list_add_tail(&en->list, &eti->extent_list);
+ spin_unlock(&eti->extent_lock);
}
-out:
+unlock_out:
write_unlock(&et->lock);
+out:
+ if (type == EX_READ && !F2FS_I(inode)->extent_tree[EX_READ])
+ set_inode_flag(inode, FI_NO_EXTENT);
}
void f2fs_init_extent_tree(struct inode *inode, struct page *ipage)
{
- __f2fs_init_extent_tree(inode, ipage);
-
- if (!F2FS_I(inode)->extent_tree)
- set_inode_flag(inode, FI_NO_EXTENT);
+ /* initialize read cache */
+ __f2fs_init_extent_tree(inode, ipage, EX_READ);
}
-static bool f2fs_lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
- struct extent_info *ei)
+static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
+ struct extent_info *ei, enum extent_type type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
- struct extent_tree *et = F2FS_I(inode)->extent_tree;
+ struct extent_tree_info *eti = &sbi->extent_tree[type];
+ struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
struct extent_node *en;
bool ret = false;
f2fs_bug_on(sbi, !et);
- trace_f2fs_lookup_extent_tree_start(inode, pgofs);
+ trace_f2fs_lookup_extent_tree_start(inode, pgofs, type);
read_lock(&et->lock);
- if (et->largest.fofs <= pgofs &&
+ if (type == EX_READ &&
+ et->largest.fofs <= pgofs &&
et->largest.fofs + et->largest.len > pgofs) {
*ei = et->largest;
ret = true;
@@ -495,23 +526,24 @@ static bool f2fs_lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
goto out;
if (en == et->cached_en)
- stat_inc_cached_node_hit(sbi);
+ stat_inc_cached_node_hit(sbi, type);
else
- stat_inc_rbtree_node_hit(sbi);
+ stat_inc_rbtree_node_hit(sbi, type);
*ei = en->ei;
- spin_lock(&sbi->extent_lock);
+ spin_lock(&eti->extent_lock);
if (!list_empty(&en->list)) {
- list_move_tail(&en->list, &sbi->extent_list);
+ list_move_tail(&en->list, &eti->extent_list);
et->cached_en = en;
}
- spin_unlock(&sbi->extent_lock);
+ spin_unlock(&eti->extent_lock);
ret = true;
out:
- stat_inc_total_hit(sbi);
+ stat_inc_total_hit(sbi, type);
read_unlock(&et->lock);
- trace_f2fs_lookup_extent_tree_end(inode, pgofs, ei);
+ if (type == EX_READ)
+ trace_f2fs_lookup_read_extent_tree_end(inode, pgofs, ei);
return ret;
}
@@ -520,18 +552,20 @@ static struct extent_node *__try_merge_extent_node(struct f2fs_sb_info *sbi,
struct extent_node *prev_ex,
struct extent_node *next_ex)
{
+ struct extent_tree_info *eti = &sbi->extent_tree[et->type];
struct extent_node *en = NULL;
- if (prev_ex && __is_back_mergeable(ei, &prev_ex->ei)) {
+ if (prev_ex && __is_back_mergeable(ei, &prev_ex->ei, et->type)) {
prev_ex->ei.len += ei->len;
ei = &prev_ex->ei;
en = prev_ex;
}
- if (next_ex && __is_front_mergeable(ei, &next_ex->ei)) {
+ if (next_ex && __is_front_mergeable(ei, &next_ex->ei, et->type)) {
next_ex->ei.fofs = ei->fofs;
- next_ex->ei.blk = ei->blk;
next_ex->ei.len += ei->len;
+ if (et->type == EX_READ)
+ next_ex->ei.blk = ei->blk;
if (en)
__release_extent_node(sbi, et, prev_ex);
@@ -543,12 +577,12 @@ static struct extent_node *__try_merge_extent_node(struct f2fs_sb_info *sbi,
__try_update_largest_extent(et, en);
- spin_lock(&sbi->extent_lock);
+ spin_lock(&eti->extent_lock);
if (!list_empty(&en->list)) {
- list_move_tail(&en->list, &sbi->extent_list);
+ list_move_tail(&en->list, &eti->extent_list);
et->cached_en = en;
}
- spin_unlock(&sbi->extent_lock);
+ spin_unlock(&eti->extent_lock);
return en;
}
@@ -558,6 +592,7 @@ static struct extent_node *__insert_extent_tree(struct f2fs_sb_info *sbi,
struct rb_node *insert_parent,
bool leftmost)
{
+ struct extent_tree_info *eti = &sbi->extent_tree[et->type];
struct rb_node **p;
struct rb_node *parent = NULL;
struct extent_node *en = NULL;
@@ -580,47 +615,50 @@ static struct extent_node *__insert_extent_tree(struct f2fs_sb_info *sbi,
__try_update_largest_extent(et, en);
/* update in global extent list */
- spin_lock(&sbi->extent_lock);
- list_add_tail(&en->list, &sbi->extent_list);
+ spin_lock(&eti->extent_lock);
+ list_add_tail(&en->list, &eti->extent_list);
et->cached_en = en;
- spin_unlock(&sbi->extent_lock);
+ spin_unlock(&eti->extent_lock);
return en;
}
-static void f2fs_update_extent_tree_range(struct inode *inode,
- pgoff_t fofs, block_t blkaddr, unsigned int len)
+static void __update_extent_tree_range(struct inode *inode,
+ struct extent_info *tei, enum extent_type type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
- struct extent_tree *et = F2FS_I(inode)->extent_tree;
+ struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
struct extent_node *en = NULL, *en1 = NULL;
struct extent_node *prev_en = NULL, *next_en = NULL;
struct extent_info ei, dei, prev;
struct rb_node **insert_p = NULL, *insert_parent = NULL;
+ unsigned int fofs = tei->fofs, len = tei->len;
unsigned int end = fofs + len;
- unsigned int pos = (unsigned int)fofs;
bool updated = false;
bool leftmost = false;
if (!et)
return;
- trace_f2fs_update_extent_tree_range(inode, fofs, blkaddr, len, 0);
-
+ if (type == EX_READ)
+ trace_f2fs_update_read_extent_tree_range(inode, fofs, len,
+ tei->blk, 0);
write_lock(&et->lock);
- if (is_inode_flag_set(inode, FI_NO_EXTENT)) {
- write_unlock(&et->lock);
- return;
- }
+ if (type == EX_READ) {
+ if (is_inode_flag_set(inode, FI_NO_EXTENT)) {
+ write_unlock(&et->lock);
+ return;
+ }
- prev = et->largest;
- dei.len = 0;
+ prev = et->largest;
+ dei.len = 0;
- /*
- * drop largest extent before lookup, in case it's already
- * been shrunk from extent tree
- */
- __drop_largest_extent(et, fofs, len);
+ /*
+ * drop largest extent before lookup, in case it's already
+ * been shrunk from extent tree
+ */
+ __drop_largest_extent(et, fofs, len);
+ }
/* 1. lookup first extent node in range [fofs, fofs + len - 1] */
en = (struct extent_node *)f2fs_lookup_rb_tree_ret(&et->root,
@@ -641,26 +679,30 @@ static void f2fs_update_extent_tree_range(struct inode *inode,
dei = en->ei;
org_end = dei.fofs + dei.len;
- f2fs_bug_on(sbi, pos >= org_end);
+ f2fs_bug_on(sbi, fofs >= org_end);
- if (pos > dei.fofs && pos - dei.fofs >= F2FS_MIN_EXTENT_LEN) {
- en->ei.len = pos - en->ei.fofs;
+ if (fofs > dei.fofs && (type != EX_READ ||
+ fofs - dei.fofs >= F2FS_MIN_EXTENT_LEN)) {
+ en->ei.len = fofs - en->ei.fofs;
prev_en = en;
parts = 1;
}
- if (end < org_end && org_end - end >= F2FS_MIN_EXTENT_LEN) {
+ if (end < org_end && (type != EX_READ ||
+ org_end - end >= F2FS_MIN_EXTENT_LEN)) {
if (parts) {
__set_extent_info(&ei,
end, org_end - end,
- end - dei.fofs + dei.blk, false);
+ end - dei.fofs + dei.blk, false,
+ type);
en1 = __insert_extent_tree(sbi, et, &ei,
NULL, NULL, true);
next_en = en1;
} else {
__set_extent_info(&en->ei,
end, en->ei.len - (end - dei.fofs),
- en->ei.blk + (end - dei.fofs), true);
+ en->ei.blk + (end - dei.fofs), true,
+ type);
next_en = en;
}
parts++;
@@ -690,9 +732,11 @@ static void f2fs_update_extent_tree_range(struct inode *inode,
en = next_en;
}
- /* 3. update extent in extent cache */
- if (blkaddr) {
- __set_extent_info(&ei, fofs, len, blkaddr, false);
+ /* 3. update extent in read extent cache */
+ BUG_ON(type != EX_READ);
+
+ if (tei->blk) {
+ __set_extent_info(&ei, fofs, len, tei->blk, false, EX_READ);
if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
__insert_extent_tree(sbi, et, &ei,
insert_p, insert_parent, leftmost);
@@ -722,19 +766,20 @@ static void f2fs_update_extent_tree_range(struct inode *inode,
}
#ifdef CONFIG_F2FS_FS_COMPRESSION
-void f2fs_update_extent_tree_range_compressed(struct inode *inode,
+void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
pgoff_t fofs, block_t blkaddr, unsigned int llen,
unsigned int c_len)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
- struct extent_tree *et = F2FS_I(inode)->extent_tree;
+ struct extent_tree *et = F2FS_I(inode)->extent_tree[EX_READ];
struct extent_node *en = NULL;
struct extent_node *prev_en = NULL, *next_en = NULL;
struct extent_info ei;
struct rb_node **insert_p = NULL, *insert_parent = NULL;
bool leftmost = false;
- trace_f2fs_update_extent_tree_range(inode, fofs, blkaddr, llen, c_len);
+ trace_f2fs_update_read_extent_tree_range(inode, fofs, llen,
+ blkaddr, c_len);
/* it is safe here to check FI_NO_EXTENT w/o et->lock in ro image */
if (is_inode_flag_set(inode, FI_NO_EXTENT))
@@ -751,7 +796,7 @@ void f2fs_update_extent_tree_range_compressed(struct inode *inode,
if (en)
goto unlock_out;
- __set_extent_info(&ei, fofs, llen, blkaddr, true);
+ __set_extent_info(&ei, fofs, llen, blkaddr, true, EX_READ);
ei.c_len = c_len;
if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
@@ -762,24 +807,43 @@ void f2fs_update_extent_tree_range_compressed(struct inode *inode,
}
#endif
-unsigned int f2fs_shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
+static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type type)
{
+ struct extent_info ei;
+
+ if (!__may_extent_tree(dn->inode, type))
+ return;
+
+ ei.fofs = f2fs_start_bidx_of_node(ofs_of_node(dn->node_page), dn->inode) +
+ dn->ofs_in_node;
+ ei.len = 1;
+
+ if (type == EX_READ) {
+ if (dn->data_blkaddr == NEW_ADDR)
+ ei.blk = NULL_ADDR;
+ else
+ ei.blk = dn->data_blkaddr;
+ }
+ __update_extent_tree_range(dn->inode, &ei, type);
+}
+
+static unsigned int __shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink,
+ enum extent_type type)
+{
+ struct extent_tree_info *eti = &sbi->extent_tree[type];
struct extent_tree *et, *next;
struct extent_node *en;
unsigned int node_cnt = 0, tree_cnt = 0;
int remained;
- if (!test_opt(sbi, READ_EXTENT_CACHE))
- return 0;
-
- if (!atomic_read(&sbi->total_zombie_tree))
+ if (!atomic_read(&eti->total_zombie_tree))
goto free_node;
- if (!mutex_trylock(&sbi->extent_tree_lock))
+ if (!mutex_trylock(&eti->extent_tree_lock))
goto out;
/* 1. remove unreferenced extent tree */
- list_for_each_entry_safe(et, next, &sbi->zombie_list, list) {
+ list_for_each_entry_safe(et, next, &eti->zombie_list, list) {
if (atomic_read(&et->node_cnt)) {
write_lock(&et->lock);
node_cnt += __free_extent_tree(sbi, et);
@@ -787,61 +851,100 @@ unsigned int f2fs_shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
}
f2fs_bug_on(sbi, atomic_read(&et->node_cnt));
list_del_init(&et->list);
- radix_tree_delete(&sbi->extent_tree_root, et->ino);
+ radix_tree_delete(&eti->extent_tree_root, et->ino);
kmem_cache_free(extent_tree_slab, et);
- atomic_dec(&sbi->total_ext_tree);
- atomic_dec(&sbi->total_zombie_tree);
+ atomic_dec(&eti->total_ext_tree);
+ atomic_dec(&eti->total_zombie_tree);
tree_cnt++;
if (node_cnt + tree_cnt >= nr_shrink)
goto unlock_out;
cond_resched();
}
- mutex_unlock(&sbi->extent_tree_lock);
+ mutex_unlock(&eti->extent_tree_lock);
free_node:
/* 2. remove LRU extent entries */
- if (!mutex_trylock(&sbi->extent_tree_lock))
+ if (!mutex_trylock(&eti->extent_tree_lock))
goto out;
remained = nr_shrink - (node_cnt + tree_cnt);
- spin_lock(&sbi->extent_lock);
+ spin_lock(&eti->extent_lock);
for (; remained > 0; remained--) {
- if (list_empty(&sbi->extent_list))
+ if (list_empty(&eti->extent_list))
break;
- en = list_first_entry(&sbi->extent_list,
+ en = list_first_entry(&eti->extent_list,
struct extent_node, list);
et = en->et;
if (!write_trylock(&et->lock)) {
/* refresh this extent node's position in extent list */
- list_move_tail(&en->list, &sbi->extent_list);
+ list_move_tail(&en->list, &eti->extent_list);
continue;
}
list_del_init(&en->list);
- spin_unlock(&sbi->extent_lock);
+ spin_unlock(&eti->extent_lock);
__detach_extent_node(sbi, et, en);
write_unlock(&et->lock);
node_cnt++;
- spin_lock(&sbi->extent_lock);
+ spin_lock(&eti->extent_lock);
}
- spin_unlock(&sbi->extent_lock);
+ spin_unlock(&eti->extent_lock);
unlock_out:
- mutex_unlock(&sbi->extent_tree_lock);
+ mutex_unlock(&eti->extent_tree_lock);
out:
- trace_f2fs_shrink_extent_tree(sbi, node_cnt, tree_cnt);
+ trace_f2fs_shrink_extent_tree(sbi, node_cnt, tree_cnt, type);
return node_cnt + tree_cnt;
}
-unsigned int f2fs_destroy_extent_node(struct inode *inode)
+/* read extent cache operations */
+bool f2fs_lookup_read_extent_cache(struct inode *inode, pgoff_t pgofs,
+ struct extent_info *ei)
+{
+ if (!__may_extent_tree(inode, EX_READ))
+ return false;
+
+ return __lookup_extent_tree(inode, pgofs, ei, EX_READ);
+}
+
+void f2fs_update_read_extent_cache(struct dnode_of_data *dn)
+{
+ return __update_extent_cache(dn, EX_READ);
+}
+
+void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
+ pgoff_t fofs, block_t blkaddr, unsigned int len)
+{
+ struct extent_info ei = {
+ .fofs = fofs,
+ .len = len,
+ .blk = blkaddr,
+ };
+
+ if (!__may_extent_tree(dn->inode, EX_READ))
+ return;
+
+ __update_extent_tree_range(dn->inode, &ei, EX_READ);
+}
+
+unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
+{
+ if (!test_opt(sbi, READ_EXTENT_CACHE))
+ return 0;
+
+ return __shrink_extent_tree(sbi, nr_shrink, EX_READ);
+}
+
+static unsigned int __destroy_extent_node(struct inode *inode,
+ enum extent_type type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
- struct extent_tree *et = F2FS_I(inode)->extent_tree;
+ struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
unsigned int node_cnt = 0;
if (!et || !atomic_read(&et->node_cnt))
@@ -854,31 +957,44 @@ unsigned int f2fs_destroy_extent_node(struct inode *inode)
return node_cnt;
}
-void f2fs_drop_extent_tree(struct inode *inode)
+void f2fs_destroy_extent_node(struct inode *inode)
+{
+ __destroy_extent_node(inode, EX_READ);
+}
+
+static void __drop_extent_tree(struct inode *inode, enum extent_type type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
- struct extent_tree *et = F2FS_I(inode)->extent_tree;
+ struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
bool updated = false;
- if (!f2fs_may_extent_tree(inode))
+ if (!__may_extent_tree(inode, type))
return;
write_lock(&et->lock);
- set_inode_flag(inode, FI_NO_EXTENT);
__free_extent_tree(sbi, et);
- if (et->largest.len) {
- et->largest.len = 0;
- updated = true;
+ if (type == EX_READ) {
+ set_inode_flag(inode, FI_NO_EXTENT);
+ if (et->largest.len) {
+ et->largest.len = 0;
+ updated = true;
+ }
}
write_unlock(&et->lock);
if (updated)
f2fs_mark_inode_dirty_sync(inode, true);
}
-void f2fs_destroy_extent_tree(struct inode *inode)
+void f2fs_drop_extent_tree(struct inode *inode)
+{
+ __drop_extent_tree(inode, EX_READ);
+}
+
+static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
- struct extent_tree *et = F2FS_I(inode)->extent_tree;
+ struct extent_tree_info *eti = &sbi->extent_tree[type];
+ struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
unsigned int node_cnt = 0;
if (!et)
@@ -886,76 +1002,49 @@ void f2fs_destroy_extent_tree(struct inode *inode)
if (inode->i_nlink && !is_bad_inode(inode) &&
atomic_read(&et->node_cnt)) {
- mutex_lock(&sbi->extent_tree_lock);
- list_add_tail(&et->list, &sbi->zombie_list);
- atomic_inc(&sbi->total_zombie_tree);
- mutex_unlock(&sbi->extent_tree_lock);
+ mutex_lock(&eti->extent_tree_lock);
+ list_add_tail(&et->list, &eti->zombie_list);
+ atomic_inc(&eti->total_zombie_tree);
+ mutex_unlock(&eti->extent_tree_lock);
return;
}
/* free all extent info belong to this extent tree */
- node_cnt = f2fs_destroy_extent_node(inode);
+ node_cnt = __destroy_extent_node(inode, type);
/* delete extent tree entry in radix tree */
- mutex_lock(&sbi->extent_tree_lock);
+ mutex_lock(&eti->extent_tree_lock);
f2fs_bug_on(sbi, atomic_read(&et->node_cnt));
- radix_tree_delete(&sbi->extent_tree_root, inode->i_ino);
+ radix_tree_delete(&eti->extent_tree_root, inode->i_ino);
kmem_cache_free(extent_tree_slab, et);
- atomic_dec(&sbi->total_ext_tree);
- mutex_unlock(&sbi->extent_tree_lock);
+ atomic_dec(&eti->total_ext_tree);
+ mutex_unlock(&eti->extent_tree_lock);
- F2FS_I(inode)->extent_tree = NULL;
+ F2FS_I(inode)->extent_tree[type] = NULL;
- trace_f2fs_destroy_extent_tree(inode, node_cnt);
+ trace_f2fs_destroy_extent_tree(inode, node_cnt, type);
}
-bool f2fs_lookup_extent_cache(struct inode *inode, pgoff_t pgofs,
- struct extent_info *ei)
-{
- if (!f2fs_may_extent_tree(inode))
- return false;
-
- return f2fs_lookup_extent_tree(inode, pgofs, ei);
-}
-
-void f2fs_update_extent_cache(struct dnode_of_data *dn)
+void f2fs_destroy_extent_tree(struct inode *inode)
{
- pgoff_t fofs;
- block_t blkaddr;
-
- if (!f2fs_may_extent_tree(dn->inode))
- return;
-
- if (dn->data_blkaddr == NEW_ADDR)
- blkaddr = NULL_ADDR;
- else
- blkaddr = dn->data_blkaddr;
-
- fofs = f2fs_start_bidx_of_node(ofs_of_node(dn->node_page), dn->inode) +
- dn->ofs_in_node;
- f2fs_update_extent_tree_range(dn->inode, fofs, blkaddr, 1);
+ __destroy_extent_tree(inode, EX_READ);
}
-void f2fs_update_extent_cache_range(struct dnode_of_data *dn,
- pgoff_t fofs, block_t blkaddr, unsigned int len)
-
+static void __init_extent_tree_info(struct extent_tree_info *eti)
{
- if (!f2fs_may_extent_tree(dn->inode))
- return;
-
- f2fs_update_extent_tree_range(dn->inode, fofs, blkaddr, len);
+ INIT_RADIX_TREE(&eti->extent_tree_root, GFP_NOIO);
+ mutex_init(&eti->extent_tree_lock);
+ INIT_LIST_HEAD(&eti->extent_list);
+ spin_lock_init(&eti->extent_lock);
+ atomic_set(&eti->total_ext_tree, 0);
+ INIT_LIST_HEAD(&eti->zombie_list);
+ atomic_set(&eti->total_zombie_tree, 0);
+ atomic_set(&eti->total_ext_node, 0);
}
void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi)
{
- INIT_RADIX_TREE(&sbi->extent_tree_root, GFP_NOIO);
- mutex_init(&sbi->extent_tree_lock);
- INIT_LIST_HEAD(&sbi->extent_list);
- spin_lock_init(&sbi->extent_lock);
- atomic_set(&sbi->total_ext_tree, 0);
- INIT_LIST_HEAD(&sbi->zombie_list);
- atomic_set(&sbi->total_zombie_tree, 0);
- atomic_set(&sbi->total_ext_node, 0);
+ __init_extent_tree_info(&sbi->extent_tree[EX_READ]);
}
int __init f2fs_create_extent_cache(void)
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 04fdf010bb77..003b04b10109 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -596,16 +596,22 @@ enum {
/* dirty segments threshold for triggering CP */
#define DEFAULT_DIRTY_THRESHOLD 4
+#define RECOVERY_MAX_RA_BLOCKS BIO_MAX_VECS
+#define RECOVERY_MIN_RA_BLOCKS 1
+
+#define F2FS_ONSTACK_PAGES 16 /* nr of onstack pages */
+
/* for in-memory extent cache entry */
#define F2FS_MIN_EXTENT_LEN 64 /* minimum extent length */
/* number of extent info in extent cache we try to shrink */
#define READ_EXTENT_CACHE_SHRINK_NUMBER 128
-#define RECOVERY_MAX_RA_BLOCKS BIO_MAX_VECS
-#define RECOVERY_MIN_RA_BLOCKS 1
-
-#define F2FS_ONSTACK_PAGES 16 /* nr of onstack pages */
+/* extent cache type */
+enum extent_type {
+ EX_READ,
+ NR_EXTENT_CACHES,
+};
struct rb_entry {
struct rb_node rb_node; /* rb node located in rb-tree */
@@ -621,10 +627,17 @@ struct rb_entry {
struct extent_info {
unsigned int fofs; /* start offset in a file */
unsigned int len; /* length of the extent */
- block_t blk; /* start block address of the extent */
+ union {
+ /* read extent_cache */
+ struct {
+ /* start block address of the extent */
+ block_t blk;
#ifdef CONFIG_F2FS_FS_COMPRESSION
- unsigned int c_len; /* physical extent length of compressed blocks */
+ /* physical extent length of compressed blocks */
+ unsigned int c_len;
#endif
+ };
+ };
};
struct extent_node {
@@ -636,13 +649,25 @@ struct extent_node {
struct extent_tree {
nid_t ino; /* inode number */
+ enum extent_type type; /* keep the extent tree type */
struct rb_root_cached root; /* root of extent info rb-tree */
struct extent_node *cached_en; /* recently accessed extent node */
- struct extent_info largest; /* largested extent info */
struct list_head list; /* to be used by sbi->zombie_list */
rwlock_t lock; /* protect extent info rb-tree */
atomic_t node_cnt; /* # of extent node in rb-tree*/
bool largest_updated; /* largest extent updated */
+ struct extent_info largest; /* largest cached extent for EX_READ */
+};
+
+struct extent_tree_info {
+ struct radix_tree_root extent_tree_root;/* cache extent cache entries */
+ struct mutex extent_tree_lock; /* locking extent radix tree */
+ struct list_head extent_list; /* lru list for shrinker */
+ spinlock_t extent_lock; /* locking extent lru list */
+ atomic_t total_ext_tree; /* extent tree count */
+ struct list_head zombie_list; /* extent zombie tree list */
+ atomic_t total_zombie_tree; /* extent zombie tree count */
+ atomic_t total_ext_node; /* extent info count */
};
/*
@@ -805,7 +830,8 @@ struct f2fs_inode_info {
struct list_head dirty_list; /* dirty list for dirs and files */
struct list_head gdirty_list; /* linked in global dirty list */
struct task_struct *atomic_write_task; /* store atomic write task */
- struct extent_tree *extent_tree; /* cached extent_tree entry */
+ struct extent_tree *extent_tree[NR_EXTENT_CACHES];
+ /* cached extent_tree entry */
struct inode *cow_inode; /* copy-on-write inode for atomic write */
/* avoid racing between foreground op and gc */
@@ -1626,14 +1652,7 @@ struct f2fs_sb_info {
struct mutex flush_lock; /* for flush exclusion */
/* for extent tree cache */
- struct radix_tree_root extent_tree_root;/* cache extent cache entries */
- struct mutex extent_tree_lock; /* locking extent radix tree */
- struct list_head extent_list; /* lru list for shrinker */
- spinlock_t extent_lock; /* locking extent lru list */
- atomic_t total_ext_tree; /* extent tree count */
- struct list_head zombie_list; /* extent zombie tree list */
- atomic_t total_zombie_tree; /* extent zombie tree count */
- atomic_t total_ext_node; /* extent info count */
+ struct extent_tree_info extent_tree[NR_EXTENT_CACHES];
/* basic filesystem units */
unsigned int log_sectors_per_block; /* log2 sectors per block */
@@ -1718,10 +1737,14 @@ struct f2fs_sb_info {
unsigned int segment_count[2]; /* # of allocated segments */
unsigned int block_count[2]; /* # of allocated blocks */
atomic_t inplace_count; /* # of inplace update */
- atomic64_t total_hit_ext; /* # of lookup extent cache */
- atomic64_t read_hit_rbtree; /* # of hit rbtree extent node */
- atomic64_t read_hit_largest; /* # of hit largest extent node */
- atomic64_t read_hit_cached; /* # of hit cached extent node */
+ /* # of lookup extent cache */
+ atomic64_t total_hit_ext[NR_EXTENT_CACHES];
+ /* # of hit rbtree extent node */
+ atomic64_t read_hit_rbtree[NR_EXTENT_CACHES];
+ /* # of hit cached extent node */
+ atomic64_t read_hit_cached[NR_EXTENT_CACHES];
+ /* # of hit largest extent node in read extent cache */
+ atomic64_t read_hit_largest;
atomic_t inline_xattr; /* # of inline_xattr inodes */
atomic_t inline_inode; /* # of inline_data inodes */
atomic_t inline_dir; /* # of inline_dentry inodes */
@@ -3823,9 +3846,17 @@ struct f2fs_stat_info {
struct f2fs_sb_info *sbi;
int all_area_segs, sit_area_segs, nat_area_segs, ssa_area_segs;
int main_area_segs, main_area_sections, main_area_zones;
- unsigned long long hit_largest, hit_cached, hit_rbtree;
- unsigned long long hit_total, total_ext;
- int ext_tree, zombie_tree, ext_node;
+ unsigned long long hit_cached[NR_EXTENT_CACHES];
+ unsigned long long hit_rbtree[NR_EXTENT_CACHES];
+ unsigned long long total_ext[NR_EXTENT_CACHES];
+ unsigned long long hit_total[NR_EXTENT_CACHES];
+ int ext_tree[NR_EXTENT_CACHES];
+ int zombie_tree[NR_EXTENT_CACHES];
+ int ext_node[NR_EXTENT_CACHES];
+ /* to count memory footprint */
+ unsigned long long ext_mem[NR_EXTENT_CACHES];
+ /* for read extent cache */
+ unsigned long long hit_largest;
int ndirty_node, ndirty_dent, ndirty_meta, ndirty_imeta;
int ndirty_data, ndirty_qdata;
unsigned int ndirty_dirs, ndirty_files, nquota_files, ndirty_all;
@@ -3884,10 +3915,10 @@ static inline struct f2fs_stat_info *F2FS_STAT(struct f2fs_sb_info *sbi)
#define stat_other_skip_bggc_count(sbi) ((sbi)->other_skip_bggc++)
#define stat_inc_dirty_inode(sbi, type) ((sbi)->ndirty_inode[type]++)
#define stat_dec_dirty_inode(sbi, type) ((sbi)->ndirty_inode[type]--)
-#define stat_inc_total_hit(sbi) (atomic64_inc(&(sbi)->total_hit_ext))
-#define stat_inc_rbtree_node_hit(sbi) (atomic64_inc(&(sbi)->read_hit_rbtree))
+#define stat_inc_total_hit(sbi, type) (atomic64_inc(&(sbi)->total_hit_ext[type]))
+#define stat_inc_rbtree_node_hit(sbi, type) (atomic64_inc(&(sbi)->read_hit_rbtree[type]))
#define stat_inc_largest_node_hit(sbi) (atomic64_inc(&(sbi)->read_hit_largest))
-#define stat_inc_cached_node_hit(sbi) (atomic64_inc(&(sbi)->read_hit_cached))
+#define stat_inc_cached_node_hit(sbi, type) (atomic64_inc(&(sbi)->read_hit_cached[type]))
#define stat_inc_inline_xattr(inode) \
do { \
if (f2fs_has_inline_xattr(inode)) \
@@ -4010,10 +4041,10 @@ void f2fs_update_sit_info(struct f2fs_sb_info *sbi);
#define stat_other_skip_bggc_count(sbi) do { } while (0)
#define stat_inc_dirty_inode(sbi, type) do { } while (0)
#define stat_dec_dirty_inode(sbi, type) do { } while (0)
-#define stat_inc_total_hit(sbi) do { } while (0)
-#define stat_inc_rbtree_node_hit(sbi) do { } while (0)
+#define stat_inc_total_hit(sbi, type) do { } while (0)
+#define stat_inc_rbtree_node_hit(sbi, type) do { } while (0)
#define stat_inc_largest_node_hit(sbi) do { } while (0)
-#define stat_inc_cached_node_hit(sbi) do { } while (0)
+#define stat_inc_cached_node_hit(sbi, type) do { } while (0)
#define stat_inc_inline_xattr(inode) do { } while (0)
#define stat_dec_inline_xattr(inode) do { } while (0)
#define stat_inc_inline_inode(inode) do { } while (0)
@@ -4119,20 +4150,23 @@ struct rb_entry *f2fs_lookup_rb_tree_ret(struct rb_root_cached *root,
bool force, bool *leftmost);
bool f2fs_check_rb_tree_consistence(struct f2fs_sb_info *sbi,
struct rb_root_cached *root, bool check_key);
-unsigned int f2fs_shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink);
void f2fs_init_extent_tree(struct inode *inode, struct page *ipage);
void f2fs_drop_extent_tree(struct inode *inode);
-unsigned int f2fs_destroy_extent_node(struct inode *inode);
+void f2fs_destroy_extent_node(struct inode *inode);
void f2fs_destroy_extent_tree(struct inode *inode);
-bool f2fs_lookup_extent_cache(struct inode *inode, pgoff_t pgofs,
- struct extent_info *ei);
-void f2fs_update_extent_cache(struct dnode_of_data *dn);
-void f2fs_update_extent_cache_range(struct dnode_of_data *dn,
- pgoff_t fofs, block_t blkaddr, unsigned int len);
void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi);
int __init f2fs_create_extent_cache(void);
void f2fs_destroy_extent_cache(void);
+/* read extent cache ops */
+bool f2fs_lookup_read_extent_cache(struct inode *inode, pgoff_t pgofs,
+ struct extent_info *ei);
+void f2fs_update_read_extent_cache(struct dnode_of_data *dn);
+void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
+ pgoff_t fofs, block_t blkaddr, unsigned int len);
+unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi,
+ int nr_shrink);
+
/*
* sysfs.c
*/
@@ -4202,7 +4236,7 @@ int f2fs_write_multi_pages(struct compress_ctx *cc,
struct writeback_control *wbc,
enum iostat_type io_type);
int f2fs_is_compressed_cluster(struct inode *inode, pgoff_t index);
-void f2fs_update_extent_tree_range_compressed(struct inode *inode,
+void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
pgoff_t fofs, block_t blkaddr, unsigned int llen,
unsigned int c_len);
int f2fs_read_multi_pages(struct compress_ctx *cc, struct bio **bio_ret,
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index ab0a0d3730f6..cbe7c24065c7 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -618,7 +618,7 @@ void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count)
*/
fofs = f2fs_start_bidx_of_node(ofs_of_node(dn->node_page),
dn->inode) + ofs;
- f2fs_update_extent_cache_range(dn, fofs, 0, len);
+ f2fs_update_read_extent_cache_range(dn, fofs, 0, len);
dec_valid_block_count(sbi, dn->inode, nr_free);
}
dn->ofs_in_node = ofs;
@@ -1496,7 +1496,7 @@ static int f2fs_do_zero_range(struct dnode_of_data *dn, pgoff_t start,
f2fs_set_data_blkaddr(dn);
}
- f2fs_update_extent_cache_range(dn, start, 0, index - start);
+ f2fs_update_read_extent_cache_range(dn, start, 0, index - start);
return ret;
}
@@ -2558,7 +2558,7 @@ static int f2fs_defragment_range(struct f2fs_sb_info *sbi,
struct f2fs_map_blocks map = { .m_next_extent = NULL,
.m_seg_type = NO_CHECK_TYPE,
.m_may_create = false };
- struct extent_info ei = {0, 0, 0};
+ struct extent_info ei = {0, };
pgoff_t pg_start, pg_end, next_pgofs;
unsigned int blk_per_seg = sbi->blocks_per_seg;
unsigned int total = 0, sec_num;
@@ -2590,7 +2590,7 @@ static int f2fs_defragment_range(struct f2fs_sb_info *sbi,
* lookup mapping info in extent cache, skip defragmenting if physical
* block addresses are continuous.
*/
- if (f2fs_lookup_extent_cache(inode, pg_start, &ei)) {
+ if (f2fs_lookup_read_extent_cache(inode, pg_start, &ei)) {
if (ei.fofs + ei.len >= pg_end)
goto out;
}
diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
index d19e26b2e875..f0c6506d8975 100644
--- a/fs/f2fs/gc.c
+++ b/fs/f2fs/gc.c
@@ -1146,7 +1146,7 @@ static int ra_data_block(struct inode *inode, pgoff_t index)
struct address_space *mapping = inode->i_mapping;
struct dnode_of_data dn;
struct page *page;
- struct extent_info ei = {0, 0, 0};
+ struct extent_info ei = {0, };
struct f2fs_io_info fio = {
.sbi = sbi,
.ino = inode->i_ino,
@@ -1164,7 +1164,7 @@ static int ra_data_block(struct inode *inode, pgoff_t index)
if (!page)
return -ENOMEM;
- if (f2fs_lookup_extent_cache(inode, index, &ei)) {
+ if (f2fs_lookup_read_extent_cache(inode, index, &ei)) {
dn.data_blkaddr = ei.blk + index - ei.fofs;
if (unlikely(!f2fs_is_valid_blkaddr(sbi, dn.data_blkaddr,
DATA_GENERIC_ENHANCE_READ))) {
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index 2c705c60019b..086f201f15a0 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -262,8 +262,8 @@ static bool sanity_check_inode(struct inode *inode, struct page *node_page)
return false;
}
- if (fi->extent_tree) {
- struct extent_info *ei = &fi->extent_tree->largest;
+ if (fi->extent_tree[EX_READ]) {
+ struct extent_info *ei = &fi->extent_tree[EX_READ]->largest;
if (ei->len &&
(!f2fs_is_valid_blkaddr(sbi, ei->blk,
@@ -607,7 +607,7 @@ struct inode *f2fs_iget_retry(struct super_block *sb, unsigned long ino)
void f2fs_update_inode(struct inode *inode, struct page *node_page)
{
struct f2fs_inode *ri;
- struct extent_tree *et = F2FS_I(inode)->extent_tree;
+ struct extent_tree *et = F2FS_I(inode)->extent_tree[EX_READ];
f2fs_wait_on_page_writeback(node_page, NODE, true, true);
set_page_dirty(node_page);
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 84b147966080..07419c3e42a5 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -86,9 +86,11 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
mem_size >>= PAGE_SHIFT;
res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
} else if (type == READ_EXTENT_CACHE) {
- mem_size = (atomic_read(&sbi->total_ext_tree) *
+ struct extent_tree_info *eti = &sbi->extent_tree[EX_READ];
+
+ mem_size = (atomic_read(&eti->total_ext_tree) *
sizeof(struct extent_tree) +
- atomic_read(&sbi->total_ext_node) *
+ atomic_read(&eti->total_ext_node) *
sizeof(struct extent_node)) >> PAGE_SHIFT;
res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
} else if (type == DISCARD_CACHE) {
@@ -859,7 +861,7 @@ int f2fs_get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode)
blkaddr = data_blkaddr(dn->inode, dn->node_page,
dn->ofs_in_node + 1);
- f2fs_update_extent_tree_range_compressed(dn->inode,
+ f2fs_update_read_extent_tree_range_compressed(dn->inode,
index, blkaddr,
F2FS_I(dn->inode)->i_cluster_size,
c_len);
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 51de358bc452..8722d1a13c17 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -450,7 +450,8 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
/* try to shrink extent cache when there is no enough memory */
if (!f2fs_available_free_memory(sbi, READ_EXTENT_CACHE))
- f2fs_shrink_extent_tree(sbi, READ_EXTENT_CACHE_SHRINK_NUMBER);
+ f2fs_shrink_read_extent_tree(sbi,
+ READ_EXTENT_CACHE_SHRINK_NUMBER);
/* check the # of cached NAT entries */
if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
diff --git a/fs/f2fs/shrinker.c b/fs/f2fs/shrinker.c
index dd3c3c7a90ec..33c490e69ae3 100644
--- a/fs/f2fs/shrinker.c
+++ b/fs/f2fs/shrinker.c
@@ -28,10 +28,13 @@ static unsigned long __count_free_nids(struct f2fs_sb_info *sbi)
return count > 0 ? count : 0;
}
-static unsigned long __count_extent_cache(struct f2fs_sb_info *sbi)
+static unsigned long __count_extent_cache(struct f2fs_sb_info *sbi,
+ enum extent_type type)
{
- return atomic_read(&sbi->total_zombie_tree) +
- atomic_read(&sbi->total_ext_node);
+ struct extent_tree_info *eti = &sbi->extent_tree[type];
+
+ return atomic_read(&eti->total_zombie_tree) +
+ atomic_read(&eti->total_ext_node);
}
unsigned long f2fs_shrink_count(struct shrinker *shrink,
@@ -53,8 +56,8 @@ unsigned long f2fs_shrink_count(struct shrinker *shrink,
}
spin_unlock(&f2fs_list_lock);
- /* count extent cache entries */
- count += __count_extent_cache(sbi);
+ /* count read extent cache entries */
+ count += __count_extent_cache(sbi, EX_READ);
/* count clean nat cache entries */
count += __count_nat_entries(sbi);
@@ -99,8 +102,8 @@ unsigned long f2fs_shrink_scan(struct shrinker *shrink,
sbi->shrinker_run_no = run_no;
- /* shrink extent cache entries */
- freed += f2fs_shrink_extent_tree(sbi, nr >> 1);
+ /* shrink read extent cache entries */
+ freed += f2fs_shrink_read_extent_tree(sbi, nr >> 1);
/* shrink clean nat cache entries */
if (freed < nr)
@@ -130,7 +133,7 @@ void f2fs_join_shrinker(struct f2fs_sb_info *sbi)
void f2fs_leave_shrinker(struct f2fs_sb_info *sbi)
{
- f2fs_shrink_extent_tree(sbi, __count_extent_cache(sbi));
+ f2fs_shrink_read_extent_tree(sbi, __count_extent_cache(sbi, EX_READ));
spin_lock(&f2fs_list_lock);
list_del_init(&sbi->s_list);
diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h
index 7fbfce498472..2bb37892d2ba 100644
--- a/include/trace/events/f2fs.h
+++ b/include/trace/events/f2fs.h
@@ -48,6 +48,7 @@ TRACE_DEFINE_ENUM(CP_DISCARD);
TRACE_DEFINE_ENUM(CP_TRIMMED);
TRACE_DEFINE_ENUM(CP_PAUSE);
TRACE_DEFINE_ENUM(CP_RESIZE);
+TRACE_DEFINE_ENUM(EX_READ);
#define show_block_type(type) \
__print_symbolic(type, \
@@ -1522,28 +1523,31 @@ TRACE_EVENT(f2fs_issue_flush,
TRACE_EVENT(f2fs_lookup_extent_tree_start,
- TP_PROTO(struct inode *inode, unsigned int pgofs),
+ TP_PROTO(struct inode *inode, unsigned int pgofs, enum extent_type type),
- TP_ARGS(inode, pgofs),
+ TP_ARGS(inode, pgofs, type),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(ino_t, ino)
__field(unsigned int, pgofs)
+ __field(enum extent_type, type)
),
TP_fast_assign(
__entry->dev = inode->i_sb->s_dev;
__entry->ino = inode->i_ino;
__entry->pgofs = pgofs;
+ __entry->type = type;
),
- TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u",
+ TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, type = %s",
show_dev_ino(__entry),
- __entry->pgofs)
+ __entry->pgofs,
+ __entry->type == EX_READ ? "Read" : "N/A")
);
-TRACE_EVENT_CONDITION(f2fs_lookup_extent_tree_end,
+TRACE_EVENT_CONDITION(f2fs_lookup_read_extent_tree_end,
TP_PROTO(struct inode *inode, unsigned int pgofs,
struct extent_info *ei),
@@ -1557,8 +1561,8 @@ TRACE_EVENT_CONDITION(f2fs_lookup_extent_tree_end,
__field(ino_t, ino)
__field(unsigned int, pgofs)
__field(unsigned int, fofs)
- __field(u32, blk)
__field(unsigned int, len)
+ __field(u32, blk)
),
TP_fast_assign(
@@ -1566,26 +1570,26 @@ TRACE_EVENT_CONDITION(f2fs_lookup_extent_tree_end,
__entry->ino = inode->i_ino;
__entry->pgofs = pgofs;
__entry->fofs = ei->fofs;
- __entry->blk = ei->blk;
__entry->len = ei->len;
+ __entry->blk = ei->blk;
),
TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, "
- "ext_info(fofs: %u, blk: %u, len: %u)",
+ "read_ext_info(fofs: %u, len: %u, blk: %u)",
show_dev_ino(__entry),
__entry->pgofs,
__entry->fofs,
- __entry->blk,
- __entry->len)
+ __entry->len,
+ __entry->blk)
);
-TRACE_EVENT(f2fs_update_extent_tree_range,
+TRACE_EVENT(f2fs_update_read_extent_tree_range,
- TP_PROTO(struct inode *inode, unsigned int pgofs, block_t blkaddr,
- unsigned int len,
+ TP_PROTO(struct inode *inode, unsigned int pgofs, unsigned int len,
+ block_t blkaddr,
unsigned int c_len),
- TP_ARGS(inode, pgofs, blkaddr, len, c_len),
+ TP_ARGS(inode, pgofs, len, blkaddr, c_len),
TP_STRUCT__entry(
__field(dev_t, dev)
@@ -1600,67 +1604,73 @@ TRACE_EVENT(f2fs_update_extent_tree_range,
__entry->dev = inode->i_sb->s_dev;
__entry->ino = inode->i_ino;
__entry->pgofs = pgofs;
- __entry->blk = blkaddr;
__entry->len = len;
+ __entry->blk = blkaddr;
__entry->c_len = c_len;
),
TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, "
- "blkaddr = %u, len = %u, "
- "c_len = %u",
+ "len = %u, blkaddr = %u, c_len = %u",
show_dev_ino(__entry),
__entry->pgofs,
- __entry->blk,
__entry->len,
+ __entry->blk,
__entry->c_len)
);
TRACE_EVENT(f2fs_shrink_extent_tree,
TP_PROTO(struct f2fs_sb_info *sbi, unsigned int node_cnt,
- unsigned int tree_cnt),
+ unsigned int tree_cnt, enum extent_type type),
- TP_ARGS(sbi, node_cnt, tree_cnt),
+ TP_ARGS(sbi, node_cnt, tree_cnt, type),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(unsigned int, node_cnt)
__field(unsigned int, tree_cnt)
+ __field(enum extent_type, type)
),
TP_fast_assign(
__entry->dev = sbi->sb->s_dev;
__entry->node_cnt = node_cnt;
__entry->tree_cnt = tree_cnt;
+ __entry->type = type;
),
- TP_printk("dev = (%d,%d), shrunk: node_cnt = %u, tree_cnt = %u",
+ TP_printk("dev = (%d,%d), shrunk: node_cnt = %u, tree_cnt = %u, type = %s",
show_dev(__entry->dev),
__entry->node_cnt,
- __entry->tree_cnt)
+ __entry->tree_cnt,
+ __entry->type == EX_READ ? "Read" : "N/A")
);
TRACE_EVENT(f2fs_destroy_extent_tree,
- TP_PROTO(struct inode *inode, unsigned int node_cnt),
+ TP_PROTO(struct inode *inode, unsigned int node_cnt,
+ enum extent_type type),
- TP_ARGS(inode, node_cnt),
+ TP_ARGS(inode, node_cnt, type),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(ino_t, ino)
__field(unsigned int, node_cnt)
+ __field(enum extent_type, type)
),
TP_fast_assign(
__entry->dev = inode->i_sb->s_dev;
__entry->ino = inode->i_ino;
__entry->node_cnt = node_cnt;
+ __entry->type = type;
),
- TP_printk("dev = (%d,%d), ino = %lu, destroyed: node_cnt = %u",
+ TP_printk("dev = (%d,%d), ino = %lu, destroyed: node_cnt = %u, type = %s",
show_dev_ino(__entry),
- __entry->node_cnt)
+ __entry->node_cnt,
+ __entry->type == EX_READ ? "Read" : "N/A")
);
DECLARE_EVENT_CLASS(f2fs_sync_dirty_inodes,
--
2.39.0.rc0.267.gcb52ba06e7-goog
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply related [flat|nested] 19+ messages in thread* Re: [f2fs-dev] [PATCH 4/6 v2] f2fs: refactor extent_cache to support for read and more
2022-12-05 18:54 ` [f2fs-dev] [PATCH 4/6] f2fs: refactor extent_cache to support for read and more Jaegeuk Kim
@ 2022-12-06 19:05 ` Jaegeuk Kim
2022-12-08 15:02 ` Chao Yu
0 siblings, 1 reply; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-06 19:05 UTC (permalink / raw)
To: linux-kernel, linux-f2fs-devel
This patch prepares extent_cache to get more use-cases.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
Change log from v1:
- fix a bug
fs/f2fs/data.c | 14 +++----
fs/f2fs/debug.c | 53 +++++++++++++++++---------
fs/f2fs/extent_cache.c | 85 +++++++++++++++++++++++++++---------------
3 files changed, 97 insertions(+), 55 deletions(-)
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 35c19248b1e2..031fa27d194e 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -1195,7 +1195,7 @@ int f2fs_get_block(struct dnode_of_data *dn, pgoff_t index)
struct extent_info ei = {0, };
struct inode *inode = dn->inode;
- if (f2fs_lookup_extent_cache(inode, index, &ei)) {
+ if (f2fs_lookup_extent_cache(inode, index, &ei, EX_READ)) {
dn->data_blkaddr = ei.blk + index - ei.fofs;
return 0;
}
@@ -1217,7 +1217,7 @@ struct page *f2fs_get_read_data_page(struct inode *inode, pgoff_t index,
if (!page)
return ERR_PTR(-ENOMEM);
- if (f2fs_lookup_extent_cache(inode, index, &ei)) {
+ if (f2fs_lookup_extent_cache(inode, index, &ei, EX_READ)) {
dn.data_blkaddr = ei.blk + index - ei.fofs;
if (!f2fs_is_valid_blkaddr(F2FS_I_SB(inode), dn.data_blkaddr,
DATA_GENERIC_ENHANCE_READ)) {
@@ -1485,7 +1485,7 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
pgofs = (pgoff_t)map->m_lblk;
end = pgofs + maxblocks;
- if (!create && f2fs_lookup_extent_cache(inode, pgofs, &ei)) {
+ if (!create && f2fs_lookup_extent_cache(inode, pgofs, &ei, EX_READ)) {
if (f2fs_lfs_mode(sbi) && flag == F2FS_GET_BLOCK_DIO &&
map->m_may_create)
goto next_dnode;
@@ -2201,7 +2201,7 @@ int f2fs_read_multi_pages(struct compress_ctx *cc, struct bio **bio_ret,
if (f2fs_cluster_is_empty(cc))
goto out;
- if (f2fs_lookup_extent_cache(inode, start_idx, &ei))
+ if (f2fs_lookup_extent_cache(inode, start_idx, &ei, EX_READ))
from_dnode = false;
if (!from_dnode)
@@ -2635,7 +2635,7 @@ int f2fs_do_write_data_page(struct f2fs_io_info *fio)
set_new_dnode(&dn, inode, NULL, NULL, 0);
if (need_inplace_update(fio) &&
- f2fs_lookup_extent_cache(inode, page->index, &ei)) {
+ f2fs_lookup_extent_cache(inode, page->index, &ei, EX_READ)) {
fio->old_blkaddr = ei.blk + page->index - ei.fofs;
if (!f2fs_is_valid_blkaddr(fio->sbi, fio->old_blkaddr,
@@ -3359,7 +3359,7 @@ static int prepare_write_begin(struct f2fs_sb_info *sbi,
} else if (locked) {
err = f2fs_get_block(&dn, index);
} else {
- if (f2fs_lookup_extent_cache(inode, index, &ei)) {
+ if (f2fs_lookup_extent_cache(inode, index, &ei, EX_READ)) {
dn.data_blkaddr = ei.blk + index - ei.fofs;
} else {
/* hole case */
@@ -3400,7 +3400,7 @@ static int __find_data_block(struct inode *inode, pgoff_t index,
set_new_dnode(&dn, inode, ipage, ipage, 0);
- if (f2fs_lookup_extent_cache(inode, index, &ei)) {
+ if (f2fs_lookup_extent_cache(inode, index, &ei, EX_READ)) {
dn.data_blkaddr = ei.blk + index - ei.fofs;
} else {
/* hole case */
diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index a216dcdf6941..71eb4c50edf6 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -72,15 +72,23 @@ static void update_general_status(struct f2fs_sb_info *sbi)
si->main_area_zones = si->main_area_sections /
le32_to_cpu(raw_super->secs_per_zone);
- /* validation check of the segment numbers */
+ /* general extent cache stats */
+ for (i = 0; i < NR_EXTENT_CACHES; i++) {
+ struct extent_tree_info *eti = &sbi->extent_tree[i];
+
+ si->hit_cached[i] = atomic64_read(&sbi->read_hit_cached[i]);
+ si->hit_rbtree[i] = atomic64_read(&sbi->read_hit_rbtree[i]);
+ si->total_ext[i] = atomic64_read(&sbi->total_hit_ext[i]);
+ si->hit_total[i] = si->hit_cached[i] + si->hit_rbtree[i];
+ si->ext_tree[i] = atomic_read(&eti->total_ext_tree);
+ si->zombie_tree[i] = atomic_read(&eti->total_zombie_tree);
+ si->ext_node[i] = atomic_read(&eti->total_ext_node);
+ }
+ /* read extent_cache only */
si->hit_largest = atomic64_read(&sbi->read_hit_largest);
- si->hit_cached = atomic64_read(&sbi->read_hit_cached);
- si->hit_rbtree = atomic64_read(&sbi->read_hit_rbtree);
- si->hit_total = si->hit_largest + si->hit_cached + si->hit_rbtree;
- si->total_ext = atomic64_read(&sbi->total_hit_ext);
- si->ext_tree = atomic_read(&sbi->total_ext_tree);
- si->zombie_tree = atomic_read(&sbi->total_zombie_tree);
- si->ext_node = atomic_read(&sbi->total_ext_node);
+ si->hit_total[EX_READ] += si->hit_largest;
+
+ /* validation check of the segment numbers */
si->ndirty_node = get_pages(sbi, F2FS_DIRTY_NODES);
si->ndirty_dent = get_pages(sbi, F2FS_DIRTY_DENTS);
si->ndirty_meta = get_pages(sbi, F2FS_DIRTY_META);
@@ -294,10 +302,16 @@ static void update_mem_info(struct f2fs_sb_info *sbi)
sizeof(struct nat_entry_set);
for (i = 0; i < MAX_INO_ENTRY; i++)
si->cache_mem += sbi->im[i].ino_num * sizeof(struct ino_entry);
- si->cache_mem += atomic_read(&sbi->total_ext_tree) *
+
+ for (i = 0; i < NR_EXTENT_CACHES; i++) {
+ struct extent_tree_info *eti = &sbi->extent_tree[i];
+
+ si->ext_mem[i] += atomic_read(&eti->total_ext_tree) *
sizeof(struct extent_tree);
- si->cache_mem += atomic_read(&sbi->total_ext_node) *
+ si->ext_mem[i] += atomic_read(&eti->total_ext_node) *
sizeof(struct extent_node);
+ si->cache_mem += si->ext_mem[i];
+ }
si->page_mem = 0;
if (sbi->node_inode) {
@@ -490,16 +504,19 @@ static int stat_show(struct seq_file *s, void *v)
si->bg_node_blks);
seq_printf(s, "BG skip : IO: %u, Other: %u\n",
si->io_skip_bggc, si->other_skip_bggc);
- seq_puts(s, "\nExtent Cache:\n");
+ seq_puts(s, "\nExtent Cache (Read):\n");
seq_printf(s, " - Hit Count: L1-1:%llu L1-2:%llu L2:%llu\n",
si->hit_largest, si->hit_cached,
- si->hit_rbtree);
+ si->hit_largest, si->hit_cached[EX_READ],
+ si->hit_rbtree[EX_READ]);
seq_printf(s, " - Hit Ratio: %llu%% (%llu / %llu)\n",
- !si->total_ext ? 0 :
- div64_u64(si->hit_total * 100, si->total_ext),
- si->hit_total, si->total_ext);
+ !si->total_ext[EX_READ] ? 0 :
+ div64_u64(si->hit_total[EX_READ] * 100,
+ si->total_ext[EX_READ]),
+ si->hit_total[EX_READ], si->total_ext[EX_READ]);
seq_printf(s, " - Inner Struct Count: tree: %d(%d), node: %d\n",
- si->ext_tree, si->zombie_tree, si->ext_node);
+ si->ext_tree[EX_READ], si->zombie_tree[EX_READ],
+ si->ext_node[EX_READ]);
seq_puts(s, "\nBalancing F2FS Async:\n");
seq_printf(s, " - DIO (R: %4d, W: %4d)\n",
si->nr_dio_read, si->nr_dio_write);
@@ -566,8 +583,10 @@ static int stat_show(struct seq_file *s, void *v)
(si->base_mem + si->cache_mem + si->page_mem) >> 10);
seq_printf(s, " - static: %llu KB\n",
si->base_mem >> 10);
- seq_printf(s, " - cached: %llu KB\n",
+ seq_printf(s, " - cached all: %llu KB\n",
si->cache_mem >> 10);
+ seq_printf(s, " - read extent cache: %llu KB\n",
+ si->ext_mem[EX_READ] >> 10);
seq_printf(s, " - paged : %llu KB\n",
si->page_mem >> 10);
}
diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
index 22a51934cd45..46dbe3629245 100644
--- a/fs/f2fs/extent_cache.c
+++ b/fs/f2fs/extent_cache.c
@@ -15,7 +15,21 @@
#include "node.h"
#include <trace/events/f2fs.h>
-static bool f2fs_may_extent_tree(struct inode *inode)
+static bool f2fs_may_read_extent_tree(struct inode *inode)
+{
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+
+ if (!test_opt(sbi, READ_EXTENT_CACHE))
+ return false;
+ if (is_inode_flag_set(inode, FI_NO_EXTENT))
+ return false;
+ if (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
+ !f2fs_sb_has_readonly(sbi))
+ return false;
+ return S_ISREG(inode->i_mode);
+}
+
+static bool f2fs_may_extent_tree(struct inode *inode, enum extent_type type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
@@ -26,18 +40,16 @@ static bool f2fs_may_extent_tree(struct inode *inode)
if (list_empty(&sbi->s_list))
return false;
- if (!test_opt(sbi, READ_EXTENT_CACHE) ||
- is_inode_flag_set(inode, FI_NO_EXTENT) ||
- (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
- !f2fs_sb_has_readonly(sbi)))
- return false;
-
- return S_ISREG(inode->i_mode);
+ if (type == EX_READ)
+ return f2fs_may_read_extent_tree(inode);
+ return false;
}
static void __try_update_largest_extent(struct extent_tree *et,
struct extent_node *en)
{
+ if (et->type != EX_READ)
+ return;
if (en->ei.len <= et->largest.len)
return;
@@ -46,28 +58,31 @@ static void __try_update_largest_extent(struct extent_tree *et,
}
static bool __is_extent_mergeable(struct extent_info *back,
- struct extent_info *front)
+ struct extent_info *front, enum extent_type type)
{
+ if (type == EX_READ) {
#ifdef CONFIG_F2FS_FS_COMPRESSION
- if (back->c_len && back->len != back->c_len)
- return false;
- if (front->c_len && front->len != front->c_len)
- return false;
+ if (back->c_len && back->len != back->c_len)
+ return false;
+ if (front->c_len && front->len != front->c_len)
+ return false;
#endif
- return (back->fofs + back->len == front->fofs &&
- back->blk + back->len == front->blk);
+ return (back->fofs + back->len == front->fofs &&
+ back->blk + back->len == front->blk);
+ }
+ return false;
}
static bool __is_back_mergeable(struct extent_info *cur,
- struct extent_info *back)
+ struct extent_info *back, enum extent_type type)
{
- return __is_extent_mergeable(back, cur);
+ return __is_extent_mergeable(back, cur, type);
}
static bool __is_front_mergeable(struct extent_info *cur,
- struct extent_info *front)
+ struct extent_info *front, enum extent_type type)
{
- return __is_extent_mergeable(cur, front);
+ return __is_extent_mergeable(cur, front, type);
}
static struct rb_entry *__lookup_rb_tree_fast(struct rb_entry *cached_re,
@@ -292,6 +307,7 @@ static struct extent_node *__attach_extent_node(struct f2fs_sb_info *sbi,
struct rb_node *parent, struct rb_node **p,
bool leftmost)
{
+ struct extent_tree_info *eti = &sbi->extent_tree[et->type];
struct extent_node *en;
en = f2fs_kmem_cache_alloc(extent_node_slab, GFP_ATOMIC, false, sbi);
@@ -305,16 +321,18 @@ static struct extent_node *__attach_extent_node(struct f2fs_sb_info *sbi,
rb_link_node(&en->rb_node, parent, p);
rb_insert_color_cached(&en->rb_node, &et->root, leftmost);
atomic_inc(&et->node_cnt);
- atomic_inc(&sbi->total_ext_node);
+ atomic_inc(&eti->total_ext_node);
return en;
}
static void __detach_extent_node(struct f2fs_sb_info *sbi,
struct extent_tree *et, struct extent_node *en)
{
+ struct extent_tree_info *eti = &sbi->extent_tree[et->type];
+
rb_erase_cached(&en->rb_node, &et->root);
atomic_dec(&et->node_cnt);
- atomic_dec(&sbi->total_ext_node);
+ atomic_dec(&eti->total_ext_node);
if (et->cached_en == en)
et->cached_en = NULL;
@@ -330,42 +348,47 @@ static void __detach_extent_node(struct f2fs_sb_info *sbi,
static void __release_extent_node(struct f2fs_sb_info *sbi,
struct extent_tree *et, struct extent_node *en)
{
- spin_lock(&sbi->extent_lock);
+ struct extent_tree_info *eti = &sbi->extent_tree[et->type];
+
+ spin_lock(&eti->extent_lock);
f2fs_bug_on(sbi, list_empty(&en->list));
list_del_init(&en->list);
- spin_unlock(&sbi->extent_lock);
+ spin_unlock(&eti->extent_lock);
__detach_extent_node(sbi, et, en);
}
-static struct extent_tree *__grab_extent_tree(struct inode *inode)
+static struct extent_tree *__grab_extent_tree(struct inode *inode,
+ enum extent_type type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+ struct extent_tree_info *eti = &sbi->extent_tree[type];
struct extent_tree *et;
nid_t ino = inode->i_ino;
- mutex_lock(&sbi->extent_tree_lock);
- et = radix_tree_lookup(&sbi->extent_tree_root, ino);
+ mutex_lock(&eti->extent_tree_lock);
+ et = radix_tree_lookup(&eti->extent_tree_root, ino);
if (!et) {
et = f2fs_kmem_cache_alloc(extent_tree_slab,
GFP_NOFS, true, NULL);
- f2fs_radix_tree_insert(&sbi->extent_tree_root, ino, et);
+ f2fs_radix_tree_insert(&eti->extent_tree_root, ino, et);
memset(et, 0, sizeof(struct extent_tree));
et->ino = ino;
+ et->type = type;
et->root = RB_ROOT_CACHED;
et->cached_en = NULL;
rwlock_init(&et->lock);
INIT_LIST_HEAD(&et->list);
atomic_set(&et->node_cnt, 0);
- atomic_inc(&sbi->total_ext_tree);
+ atomic_inc(&eti->total_ext_tree);
} else {
- atomic_dec(&sbi->total_zombie_tree);
+ atomic_dec(&eti->total_zombie_tree);
list_del_init(&et->list);
}
- mutex_unlock(&sbi->extent_tree_lock);
+ mutex_unlock(&eti->extent_tree_lock);
/* never died until evict_inode */
- F2FS_I(inode)->extent_tree = et;
+ F2FS_I(inode)->extent_tree[type] = et;
return et;
}
--
2.38.1.584.g0f3c55d4c2-goog
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply related [flat|nested] 19+ messages in thread
* [f2fs-dev] [PATCH 5/6] f2fs: allocate the extent_cache by default
2022-12-05 18:54 [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly Jaegeuk Kim
` (2 preceding siblings ...)
2022-12-05 18:54 ` [f2fs-dev] [PATCH 4/6] f2fs: refactor extent_cache to support for read and more Jaegeuk Kim
@ 2022-12-05 18:54 ` Jaegeuk Kim
2022-12-05 18:54 ` [f2fs-dev] [PATCH 6/6] f2fs: add block_age-based extent cache Jaegeuk Kim
2022-12-08 14:43 ` [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly Chao Yu
5 siblings, 0 replies; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-05 18:54 UTC (permalink / raw)
To: linux-kernel, linux-f2fs-devel; +Cc: Jaegeuk Kim
Let's allocate it to remove the runtime complexity.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
fs/f2fs/extent_cache.c | 38 +++++++++++++++++++-------------------
fs/f2fs/f2fs.h | 3 ++-
fs/f2fs/inode.c | 6 ++++--
fs/f2fs/namei.c | 4 ++--
4 files changed, 27 insertions(+), 24 deletions(-)
diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
index 654a14ab8977..305f969e3ad1 100644
--- a/fs/f2fs/extent_cache.c
+++ b/fs/f2fs/extent_cache.c
@@ -47,20 +47,23 @@ static bool __may_read_extent_tree(struct inode *inode)
return S_ISREG(inode->i_mode);
}
-static bool __may_extent_tree(struct inode *inode, enum extent_type type)
+static bool __init_may_extent_tree(struct inode *inode, enum extent_type type)
{
- struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+ if (type == EX_READ)
+ return __may_read_extent_tree(inode);
+ return false;
+}
+static bool __may_extent_tree(struct inode *inode, enum extent_type type)
+{
/*
* for recovered files during mount do not create extents
* if shrinker is not registered.
*/
- if (list_empty(&sbi->s_list))
+ if (list_empty(&F2FS_I_SB(inode)->s_list))
return false;
- if (type == EX_READ)
- return __may_read_extent_tree(inode);
- return false;
+ return __init_may_extent_tree(inode, type);
}
static void __try_update_largest_extent(struct extent_tree *et,
@@ -439,20 +442,18 @@ static void __drop_largest_extent(struct extent_tree *et,
}
}
-/* return true, if inode page is changed */
-static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage,
- enum extent_type type)
+void f2fs_init_read_extent_tree(struct inode *inode, struct page *ipage)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
- struct extent_tree_info *eti = &sbi->extent_tree[type];
- struct f2fs_extent *i_ext = ipage ? &F2FS_INODE(ipage)->i_ext : NULL;
+ struct extent_tree_info *eti = &sbi->extent_tree[EX_READ];
+ struct f2fs_extent *i_ext = &F2FS_INODE(ipage)->i_ext;
struct extent_tree *et;
struct extent_node *en;
struct extent_info ei;
- if (!__may_extent_tree(inode, type)) {
+ if (!__may_extent_tree(inode, EX_READ)) {
/* drop largest read extent */
- if (type == EX_READ && i_ext && i_ext->len) {
+ if (i_ext && i_ext->len) {
f2fs_wait_on_page_writeback(ipage, NODE, true, true);
i_ext->len = 0;
set_page_dirty(ipage);
@@ -460,13 +461,11 @@ static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage,
goto out;
}
- et = __grab_extent_tree(inode, type);
+ et = __grab_extent_tree(inode, EX_READ);
if (!i_ext || !i_ext->len)
goto out;
- BUG_ON(type != EX_READ);
-
get_read_extent_info(&ei, i_ext);
write_lock(&et->lock);
@@ -486,14 +485,15 @@ static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage,
unlock_out:
write_unlock(&et->lock);
out:
- if (type == EX_READ && !F2FS_I(inode)->extent_tree[EX_READ])
+ if (!F2FS_I(inode)->extent_tree[EX_READ])
set_inode_flag(inode, FI_NO_EXTENT);
}
-void f2fs_init_extent_tree(struct inode *inode, struct page *ipage)
+void f2fs_init_extent_tree(struct inode *inode)
{
/* initialize read cache */
- __f2fs_init_extent_tree(inode, ipage, EX_READ);
+ if (__init_may_extent_tree(inode, EX_READ))
+ __grab_extent_tree(inode, EX_READ);
}
static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 003b04b10109..ff940cba4600 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -4150,7 +4150,7 @@ struct rb_entry *f2fs_lookup_rb_tree_ret(struct rb_root_cached *root,
bool force, bool *leftmost);
bool f2fs_check_rb_tree_consistence(struct f2fs_sb_info *sbi,
struct rb_root_cached *root, bool check_key);
-void f2fs_init_extent_tree(struct inode *inode, struct page *ipage);
+void f2fs_init_extent_tree(struct inode *inode);
void f2fs_drop_extent_tree(struct inode *inode);
void f2fs_destroy_extent_node(struct inode *inode);
void f2fs_destroy_extent_tree(struct inode *inode);
@@ -4159,6 +4159,7 @@ int __init f2fs_create_extent_cache(void);
void f2fs_destroy_extent_cache(void);
/* read extent cache ops */
+void f2fs_init_read_extent_tree(struct inode *inode, struct page *ipage);
bool f2fs_lookup_read_extent_cache(struct inode *inode, pgoff_t pgofs,
struct extent_info *ei);
void f2fs_update_read_extent_cache(struct dnode_of_data *dn);
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index 086f201f15a0..c845c16f97d0 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -392,8 +392,6 @@ static int do_read_inode(struct inode *inode)
fi->i_pino = le32_to_cpu(ri->i_pino);
fi->i_dir_level = ri->i_dir_level;
- f2fs_init_extent_tree(inode, node_page);
-
get_inline_info(inode, ri);
fi->i_extra_isize = f2fs_has_extra_attr(inode) ?
@@ -479,6 +477,10 @@ static int do_read_inode(struct inode *inode)
}
init_idisk_time(inode);
+
+ /* Need all the flag bits */
+ f2fs_init_read_extent_tree(inode, node_page);
+
f2fs_put_page(node_page, 1);
stat_inc_inline_xattr(inode);
diff --git a/fs/f2fs/namei.c b/fs/f2fs/namei.c
index 58a91ce8fe08..46de782c2baa 100644
--- a/fs/f2fs/namei.c
+++ b/fs/f2fs/namei.c
@@ -284,8 +284,6 @@ static struct inode *f2fs_new_inode(struct user_namespace *mnt_userns,
}
F2FS_I(inode)->i_inline_xattr_size = xattr_size;
- f2fs_init_extent_tree(inode, NULL);
-
F2FS_I(inode)->i_flags =
f2fs_mask_flags(mode, F2FS_I(dir)->i_flags & F2FS_FL_INHERITED);
@@ -311,6 +309,8 @@ static struct inode *f2fs_new_inode(struct user_namespace *mnt_userns,
f2fs_set_inode_flags(inode);
+ f2fs_init_extent_tree(inode);
+
trace_f2fs_new_inode(inode, 0);
return inode;
--
2.39.0.rc0.267.gcb52ba06e7-goog
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply related [flat|nested] 19+ messages in thread* [f2fs-dev] [PATCH 6/6] f2fs: add block_age-based extent cache
2022-12-05 18:54 [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly Jaegeuk Kim
` (3 preceding siblings ...)
2022-12-05 18:54 ` [f2fs-dev] [PATCH 5/6] f2fs: allocate the extent_cache by default Jaegeuk Kim
@ 2022-12-05 18:54 ` Jaegeuk Kim
2022-12-06 19:06 ` [f2fs-dev] [PATCH 6/6 v2] " Jaegeuk Kim
2022-12-16 12:13 ` [f2fs-dev] [PATCH 6/6] " qixiaoyu
2022-12-08 14:43 ` [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly Chao Yu
5 siblings, 2 replies; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-05 18:54 UTC (permalink / raw)
To: linux-kernel, linux-f2fs-devel; +Cc: Jaegeuk Kim, qixiaoyu1, xiongping1
This patch introduces a runtime hot/cold data separation method
for f2fs, in order to improve the accuracy for data temperature
classification, reduce the garbage collection overhead after
long-term data updates.
Enhanced hot/cold data separation can record data block update
frequency as "age" of the extent per inode, and take use of the age
info to indicate better temperature type for data block allocation:
- It records total data blocks allocated since mount;
- When file extent has been updated, it calculate the count of data
blocks allocated since last update as the age of the extent;
- Before the data block allocated, it searches for the age info and
chooses the suitable segment for allocation.
Test and result:
- Prepare: create about 30000 files
* 3% for cold files (with cold file extension like .apk, from 3M to 10M)
* 50% for warm files (with random file extension like .FcDxq, from 1K
to 4M)
* 47% for hot files (with hot file extension like .db, from 1K to 256K)
- create(5%)/random update(90%)/delete(5%) the files
* total write amount is about 70G
* fsync will be called for .db files, and buffered write will be used
for other files
The storage of test device is large enough(128G) so that it will not
switch to SSR mode during the test.
Benefit: dirty segment count increment reduce about 14%
- before: Dirty +21110
- after: Dirty +18286
Signed-off-by: qixiaoyu1 <qixiaoyu1@xiaomi.com>
Signed-off-by: xiongping1 <xiongping1@xiaomi.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
Documentation/ABI/testing/sysfs-fs-f2fs | 14 ++
Documentation/filesystems/f2fs.rst | 4 +
fs/f2fs/debug.c | 21 +++
fs/f2fs/extent_cache.c | 182 +++++++++++++++++++++++-
fs/f2fs/f2fs.h | 38 +++++
fs/f2fs/file.c | 1 +
fs/f2fs/inode.c | 1 +
fs/f2fs/node.c | 10 +-
fs/f2fs/node.h | 1 +
fs/f2fs/segment.c | 33 +++++
fs/f2fs/shrinker.c | 10 +-
fs/f2fs/super.c | 14 ++
fs/f2fs/sysfs.c | 24 ++++
include/trace/events/f2fs.h | 86 ++++++++++-
14 files changed, 429 insertions(+), 10 deletions(-)
diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs
index 84a009aab1a1..9e3756625a81 100644
--- a/Documentation/ABI/testing/sysfs-fs-f2fs
+++ b/Documentation/ABI/testing/sysfs-fs-f2fs
@@ -655,3 +655,17 @@ Description: When space utilization exceeds this, do background DISCARD aggressi
Does DISCARD forcibly in a period of given min_discard_issue_time when the number
of discards is not 0 and set discard granularity to 1.
Default: 80
+
+What: /sys/fs/f2fs/<disk>/hot_data_age_threshold
+Date: November 2022
+Contact: "Ping Xiong" <xiongping1@xiaomi.com>
+Description: When DATA SEPARATION is on, it controls the age threshold to indicate
+ the data blocks as hot. By default it was initialized as 262144 blocks
+ (equals to 1GB).
+
+What: /sys/fs/f2fs/<disk>/warm_data_age_threshold
+Date: November 2022
+Contact: "Ping Xiong" <xiongping1@xiaomi.com>
+Description: When DATA SEPARATION is on, it controls the age threshold to indicate
+ the data blocks as warm. By default it was initialized as 2621440 blocks
+ (equals to 10GB).
diff --git a/Documentation/filesystems/f2fs.rst b/Documentation/filesystems/f2fs.rst
index 67e1f3e86f32..220f3e0d3f55 100644
--- a/Documentation/filesystems/f2fs.rst
+++ b/Documentation/filesystems/f2fs.rst
@@ -347,6 +347,10 @@ memory=%s Control memory mode. This supports "normal" and "low" modes.
Because of the nature of low memory devices, in this mode, f2fs
will try to save memory sometimes by sacrificing performance.
"normal" mode is the default mode and same as before.
+age_extent_cache Enable an age extent cache based on rb-tree. It records
+ data block update frequency of the extent per inode, in
+ order to provide better temperature hints for data block
+ allocation.
======================== ============================================================
Debugfs Entries
diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index a9baa121d829..8f1ef742551f 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -88,6 +88,9 @@ static void update_general_status(struct f2fs_sb_info *sbi)
si->hit_largest = atomic64_read(&sbi->read_hit_largest);
si->hit_total[EX_READ] += si->hit_largest;
+ /* block age extent_cache only */
+ si->allocated_data_blocks = atomic64_read(&sbi->allocated_data_blocks);
+
/* validation check of the segment numbers */
si->ndirty_node = get_pages(sbi, F2FS_DIRTY_NODES);
si->ndirty_dent = get_pages(sbi, F2FS_DIRTY_DENTS);
@@ -516,6 +519,22 @@ static int stat_show(struct seq_file *s, void *v)
seq_printf(s, " - Inner Struct Count: tree: %d(%d), node: %d\n",
si->ext_tree[EX_READ], si->zombie_tree[EX_READ],
si->ext_node[EX_READ]);
+ seq_puts(s, "\nExtent Cache (Block Age):\n");
+ seq_printf(s, " - Allocated Data Blocks: %llu\n",
+ si->allocated_data_blocks);
+ seq_printf(s, " - Hit Count: L1:%llu L2:%llu\n",
+ si->hit_cached[EX_BLOCK_AGE],
+ si->hit_rbtree[EX_BLOCK_AGE]);
+ seq_printf(s, " - Hit Ratio: %llu%% (%llu / %llu)\n",
+ !si->total_ext[EX_BLOCK_AGE] ? 0 :
+ div64_u64(si->hit_total[EX_BLOCK_AGE] * 100,
+ si->total_ext[EX_BLOCK_AGE]),
+ si->hit_total[EX_BLOCK_AGE],
+ si->total_ext[EX_BLOCK_AGE]);
+ seq_printf(s, " - Inner Struct Count: tree: %d(%d), node: %d\n",
+ si->ext_tree[EX_BLOCK_AGE],
+ si->zombie_tree[EX_BLOCK_AGE],
+ si->ext_node[EX_BLOCK_AGE]);
seq_puts(s, "\nBalancing F2FS Async:\n");
seq_printf(s, " - DIO (R: %4d, W: %4d)\n",
si->nr_dio_read, si->nr_dio_write);
@@ -586,6 +605,8 @@ static int stat_show(struct seq_file *s, void *v)
si->cache_mem >> 10);
seq_printf(s, " - read extent cache: %llu KB\n",
si->ext_mem[EX_READ] >> 10);
+ seq_printf(s, " - block age extent cache: %llu KB\n",
+ si->ext_mem[EX_BLOCK_AGE] >> 10);
seq_printf(s, " - paged : %llu KB\n",
si->page_mem >> 10);
}
diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
index 305f969e3ad1..2fc675c45606 100644
--- a/fs/f2fs/extent_cache.c
+++ b/fs/f2fs/extent_cache.c
@@ -6,6 +6,10 @@
* Copyright (c) 2015 Samsung Electronics
* Authors: Jaegeuk Kim <jaegeuk@kernel.org>
* Chao Yu <chao2.yu@samsung.com>
+ *
+ * block_age-based extent cache added by:
+ * Copyright (c) 2022 xiaomi Co., Ltd.
+ * http://www.xiaomi.com/
*/
#include <linux/fs.h>
@@ -18,6 +22,7 @@
static void __set_extent_info(struct extent_info *ei,
unsigned int fofs, unsigned int len,
block_t blk, bool keep_clen,
+ unsigned long age, unsigned long last_blocks,
enum extent_type type)
{
ei->fofs = fofs;
@@ -30,6 +35,9 @@ static void __set_extent_info(struct extent_info *ei,
#ifdef CONFIG_F2FS_FS_COMPRESSION
ei->c_len = 0;
#endif
+ } else if (type == EX_BLOCK_AGE) {
+ ei->age = age;
+ ei->last_blocks = last_blocks;
}
}
@@ -47,10 +55,27 @@ static bool __may_read_extent_tree(struct inode *inode)
return S_ISREG(inode->i_mode);
}
+static bool __may_age_extent_tree(struct inode *inode)
+{
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+
+ if (!test_opt(sbi, AGE_EXTENT_CACHE))
+ return false;
+ /* don't cache block age info for cold file */
+ if (is_inode_flag_set(inode, FI_COMPRESSED_FILE))
+ return false;
+ if (file_is_cold(inode))
+ return false;
+
+ return S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode);
+}
+
static bool __init_may_extent_tree(struct inode *inode, enum extent_type type)
{
if (type == EX_READ)
return __may_read_extent_tree(inode);
+ else if (type == EX_BLOCK_AGE)
+ return __may_age_extent_tree(inode);
return false;
}
@@ -90,6 +115,11 @@ static bool __is_extent_mergeable(struct extent_info *back,
#endif
return (back->fofs + back->len == front->fofs &&
back->blk + back->len == front->blk);
+ } else if (type == EX_BLOCK_AGE) {
+ return (back->fofs + back->len == front->fofs &&
+ abs(back->age - front->age) <= SAME_AGE_REGION &&
+ abs(back->last_blocks - front->last_blocks) <=
+ SAME_AGE_REGION);
}
return false;
}
@@ -489,11 +519,22 @@ void f2fs_init_read_extent_tree(struct inode *inode, struct page *ipage)
set_inode_flag(inode, FI_NO_EXTENT);
}
+void f2fs_init_age_extent_tree(struct inode *inode)
+{
+ if (!__init_may_extent_tree(inode, EX_BLOCK_AGE))
+ return;
+ __grab_extent_tree(inode, EX_BLOCK_AGE);
+}
+
void f2fs_init_extent_tree(struct inode *inode)
{
/* initialize read cache */
if (__init_may_extent_tree(inode, EX_READ))
__grab_extent_tree(inode, EX_READ);
+
+ /* initialize block age cache */
+ if (__init_may_extent_tree(inode, EX_BLOCK_AGE))
+ __grab_extent_tree(inode, EX_BLOCK_AGE);
}
static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
@@ -544,6 +585,8 @@ static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
if (type == EX_READ)
trace_f2fs_lookup_read_extent_tree_end(inode, pgofs, ei);
+ else if (type == EX_BLOCK_AGE)
+ trace_f2fs_lookup_age_extent_tree_end(inode, pgofs, ei);
return ret;
}
@@ -642,6 +685,10 @@ static void __update_extent_tree_range(struct inode *inode,
if (type == EX_READ)
trace_f2fs_update_read_extent_tree_range(inode, fofs, len,
tei->blk, 0);
+ else if (type == EX_BLOCK_AGE)
+ trace_f2fs_update_age_extent_tree_range(inode, fofs, len,
+ tei->age, tei->last_blocks);
+
write_lock(&et->lock);
if (type == EX_READ) {
@@ -694,6 +741,7 @@ static void __update_extent_tree_range(struct inode *inode,
__set_extent_info(&ei,
end, org_end - end,
end - dei.fofs + dei.blk, false,
+ dei.age, dei.last_blocks,
type);
en1 = __insert_extent_tree(sbi, et, &ei,
NULL, NULL, true);
@@ -702,6 +750,7 @@ static void __update_extent_tree_range(struct inode *inode,
__set_extent_info(&en->ei,
end, en->ei.len - (end - dei.fofs),
en->ei.blk + (end - dei.fofs), true,
+ dei.age, dei.last_blocks,
type);
next_en = en;
}
@@ -732,11 +781,15 @@ static void __update_extent_tree_range(struct inode *inode,
en = next_en;
}
+ if (type == EX_BLOCK_AGE)
+ goto update_age_extent_cache;
+
/* 3. update extent in read extent cache */
BUG_ON(type != EX_READ);
if (tei->blk) {
- __set_extent_info(&ei, fofs, len, tei->blk, false, EX_READ);
+ __set_extent_info(&ei, fofs, len, tei->blk, false,
+ 0, 0, EX_READ);
if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
__insert_extent_tree(sbi, et, &ei,
insert_p, insert_parent, leftmost);
@@ -758,7 +811,17 @@ static void __update_extent_tree_range(struct inode *inode,
et->largest_updated = false;
updated = true;
}
+ goto out_read_extent_cache;
+update_age_extent_cache:
+ if (!tei->last_blocks)
+ goto out_read_extent_cache;
+ __set_extent_info(&ei, fofs, len, 0, false,
+ tei->age, tei->last_blocks, EX_BLOCK_AGE);
+ if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
+ __insert_extent_tree(sbi, et, &ei,
+ insert_p, insert_parent, leftmost);
+out_read_extent_cache:
write_unlock(&et->lock);
if (updated)
@@ -796,7 +859,7 @@ void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
if (en)
goto unlock_out;
- __set_extent_info(&ei, fofs, llen, blkaddr, true, EX_READ);
+ __set_extent_info(&ei, fofs, llen, blkaddr, true, 0, 0, EX_READ);
ei.c_len = c_len;
if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
@@ -807,6 +870,71 @@ void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
}
#endif
+static unsigned long long __calculate_block_age(unsigned long long new,
+ unsigned long long old)
+{
+ if (new >= old)
+ return new - (new - old) * LAST_AGE_WEIGHT / 100;
+ else
+ return new + (old - new) * LAST_AGE_WEIGHT / 100;
+}
+
+/* This returns a new age and allocated blocks in ei */
+static int __get_new_block_age(struct inode *inode, struct extent_info *ei)
+{
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+ loff_t f_size = i_size_read(inode);
+ unsigned long long cur_blocks =
+ atomic64_read(&sbi->allocated_data_blocks);
+
+ /*
+ * When I/O is not aligned to a PAGE_SIZE, update will happen to the last
+ * file block even in seq write. So don't record age for newly last file
+ * block here.
+ */
+ if ((f_size >> PAGE_SHIFT) == ei->fofs && f_size & (PAGE_SIZE - 1) &&
+ ei->blk == NEW_ADDR)
+ return -EINVAL;
+
+ if (__lookup_extent_tree(inode, ei->fofs, ei, EX_BLOCK_AGE)) {
+ unsigned long long cur_age;
+
+ if (cur_blocks >= ei->last_blocks)
+ cur_age = cur_blocks - ei->last_blocks;
+ else
+ /* allocated_data_blocks overflow */
+ cur_age = ULLONG_MAX - ei->last_blocks + cur_blocks;
+
+ if (ei->age)
+ ei->age = __calculate_block_age(cur_age, ei->age);
+ else
+ ei->age = cur_age;
+ ei->last_blocks = cur_blocks;
+ WARN_ON(ei->age > cur_blocks);
+ return 0;
+ }
+
+ f2fs_bug_on(sbi, ei->blk == NULL_ADDR);
+
+ /* the data block was allocated for the first time */
+ if (ei->blk == NEW_ADDR)
+ goto out;
+
+ if (__is_valid_data_blkaddr(ei->blk) &&
+ !f2fs_is_valid_blkaddr(sbi, ei->blk, DATA_GENERIC_ENHANCE)) {
+ f2fs_bug_on(sbi, 1);
+ return -EINVAL;
+ }
+out:
+ /*
+ * init block age with zero, this can happen when the block age extent
+ * was reclaimed due to memory constraint or system reboot
+ */
+ ei->age = 0;
+ ei->last_blocks = cur_blocks;
+ return 0;
+}
+
static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type type)
{
struct extent_info ei;
@@ -823,6 +951,10 @@ static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type typ
ei.blk = NULL_ADDR;
else
ei.blk = dn->data_blkaddr;
+ } else if (type == EX_BLOCK_AGE) {
+ ei.blk = dn->data_blkaddr;
+ if (__get_new_block_age(dn->inode, &ei))
+ return;
}
__update_extent_tree_range(dn->inode, &ei, type);
}
@@ -940,6 +1072,43 @@ unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi, int nr_shrin
return __shrink_extent_tree(sbi, nr_shrink, EX_READ);
}
+/* block age extent cache operations */
+bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
+ struct extent_info *ei)
+{
+ if (!__may_extent_tree(inode, EX_BLOCK_AGE))
+ return false;
+
+ return __lookup_extent_tree(inode, pgofs, ei, EX_BLOCK_AGE);
+}
+
+void f2fs_update_age_extent_cache(struct dnode_of_data *dn)
+{
+ return __update_extent_cache(dn, EX_BLOCK_AGE);
+}
+
+void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
+ pgoff_t fofs, unsigned int len)
+{
+ struct extent_info ei = {
+ .fofs = fofs,
+ .len = len,
+ };
+
+ if (!__may_extent_tree(dn->inode, EX_BLOCK_AGE))
+ return;
+
+ __update_extent_tree_range(dn->inode, &ei, EX_BLOCK_AGE);
+}
+
+unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
+{
+ if (!test_opt(sbi, AGE_EXTENT_CACHE))
+ return 0;
+
+ return __shrink_extent_tree(sbi, nr_shrink, EX_BLOCK_AGE);
+}
+
static unsigned int __destroy_extent_node(struct inode *inode,
enum extent_type type)
{
@@ -960,6 +1129,7 @@ static unsigned int __destroy_extent_node(struct inode *inode,
void f2fs_destroy_extent_node(struct inode *inode)
{
__destroy_extent_node(inode, EX_READ);
+ __destroy_extent_node(inode, EX_BLOCK_AGE);
}
static void __drop_extent_tree(struct inode *inode, enum extent_type type)
@@ -988,6 +1158,7 @@ static void __drop_extent_tree(struct inode *inode, enum extent_type type)
void f2fs_drop_extent_tree(struct inode *inode)
{
__drop_extent_tree(inode, EX_READ);
+ __drop_extent_tree(inode, EX_BLOCK_AGE);
}
static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
@@ -1028,6 +1199,7 @@ static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
void f2fs_destroy_extent_tree(struct inode *inode)
{
__destroy_extent_tree(inode, EX_READ);
+ __destroy_extent_tree(inode, EX_BLOCK_AGE);
}
static void __init_extent_tree_info(struct extent_tree_info *eti)
@@ -1045,6 +1217,12 @@ static void __init_extent_tree_info(struct extent_tree_info *eti)
void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi)
{
__init_extent_tree_info(&sbi->extent_tree[EX_READ]);
+ __init_extent_tree_info(&sbi->extent_tree[EX_BLOCK_AGE]);
+
+ /* initialize for block age extents */
+ atomic64_set(&sbi->allocated_data_blocks, 0);
+ sbi->hot_data_age_threshold = DEF_HOT_DATA_AGE_THRESHOLD;
+ sbi->warm_data_age_threshold = DEF_WARM_DATA_AGE_THRESHOLD;
}
int __init f2fs_create_extent_cache(void)
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index ff940cba4600..eb71edcf70de 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -107,6 +107,7 @@ extern const char *f2fs_fault_name[FAULT_MAX];
#define F2FS_MOUNT_MERGE_CHECKPOINT 0x10000000
#define F2FS_MOUNT_GC_MERGE 0x20000000
#define F2FS_MOUNT_COMPRESS_CACHE 0x40000000
+#define F2FS_MOUNT_AGE_EXTENT_CACHE 0x80000000
#define F2FS_OPTION(sbi) ((sbi)->mount_opt)
#define clear_opt(sbi, option) (F2FS_OPTION(sbi).opt &= ~F2FS_MOUNT_##option)
@@ -607,9 +608,22 @@ enum {
/* number of extent info in extent cache we try to shrink */
#define READ_EXTENT_CACHE_SHRINK_NUMBER 128
+/* number of age extent info in extent cache we try to shrink */
+#define AGE_EXTENT_CACHE_SHRINK_NUMBER 128
+#define LAST_AGE_WEIGHT 30
+#define SAME_AGE_REGION 1024
+
+/*
+ * Define data block with age less than 1GB as hot data
+ * define data block with age less than 10GB but more than 1GB as warm data
+ */
+#define DEF_HOT_DATA_AGE_THRESHOLD 262144
+#define DEF_WARM_DATA_AGE_THRESHOLD 2621440
+
/* extent cache type */
enum extent_type {
EX_READ,
+ EX_BLOCK_AGE,
NR_EXTENT_CACHES,
};
@@ -637,6 +651,13 @@ struct extent_info {
unsigned int c_len;
#endif
};
+ /* block age extent_cache */
+ struct {
+ /* block age of the extent */
+ unsigned long long age;
+ /* last total blocks allocated */
+ unsigned long long last_blocks;
+ };
};
};
@@ -1653,6 +1674,11 @@ struct f2fs_sb_info {
/* for extent tree cache */
struct extent_tree_info extent_tree[NR_EXTENT_CACHES];
+ atomic64_t allocated_data_blocks; /* for block age extent_cache */
+
+ /* The threshold used for hot and warm data seperation*/
+ unsigned int hot_data_age_threshold;
+ unsigned int warm_data_age_threshold;
/* basic filesystem units */
unsigned int log_sectors_per_block; /* log2 sectors per block */
@@ -3857,6 +3883,8 @@ struct f2fs_stat_info {
unsigned long long ext_mem[NR_EXTENT_CACHES];
/* for read extent cache */
unsigned long long hit_largest;
+ /* for block age extent cache */
+ 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;
@@ -4168,6 +4196,16 @@ void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi,
int nr_shrink);
+/* block age extent cache ops */
+void f2fs_init_age_extent_tree(struct inode *inode);
+bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
+ struct extent_info *ei);
+void f2fs_update_age_extent_cache(struct dnode_of_data *dn);
+void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
+ pgoff_t fofs, unsigned int len);
+unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi,
+ int nr_shrink);
+
/*
* sysfs.c
*/
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index cbe7c24065c7..56c23b5e9d65 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -619,6 +619,7 @@ void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count)
fofs = f2fs_start_bidx_of_node(ofs_of_node(dn->node_page),
dn->inode) + ofs;
f2fs_update_read_extent_cache_range(dn, fofs, 0, len);
+ f2fs_update_age_extent_cache_range(dn, fofs, nr_free);
dec_valid_block_count(sbi, dn->inode, nr_free);
}
dn->ofs_in_node = ofs;
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index c845c16f97d0..ff6cf66ed46b 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -480,6 +480,7 @@ static int do_read_inode(struct inode *inode)
/* Need all the flag bits */
f2fs_init_read_extent_tree(inode, node_page);
+ f2fs_init_age_extent_tree(inode);
f2fs_put_page(node_page, 1);
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 07419c3e42a5..dde4c0458704 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -60,7 +60,7 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
avail_ram = val.totalram - val.totalhigh;
/*
- * give 25%, 25%, 50%, 50%, 50% memory for each components respectively
+ * give 25%, 25%, 50%, 50%, 25%, 25% memory for each components respectively
*/
if (type == FREE_NIDS) {
mem_size = (nm_i->nid_cnt[FREE_NID] *
@@ -85,14 +85,16 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
sizeof(struct ino_entry);
mem_size >>= PAGE_SHIFT;
res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
- } else if (type == READ_EXTENT_CACHE) {
- struct extent_tree_info *eti = &sbi->extent_tree[EX_READ];
+ } else if (type == READ_EXTENT_CACHE || type == AGE_EXTENT_CACHE) {
+ enum extent_type etype = type == READ_EXTENT_CACHE ?
+ EX_READ : EX_BLOCK_AGE;
+ struct extent_tree_info *eti = &sbi->extent_tree[etype];
mem_size = (atomic_read(&eti->total_ext_tree) *
sizeof(struct extent_tree) +
atomic_read(&eti->total_ext_node) *
sizeof(struct extent_node)) >> PAGE_SHIFT;
- res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
+ res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 2);
} else if (type == DISCARD_CACHE) {
mem_size = (atomic_read(&dcc->discard_cmd_cnt) *
sizeof(struct discard_cmd)) >> PAGE_SHIFT;
diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
index 0aa48704c77a..99454d46a939 100644
--- a/fs/f2fs/node.h
+++ b/fs/f2fs/node.h
@@ -147,6 +147,7 @@ enum mem_type {
DIRTY_DENTS, /* indicates dirty dentry pages */
INO_ENTRIES, /* indicates inode entries */
READ_EXTENT_CACHE, /* indicates read extent cache */
+ AGE_EXTENT_CACHE, /* indicates age extent cache */
DISCARD_CACHE, /* indicates memory of cached discard cmds */
COMPRESS_PAGE, /* indicates memory of cached compressed pages */
BASE_CHECK, /* check kernel status */
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 8722d1a13c17..dee712f7225f 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -453,6 +453,11 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
f2fs_shrink_read_extent_tree(sbi,
READ_EXTENT_CACHE_SHRINK_NUMBER);
+ /* try to shrink age extent cache when there is no enough memory */
+ if (!f2fs_available_free_memory(sbi, AGE_EXTENT_CACHE))
+ f2fs_shrink_age_extent_tree(sbi,
+ AGE_EXTENT_CACHE_SHRINK_NUMBER);
+
/* check the # of cached NAT entries */
if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
f2fs_try_to_free_nats(sbi, NAT_ENTRY_PER_BLOCK);
@@ -3151,10 +3156,28 @@ static int __get_segment_type_4(struct f2fs_io_info *fio)
}
}
+static int __get_age_segment_type(struct inode *inode, pgoff_t pgofs)
+{
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+ struct extent_info ei;
+
+ if (f2fs_lookup_age_extent_cache(inode, pgofs, &ei)) {
+ if (!ei.age)
+ return NO_CHECK_TYPE;
+ if (ei.age <= sbi->hot_data_age_threshold)
+ return CURSEG_HOT_DATA;
+ if (ei.age <= sbi->warm_data_age_threshold)
+ return CURSEG_WARM_DATA;
+ return CURSEG_COLD_DATA;
+ }
+ return NO_CHECK_TYPE;
+}
+
static int __get_segment_type_6(struct f2fs_io_info *fio)
{
if (fio->type == DATA) {
struct inode *inode = fio->page->mapping->host;
+ int type;
if (is_inode_flag_set(inode, FI_ALIGNED_WRITE))
return CURSEG_COLD_DATA_PINNED;
@@ -3169,6 +3192,11 @@ static int __get_segment_type_6(struct f2fs_io_info *fio)
}
if (file_is_cold(inode) || f2fs_need_compress_data(inode))
return CURSEG_COLD_DATA;
+
+ type = __get_age_segment_type(inode, fio->page->index);
+ if (type != NO_CHECK_TYPE)
+ return type;
+
if (file_is_hot(inode) ||
is_inode_flag_set(inode, FI_HOT_DATA) ||
f2fs_is_cow_file(inode))
@@ -3287,6 +3315,9 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr));
locate_dirty_segment(sbi, GET_SEGNO(sbi, *new_blkaddr));
+ if (IS_DATASEG(type))
+ atomic64_inc(&sbi->allocated_data_blocks);
+
up_write(&sit_i->sentry_lock);
if (page && IS_NODESEG(type)) {
@@ -3414,6 +3445,8 @@ void f2fs_outplace_write_data(struct dnode_of_data *dn,
struct f2fs_summary sum;
f2fs_bug_on(sbi, dn->data_blkaddr == NULL_ADDR);
+ if (fio->io_type == FS_DATA_IO || fio->io_type == FS_CP_DATA_IO)
+ f2fs_update_age_extent_cache(dn);
set_summary(&sum, dn->nid, dn->ofs_in_node, fio->version);
do_write_page(&sum, fio);
f2fs_update_data_blkaddr(dn, fio->new_blkaddr);
diff --git a/fs/f2fs/shrinker.c b/fs/f2fs/shrinker.c
index 33c490e69ae3..83d6fb97dcae 100644
--- a/fs/f2fs/shrinker.c
+++ b/fs/f2fs/shrinker.c
@@ -59,6 +59,9 @@ unsigned long f2fs_shrink_count(struct shrinker *shrink,
/* count read extent cache entries */
count += __count_extent_cache(sbi, EX_READ);
+ /* count block age extent cache entries */
+ count += __count_extent_cache(sbi, EX_BLOCK_AGE);
+
/* count clean nat cache entries */
count += __count_nat_entries(sbi);
@@ -102,8 +105,11 @@ unsigned long f2fs_shrink_scan(struct shrinker *shrink,
sbi->shrinker_run_no = run_no;
+ /* shrink extent cache entries */
+ freed += f2fs_shrink_age_extent_tree(sbi, nr >> 2);
+
/* shrink read extent cache entries */
- freed += f2fs_shrink_read_extent_tree(sbi, nr >> 1);
+ freed += f2fs_shrink_read_extent_tree(sbi, nr >> 2);
/* shrink clean nat cache entries */
if (freed < nr)
@@ -134,6 +140,8 @@ void f2fs_join_shrinker(struct f2fs_sb_info *sbi)
void f2fs_leave_shrinker(struct f2fs_sb_info *sbi)
{
f2fs_shrink_read_extent_tree(sbi, __count_extent_cache(sbi, EX_READ));
+ f2fs_shrink_age_extent_tree(sbi,
+ __count_extent_cache(sbi, EX_BLOCK_AGE));
spin_lock(&f2fs_list_lock);
list_del_init(&sbi->s_list);
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 10bd03bbefec..5bdab376b852 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -163,6 +163,7 @@ enum {
Opt_nogc_merge,
Opt_discard_unit,
Opt_memory_mode,
+ Opt_age_extent_cache,
Opt_err,
};
@@ -241,6 +242,7 @@ static match_table_t f2fs_tokens = {
{Opt_nogc_merge, "nogc_merge"},
{Opt_discard_unit, "discard_unit=%s"},
{Opt_memory_mode, "memory=%s"},
+ {Opt_age_extent_cache, "age_extent_cache"},
{Opt_err, NULL},
};
@@ -1257,6 +1259,9 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
}
kfree(name);
break;
+ case Opt_age_extent_cache:
+ set_opt(sbi, AGE_EXTENT_CACHE);
+ break;
default:
f2fs_err(sbi, "Unrecognized mount option \"%s\" or missing value",
p);
@@ -1958,6 +1963,8 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
seq_puts(seq, ",read_extent_cache");
else
seq_puts(seq, ",no_read_extent_cache");
+ if (test_opt(sbi, AGE_EXTENT_CACHE))
+ seq_puts(seq, ",age_extent_cache");
if (test_opt(sbi, DATA_FLUSH))
seq_puts(seq, ",data_flush");
@@ -2219,6 +2226,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
bool need_restart_flush = false, need_stop_flush = false;
bool need_restart_discard = false, need_stop_discard = false;
bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE);
+ bool no_age_extent_cache = !test_opt(sbi, AGE_EXTENT_CACHE);
bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT);
bool no_io_align = !F2FS_IO_ALIGNED(sbi);
bool no_atgc = !test_opt(sbi, ATGC);
@@ -2313,6 +2321,12 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
f2fs_warn(sbi, "switch extent_cache option is not allowed");
goto restore_opts;
}
+ /* disallow enable/disable age extent_cache dynamically */
+ if (no_age_extent_cache == !!test_opt(sbi, AGE_EXTENT_CACHE)) {
+ err = -EINVAL;
+ f2fs_warn(sbi, "switch age_extent_cache option is not allowed");
+ goto restore_opts;
+ }
if (no_io_align == !!F2FS_IO_ALIGNED(sbi)) {
err = -EINVAL;
diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
index a4745d596310..2ab215110596 100644
--- a/fs/f2fs/sysfs.c
+++ b/fs/f2fs/sysfs.c
@@ -668,6 +668,24 @@ static ssize_t __sbi_store(struct f2fs_attr *a,
return count;
}
+ if (!strcmp(a->attr.name, "hot_data_age_threshold")) {
+ if (t == 0 || t >= sbi->warm_data_age_threshold)
+ return -EINVAL;
+ if (t == *ui)
+ return count;
+ *ui = (unsigned int)t;
+ return count;
+ }
+
+ if (!strcmp(a->attr.name, "warm_data_age_threshold")) {
+ if (t == 0 || t <= sbi->hot_data_age_threshold)
+ return -EINVAL;
+ if (t == *ui)
+ return count;
+ *ui = (unsigned int)t;
+ return count;
+ }
+
*ui = (unsigned int)t;
return count;
@@ -923,6 +941,10 @@ F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, peak_atomic_write, peak_atomic_write);
F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, committed_atomic_block, committed_atomic_block);
F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, revoked_atomic_block, revoked_atomic_block);
+/* For block age extent cache */
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, hot_data_age_threshold, hot_data_age_threshold);
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, warm_data_age_threshold, warm_data_age_threshold);
+
#define ATTR_LIST(name) (&f2fs_attr_##name.attr)
static struct attribute *f2fs_attrs[] = {
ATTR_LIST(gc_urgent_sleep_time),
@@ -1018,6 +1040,8 @@ static struct attribute *f2fs_attrs[] = {
ATTR_LIST(peak_atomic_write),
ATTR_LIST(committed_atomic_block),
ATTR_LIST(revoked_atomic_block),
+ ATTR_LIST(hot_data_age_threshold),
+ ATTR_LIST(warm_data_age_threshold),
NULL,
};
ATTRIBUTE_GROUPS(f2fs);
diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h
index 2bb37892d2ba..31d994e6b4ca 100644
--- a/include/trace/events/f2fs.h
+++ b/include/trace/events/f2fs.h
@@ -49,6 +49,7 @@ TRACE_DEFINE_ENUM(CP_TRIMMED);
TRACE_DEFINE_ENUM(CP_PAUSE);
TRACE_DEFINE_ENUM(CP_RESIZE);
TRACE_DEFINE_ENUM(EX_READ);
+TRACE_DEFINE_ENUM(EX_BLOCK_AGE);
#define show_block_type(type) \
__print_symbolic(type, \
@@ -155,6 +156,11 @@ TRACE_DEFINE_ENUM(EX_READ);
{ COMPRESS_ZSTD, "ZSTD" }, \
{ COMPRESS_LZORLE, "LZO-RLE" })
+#define show_extent_type(type) \
+ __print_symbolic(type, \
+ { EX_READ, "Read" }, \
+ { EX_BLOCK_AGE, "Block Age" })
+
struct f2fs_sb_info;
struct f2fs_io_info;
struct extent_info;
@@ -1544,7 +1550,7 @@ TRACE_EVENT(f2fs_lookup_extent_tree_start,
TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, type = %s",
show_dev_ino(__entry),
__entry->pgofs,
- __entry->type == EX_READ ? "Read" : "N/A")
+ show_extent_type(__entry->type))
);
TRACE_EVENT_CONDITION(f2fs_lookup_read_extent_tree_end,
@@ -1583,6 +1589,45 @@ TRACE_EVENT_CONDITION(f2fs_lookup_read_extent_tree_end,
__entry->blk)
);
+TRACE_EVENT_CONDITION(f2fs_lookup_age_extent_tree_end,
+
+ TP_PROTO(struct inode *inode, unsigned int pgofs,
+ struct extent_info *ei),
+
+ TP_ARGS(inode, pgofs, ei),
+
+ TP_CONDITION(ei),
+
+ TP_STRUCT__entry(
+ __field(dev_t, dev)
+ __field(ino_t, ino)
+ __field(unsigned int, pgofs)
+ __field(unsigned int, fofs)
+ __field(unsigned int, len)
+ __field(unsigned long long, age)
+ __field(unsigned long long, blocks)
+ ),
+
+ TP_fast_assign(
+ __entry->dev = inode->i_sb->s_dev;
+ __entry->ino = inode->i_ino;
+ __entry->pgofs = pgofs;
+ __entry->fofs = ei->fofs;
+ __entry->len = ei->len;
+ __entry->age = ei->age;
+ __entry->blocks = ei->last_blocks;
+ ),
+
+ TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, "
+ "age_ext_info(fofs: %u, len: %u, age: %llu, blocks: %llu)",
+ show_dev_ino(__entry),
+ __entry->pgofs,
+ __entry->fofs,
+ __entry->len,
+ __entry->age,
+ __entry->blocks)
+);
+
TRACE_EVENT(f2fs_update_read_extent_tree_range,
TP_PROTO(struct inode *inode, unsigned int pgofs, unsigned int len,
@@ -1618,6 +1663,41 @@ TRACE_EVENT(f2fs_update_read_extent_tree_range,
__entry->c_len)
);
+TRACE_EVENT(f2fs_update_age_extent_tree_range,
+
+ TP_PROTO(struct inode *inode, unsigned int pgofs, unsigned int len,
+ unsigned long long age,
+ unsigned long long last_blks),
+
+ TP_ARGS(inode, pgofs, len, age, last_blks),
+
+ TP_STRUCT__entry(
+ __field(dev_t, dev)
+ __field(ino_t, ino)
+ __field(unsigned int, pgofs)
+ __field(unsigned int, len)
+ __field(unsigned long long, age)
+ __field(unsigned long long, blocks)
+ ),
+
+ TP_fast_assign(
+ __entry->dev = inode->i_sb->s_dev;
+ __entry->ino = inode->i_ino;
+ __entry->pgofs = pgofs;
+ __entry->len = len;
+ __entry->age = age;
+ __entry->blocks = last_blks;
+ ),
+
+ TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, "
+ "len = %u, age = %llu, blocks = %llu",
+ show_dev_ino(__entry),
+ __entry->pgofs,
+ __entry->len,
+ __entry->age,
+ __entry->blocks)
+);
+
TRACE_EVENT(f2fs_shrink_extent_tree,
TP_PROTO(struct f2fs_sb_info *sbi, unsigned int node_cnt,
@@ -1643,7 +1723,7 @@ TRACE_EVENT(f2fs_shrink_extent_tree,
show_dev(__entry->dev),
__entry->node_cnt,
__entry->tree_cnt,
- __entry->type == EX_READ ? "Read" : "N/A")
+ show_extent_type(__entry->type))
);
TRACE_EVENT(f2fs_destroy_extent_tree,
@@ -1670,7 +1750,7 @@ TRACE_EVENT(f2fs_destroy_extent_tree,
TP_printk("dev = (%d,%d), ino = %lu, destroyed: node_cnt = %u, type = %s",
show_dev_ino(__entry),
__entry->node_cnt,
- __entry->type == EX_READ ? "Read" : "N/A")
+ show_extent_type(__entry->type))
);
DECLARE_EVENT_CLASS(f2fs_sync_dirty_inodes,
--
2.39.0.rc0.267.gcb52ba06e7-goog
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply related [flat|nested] 19+ messages in thread* Re: [f2fs-dev] [PATCH 6/6 v2] f2fs: add block_age-based extent cache
2022-12-05 18:54 ` [f2fs-dev] [PATCH 6/6] f2fs: add block_age-based extent cache Jaegeuk Kim
@ 2022-12-06 19:06 ` Jaegeuk Kim
2022-12-16 12:13 ` [f2fs-dev] [PATCH 6/6] " qixiaoyu
1 sibling, 0 replies; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-06 19:06 UTC (permalink / raw)
To: linux-kernel, linux-f2fs-devel; +Cc: xiongping1, qixiaoyu1
This patch introduces a runtime hot/cold data separation method
for f2fs, in order to improve the accuracy for data temperature
classification, reduce the garbage collection overhead after
long-term data updates.
Enhanced hot/cold data separation can record data block update
frequency as "age" of the extent per inode, and take use of the age
info to indicate better temperature type for data block allocation:
- It records total data blocks allocated since mount;
- When file extent has been updated, it calculate the count of data
blocks allocated since last update as the age of the extent;
- Before the data block allocated, it searches for the age info and
chooses the suitable segment for allocation.
Test and result:
- Prepare: create about 30000 files
* 3% for cold files (with cold file extension like .apk, from 3M to 10M)
* 50% for warm files (with random file extension like .FcDxq, from 1K
to 4M)
* 47% for hot files (with hot file extension like .db, from 1K to 256K)
- create(5%)/random update(90%)/delete(5%) the files
* total write amount is about 70G
* fsync will be called for .db files, and buffered write will be used
for other files
The storage of test device is large enough(128G) so that it will not
switch to SSR mode during the test.
Benefit: dirty segment count increment reduce about 14%
- before: Dirty +21110
- after: Dirty +18286
Signed-off-by: qixiaoyu1 <qixiaoyu1@xiaomi.com>
Signed-off-by: xiongping1 <xiongping1@xiaomi.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
Change log from v1:
- fix udiv
Documentation/ABI/testing/sysfs-fs-f2fs | 14 ++
Documentation/filesystems/f2fs.rst | 4 +
fs/f2fs/debug.c | 21 +++
fs/f2fs/extent_cache.c | 183 +++++++++++++++++++++++-
fs/f2fs/f2fs.h | 38 +++++
fs/f2fs/file.c | 1 +
fs/f2fs/inode.c | 1 +
fs/f2fs/node.c | 10 +-
fs/f2fs/node.h | 1 +
fs/f2fs/segment.c | 33 +++++
fs/f2fs/shrinker.c | 10 +-
fs/f2fs/super.c | 14 ++
fs/f2fs/sysfs.c | 24 ++++
include/trace/events/f2fs.h | 86 ++++++++++-
14 files changed, 430 insertions(+), 10 deletions(-)
diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs
index 84a009aab1a1..9e3756625a81 100644
--- a/Documentation/ABI/testing/sysfs-fs-f2fs
+++ b/Documentation/ABI/testing/sysfs-fs-f2fs
@@ -655,3 +655,17 @@ Description: When space utilization exceeds this, do background DISCARD aggressi
Does DISCARD forcibly in a period of given min_discard_issue_time when the number
of discards is not 0 and set discard granularity to 1.
Default: 80
+
+What: /sys/fs/f2fs/<disk>/hot_data_age_threshold
+Date: November 2022
+Contact: "Ping Xiong" <xiongping1@xiaomi.com>
+Description: When DATA SEPARATION is on, it controls the age threshold to indicate
+ the data blocks as hot. By default it was initialized as 262144 blocks
+ (equals to 1GB).
+
+What: /sys/fs/f2fs/<disk>/warm_data_age_threshold
+Date: November 2022
+Contact: "Ping Xiong" <xiongping1@xiaomi.com>
+Description: When DATA SEPARATION is on, it controls the age threshold to indicate
+ the data blocks as warm. By default it was initialized as 2621440 blocks
+ (equals to 10GB).
diff --git a/Documentation/filesystems/f2fs.rst b/Documentation/filesystems/f2fs.rst
index 67e1f3e86f32..220f3e0d3f55 100644
--- a/Documentation/filesystems/f2fs.rst
+++ b/Documentation/filesystems/f2fs.rst
@@ -347,6 +347,10 @@ memory=%s Control memory mode. This supports "normal" and "low" modes.
Because of the nature of low memory devices, in this mode, f2fs
will try to save memory sometimes by sacrificing performance.
"normal" mode is the default mode and same as before.
+age_extent_cache Enable an age extent cache based on rb-tree. It records
+ data block update frequency of the extent per inode, in
+ order to provide better temperature hints for data block
+ allocation.
======================== ============================================================
Debugfs Entries
diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index a9baa121d829..8f1ef742551f 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -88,6 +88,9 @@ static void update_general_status(struct f2fs_sb_info *sbi)
si->hit_largest = atomic64_read(&sbi->read_hit_largest);
si->hit_total[EX_READ] += si->hit_largest;
+ /* block age extent_cache only */
+ si->allocated_data_blocks = atomic64_read(&sbi->allocated_data_blocks);
+
/* validation check of the segment numbers */
si->ndirty_node = get_pages(sbi, F2FS_DIRTY_NODES);
si->ndirty_dent = get_pages(sbi, F2FS_DIRTY_DENTS);
@@ -516,6 +519,22 @@ static int stat_show(struct seq_file *s, void *v)
seq_printf(s, " - Inner Struct Count: tree: %d(%d), node: %d\n",
si->ext_tree[EX_READ], si->zombie_tree[EX_READ],
si->ext_node[EX_READ]);
+ seq_puts(s, "\nExtent Cache (Block Age):\n");
+ seq_printf(s, " - Allocated Data Blocks: %llu\n",
+ si->allocated_data_blocks);
+ seq_printf(s, " - Hit Count: L1:%llu L2:%llu\n",
+ si->hit_cached[EX_BLOCK_AGE],
+ si->hit_rbtree[EX_BLOCK_AGE]);
+ seq_printf(s, " - Hit Ratio: %llu%% (%llu / %llu)\n",
+ !si->total_ext[EX_BLOCK_AGE] ? 0 :
+ div64_u64(si->hit_total[EX_BLOCK_AGE] * 100,
+ si->total_ext[EX_BLOCK_AGE]),
+ si->hit_total[EX_BLOCK_AGE],
+ si->total_ext[EX_BLOCK_AGE]);
+ seq_printf(s, " - Inner Struct Count: tree: %d(%d), node: %d\n",
+ si->ext_tree[EX_BLOCK_AGE],
+ si->zombie_tree[EX_BLOCK_AGE],
+ si->ext_node[EX_BLOCK_AGE]);
seq_puts(s, "\nBalancing F2FS Async:\n");
seq_printf(s, " - DIO (R: %4d, W: %4d)\n",
si->nr_dio_read, si->nr_dio_write);
@@ -586,6 +605,8 @@ static int stat_show(struct seq_file *s, void *v)
si->cache_mem >> 10);
seq_printf(s, " - read extent cache: %llu KB\n",
si->ext_mem[EX_READ] >> 10);
+ seq_printf(s, " - block age extent cache: %llu KB\n",
+ si->ext_mem[EX_BLOCK_AGE] >> 10);
seq_printf(s, " - paged : %llu KB\n",
si->page_mem >> 10);
}
diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
index 305f969e3ad1..1bd38a78ebba 100644
--- a/fs/f2fs/extent_cache.c
+++ b/fs/f2fs/extent_cache.c
@@ -6,6 +6,10 @@
* Copyright (c) 2015 Samsung Electronics
* Authors: Jaegeuk Kim <jaegeuk@kernel.org>
* Chao Yu <chao2.yu@samsung.com>
+ *
+ * block_age-based extent cache added by:
+ * Copyright (c) 2022 xiaomi Co., Ltd.
+ * http://www.xiaomi.com/
*/
#include <linux/fs.h>
@@ -18,6 +22,7 @@
static void __set_extent_info(struct extent_info *ei,
unsigned int fofs, unsigned int len,
block_t blk, bool keep_clen,
+ unsigned long age, unsigned long last_blocks,
enum extent_type type)
{
ei->fofs = fofs;
@@ -30,6 +35,9 @@ static void __set_extent_info(struct extent_info *ei,
#ifdef CONFIG_F2FS_FS_COMPRESSION
ei->c_len = 0;
#endif
+ } else if (type == EX_BLOCK_AGE) {
+ ei->age = age;
+ ei->last_blocks = last_blocks;
}
}
@@ -47,10 +55,27 @@ static bool __may_read_extent_tree(struct inode *inode)
return S_ISREG(inode->i_mode);
}
+static bool __may_age_extent_tree(struct inode *inode)
+{
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+
+ if (!test_opt(sbi, AGE_EXTENT_CACHE))
+ return false;
+ /* don't cache block age info for cold file */
+ if (is_inode_flag_set(inode, FI_COMPRESSED_FILE))
+ return false;
+ if (file_is_cold(inode))
+ return false;
+
+ return S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode);
+}
+
static bool __init_may_extent_tree(struct inode *inode, enum extent_type type)
{
if (type == EX_READ)
return __may_read_extent_tree(inode);
+ else if (type == EX_BLOCK_AGE)
+ return __may_age_extent_tree(inode);
return false;
}
@@ -90,6 +115,11 @@ static bool __is_extent_mergeable(struct extent_info *back,
#endif
return (back->fofs + back->len == front->fofs &&
back->blk + back->len == front->blk);
+ } else if (type == EX_BLOCK_AGE) {
+ return (back->fofs + back->len == front->fofs &&
+ abs(back->age - front->age) <= SAME_AGE_REGION &&
+ abs(back->last_blocks - front->last_blocks) <=
+ SAME_AGE_REGION);
}
return false;
}
@@ -489,11 +519,22 @@ void f2fs_init_read_extent_tree(struct inode *inode, struct page *ipage)
set_inode_flag(inode, FI_NO_EXTENT);
}
+void f2fs_init_age_extent_tree(struct inode *inode)
+{
+ if (!__init_may_extent_tree(inode, EX_BLOCK_AGE))
+ return;
+ __grab_extent_tree(inode, EX_BLOCK_AGE);
+}
+
void f2fs_init_extent_tree(struct inode *inode)
{
/* initialize read cache */
if (__init_may_extent_tree(inode, EX_READ))
__grab_extent_tree(inode, EX_READ);
+
+ /* initialize block age cache */
+ if (__init_may_extent_tree(inode, EX_BLOCK_AGE))
+ __grab_extent_tree(inode, EX_BLOCK_AGE);
}
static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
@@ -544,6 +585,8 @@ static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
if (type == EX_READ)
trace_f2fs_lookup_read_extent_tree_end(inode, pgofs, ei);
+ else if (type == EX_BLOCK_AGE)
+ trace_f2fs_lookup_age_extent_tree_end(inode, pgofs, ei);
return ret;
}
@@ -642,6 +685,10 @@ static void __update_extent_tree_range(struct inode *inode,
if (type == EX_READ)
trace_f2fs_update_read_extent_tree_range(inode, fofs, len,
tei->blk, 0);
+ else if (type == EX_BLOCK_AGE)
+ trace_f2fs_update_age_extent_tree_range(inode, fofs, len,
+ tei->age, tei->last_blocks);
+
write_lock(&et->lock);
if (type == EX_READ) {
@@ -694,6 +741,7 @@ static void __update_extent_tree_range(struct inode *inode,
__set_extent_info(&ei,
end, org_end - end,
end - dei.fofs + dei.blk, false,
+ dei.age, dei.last_blocks,
type);
en1 = __insert_extent_tree(sbi, et, &ei,
NULL, NULL, true);
@@ -702,6 +750,7 @@ static void __update_extent_tree_range(struct inode *inode,
__set_extent_info(&en->ei,
end, en->ei.len - (end - dei.fofs),
en->ei.blk + (end - dei.fofs), true,
+ dei.age, dei.last_blocks,
type);
next_en = en;
}
@@ -732,11 +781,15 @@ static void __update_extent_tree_range(struct inode *inode,
en = next_en;
}
+ if (type == EX_BLOCK_AGE)
+ goto update_age_extent_cache;
+
/* 3. update extent in read extent cache */
BUG_ON(type != EX_READ);
if (tei->blk) {
- __set_extent_info(&ei, fofs, len, tei->blk, false, EX_READ);
+ __set_extent_info(&ei, fofs, len, tei->blk, false,
+ 0, 0, EX_READ);
if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
__insert_extent_tree(sbi, et, &ei,
insert_p, insert_parent, leftmost);
@@ -758,7 +811,17 @@ static void __update_extent_tree_range(struct inode *inode,
et->largest_updated = false;
updated = true;
}
+ goto out_read_extent_cache;
+update_age_extent_cache:
+ if (!tei->last_blocks)
+ goto out_read_extent_cache;
+ __set_extent_info(&ei, fofs, len, 0, false,
+ tei->age, tei->last_blocks, EX_BLOCK_AGE);
+ if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
+ __insert_extent_tree(sbi, et, &ei,
+ insert_p, insert_parent, leftmost);
+out_read_extent_cache:
write_unlock(&et->lock);
if (updated)
@@ -796,7 +859,7 @@ void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
if (en)
goto unlock_out;
- __set_extent_info(&ei, fofs, llen, blkaddr, true, EX_READ);
+ __set_extent_info(&ei, fofs, llen, blkaddr, true, 0, 0, EX_READ);
ei.c_len = c_len;
if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
@@ -807,6 +870,72 @@ void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
}
#endif
+static unsigned long long __calculate_block_age(unsigned long long new,
+ unsigned long long old)
+{
+ unsigned long long diff;
+
+ diff = (new >= old) ? new - (new - old) : new + (old - new);
+
+ return div_u64(diff * LAST_AGE_WEIGHT, 100);
+}
+
+/* This returns a new age and allocated blocks in ei */
+static int __get_new_block_age(struct inode *inode, struct extent_info *ei)
+{
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+ loff_t f_size = i_size_read(inode);
+ unsigned long long cur_blocks =
+ atomic64_read(&sbi->allocated_data_blocks);
+
+ /*
+ * When I/O is not aligned to a PAGE_SIZE, update will happen to the last
+ * file block even in seq write. So don't record age for newly last file
+ * block here.
+ */
+ if ((f_size >> PAGE_SHIFT) == ei->fofs && f_size & (PAGE_SIZE - 1) &&
+ ei->blk == NEW_ADDR)
+ return -EINVAL;
+
+ if (__lookup_extent_tree(inode, ei->fofs, ei, EX_BLOCK_AGE)) {
+ unsigned long long cur_age;
+
+ if (cur_blocks >= ei->last_blocks)
+ cur_age = cur_blocks - ei->last_blocks;
+ else
+ /* allocated_data_blocks overflow */
+ cur_age = ULLONG_MAX - ei->last_blocks + cur_blocks;
+
+ if (ei->age)
+ ei->age = __calculate_block_age(cur_age, ei->age);
+ else
+ ei->age = cur_age;
+ ei->last_blocks = cur_blocks;
+ WARN_ON(ei->age > cur_blocks);
+ return 0;
+ }
+
+ f2fs_bug_on(sbi, ei->blk == NULL_ADDR);
+
+ /* the data block was allocated for the first time */
+ if (ei->blk == NEW_ADDR)
+ goto out;
+
+ if (__is_valid_data_blkaddr(ei->blk) &&
+ !f2fs_is_valid_blkaddr(sbi, ei->blk, DATA_GENERIC_ENHANCE)) {
+ f2fs_bug_on(sbi, 1);
+ return -EINVAL;
+ }
+out:
+ /*
+ * init block age with zero, this can happen when the block age extent
+ * was reclaimed due to memory constraint or system reboot
+ */
+ ei->age = 0;
+ ei->last_blocks = cur_blocks;
+ return 0;
+}
+
static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type type)
{
struct extent_info ei;
@@ -823,6 +952,10 @@ static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type typ
ei.blk = NULL_ADDR;
else
ei.blk = dn->data_blkaddr;
+ } else if (type == EX_BLOCK_AGE) {
+ ei.blk = dn->data_blkaddr;
+ if (__get_new_block_age(dn->inode, &ei))
+ return;
}
__update_extent_tree_range(dn->inode, &ei, type);
}
@@ -940,6 +1073,43 @@ unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi, int nr_shrin
return __shrink_extent_tree(sbi, nr_shrink, EX_READ);
}
+/* block age extent cache operations */
+bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
+ struct extent_info *ei)
+{
+ if (!__may_extent_tree(inode, EX_BLOCK_AGE))
+ return false;
+
+ return __lookup_extent_tree(inode, pgofs, ei, EX_BLOCK_AGE);
+}
+
+void f2fs_update_age_extent_cache(struct dnode_of_data *dn)
+{
+ return __update_extent_cache(dn, EX_BLOCK_AGE);
+}
+
+void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
+ pgoff_t fofs, unsigned int len)
+{
+ struct extent_info ei = {
+ .fofs = fofs,
+ .len = len,
+ };
+
+ if (!__may_extent_tree(dn->inode, EX_BLOCK_AGE))
+ return;
+
+ __update_extent_tree_range(dn->inode, &ei, EX_BLOCK_AGE);
+}
+
+unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
+{
+ if (!test_opt(sbi, AGE_EXTENT_CACHE))
+ return 0;
+
+ return __shrink_extent_tree(sbi, nr_shrink, EX_BLOCK_AGE);
+}
+
static unsigned int __destroy_extent_node(struct inode *inode,
enum extent_type type)
{
@@ -960,6 +1130,7 @@ static unsigned int __destroy_extent_node(struct inode *inode,
void f2fs_destroy_extent_node(struct inode *inode)
{
__destroy_extent_node(inode, EX_READ);
+ __destroy_extent_node(inode, EX_BLOCK_AGE);
}
static void __drop_extent_tree(struct inode *inode, enum extent_type type)
@@ -988,6 +1159,7 @@ static void __drop_extent_tree(struct inode *inode, enum extent_type type)
void f2fs_drop_extent_tree(struct inode *inode)
{
__drop_extent_tree(inode, EX_READ);
+ __drop_extent_tree(inode, EX_BLOCK_AGE);
}
static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
@@ -1028,6 +1200,7 @@ static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
void f2fs_destroy_extent_tree(struct inode *inode)
{
__destroy_extent_tree(inode, EX_READ);
+ __destroy_extent_tree(inode, EX_BLOCK_AGE);
}
static void __init_extent_tree_info(struct extent_tree_info *eti)
@@ -1045,6 +1218,12 @@ static void __init_extent_tree_info(struct extent_tree_info *eti)
void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi)
{
__init_extent_tree_info(&sbi->extent_tree[EX_READ]);
+ __init_extent_tree_info(&sbi->extent_tree[EX_BLOCK_AGE]);
+
+ /* initialize for block age extents */
+ atomic64_set(&sbi->allocated_data_blocks, 0);
+ sbi->hot_data_age_threshold = DEF_HOT_DATA_AGE_THRESHOLD;
+ sbi->warm_data_age_threshold = DEF_WARM_DATA_AGE_THRESHOLD;
}
int __init f2fs_create_extent_cache(void)
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index ec52e06f8e61..e8953c3dc81a 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -107,6 +107,7 @@ extern const char *f2fs_fault_name[FAULT_MAX];
#define F2FS_MOUNT_MERGE_CHECKPOINT 0x10000000
#define F2FS_MOUNT_GC_MERGE 0x20000000
#define F2FS_MOUNT_COMPRESS_CACHE 0x40000000
+#define F2FS_MOUNT_AGE_EXTENT_CACHE 0x80000000
#define F2FS_OPTION(sbi) ((sbi)->mount_opt)
#define clear_opt(sbi, option) (F2FS_OPTION(sbi).opt &= ~F2FS_MOUNT_##option)
@@ -607,9 +608,22 @@ enum {
/* number of extent info in extent cache we try to shrink */
#define READ_EXTENT_CACHE_SHRINK_NUMBER 128
+/* number of age extent info in extent cache we try to shrink */
+#define AGE_EXTENT_CACHE_SHRINK_NUMBER 128
+#define LAST_AGE_WEIGHT 30
+#define SAME_AGE_REGION 1024
+
+/*
+ * Define data block with age less than 1GB as hot data
+ * define data block with age less than 10GB but more than 1GB as warm data
+ */
+#define DEF_HOT_DATA_AGE_THRESHOLD 262144
+#define DEF_WARM_DATA_AGE_THRESHOLD 2621440
+
/* extent cache type */
enum extent_type {
EX_READ,
+ EX_BLOCK_AGE,
NR_EXTENT_CACHES,
};
@@ -637,6 +651,13 @@ struct extent_info {
unsigned int c_len;
#endif
};
+ /* block age extent_cache */
+ struct {
+ /* block age of the extent */
+ unsigned long long age;
+ /* last total blocks allocated */
+ unsigned long long last_blocks;
+ };
};
};
@@ -1653,6 +1674,11 @@ struct f2fs_sb_info {
/* for extent tree cache */
struct extent_tree_info extent_tree[NR_EXTENT_CACHES];
+ atomic64_t allocated_data_blocks; /* for block age extent_cache */
+
+ /* The threshold used for hot and warm data seperation*/
+ unsigned int hot_data_age_threshold;
+ unsigned int warm_data_age_threshold;
/* basic filesystem units */
unsigned int log_sectors_per_block; /* log2 sectors per block */
@@ -3857,6 +3883,8 @@ struct f2fs_stat_info {
unsigned long long ext_mem[NR_EXTENT_CACHES];
/* for read extent cache */
unsigned long long hit_largest;
+ /* for block age extent cache */
+ 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;
@@ -4168,6 +4196,16 @@ void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi,
int nr_shrink);
+/* block age extent cache ops */
+void f2fs_init_age_extent_tree(struct inode *inode);
+bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
+ struct extent_info *ei);
+void f2fs_update_age_extent_cache(struct dnode_of_data *dn);
+void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
+ pgoff_t fofs, unsigned int len);
+unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi,
+ int nr_shrink);
+
/*
* sysfs.c
*/
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index cbe7c24065c7..56c23b5e9d65 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -619,6 +619,7 @@ void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count)
fofs = f2fs_start_bidx_of_node(ofs_of_node(dn->node_page),
dn->inode) + ofs;
f2fs_update_read_extent_cache_range(dn, fofs, 0, len);
+ f2fs_update_age_extent_cache_range(dn, fofs, nr_free);
dec_valid_block_count(sbi, dn->inode, nr_free);
}
dn->ofs_in_node = ofs;
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index c845c16f97d0..ff6cf66ed46b 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -480,6 +480,7 @@ static int do_read_inode(struct inode *inode)
/* Need all the flag bits */
f2fs_init_read_extent_tree(inode, node_page);
+ f2fs_init_age_extent_tree(inode);
f2fs_put_page(node_page, 1);
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 07419c3e42a5..dde4c0458704 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -60,7 +60,7 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
avail_ram = val.totalram - val.totalhigh;
/*
- * give 25%, 25%, 50%, 50%, 50% memory for each components respectively
+ * give 25%, 25%, 50%, 50%, 25%, 25% memory for each components respectively
*/
if (type == FREE_NIDS) {
mem_size = (nm_i->nid_cnt[FREE_NID] *
@@ -85,14 +85,16 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
sizeof(struct ino_entry);
mem_size >>= PAGE_SHIFT;
res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
- } else if (type == READ_EXTENT_CACHE) {
- struct extent_tree_info *eti = &sbi->extent_tree[EX_READ];
+ } else if (type == READ_EXTENT_CACHE || type == AGE_EXTENT_CACHE) {
+ enum extent_type etype = type == READ_EXTENT_CACHE ?
+ EX_READ : EX_BLOCK_AGE;
+ struct extent_tree_info *eti = &sbi->extent_tree[etype];
mem_size = (atomic_read(&eti->total_ext_tree) *
sizeof(struct extent_tree) +
atomic_read(&eti->total_ext_node) *
sizeof(struct extent_node)) >> PAGE_SHIFT;
- res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
+ res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 2);
} else if (type == DISCARD_CACHE) {
mem_size = (atomic_read(&dcc->discard_cmd_cnt) *
sizeof(struct discard_cmd)) >> PAGE_SHIFT;
diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
index 0aa48704c77a..99454d46a939 100644
--- a/fs/f2fs/node.h
+++ b/fs/f2fs/node.h
@@ -147,6 +147,7 @@ enum mem_type {
DIRTY_DENTS, /* indicates dirty dentry pages */
INO_ENTRIES, /* indicates inode entries */
READ_EXTENT_CACHE, /* indicates read extent cache */
+ AGE_EXTENT_CACHE, /* indicates age extent cache */
DISCARD_CACHE, /* indicates memory of cached discard cmds */
COMPRESS_PAGE, /* indicates memory of cached compressed pages */
BASE_CHECK, /* check kernel status */
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 8722d1a13c17..dee712f7225f 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -453,6 +453,11 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
f2fs_shrink_read_extent_tree(sbi,
READ_EXTENT_CACHE_SHRINK_NUMBER);
+ /* try to shrink age extent cache when there is no enough memory */
+ if (!f2fs_available_free_memory(sbi, AGE_EXTENT_CACHE))
+ f2fs_shrink_age_extent_tree(sbi,
+ AGE_EXTENT_CACHE_SHRINK_NUMBER);
+
/* check the # of cached NAT entries */
if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
f2fs_try_to_free_nats(sbi, NAT_ENTRY_PER_BLOCK);
@@ -3151,10 +3156,28 @@ static int __get_segment_type_4(struct f2fs_io_info *fio)
}
}
+static int __get_age_segment_type(struct inode *inode, pgoff_t pgofs)
+{
+ struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+ struct extent_info ei;
+
+ if (f2fs_lookup_age_extent_cache(inode, pgofs, &ei)) {
+ if (!ei.age)
+ return NO_CHECK_TYPE;
+ if (ei.age <= sbi->hot_data_age_threshold)
+ return CURSEG_HOT_DATA;
+ if (ei.age <= sbi->warm_data_age_threshold)
+ return CURSEG_WARM_DATA;
+ return CURSEG_COLD_DATA;
+ }
+ return NO_CHECK_TYPE;
+}
+
static int __get_segment_type_6(struct f2fs_io_info *fio)
{
if (fio->type == DATA) {
struct inode *inode = fio->page->mapping->host;
+ int type;
if (is_inode_flag_set(inode, FI_ALIGNED_WRITE))
return CURSEG_COLD_DATA_PINNED;
@@ -3169,6 +3192,11 @@ static int __get_segment_type_6(struct f2fs_io_info *fio)
}
if (file_is_cold(inode) || f2fs_need_compress_data(inode))
return CURSEG_COLD_DATA;
+
+ type = __get_age_segment_type(inode, fio->page->index);
+ if (type != NO_CHECK_TYPE)
+ return type;
+
if (file_is_hot(inode) ||
is_inode_flag_set(inode, FI_HOT_DATA) ||
f2fs_is_cow_file(inode))
@@ -3287,6 +3315,9 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr));
locate_dirty_segment(sbi, GET_SEGNO(sbi, *new_blkaddr));
+ if (IS_DATASEG(type))
+ atomic64_inc(&sbi->allocated_data_blocks);
+
up_write(&sit_i->sentry_lock);
if (page && IS_NODESEG(type)) {
@@ -3414,6 +3445,8 @@ void f2fs_outplace_write_data(struct dnode_of_data *dn,
struct f2fs_summary sum;
f2fs_bug_on(sbi, dn->data_blkaddr == NULL_ADDR);
+ if (fio->io_type == FS_DATA_IO || fio->io_type == FS_CP_DATA_IO)
+ f2fs_update_age_extent_cache(dn);
set_summary(&sum, dn->nid, dn->ofs_in_node, fio->version);
do_write_page(&sum, fio);
f2fs_update_data_blkaddr(dn, fio->new_blkaddr);
diff --git a/fs/f2fs/shrinker.c b/fs/f2fs/shrinker.c
index 33c490e69ae3..83d6fb97dcae 100644
--- a/fs/f2fs/shrinker.c
+++ b/fs/f2fs/shrinker.c
@@ -59,6 +59,9 @@ unsigned long f2fs_shrink_count(struct shrinker *shrink,
/* count read extent cache entries */
count += __count_extent_cache(sbi, EX_READ);
+ /* count block age extent cache entries */
+ count += __count_extent_cache(sbi, EX_BLOCK_AGE);
+
/* count clean nat cache entries */
count += __count_nat_entries(sbi);
@@ -102,8 +105,11 @@ unsigned long f2fs_shrink_scan(struct shrinker *shrink,
sbi->shrinker_run_no = run_no;
+ /* shrink extent cache entries */
+ freed += f2fs_shrink_age_extent_tree(sbi, nr >> 2);
+
/* shrink read extent cache entries */
- freed += f2fs_shrink_read_extent_tree(sbi, nr >> 1);
+ freed += f2fs_shrink_read_extent_tree(sbi, nr >> 2);
/* shrink clean nat cache entries */
if (freed < nr)
@@ -134,6 +140,8 @@ void f2fs_join_shrinker(struct f2fs_sb_info *sbi)
void f2fs_leave_shrinker(struct f2fs_sb_info *sbi)
{
f2fs_shrink_read_extent_tree(sbi, __count_extent_cache(sbi, EX_READ));
+ f2fs_shrink_age_extent_tree(sbi,
+ __count_extent_cache(sbi, EX_BLOCK_AGE));
spin_lock(&f2fs_list_lock);
list_del_init(&sbi->s_list);
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 10bd03bbefec..5bdab376b852 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -163,6 +163,7 @@ enum {
Opt_nogc_merge,
Opt_discard_unit,
Opt_memory_mode,
+ Opt_age_extent_cache,
Opt_err,
};
@@ -241,6 +242,7 @@ static match_table_t f2fs_tokens = {
{Opt_nogc_merge, "nogc_merge"},
{Opt_discard_unit, "discard_unit=%s"},
{Opt_memory_mode, "memory=%s"},
+ {Opt_age_extent_cache, "age_extent_cache"},
{Opt_err, NULL},
};
@@ -1257,6 +1259,9 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
}
kfree(name);
break;
+ case Opt_age_extent_cache:
+ set_opt(sbi, AGE_EXTENT_CACHE);
+ break;
default:
f2fs_err(sbi, "Unrecognized mount option \"%s\" or missing value",
p);
@@ -1958,6 +1963,8 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
seq_puts(seq, ",read_extent_cache");
else
seq_puts(seq, ",no_read_extent_cache");
+ if (test_opt(sbi, AGE_EXTENT_CACHE))
+ seq_puts(seq, ",age_extent_cache");
if (test_opt(sbi, DATA_FLUSH))
seq_puts(seq, ",data_flush");
@@ -2219,6 +2226,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
bool need_restart_flush = false, need_stop_flush = false;
bool need_restart_discard = false, need_stop_discard = false;
bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE);
+ bool no_age_extent_cache = !test_opt(sbi, AGE_EXTENT_CACHE);
bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT);
bool no_io_align = !F2FS_IO_ALIGNED(sbi);
bool no_atgc = !test_opt(sbi, ATGC);
@@ -2313,6 +2321,12 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
f2fs_warn(sbi, "switch extent_cache option is not allowed");
goto restore_opts;
}
+ /* disallow enable/disable age extent_cache dynamically */
+ if (no_age_extent_cache == !!test_opt(sbi, AGE_EXTENT_CACHE)) {
+ err = -EINVAL;
+ f2fs_warn(sbi, "switch age_extent_cache option is not allowed");
+ goto restore_opts;
+ }
if (no_io_align == !!F2FS_IO_ALIGNED(sbi)) {
err = -EINVAL;
diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
index a4745d596310..2ab215110596 100644
--- a/fs/f2fs/sysfs.c
+++ b/fs/f2fs/sysfs.c
@@ -668,6 +668,24 @@ static ssize_t __sbi_store(struct f2fs_attr *a,
return count;
}
+ if (!strcmp(a->attr.name, "hot_data_age_threshold")) {
+ if (t == 0 || t >= sbi->warm_data_age_threshold)
+ return -EINVAL;
+ if (t == *ui)
+ return count;
+ *ui = (unsigned int)t;
+ return count;
+ }
+
+ if (!strcmp(a->attr.name, "warm_data_age_threshold")) {
+ if (t == 0 || t <= sbi->hot_data_age_threshold)
+ return -EINVAL;
+ if (t == *ui)
+ return count;
+ *ui = (unsigned int)t;
+ return count;
+ }
+
*ui = (unsigned int)t;
return count;
@@ -923,6 +941,10 @@ F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, peak_atomic_write, peak_atomic_write);
F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, committed_atomic_block, committed_atomic_block);
F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, revoked_atomic_block, revoked_atomic_block);
+/* For block age extent cache */
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, hot_data_age_threshold, hot_data_age_threshold);
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, warm_data_age_threshold, warm_data_age_threshold);
+
#define ATTR_LIST(name) (&f2fs_attr_##name.attr)
static struct attribute *f2fs_attrs[] = {
ATTR_LIST(gc_urgent_sleep_time),
@@ -1018,6 +1040,8 @@ static struct attribute *f2fs_attrs[] = {
ATTR_LIST(peak_atomic_write),
ATTR_LIST(committed_atomic_block),
ATTR_LIST(revoked_atomic_block),
+ ATTR_LIST(hot_data_age_threshold),
+ ATTR_LIST(warm_data_age_threshold),
NULL,
};
ATTRIBUTE_GROUPS(f2fs);
diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h
index 2bb37892d2ba..31d994e6b4ca 100644
--- a/include/trace/events/f2fs.h
+++ b/include/trace/events/f2fs.h
@@ -49,6 +49,7 @@ TRACE_DEFINE_ENUM(CP_TRIMMED);
TRACE_DEFINE_ENUM(CP_PAUSE);
TRACE_DEFINE_ENUM(CP_RESIZE);
TRACE_DEFINE_ENUM(EX_READ);
+TRACE_DEFINE_ENUM(EX_BLOCK_AGE);
#define show_block_type(type) \
__print_symbolic(type, \
@@ -155,6 +156,11 @@ TRACE_DEFINE_ENUM(EX_READ);
{ COMPRESS_ZSTD, "ZSTD" }, \
{ COMPRESS_LZORLE, "LZO-RLE" })
+#define show_extent_type(type) \
+ __print_symbolic(type, \
+ { EX_READ, "Read" }, \
+ { EX_BLOCK_AGE, "Block Age" })
+
struct f2fs_sb_info;
struct f2fs_io_info;
struct extent_info;
@@ -1544,7 +1550,7 @@ TRACE_EVENT(f2fs_lookup_extent_tree_start,
TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, type = %s",
show_dev_ino(__entry),
__entry->pgofs,
- __entry->type == EX_READ ? "Read" : "N/A")
+ show_extent_type(__entry->type))
);
TRACE_EVENT_CONDITION(f2fs_lookup_read_extent_tree_end,
@@ -1583,6 +1589,45 @@ TRACE_EVENT_CONDITION(f2fs_lookup_read_extent_tree_end,
__entry->blk)
);
+TRACE_EVENT_CONDITION(f2fs_lookup_age_extent_tree_end,
+
+ TP_PROTO(struct inode *inode, unsigned int pgofs,
+ struct extent_info *ei),
+
+ TP_ARGS(inode, pgofs, ei),
+
+ TP_CONDITION(ei),
+
+ TP_STRUCT__entry(
+ __field(dev_t, dev)
+ __field(ino_t, ino)
+ __field(unsigned int, pgofs)
+ __field(unsigned int, fofs)
+ __field(unsigned int, len)
+ __field(unsigned long long, age)
+ __field(unsigned long long, blocks)
+ ),
+
+ TP_fast_assign(
+ __entry->dev = inode->i_sb->s_dev;
+ __entry->ino = inode->i_ino;
+ __entry->pgofs = pgofs;
+ __entry->fofs = ei->fofs;
+ __entry->len = ei->len;
+ __entry->age = ei->age;
+ __entry->blocks = ei->last_blocks;
+ ),
+
+ TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, "
+ "age_ext_info(fofs: %u, len: %u, age: %llu, blocks: %llu)",
+ show_dev_ino(__entry),
+ __entry->pgofs,
+ __entry->fofs,
+ __entry->len,
+ __entry->age,
+ __entry->blocks)
+);
+
TRACE_EVENT(f2fs_update_read_extent_tree_range,
TP_PROTO(struct inode *inode, unsigned int pgofs, unsigned int len,
@@ -1618,6 +1663,41 @@ TRACE_EVENT(f2fs_update_read_extent_tree_range,
__entry->c_len)
);
+TRACE_EVENT(f2fs_update_age_extent_tree_range,
+
+ TP_PROTO(struct inode *inode, unsigned int pgofs, unsigned int len,
+ unsigned long long age,
+ unsigned long long last_blks),
+
+ TP_ARGS(inode, pgofs, len, age, last_blks),
+
+ TP_STRUCT__entry(
+ __field(dev_t, dev)
+ __field(ino_t, ino)
+ __field(unsigned int, pgofs)
+ __field(unsigned int, len)
+ __field(unsigned long long, age)
+ __field(unsigned long long, blocks)
+ ),
+
+ TP_fast_assign(
+ __entry->dev = inode->i_sb->s_dev;
+ __entry->ino = inode->i_ino;
+ __entry->pgofs = pgofs;
+ __entry->len = len;
+ __entry->age = age;
+ __entry->blocks = last_blks;
+ ),
+
+ TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, "
+ "len = %u, age = %llu, blocks = %llu",
+ show_dev_ino(__entry),
+ __entry->pgofs,
+ __entry->len,
+ __entry->age,
+ __entry->blocks)
+);
+
TRACE_EVENT(f2fs_shrink_extent_tree,
TP_PROTO(struct f2fs_sb_info *sbi, unsigned int node_cnt,
@@ -1643,7 +1723,7 @@ TRACE_EVENT(f2fs_shrink_extent_tree,
show_dev(__entry->dev),
__entry->node_cnt,
__entry->tree_cnt,
- __entry->type == EX_READ ? "Read" : "N/A")
+ show_extent_type(__entry->type))
);
TRACE_EVENT(f2fs_destroy_extent_tree,
@@ -1670,7 +1750,7 @@ TRACE_EVENT(f2fs_destroy_extent_tree,
TP_printk("dev = (%d,%d), ino = %lu, destroyed: node_cnt = %u, type = %s",
show_dev_ino(__entry),
__entry->node_cnt,
- __entry->type == EX_READ ? "Read" : "N/A")
+ show_extent_type(__entry->type))
);
DECLARE_EVENT_CLASS(f2fs_sync_dirty_inodes,
--
2.39.0.rc0.267.gcb52ba06e7-goog
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply related [flat|nested] 19+ messages in thread* Re: [f2fs-dev] [PATCH 6/6] f2fs: add block_age-based extent cache
2022-12-05 18:54 ` [f2fs-dev] [PATCH 6/6] f2fs: add block_age-based extent cache Jaegeuk Kim
2022-12-06 19:06 ` [f2fs-dev] [PATCH 6/6 v2] " Jaegeuk Kim
@ 2022-12-16 12:13 ` qixiaoyu
2022-12-17 0:50 ` Jaegeuk Kim
1 sibling, 1 reply; 19+ messages in thread
From: qixiaoyu @ 2022-12-16 12:13 UTC (permalink / raw)
To: Jaegeuk Kim; +Cc: xiongping1, qixiaoyu1, linux-kernel, linux-f2fs-devel
On Mon, Dec 05, 2022 at 10:54:33AM -0800, Jaegeuk Kim wrote:
Hi Jaegeuk,
Thank you for your patchset!
>
> static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
> @@ -544,6 +585,8 @@ static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
>
> if (type == EX_READ)
> trace_f2fs_lookup_read_extent_tree_end(inode, pgofs, ei);
> + else if (type == EX_BLOCK_AGE)
> + trace_f2fs_lookup_age_extent_tree_end(inode, pgofs, ei);
> return ret;
> }
>
If (!en), ei->age may be uninitialized, which makes the trace output confused.
> +/* This returns a new age and allocated blocks in ei */
> +static int __get_new_block_age(struct inode *inode, struct extent_info *ei)
> +{
> + struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
> + loff_t f_size = i_size_read(inode);
> + unsigned long long cur_blocks =
> + atomic64_read(&sbi->allocated_data_blocks);
> +
> + /*
> + * When I/O is not aligned to a PAGE_SIZE, update will happen to the last
> + * file block even in seq write. So don't record age for newly last file
> + * block here.
> + */
> + if ((f_size >> PAGE_SHIFT) == ei->fofs && f_size & (PAGE_SIZE - 1) &&
> + ei->blk == NEW_ADDR)
> + return -EINVAL;
> +
> + if (__lookup_extent_tree(inode, ei->fofs, ei, EX_BLOCK_AGE)) {
> + unsigned long long cur_age;
> +
> + if (cur_blocks >= ei->last_blocks)
> + cur_age = cur_blocks - ei->last_blocks;
> + else
> + /* allocated_data_blocks overflow */
> + cur_age = ULLONG_MAX - ei->last_blocks + cur_blocks;
> +
> + if (ei->age)
> + ei->age = __calculate_block_age(cur_age, ei->age);
> + else
> + ei->age = cur_age;
> + ei->last_blocks = cur_blocks;
> + WARN_ON(ei->age > cur_blocks);
> + return 0;
> + }
> +
> + f2fs_bug_on(sbi, ei->blk == NULL_ADDR);
> +
> + /* the data block was allocated for the first time */
> + if (ei->blk == NEW_ADDR)
> + goto out;
> +
> + if (__is_valid_data_blkaddr(ei->blk) &&
> + !f2fs_is_valid_blkaddr(sbi, ei->blk, DATA_GENERIC_ENHANCE)) {
> + f2fs_bug_on(sbi, 1);
> + return -EINVAL;
> + }
> +out:
> + /*
> + * init block age with zero, this can happen when the block age extent
> + * was reclaimed due to memory constraint or system reboot
> + */
> + ei->age = 0;
> + ei->last_blocks = cur_blocks;
> + return 0;
> +}
> +
> static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type type)
> {
> struct extent_info ei;
> @@ -823,6 +951,10 @@ static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type typ
> ei.blk = NULL_ADDR;
> else
> ei.blk = dn->data_blkaddr;
> + } else if (type == EX_BLOCK_AGE) {
> + ei.blk = dn->data_blkaddr;
> + if (__get_new_block_age(dn->inode, &ei))
> + return;
> }
> __update_extent_tree_range(dn->inode, &ei, type);
> }
Here, the whole extent will be updated to the same new age in __get_new_block_age
and __update_extent_tree_range.
For example, if we create a new file with 10 blocks and then update fofs=3, we
will get:
fofs: 0, len: 10, age: 1000
But in our design, we expect to split the large age extent cache to record the
updated block:
fofs: 0, len: 3, age: 0
fofs: 3, len: 1, age: 1000
fofs: 4, len: 6, age: 0
Some blocks in the file can be updated more frequently than others and we want
distinguish them by the block age extent cache, currently the code breaks it.
Moreover, in the current code, if we update more than one block of the file at
once, the age of all extent will soon be close to 0, which is not expected.
f2fs_lookup_age_extent_tree_end: dev = (254,7), ino = 34547, pgofs = 2, age_ext_info(fofs: 0, len: 325, age: 172156, blocks: 861203)
f2fs_update_age_extent_tree_range: dev = (254,7), ino = 34547, pgofs = 0, len = 325, age = 51647, blocks = 861204
f2fs_lookup_age_extent_tree_end: dev = (254,7), ino = 34547, pgofs = 4, age_ext_info(fofs: 0, len: 325, age: 51647, blocks: 861204)
f2fs_update_age_extent_tree_range: dev = (254,7), ino = 34547, pgofs = 0, len = 325, age = 15494, blocks = 861205
f2fs_lookup_age_extent_tree_end: dev = (254,7), ino = 34547, pgofs = 6, age_ext_info(fofs: 0, len: 325, age: 15494, blocks: 861205)
f2fs_update_age_extent_tree_range: dev = (254,7), ino = 34547, pgofs = 0, len = 325, age = 4648, blocks = 861206
f2fs_lookup_age_extent_tree_end: dev = (254,7), ino = 34547, pgofs = 8, age_ext_info(fofs: 0, len: 325, age: 4648, blocks: 861206)
f2fs_update_age_extent_tree_range: dev = (254,7), ino = 34547, pgofs = 0, len = 325, age = 1395, blocks = 861207
> @@ -940,6 +1072,43 @@ unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi, int nr_shrin
> return __shrink_extent_tree(sbi, nr_shrink, EX_READ);
> }
>
> +/* block age extent cache operations */
> +bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
> + struct extent_info *ei)
> +{
> + if (!__may_extent_tree(inode, EX_BLOCK_AGE))
> + return false;
> +
> + return __lookup_extent_tree(inode, pgofs, ei, EX_BLOCK_AGE);
> +}
> +
> +void f2fs_update_age_extent_cache(struct dnode_of_data *dn)
> +{
> + return __update_extent_cache(dn, EX_BLOCK_AGE);
> +}
> +
> +void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
> + pgoff_t fofs, unsigned int len)
> +{
> + struct extent_info ei = {
> + .fofs = fofs,
> + .len = len,
> + };
> +
> + if (!__may_extent_tree(dn->inode, EX_BLOCK_AGE))
> + return;
> +
> + __update_extent_tree_range(dn->inode, &ei, EX_BLOCK_AGE);
> +}
> +
> +unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
> +{
> + if (!test_opt(sbi, AGE_EXTENT_CACHE))
> + return 0;
> +
> + return __shrink_extent_tree(sbi, nr_shrink, EX_BLOCK_AGE);
> +}
> +
> static unsigned int __destroy_extent_node(struct inode *inode,
> enum extent_type type)
> {
> @@ -960,6 +1129,7 @@ static unsigned int __destroy_extent_node(struct inode *inode,
> void f2fs_destroy_extent_node(struct inode *inode)
> {
> __destroy_extent_node(inode, EX_READ);
> + __destroy_extent_node(inode, EX_BLOCK_AGE);
> }
>
> static void __drop_extent_tree(struct inode *inode, enum extent_type type)
> @@ -988,6 +1158,7 @@ static void __drop_extent_tree(struct inode *inode, enum extent_type type)
> void f2fs_drop_extent_tree(struct inode *inode)
> {
> __drop_extent_tree(inode, EX_READ);
> + __drop_extent_tree(inode, EX_BLOCK_AGE);
> }
>
> static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
> @@ -1028,6 +1199,7 @@ static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
> void f2fs_destroy_extent_tree(struct inode *inode)
> {
> __destroy_extent_tree(inode, EX_READ);
> + __destroy_extent_tree(inode, EX_BLOCK_AGE);
> }
>
> static void __init_extent_tree_info(struct extent_tree_info *eti)
> @@ -1045,6 +1217,12 @@ static void __init_extent_tree_info(struct extent_tree_info *eti)
> void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi)
> {
> __init_extent_tree_info(&sbi->extent_tree[EX_READ]);
> + __init_extent_tree_info(&sbi->extent_tree[EX_BLOCK_AGE]);
> +
> + /* initialize for block age extents */
> + atomic64_set(&sbi->allocated_data_blocks, 0);
> + sbi->hot_data_age_threshold = DEF_HOT_DATA_AGE_THRESHOLD;
> + sbi->warm_data_age_threshold = DEF_WARM_DATA_AGE_THRESHOLD;
> }
>
> int __init f2fs_create_extent_cache(void)
> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> index ff940cba4600..eb71edcf70de 100644
> --- a/fs/f2fs/f2fs.h
> +++ b/fs/f2fs/f2fs.h
> @@ -107,6 +107,7 @@ extern const char *f2fs_fault_name[FAULT_MAX];
> #define F2FS_MOUNT_MERGE_CHECKPOINT 0x10000000
> #define F2FS_MOUNT_GC_MERGE 0x20000000
> #define F2FS_MOUNT_COMPRESS_CACHE 0x40000000
> +#define F2FS_MOUNT_AGE_EXTENT_CACHE 0x80000000
>
> #define F2FS_OPTION(sbi) ((sbi)->mount_opt)
> #define clear_opt(sbi, option) (F2FS_OPTION(sbi).opt &= ~F2FS_MOUNT_##option)
> @@ -607,9 +608,22 @@ enum {
> /* number of extent info in extent cache we try to shrink */
> #define READ_EXTENT_CACHE_SHRINK_NUMBER 128
>
> +/* number of age extent info in extent cache we try to shrink */
> +#define AGE_EXTENT_CACHE_SHRINK_NUMBER 128
> +#define LAST_AGE_WEIGHT 30
> +#define SAME_AGE_REGION 1024
> +
> +/*
> + * Define data block with age less than 1GB as hot data
> + * define data block with age less than 10GB but more than 1GB as warm data
> + */
> +#define DEF_HOT_DATA_AGE_THRESHOLD 262144
> +#define DEF_WARM_DATA_AGE_THRESHOLD 2621440
> +
> /* extent cache type */
> enum extent_type {
> EX_READ,
> + EX_BLOCK_AGE,
> NR_EXTENT_CACHES,
> };
>
> @@ -637,6 +651,13 @@ struct extent_info {
> unsigned int c_len;
> #endif
> };
> + /* block age extent_cache */
> + struct {
> + /* block age of the extent */
> + unsigned long long age;
> + /* last total blocks allocated */
> + unsigned long long last_blocks;
> + };
> };
> };
>
> @@ -1653,6 +1674,11 @@ struct f2fs_sb_info {
>
> /* for extent tree cache */
> struct extent_tree_info extent_tree[NR_EXTENT_CACHES];
> + atomic64_t allocated_data_blocks; /* for block age extent_cache */
> +
> + /* The threshold used for hot and warm data seperation*/
> + unsigned int hot_data_age_threshold;
> + unsigned int warm_data_age_threshold;
>
> /* basic filesystem units */
> unsigned int log_sectors_per_block; /* log2 sectors per block */
> @@ -3857,6 +3883,8 @@ struct f2fs_stat_info {
> unsigned long long ext_mem[NR_EXTENT_CACHES];
> /* for read extent cache */
> unsigned long long hit_largest;
> + /* for block age extent cache */
> + 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;
> @@ -4168,6 +4196,16 @@ void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
> unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi,
> int nr_shrink);
>
> +/* block age extent cache ops */
> +void f2fs_init_age_extent_tree(struct inode *inode);
> +bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
> + struct extent_info *ei);
> +void f2fs_update_age_extent_cache(struct dnode_of_data *dn);
> +void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
> + pgoff_t fofs, unsigned int len);
> +unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi,
> + int nr_shrink);
> +
> /*
> * sysfs.c
> */
> diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
> index cbe7c24065c7..56c23b5e9d65 100644
> --- a/fs/f2fs/file.c
> +++ b/fs/f2fs/file.c
> @@ -619,6 +619,7 @@ void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count)
> fofs = f2fs_start_bidx_of_node(ofs_of_node(dn->node_page),
> dn->inode) + ofs;
> f2fs_update_read_extent_cache_range(dn, fofs, 0, len);
> + f2fs_update_age_extent_cache_range(dn, fofs, nr_free);
> dec_valid_block_count(sbi, dn->inode, nr_free);
> }
> dn->ofs_in_node = ofs;
> diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> index c845c16f97d0..ff6cf66ed46b 100644
> --- a/fs/f2fs/inode.c
> +++ b/fs/f2fs/inode.c
> @@ -480,6 +480,7 @@ static int do_read_inode(struct inode *inode)
>
> /* Need all the flag bits */
> f2fs_init_read_extent_tree(inode, node_page);
> + f2fs_init_age_extent_tree(inode);
>
> f2fs_put_page(node_page, 1);
>
> diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
> index 07419c3e42a5..dde4c0458704 100644
> --- a/fs/f2fs/node.c
> +++ b/fs/f2fs/node.c
> @@ -60,7 +60,7 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
> avail_ram = val.totalram - val.totalhigh;
>
> /*
> - * give 25%, 25%, 50%, 50%, 50% memory for each components respectively
> + * give 25%, 25%, 50%, 50%, 25%, 25% memory for each components respectively
> */
> if (type == FREE_NIDS) {
> mem_size = (nm_i->nid_cnt[FREE_NID] *
> @@ -85,14 +85,16 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
> sizeof(struct ino_entry);
> mem_size >>= PAGE_SHIFT;
> res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
> - } else if (type == READ_EXTENT_CACHE) {
> - struct extent_tree_info *eti = &sbi->extent_tree[EX_READ];
> + } else if (type == READ_EXTENT_CACHE || type == AGE_EXTENT_CACHE) {
> + enum extent_type etype = type == READ_EXTENT_CACHE ?
> + EX_READ : EX_BLOCK_AGE;
> + struct extent_tree_info *eti = &sbi->extent_tree[etype];
>
> mem_size = (atomic_read(&eti->total_ext_tree) *
> sizeof(struct extent_tree) +
> atomic_read(&eti->total_ext_node) *
> sizeof(struct extent_node)) >> PAGE_SHIFT;
> - res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
> + res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 2);
> } else if (type == DISCARD_CACHE) {
> mem_size = (atomic_read(&dcc->discard_cmd_cnt) *
> sizeof(struct discard_cmd)) >> PAGE_SHIFT;
> diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
> index 0aa48704c77a..99454d46a939 100644
> --- a/fs/f2fs/node.h
> +++ b/fs/f2fs/node.h
> @@ -147,6 +147,7 @@ enum mem_type {
> DIRTY_DENTS, /* indicates dirty dentry pages */
> INO_ENTRIES, /* indicates inode entries */
> READ_EXTENT_CACHE, /* indicates read extent cache */
> + AGE_EXTENT_CACHE, /* indicates age extent cache */
> DISCARD_CACHE, /* indicates memory of cached discard cmds */
> COMPRESS_PAGE, /* indicates memory of cached compressed pages */
> BASE_CHECK, /* check kernel status */
> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
> index 8722d1a13c17..dee712f7225f 100644
> --- a/fs/f2fs/segment.c
> +++ b/fs/f2fs/segment.c
> @@ -453,6 +453,11 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
> f2fs_shrink_read_extent_tree(sbi,
> READ_EXTENT_CACHE_SHRINK_NUMBER);
>
> + /* try to shrink age extent cache when there is no enough memory */
> + if (!f2fs_available_free_memory(sbi, AGE_EXTENT_CACHE))
> + f2fs_shrink_age_extent_tree(sbi,
> + AGE_EXTENT_CACHE_SHRINK_NUMBER);
> +
> /* check the # of cached NAT entries */
> if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
> f2fs_try_to_free_nats(sbi, NAT_ENTRY_PER_BLOCK);
> @@ -3151,10 +3156,28 @@ static int __get_segment_type_4(struct f2fs_io_info *fio)
> }
> }
>
> +static int __get_age_segment_type(struct inode *inode, pgoff_t pgofs)
> +{
> + struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
> + struct extent_info ei;
> +
> + if (f2fs_lookup_age_extent_cache(inode, pgofs, &ei)) {
> + if (!ei.age)
> + return NO_CHECK_TYPE;
> + if (ei.age <= sbi->hot_data_age_threshold)
> + return CURSEG_HOT_DATA;
> + if (ei.age <= sbi->warm_data_age_threshold)
> + return CURSEG_WARM_DATA;
> + return CURSEG_COLD_DATA;
> + }
> + return NO_CHECK_TYPE;
> +}
> +
> static int __get_segment_type_6(struct f2fs_io_info *fio)
> {
> if (fio->type == DATA) {
> struct inode *inode = fio->page->mapping->host;
> + int type;
>
> if (is_inode_flag_set(inode, FI_ALIGNED_WRITE))
> return CURSEG_COLD_DATA_PINNED;
> @@ -3169,6 +3192,11 @@ static int __get_segment_type_6(struct f2fs_io_info *fio)
> }
> if (file_is_cold(inode) || f2fs_need_compress_data(inode))
> return CURSEG_COLD_DATA;
> +
> + type = __get_age_segment_type(inode, fio->page->index);
> + if (type != NO_CHECK_TYPE)
> + return type;
> +
> if (file_is_hot(inode) ||
> is_inode_flag_set(inode, FI_HOT_DATA) ||
> f2fs_is_cow_file(inode))
> @@ -3287,6 +3315,9 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
> locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr));
> locate_dirty_segment(sbi, GET_SEGNO(sbi, *new_blkaddr));
>
> + if (IS_DATASEG(type))
> + atomic64_inc(&sbi->allocated_data_blocks);
> +
> up_write(&sit_i->sentry_lock);
>
> if (page && IS_NODESEG(type)) {
> @@ -3414,6 +3445,8 @@ void f2fs_outplace_write_data(struct dnode_of_data *dn,
> struct f2fs_summary sum;
>
> f2fs_bug_on(sbi, dn->data_blkaddr == NULL_ADDR);
> + if (fio->io_type == FS_DATA_IO || fio->io_type == FS_CP_DATA_IO)
> + f2fs_update_age_extent_cache(dn);
> set_summary(&sum, dn->nid, dn->ofs_in_node, fio->version);
> do_write_page(&sum, fio);
> f2fs_update_data_blkaddr(dn, fio->new_blkaddr);
> diff --git a/fs/f2fs/shrinker.c b/fs/f2fs/shrinker.c
> index 33c490e69ae3..83d6fb97dcae 100644
> --- a/fs/f2fs/shrinker.c
> +++ b/fs/f2fs/shrinker.c
> @@ -59,6 +59,9 @@ unsigned long f2fs_shrink_count(struct shrinker *shrink,
> /* count read extent cache entries */
> count += __count_extent_cache(sbi, EX_READ);
>
> + /* count block age extent cache entries */
> + count += __count_extent_cache(sbi, EX_BLOCK_AGE);
> +
> /* count clean nat cache entries */
> count += __count_nat_entries(sbi);
>
> @@ -102,8 +105,11 @@ unsigned long f2fs_shrink_scan(struct shrinker *shrink,
>
> sbi->shrinker_run_no = run_no;
>
> + /* shrink extent cache entries */
> + freed += f2fs_shrink_age_extent_tree(sbi, nr >> 2);
> +
> /* shrink read extent cache entries */
> - freed += f2fs_shrink_read_extent_tree(sbi, nr >> 1);
> + freed += f2fs_shrink_read_extent_tree(sbi, nr >> 2);
>
> /* shrink clean nat cache entries */
> if (freed < nr)
> @@ -134,6 +140,8 @@ void f2fs_join_shrinker(struct f2fs_sb_info *sbi)
> void f2fs_leave_shrinker(struct f2fs_sb_info *sbi)
> {
> f2fs_shrink_read_extent_tree(sbi, __count_extent_cache(sbi, EX_READ));
> + f2fs_shrink_age_extent_tree(sbi,
> + __count_extent_cache(sbi, EX_BLOCK_AGE));
>
> spin_lock(&f2fs_list_lock);
> list_del_init(&sbi->s_list);
> diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> index 10bd03bbefec..5bdab376b852 100644
> --- a/fs/f2fs/super.c
> +++ b/fs/f2fs/super.c
> @@ -163,6 +163,7 @@ enum {
> Opt_nogc_merge,
> Opt_discard_unit,
> Opt_memory_mode,
> + Opt_age_extent_cache,
> Opt_err,
> };
>
> @@ -241,6 +242,7 @@ static match_table_t f2fs_tokens = {
> {Opt_nogc_merge, "nogc_merge"},
> {Opt_discard_unit, "discard_unit=%s"},
> {Opt_memory_mode, "memory=%s"},
> + {Opt_age_extent_cache, "age_extent_cache"},
> {Opt_err, NULL},
> };
>
> @@ -1257,6 +1259,9 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
> }
> kfree(name);
> break;
> + case Opt_age_extent_cache:
> + set_opt(sbi, AGE_EXTENT_CACHE);
> + break;
> default:
> f2fs_err(sbi, "Unrecognized mount option \"%s\" or missing value",
> p);
> @@ -1958,6 +1963,8 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
> seq_puts(seq, ",read_extent_cache");
> else
> seq_puts(seq, ",no_read_extent_cache");
> + if (test_opt(sbi, AGE_EXTENT_CACHE))
> + seq_puts(seq, ",age_extent_cache");
> if (test_opt(sbi, DATA_FLUSH))
> seq_puts(seq, ",data_flush");
>
> @@ -2219,6 +2226,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
> bool need_restart_flush = false, need_stop_flush = false;
> bool need_restart_discard = false, need_stop_discard = false;
> bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE);
> + bool no_age_extent_cache = !test_opt(sbi, AGE_EXTENT_CACHE);
> bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT);
> bool no_io_align = !F2FS_IO_ALIGNED(sbi);
> bool no_atgc = !test_opt(sbi, ATGC);
> @@ -2313,6 +2321,12 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
> f2fs_warn(sbi, "switch extent_cache option is not allowed");
> goto restore_opts;
> }
> + /* disallow enable/disable age extent_cache dynamically */
> + if (no_age_extent_cache == !!test_opt(sbi, AGE_EXTENT_CACHE)) {
> + err = -EINVAL;
> + f2fs_warn(sbi, "switch age_extent_cache option is not allowed");
> + goto restore_opts;
> + }
>
> if (no_io_align == !!F2FS_IO_ALIGNED(sbi)) {
> err = -EINVAL;
> diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
> index a4745d596310..2ab215110596 100644
> --- a/fs/f2fs/sysfs.c
> +++ b/fs/f2fs/sysfs.c
> @@ -668,6 +668,24 @@ static ssize_t __sbi_store(struct f2fs_attr *a,
> return count;
> }
>
> + if (!strcmp(a->attr.name, "hot_data_age_threshold")) {
> + if (t == 0 || t >= sbi->warm_data_age_threshold)
> + return -EINVAL;
> + if (t == *ui)
> + return count;
> + *ui = (unsigned int)t;
> + return count;
> + }
> +
> + if (!strcmp(a->attr.name, "warm_data_age_threshold")) {
> + if (t == 0 || t <= sbi->hot_data_age_threshold)
> + return -EINVAL;
> + if (t == *ui)
> + return count;
> + *ui = (unsigned int)t;
> + return count;
> + }
> +
> *ui = (unsigned int)t;
>
> return count;
> @@ -923,6 +941,10 @@ F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, peak_atomic_write, peak_atomic_write);
> F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, committed_atomic_block, committed_atomic_block);
> F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, revoked_atomic_block, revoked_atomic_block);
>
> +/* For block age extent cache */
> +F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, hot_data_age_threshold, hot_data_age_threshold);
> +F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, warm_data_age_threshold, warm_data_age_threshold);
> +
> #define ATTR_LIST(name) (&f2fs_attr_##name.attr)
> static struct attribute *f2fs_attrs[] = {
> ATTR_LIST(gc_urgent_sleep_time),
> @@ -1018,6 +1040,8 @@ static struct attribute *f2fs_attrs[] = {
> ATTR_LIST(peak_atomic_write),
> ATTR_LIST(committed_atomic_block),
> ATTR_LIST(revoked_atomic_block),
> + ATTR_LIST(hot_data_age_threshold),
> + ATTR_LIST(warm_data_age_threshold),
> NULL,
> };
> ATTRIBUTE_GROUPS(f2fs);
> diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h
> index 2bb37892d2ba..31d994e6b4ca 100644
> --- a/include/trace/events/f2fs.h
> +++ b/include/trace/events/f2fs.h
> @@ -49,6 +49,7 @@ TRACE_DEFINE_ENUM(CP_TRIMMED);
> TRACE_DEFINE_ENUM(CP_PAUSE);
> TRACE_DEFINE_ENUM(CP_RESIZE);
> TRACE_DEFINE_ENUM(EX_READ);
> +TRACE_DEFINE_ENUM(EX_BLOCK_AGE);
>
> #define show_block_type(type) \
> __print_symbolic(type, \
> @@ -155,6 +156,11 @@ TRACE_DEFINE_ENUM(EX_READ);
> { COMPRESS_ZSTD, "ZSTD" }, \
> { COMPRESS_LZORLE, "LZO-RLE" })
>
> +#define show_extent_type(type) \
> + __print_symbolic(type, \
> + { EX_READ, "Read" }, \
> + { EX_BLOCK_AGE, "Block Age" })
> +
> struct f2fs_sb_info;
> struct f2fs_io_info;
> struct extent_info;
> @@ -1544,7 +1550,7 @@ TRACE_EVENT(f2fs_lookup_extent_tree_start,
> TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, type = %s",
> show_dev_ino(__entry),
> __entry->pgofs,
> - __entry->type == EX_READ ? "Read" : "N/A")
> + show_extent_type(__entry->type))
> );
>
> TRACE_EVENT_CONDITION(f2fs_lookup_read_extent_tree_end,
> @@ -1583,6 +1589,45 @@ TRACE_EVENT_CONDITION(f2fs_lookup_read_extent_tree_end,
> __entry->blk)
> );
>
> +TRACE_EVENT_CONDITION(f2fs_lookup_age_extent_tree_end,
> +
> + TP_PROTO(struct inode *inode, unsigned int pgofs,
> + struct extent_info *ei),
> +
> + TP_ARGS(inode, pgofs, ei),
> +
> + TP_CONDITION(ei),
> +
> + TP_STRUCT__entry(
> + __field(dev_t, dev)
> + __field(ino_t, ino)
> + __field(unsigned int, pgofs)
> + __field(unsigned int, fofs)
> + __field(unsigned int, len)
> + __field(unsigned long long, age)
> + __field(unsigned long long, blocks)
> + ),
> +
> + TP_fast_assign(
> + __entry->dev = inode->i_sb->s_dev;
> + __entry->ino = inode->i_ino;
> + __entry->pgofs = pgofs;
> + __entry->fofs = ei->fofs;
> + __entry->len = ei->len;
> + __entry->age = ei->age;
> + __entry->blocks = ei->last_blocks;
> + ),
> +
> + TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, "
> + "age_ext_info(fofs: %u, len: %u, age: %llu, blocks: %llu)",
> + show_dev_ino(__entry),
> + __entry->pgofs,
> + __entry->fofs,
> + __entry->len,
> + __entry->age,
> + __entry->blocks)
> +);
> +
> TRACE_EVENT(f2fs_update_read_extent_tree_range,
>
> TP_PROTO(struct inode *inode, unsigned int pgofs, unsigned int len,
> @@ -1618,6 +1663,41 @@ TRACE_EVENT(f2fs_update_read_extent_tree_range,
> __entry->c_len)
> );
>
> +TRACE_EVENT(f2fs_update_age_extent_tree_range,
> +
> + TP_PROTO(struct inode *inode, unsigned int pgofs, unsigned int len,
> + unsigned long long age,
> + unsigned long long last_blks),
> +
> + TP_ARGS(inode, pgofs, len, age, last_blks),
> +
> + TP_STRUCT__entry(
> + __field(dev_t, dev)
> + __field(ino_t, ino)
> + __field(unsigned int, pgofs)
> + __field(unsigned int, len)
> + __field(unsigned long long, age)
> + __field(unsigned long long, blocks)
> + ),
> +
> + TP_fast_assign(
> + __entry->dev = inode->i_sb->s_dev;
> + __entry->ino = inode->i_ino;
> + __entry->pgofs = pgofs;
> + __entry->len = len;
> + __entry->age = age;
> + __entry->blocks = last_blks;
> + ),
> +
> + TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, "
> + "len = %u, age = %llu, blocks = %llu",
> + show_dev_ino(__entry),
> + __entry->pgofs,
> + __entry->len,
> + __entry->age,
> + __entry->blocks)
> +);
> +
> TRACE_EVENT(f2fs_shrink_extent_tree,
>
> TP_PROTO(struct f2fs_sb_info *sbi, unsigned int node_cnt,
> @@ -1643,7 +1723,7 @@ TRACE_EVENT(f2fs_shrink_extent_tree,
> show_dev(__entry->dev),
> __entry->node_cnt,
> __entry->tree_cnt,
> - __entry->type == EX_READ ? "Read" : "N/A")
> + show_extent_type(__entry->type))
> );
>
> TRACE_EVENT(f2fs_destroy_extent_tree,
> @@ -1670,7 +1750,7 @@ TRACE_EVENT(f2fs_destroy_extent_tree,
> TP_printk("dev = (%d,%d), ino = %lu, destroyed: node_cnt = %u, type = %s",
> show_dev_ino(__entry),
> __entry->node_cnt,
> - __entry->type == EX_READ ? "Read" : "N/A")
> + show_extent_type(__entry->type))
> );
>
> DECLARE_EVENT_CLASS(f2fs_sync_dirty_inodes,
> --
> 2.39.0.rc0.267.gcb52ba06e7-goog
>
>
>
> _______________________________________________
> Linux-f2fs-devel mailing list
> Linux-f2fs-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [f2fs-dev] [PATCH 6/6] f2fs: add block_age-based extent cache
2022-12-16 12:13 ` [f2fs-dev] [PATCH 6/6] " qixiaoyu
@ 2022-12-17 0:50 ` Jaegeuk Kim
0 siblings, 0 replies; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-17 0:50 UTC (permalink / raw)
To: qixiaoyu; +Cc: xiongping1, qixiaoyu1, linux-kernel, linux-f2fs-devel
On 12/16, qixiaoyu wrote:
> On Mon, Dec 05, 2022 at 10:54:33AM -0800, Jaegeuk Kim wrote:
>
> Hi Jaegeuk,
>
> Thank you for your patchset!
>
> >
> > static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
> > @@ -544,6 +585,8 @@ static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
> >
> > if (type == EX_READ)
> > trace_f2fs_lookup_read_extent_tree_end(inode, pgofs, ei);
> > + else if (type == EX_BLOCK_AGE)
> > + trace_f2fs_lookup_age_extent_tree_end(inode, pgofs, ei);
> > return ret;
> > }
> >
>
> If (!en), ei->age may be uninitialized, which makes the trace output confused.
>
> > +/* This returns a new age and allocated blocks in ei */
> > +static int __get_new_block_age(struct inode *inode, struct extent_info *ei)
> > +{
> > + struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
> > + loff_t f_size = i_size_read(inode);
> > + unsigned long long cur_blocks =
> > + atomic64_read(&sbi->allocated_data_blocks);
> > +
> > + /*
> > + * When I/O is not aligned to a PAGE_SIZE, update will happen to the last
> > + * file block even in seq write. So don't record age for newly last file
> > + * block here.
> > + */
> > + if ((f_size >> PAGE_SHIFT) == ei->fofs && f_size & (PAGE_SIZE - 1) &&
> > + ei->blk == NEW_ADDR)
> > + return -EINVAL;
> > +
> > + if (__lookup_extent_tree(inode, ei->fofs, ei, EX_BLOCK_AGE)) {
> > + unsigned long long cur_age;
> > +
> > + if (cur_blocks >= ei->last_blocks)
> > + cur_age = cur_blocks - ei->last_blocks;
> > + else
> > + /* allocated_data_blocks overflow */
> > + cur_age = ULLONG_MAX - ei->last_blocks + cur_blocks;
> > +
> > + if (ei->age)
> > + ei->age = __calculate_block_age(cur_age, ei->age);
> > + else
> > + ei->age = cur_age;
> > + ei->last_blocks = cur_blocks;
> > + WARN_ON(ei->age > cur_blocks);
> > + return 0;
> > + }
> > +
> > + f2fs_bug_on(sbi, ei->blk == NULL_ADDR);
> > +
> > + /* the data block was allocated for the first time */
> > + if (ei->blk == NEW_ADDR)
> > + goto out;
> > +
> > + if (__is_valid_data_blkaddr(ei->blk) &&
> > + !f2fs_is_valid_blkaddr(sbi, ei->blk, DATA_GENERIC_ENHANCE)) {
> > + f2fs_bug_on(sbi, 1);
> > + return -EINVAL;
> > + }
> > +out:
> > + /*
> > + * init block age with zero, this can happen when the block age extent
> > + * was reclaimed due to memory constraint or system reboot
> > + */
> > + ei->age = 0;
> > + ei->last_blocks = cur_blocks;
> > + return 0;
> > +}
> > +
> > static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type type)
> > {
> > struct extent_info ei;
> > @@ -823,6 +951,10 @@ static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type typ
> > ei.blk = NULL_ADDR;
> > else
> > ei.blk = dn->data_blkaddr;
> > + } else if (type == EX_BLOCK_AGE) {
> > + ei.blk = dn->data_blkaddr;
> > + if (__get_new_block_age(dn->inode, &ei))
> > + return;
> > }
> > __update_extent_tree_range(dn->inode, &ei, type);
> > }
>
> Here, the whole extent will be updated to the same new age in __get_new_block_age
> and __update_extent_tree_range.
>
> For example, if we create a new file with 10 blocks and then update fofs=3, we
> will get:
> fofs: 0, len: 10, age: 1000
> But in our design, we expect to split the large age extent cache to record the
> updated block:
> fofs: 0, len: 3, age: 0
> fofs: 3, len: 1, age: 1000
> fofs: 4, len: 6, age: 0
>
> Some blocks in the file can be updated more frequently than others and we want
> distinguish them by the block age extent cache, currently the code breaks it.
Could you please check this patch series?
https://lore.kernel.org/linux-f2fs-devel/20221217004659.2092427-1-jaegeuk@kernel.org/T/#t
>
> Moreover, in the current code, if we update more than one block of the file at
> once, the age of all extent will soon be close to 0, which is not expected.
>
> f2fs_lookup_age_extent_tree_end: dev = (254,7), ino = 34547, pgofs = 2, age_ext_info(fofs: 0, len: 325, age: 172156, blocks: 861203)
> f2fs_update_age_extent_tree_range: dev = (254,7), ino = 34547, pgofs = 0, len = 325, age = 51647, blocks = 861204
> f2fs_lookup_age_extent_tree_end: dev = (254,7), ino = 34547, pgofs = 4, age_ext_info(fofs: 0, len: 325, age: 51647, blocks: 861204)
> f2fs_update_age_extent_tree_range: dev = (254,7), ino = 34547, pgofs = 0, len = 325, age = 15494, blocks = 861205
> f2fs_lookup_age_extent_tree_end: dev = (254,7), ino = 34547, pgofs = 6, age_ext_info(fofs: 0, len: 325, age: 15494, blocks: 861205)
> f2fs_update_age_extent_tree_range: dev = (254,7), ino = 34547, pgofs = 0, len = 325, age = 4648, blocks = 861206
> f2fs_lookup_age_extent_tree_end: dev = (254,7), ino = 34547, pgofs = 8, age_ext_info(fofs: 0, len: 325, age: 4648, blocks: 861206)
> f2fs_update_age_extent_tree_range: dev = (254,7), ino = 34547, pgofs = 0, len = 325, age = 1395, blocks = 861207
>
> > @@ -940,6 +1072,43 @@ unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi, int nr_shrin
> > return __shrink_extent_tree(sbi, nr_shrink, EX_READ);
> > }
> >
> > +/* block age extent cache operations */
> > +bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
> > + struct extent_info *ei)
> > +{
> > + if (!__may_extent_tree(inode, EX_BLOCK_AGE))
> > + return false;
> > +
> > + return __lookup_extent_tree(inode, pgofs, ei, EX_BLOCK_AGE);
> > +}
> > +
> > +void f2fs_update_age_extent_cache(struct dnode_of_data *dn)
> > +{
> > + return __update_extent_cache(dn, EX_BLOCK_AGE);
> > +}
> > +
> > +void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
> > + pgoff_t fofs, unsigned int len)
> > +{
> > + struct extent_info ei = {
> > + .fofs = fofs,
> > + .len = len,
> > + };
> > +
> > + if (!__may_extent_tree(dn->inode, EX_BLOCK_AGE))
> > + return;
> > +
> > + __update_extent_tree_range(dn->inode, &ei, EX_BLOCK_AGE);
> > +}
> > +
> > +unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
> > +{
> > + if (!test_opt(sbi, AGE_EXTENT_CACHE))
> > + return 0;
> > +
> > + return __shrink_extent_tree(sbi, nr_shrink, EX_BLOCK_AGE);
> > +}
> > +
> > static unsigned int __destroy_extent_node(struct inode *inode,
> > enum extent_type type)
> > {
> > @@ -960,6 +1129,7 @@ static unsigned int __destroy_extent_node(struct inode *inode,
> > void f2fs_destroy_extent_node(struct inode *inode)
> > {
> > __destroy_extent_node(inode, EX_READ);
> > + __destroy_extent_node(inode, EX_BLOCK_AGE);
> > }
> >
> > static void __drop_extent_tree(struct inode *inode, enum extent_type type)
> > @@ -988,6 +1158,7 @@ static void __drop_extent_tree(struct inode *inode, enum extent_type type)
> > void f2fs_drop_extent_tree(struct inode *inode)
> > {
> > __drop_extent_tree(inode, EX_READ);
> > + __drop_extent_tree(inode, EX_BLOCK_AGE);
> > }
> >
> > static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
> > @@ -1028,6 +1199,7 @@ static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
> > void f2fs_destroy_extent_tree(struct inode *inode)
> > {
> > __destroy_extent_tree(inode, EX_READ);
> > + __destroy_extent_tree(inode, EX_BLOCK_AGE);
> > }
> >
> > static void __init_extent_tree_info(struct extent_tree_info *eti)
> > @@ -1045,6 +1217,12 @@ static void __init_extent_tree_info(struct extent_tree_info *eti)
> > void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi)
> > {
> > __init_extent_tree_info(&sbi->extent_tree[EX_READ]);
> > + __init_extent_tree_info(&sbi->extent_tree[EX_BLOCK_AGE]);
> > +
> > + /* initialize for block age extents */
> > + atomic64_set(&sbi->allocated_data_blocks, 0);
> > + sbi->hot_data_age_threshold = DEF_HOT_DATA_AGE_THRESHOLD;
> > + sbi->warm_data_age_threshold = DEF_WARM_DATA_AGE_THRESHOLD;
> > }
> >
> > int __init f2fs_create_extent_cache(void)
> > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> > index ff940cba4600..eb71edcf70de 100644
> > --- a/fs/f2fs/f2fs.h
> > +++ b/fs/f2fs/f2fs.h
> > @@ -107,6 +107,7 @@ extern const char *f2fs_fault_name[FAULT_MAX];
> > #define F2FS_MOUNT_MERGE_CHECKPOINT 0x10000000
> > #define F2FS_MOUNT_GC_MERGE 0x20000000
> > #define F2FS_MOUNT_COMPRESS_CACHE 0x40000000
> > +#define F2FS_MOUNT_AGE_EXTENT_CACHE 0x80000000
> >
> > #define F2FS_OPTION(sbi) ((sbi)->mount_opt)
> > #define clear_opt(sbi, option) (F2FS_OPTION(sbi).opt &= ~F2FS_MOUNT_##option)
> > @@ -607,9 +608,22 @@ enum {
> > /* number of extent info in extent cache we try to shrink */
> > #define READ_EXTENT_CACHE_SHRINK_NUMBER 128
> >
> > +/* number of age extent info in extent cache we try to shrink */
> > +#define AGE_EXTENT_CACHE_SHRINK_NUMBER 128
> > +#define LAST_AGE_WEIGHT 30
> > +#define SAME_AGE_REGION 1024
> > +
> > +/*
> > + * Define data block with age less than 1GB as hot data
> > + * define data block with age less than 10GB but more than 1GB as warm data
> > + */
> > +#define DEF_HOT_DATA_AGE_THRESHOLD 262144
> > +#define DEF_WARM_DATA_AGE_THRESHOLD 2621440
> > +
> > /* extent cache type */
> > enum extent_type {
> > EX_READ,
> > + EX_BLOCK_AGE,
> > NR_EXTENT_CACHES,
> > };
> >
> > @@ -637,6 +651,13 @@ struct extent_info {
> > unsigned int c_len;
> > #endif
> > };
> > + /* block age extent_cache */
> > + struct {
> > + /* block age of the extent */
> > + unsigned long long age;
> > + /* last total blocks allocated */
> > + unsigned long long last_blocks;
> > + };
> > };
> > };
> >
> > @@ -1653,6 +1674,11 @@ struct f2fs_sb_info {
> >
> > /* for extent tree cache */
> > struct extent_tree_info extent_tree[NR_EXTENT_CACHES];
> > + atomic64_t allocated_data_blocks; /* for block age extent_cache */
> > +
> > + /* The threshold used for hot and warm data seperation*/
> > + unsigned int hot_data_age_threshold;
> > + unsigned int warm_data_age_threshold;
> >
> > /* basic filesystem units */
> > unsigned int log_sectors_per_block; /* log2 sectors per block */
> > @@ -3857,6 +3883,8 @@ struct f2fs_stat_info {
> > unsigned long long ext_mem[NR_EXTENT_CACHES];
> > /* for read extent cache */
> > unsigned long long hit_largest;
> > + /* for block age extent cache */
> > + 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;
> > @@ -4168,6 +4196,16 @@ void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
> > unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi,
> > int nr_shrink);
> >
> > +/* block age extent cache ops */
> > +void f2fs_init_age_extent_tree(struct inode *inode);
> > +bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
> > + struct extent_info *ei);
> > +void f2fs_update_age_extent_cache(struct dnode_of_data *dn);
> > +void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
> > + pgoff_t fofs, unsigned int len);
> > +unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi,
> > + int nr_shrink);
> > +
> > /*
> > * sysfs.c
> > */
> > diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
> > index cbe7c24065c7..56c23b5e9d65 100644
> > --- a/fs/f2fs/file.c
> > +++ b/fs/f2fs/file.c
> > @@ -619,6 +619,7 @@ void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count)
> > fofs = f2fs_start_bidx_of_node(ofs_of_node(dn->node_page),
> > dn->inode) + ofs;
> > f2fs_update_read_extent_cache_range(dn, fofs, 0, len);
> > + f2fs_update_age_extent_cache_range(dn, fofs, nr_free);
> > dec_valid_block_count(sbi, dn->inode, nr_free);
> > }
> > dn->ofs_in_node = ofs;
> > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> > index c845c16f97d0..ff6cf66ed46b 100644
> > --- a/fs/f2fs/inode.c
> > +++ b/fs/f2fs/inode.c
> > @@ -480,6 +480,7 @@ static int do_read_inode(struct inode *inode)
> >
> > /* Need all the flag bits */
> > f2fs_init_read_extent_tree(inode, node_page);
> > + f2fs_init_age_extent_tree(inode);
> >
> > f2fs_put_page(node_page, 1);
> >
> > diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
> > index 07419c3e42a5..dde4c0458704 100644
> > --- a/fs/f2fs/node.c
> > +++ b/fs/f2fs/node.c
> > @@ -60,7 +60,7 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
> > avail_ram = val.totalram - val.totalhigh;
> >
> > /*
> > - * give 25%, 25%, 50%, 50%, 50% memory for each components respectively
> > + * give 25%, 25%, 50%, 50%, 25%, 25% memory for each components respectively
> > */
> > if (type == FREE_NIDS) {
> > mem_size = (nm_i->nid_cnt[FREE_NID] *
> > @@ -85,14 +85,16 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
> > sizeof(struct ino_entry);
> > mem_size >>= PAGE_SHIFT;
> > res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
> > - } else if (type == READ_EXTENT_CACHE) {
> > - struct extent_tree_info *eti = &sbi->extent_tree[EX_READ];
> > + } else if (type == READ_EXTENT_CACHE || type == AGE_EXTENT_CACHE) {
> > + enum extent_type etype = type == READ_EXTENT_CACHE ?
> > + EX_READ : EX_BLOCK_AGE;
> > + struct extent_tree_info *eti = &sbi->extent_tree[etype];
> >
> > mem_size = (atomic_read(&eti->total_ext_tree) *
> > sizeof(struct extent_tree) +
> > atomic_read(&eti->total_ext_node) *
> > sizeof(struct extent_node)) >> PAGE_SHIFT;
> > - res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
> > + res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 2);
> > } else if (type == DISCARD_CACHE) {
> > mem_size = (atomic_read(&dcc->discard_cmd_cnt) *
> > sizeof(struct discard_cmd)) >> PAGE_SHIFT;
> > diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
> > index 0aa48704c77a..99454d46a939 100644
> > --- a/fs/f2fs/node.h
> > +++ b/fs/f2fs/node.h
> > @@ -147,6 +147,7 @@ enum mem_type {
> > DIRTY_DENTS, /* indicates dirty dentry pages */
> > INO_ENTRIES, /* indicates inode entries */
> > READ_EXTENT_CACHE, /* indicates read extent cache */
> > + AGE_EXTENT_CACHE, /* indicates age extent cache */
> > DISCARD_CACHE, /* indicates memory of cached discard cmds */
> > COMPRESS_PAGE, /* indicates memory of cached compressed pages */
> > BASE_CHECK, /* check kernel status */
> > diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
> > index 8722d1a13c17..dee712f7225f 100644
> > --- a/fs/f2fs/segment.c
> > +++ b/fs/f2fs/segment.c
> > @@ -453,6 +453,11 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
> > f2fs_shrink_read_extent_tree(sbi,
> > READ_EXTENT_CACHE_SHRINK_NUMBER);
> >
> > + /* try to shrink age extent cache when there is no enough memory */
> > + if (!f2fs_available_free_memory(sbi, AGE_EXTENT_CACHE))
> > + f2fs_shrink_age_extent_tree(sbi,
> > + AGE_EXTENT_CACHE_SHRINK_NUMBER);
> > +
> > /* check the # of cached NAT entries */
> > if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
> > f2fs_try_to_free_nats(sbi, NAT_ENTRY_PER_BLOCK);
> > @@ -3151,10 +3156,28 @@ static int __get_segment_type_4(struct f2fs_io_info *fio)
> > }
> > }
> >
> > +static int __get_age_segment_type(struct inode *inode, pgoff_t pgofs)
> > +{
> > + struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
> > + struct extent_info ei;
> > +
> > + if (f2fs_lookup_age_extent_cache(inode, pgofs, &ei)) {
> > + if (!ei.age)
> > + return NO_CHECK_TYPE;
> > + if (ei.age <= sbi->hot_data_age_threshold)
> > + return CURSEG_HOT_DATA;
> > + if (ei.age <= sbi->warm_data_age_threshold)
> > + return CURSEG_WARM_DATA;
> > + return CURSEG_COLD_DATA;
> > + }
> > + return NO_CHECK_TYPE;
> > +}
> > +
> > static int __get_segment_type_6(struct f2fs_io_info *fio)
> > {
> > if (fio->type == DATA) {
> > struct inode *inode = fio->page->mapping->host;
> > + int type;
> >
> > if (is_inode_flag_set(inode, FI_ALIGNED_WRITE))
> > return CURSEG_COLD_DATA_PINNED;
> > @@ -3169,6 +3192,11 @@ static int __get_segment_type_6(struct f2fs_io_info *fio)
> > }
> > if (file_is_cold(inode) || f2fs_need_compress_data(inode))
> > return CURSEG_COLD_DATA;
> > +
> > + type = __get_age_segment_type(inode, fio->page->index);
> > + if (type != NO_CHECK_TYPE)
> > + return type;
> > +
> > if (file_is_hot(inode) ||
> > is_inode_flag_set(inode, FI_HOT_DATA) ||
> > f2fs_is_cow_file(inode))
> > @@ -3287,6 +3315,9 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
> > locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr));
> > locate_dirty_segment(sbi, GET_SEGNO(sbi, *new_blkaddr));
> >
> > + if (IS_DATASEG(type))
> > + atomic64_inc(&sbi->allocated_data_blocks);
> > +
> > up_write(&sit_i->sentry_lock);
> >
> > if (page && IS_NODESEG(type)) {
> > @@ -3414,6 +3445,8 @@ void f2fs_outplace_write_data(struct dnode_of_data *dn,
> > struct f2fs_summary sum;
> >
> > f2fs_bug_on(sbi, dn->data_blkaddr == NULL_ADDR);
> > + if (fio->io_type == FS_DATA_IO || fio->io_type == FS_CP_DATA_IO)
> > + f2fs_update_age_extent_cache(dn);
> > set_summary(&sum, dn->nid, dn->ofs_in_node, fio->version);
> > do_write_page(&sum, fio);
> > f2fs_update_data_blkaddr(dn, fio->new_blkaddr);
> > diff --git a/fs/f2fs/shrinker.c b/fs/f2fs/shrinker.c
> > index 33c490e69ae3..83d6fb97dcae 100644
> > --- a/fs/f2fs/shrinker.c
> > +++ b/fs/f2fs/shrinker.c
> > @@ -59,6 +59,9 @@ unsigned long f2fs_shrink_count(struct shrinker *shrink,
> > /* count read extent cache entries */
> > count += __count_extent_cache(sbi, EX_READ);
> >
> > + /* count block age extent cache entries */
> > + count += __count_extent_cache(sbi, EX_BLOCK_AGE);
> > +
> > /* count clean nat cache entries */
> > count += __count_nat_entries(sbi);
> >
> > @@ -102,8 +105,11 @@ unsigned long f2fs_shrink_scan(struct shrinker *shrink,
> >
> > sbi->shrinker_run_no = run_no;
> >
> > + /* shrink extent cache entries */
> > + freed += f2fs_shrink_age_extent_tree(sbi, nr >> 2);
> > +
> > /* shrink read extent cache entries */
> > - freed += f2fs_shrink_read_extent_tree(sbi, nr >> 1);
> > + freed += f2fs_shrink_read_extent_tree(sbi, nr >> 2);
> >
> > /* shrink clean nat cache entries */
> > if (freed < nr)
> > @@ -134,6 +140,8 @@ void f2fs_join_shrinker(struct f2fs_sb_info *sbi)
> > void f2fs_leave_shrinker(struct f2fs_sb_info *sbi)
> > {
> > f2fs_shrink_read_extent_tree(sbi, __count_extent_cache(sbi, EX_READ));
> > + f2fs_shrink_age_extent_tree(sbi,
> > + __count_extent_cache(sbi, EX_BLOCK_AGE));
> >
> > spin_lock(&f2fs_list_lock);
> > list_del_init(&sbi->s_list);
> > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> > index 10bd03bbefec..5bdab376b852 100644
> > --- a/fs/f2fs/super.c
> > +++ b/fs/f2fs/super.c
> > @@ -163,6 +163,7 @@ enum {
> > Opt_nogc_merge,
> > Opt_discard_unit,
> > Opt_memory_mode,
> > + Opt_age_extent_cache,
> > Opt_err,
> > };
> >
> > @@ -241,6 +242,7 @@ static match_table_t f2fs_tokens = {
> > {Opt_nogc_merge, "nogc_merge"},
> > {Opt_discard_unit, "discard_unit=%s"},
> > {Opt_memory_mode, "memory=%s"},
> > + {Opt_age_extent_cache, "age_extent_cache"},
> > {Opt_err, NULL},
> > };
> >
> > @@ -1257,6 +1259,9 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
> > }
> > kfree(name);
> > break;
> > + case Opt_age_extent_cache:
> > + set_opt(sbi, AGE_EXTENT_CACHE);
> > + break;
> > default:
> > f2fs_err(sbi, "Unrecognized mount option \"%s\" or missing value",
> > p);
> > @@ -1958,6 +1963,8 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
> > seq_puts(seq, ",read_extent_cache");
> > else
> > seq_puts(seq, ",no_read_extent_cache");
> > + if (test_opt(sbi, AGE_EXTENT_CACHE))
> > + seq_puts(seq, ",age_extent_cache");
> > if (test_opt(sbi, DATA_FLUSH))
> > seq_puts(seq, ",data_flush");
> >
> > @@ -2219,6 +2226,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
> > bool need_restart_flush = false, need_stop_flush = false;
> > bool need_restart_discard = false, need_stop_discard = false;
> > bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE);
> > + bool no_age_extent_cache = !test_opt(sbi, AGE_EXTENT_CACHE);
> > bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT);
> > bool no_io_align = !F2FS_IO_ALIGNED(sbi);
> > bool no_atgc = !test_opt(sbi, ATGC);
> > @@ -2313,6 +2321,12 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
> > f2fs_warn(sbi, "switch extent_cache option is not allowed");
> > goto restore_opts;
> > }
> > + /* disallow enable/disable age extent_cache dynamically */
> > + if (no_age_extent_cache == !!test_opt(sbi, AGE_EXTENT_CACHE)) {
> > + err = -EINVAL;
> > + f2fs_warn(sbi, "switch age_extent_cache option is not allowed");
> > + goto restore_opts;
> > + }
> >
> > if (no_io_align == !!F2FS_IO_ALIGNED(sbi)) {
> > err = -EINVAL;
> > diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
> > index a4745d596310..2ab215110596 100644
> > --- a/fs/f2fs/sysfs.c
> > +++ b/fs/f2fs/sysfs.c
> > @@ -668,6 +668,24 @@ static ssize_t __sbi_store(struct f2fs_attr *a,
> > return count;
> > }
> >
> > + if (!strcmp(a->attr.name, "hot_data_age_threshold")) {
> > + if (t == 0 || t >= sbi->warm_data_age_threshold)
> > + return -EINVAL;
> > + if (t == *ui)
> > + return count;
> > + *ui = (unsigned int)t;
> > + return count;
> > + }
> > +
> > + if (!strcmp(a->attr.name, "warm_data_age_threshold")) {
> > + if (t == 0 || t <= sbi->hot_data_age_threshold)
> > + return -EINVAL;
> > + if (t == *ui)
> > + return count;
> > + *ui = (unsigned int)t;
> > + return count;
> > + }
> > +
> > *ui = (unsigned int)t;
> >
> > return count;
> > @@ -923,6 +941,10 @@ F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, peak_atomic_write, peak_atomic_write);
> > F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, committed_atomic_block, committed_atomic_block);
> > F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, revoked_atomic_block, revoked_atomic_block);
> >
> > +/* For block age extent cache */
> > +F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, hot_data_age_threshold, hot_data_age_threshold);
> > +F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, warm_data_age_threshold, warm_data_age_threshold);
> > +
> > #define ATTR_LIST(name) (&f2fs_attr_##name.attr)
> > static struct attribute *f2fs_attrs[] = {
> > ATTR_LIST(gc_urgent_sleep_time),
> > @@ -1018,6 +1040,8 @@ static struct attribute *f2fs_attrs[] = {
> > ATTR_LIST(peak_atomic_write),
> > ATTR_LIST(committed_atomic_block),
> > ATTR_LIST(revoked_atomic_block),
> > + ATTR_LIST(hot_data_age_threshold),
> > + ATTR_LIST(warm_data_age_threshold),
> > NULL,
> > };
> > ATTRIBUTE_GROUPS(f2fs);
> > diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h
> > index 2bb37892d2ba..31d994e6b4ca 100644
> > --- a/include/trace/events/f2fs.h
> > +++ b/include/trace/events/f2fs.h
> > @@ -49,6 +49,7 @@ TRACE_DEFINE_ENUM(CP_TRIMMED);
> > TRACE_DEFINE_ENUM(CP_PAUSE);
> > TRACE_DEFINE_ENUM(CP_RESIZE);
> > TRACE_DEFINE_ENUM(EX_READ);
> > +TRACE_DEFINE_ENUM(EX_BLOCK_AGE);
> >
> > #define show_block_type(type) \
> > __print_symbolic(type, \
> > @@ -155,6 +156,11 @@ TRACE_DEFINE_ENUM(EX_READ);
> > { COMPRESS_ZSTD, "ZSTD" }, \
> > { COMPRESS_LZORLE, "LZO-RLE" })
> >
> > +#define show_extent_type(type) \
> > + __print_symbolic(type, \
> > + { EX_READ, "Read" }, \
> > + { EX_BLOCK_AGE, "Block Age" })
> > +
> > struct f2fs_sb_info;
> > struct f2fs_io_info;
> > struct extent_info;
> > @@ -1544,7 +1550,7 @@ TRACE_EVENT(f2fs_lookup_extent_tree_start,
> > TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, type = %s",
> > show_dev_ino(__entry),
> > __entry->pgofs,
> > - __entry->type == EX_READ ? "Read" : "N/A")
> > + show_extent_type(__entry->type))
> > );
> >
> > TRACE_EVENT_CONDITION(f2fs_lookup_read_extent_tree_end,
> > @@ -1583,6 +1589,45 @@ TRACE_EVENT_CONDITION(f2fs_lookup_read_extent_tree_end,
> > __entry->blk)
> > );
> >
> > +TRACE_EVENT_CONDITION(f2fs_lookup_age_extent_tree_end,
> > +
> > + TP_PROTO(struct inode *inode, unsigned int pgofs,
> > + struct extent_info *ei),
> > +
> > + TP_ARGS(inode, pgofs, ei),
> > +
> > + TP_CONDITION(ei),
> > +
> > + TP_STRUCT__entry(
> > + __field(dev_t, dev)
> > + __field(ino_t, ino)
> > + __field(unsigned int, pgofs)
> > + __field(unsigned int, fofs)
> > + __field(unsigned int, len)
> > + __field(unsigned long long, age)
> > + __field(unsigned long long, blocks)
> > + ),
> > +
> > + TP_fast_assign(
> > + __entry->dev = inode->i_sb->s_dev;
> > + __entry->ino = inode->i_ino;
> > + __entry->pgofs = pgofs;
> > + __entry->fofs = ei->fofs;
> > + __entry->len = ei->len;
> > + __entry->age = ei->age;
> > + __entry->blocks = ei->last_blocks;
> > + ),
> > +
> > + TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, "
> > + "age_ext_info(fofs: %u, len: %u, age: %llu, blocks: %llu)",
> > + show_dev_ino(__entry),
> > + __entry->pgofs,
> > + __entry->fofs,
> > + __entry->len,
> > + __entry->age,
> > + __entry->blocks)
> > +);
> > +
> > TRACE_EVENT(f2fs_update_read_extent_tree_range,
> >
> > TP_PROTO(struct inode *inode, unsigned int pgofs, unsigned int len,
> > @@ -1618,6 +1663,41 @@ TRACE_EVENT(f2fs_update_read_extent_tree_range,
> > __entry->c_len)
> > );
> >
> > +TRACE_EVENT(f2fs_update_age_extent_tree_range,
> > +
> > + TP_PROTO(struct inode *inode, unsigned int pgofs, unsigned int len,
> > + unsigned long long age,
> > + unsigned long long last_blks),
> > +
> > + TP_ARGS(inode, pgofs, len, age, last_blks),
> > +
> > + TP_STRUCT__entry(
> > + __field(dev_t, dev)
> > + __field(ino_t, ino)
> > + __field(unsigned int, pgofs)
> > + __field(unsigned int, len)
> > + __field(unsigned long long, age)
> > + __field(unsigned long long, blocks)
> > + ),
> > +
> > + TP_fast_assign(
> > + __entry->dev = inode->i_sb->s_dev;
> > + __entry->ino = inode->i_ino;
> > + __entry->pgofs = pgofs;
> > + __entry->len = len;
> > + __entry->age = age;
> > + __entry->blocks = last_blks;
> > + ),
> > +
> > + TP_printk("dev = (%d,%d), ino = %lu, pgofs = %u, "
> > + "len = %u, age = %llu, blocks = %llu",
> > + show_dev_ino(__entry),
> > + __entry->pgofs,
> > + __entry->len,
> > + __entry->age,
> > + __entry->blocks)
> > +);
> > +
> > TRACE_EVENT(f2fs_shrink_extent_tree,
> >
> > TP_PROTO(struct f2fs_sb_info *sbi, unsigned int node_cnt,
> > @@ -1643,7 +1723,7 @@ TRACE_EVENT(f2fs_shrink_extent_tree,
> > show_dev(__entry->dev),
> > __entry->node_cnt,
> > __entry->tree_cnt,
> > - __entry->type == EX_READ ? "Read" : "N/A")
> > + show_extent_type(__entry->type))
> > );
> >
> > TRACE_EVENT(f2fs_destroy_extent_tree,
> > @@ -1670,7 +1750,7 @@ TRACE_EVENT(f2fs_destroy_extent_tree,
> > TP_printk("dev = (%d,%d), ino = %lu, destroyed: node_cnt = %u, type = %s",
> > show_dev_ino(__entry),
> > __entry->node_cnt,
> > - __entry->type == EX_READ ? "Read" : "N/A")
> > + show_extent_type(__entry->type))
> > );
> >
> > DECLARE_EVENT_CLASS(f2fs_sync_dirty_inodes,
> > --
> > 2.39.0.rc0.267.gcb52ba06e7-goog
> >
> >
> >
> > _______________________________________________
> > Linux-f2fs-devel mailing list
> > Linux-f2fs-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly
2022-12-05 18:54 [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly Jaegeuk Kim
` (4 preceding siblings ...)
2022-12-05 18:54 ` [f2fs-dev] [PATCH 6/6] f2fs: add block_age-based extent cache Jaegeuk Kim
@ 2022-12-08 14:43 ` Chao Yu
2022-12-08 17:29 ` Jaegeuk Kim
2022-12-09 19:40 ` [f2fs-dev] [PATCH 1/6 v2] " Jaegeuk Kim
5 siblings, 2 replies; 19+ messages in thread
From: Chao Yu @ 2022-12-08 14:43 UTC (permalink / raw)
To: Jaegeuk Kim, linux-kernel, linux-f2fs-devel
On 2022/12/6 2:54, Jaegeuk Kim wrote:
> Let's descrbie it's read extent cache.
>
> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
> ---
> fs/f2fs/extent_cache.c | 4 ++--
> fs/f2fs/f2fs.h | 10 +++++-----
> fs/f2fs/inode.c | 2 +-
> fs/f2fs/node.c | 2 +-
> fs/f2fs/node.h | 2 +-
> fs/f2fs/segment.c | 4 ++--
> fs/f2fs/super.c | 16 ++++++++--------
> 7 files changed, 20 insertions(+), 20 deletions(-)
>
> diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
> index 932c070173b9..8cd87aee0292 100644
> --- a/fs/f2fs/extent_cache.c
> +++ b/fs/f2fs/extent_cache.c
> @@ -383,7 +383,7 @@ static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage)
> if (!i_ext || !i_ext->len)
> return;
>
> - get_extent_info(&ei, i_ext);
> + get_read_extent_info(&ei, i_ext);
>
> write_lock(&et->lock);
> if (atomic_read(&et->node_cnt))
> @@ -710,7 +710,7 @@ unsigned int f2fs_shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
> unsigned int node_cnt = 0, tree_cnt = 0;
> int remained;
>
> - if (!test_opt(sbi, EXTENT_CACHE))
> + if (!test_opt(sbi, READ_EXTENT_CACHE))
> return 0;
>
> if (!atomic_read(&sbi->total_zombie_tree))
> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> index eb8c27c4e5fc..1c39f8145b61 100644
> --- a/fs/f2fs/f2fs.h
> +++ b/fs/f2fs/f2fs.h
> @@ -92,7 +92,7 @@ extern const char *f2fs_fault_name[FAULT_MAX];
> #define F2FS_MOUNT_FLUSH_MERGE 0x00000400
> #define F2FS_MOUNT_NOBARRIER 0x00000800
> #define F2FS_MOUNT_FASTBOOT 0x00001000
> -#define F2FS_MOUNT_EXTENT_CACHE 0x00002000
> +#define F2FS_MOUNT_READ_EXTENT_CACHE 0x00002000
> #define F2FS_MOUNT_DATA_FLUSH 0x00008000
> #define F2FS_MOUNT_FAULT_INJECTION 0x00010000
> #define F2FS_MOUNT_USRQUOTA 0x00080000
> @@ -600,7 +600,7 @@ enum {
> #define F2FS_MIN_EXTENT_LEN 64 /* minimum extent length */
>
> /* number of extent info in extent cache we try to shrink */
> -#define EXTENT_CACHE_SHRINK_NUMBER 128
> +#define READ_EXTENT_CACHE_SHRINK_NUMBER 128
>
> #define RECOVERY_MAX_RA_BLOCKS BIO_MAX_VECS
> #define RECOVERY_MIN_RA_BLOCKS 1
> @@ -830,7 +830,7 @@ struct f2fs_inode_info {
> loff_t original_i_size; /* original i_size before atomic write */
> };
>
> -static inline void get_extent_info(struct extent_info *ext,
> +static inline void get_read_extent_info(struct extent_info *ext,
> struct f2fs_extent *i_ext)
> {
> ext->fofs = le32_to_cpu(i_ext->fofs);
> @@ -838,7 +838,7 @@ static inline void get_extent_info(struct extent_info *ext,
> ext->len = le32_to_cpu(i_ext->len);
> }
>
> -static inline void set_raw_extent(struct extent_info *ext,
> +static inline void set_raw_read_extent(struct extent_info *ext,
> struct f2fs_extent *i_ext)
> {
> i_ext->fofs = cpu_to_le32(ext->fofs);
> @@ -4407,7 +4407,7 @@ static inline bool f2fs_may_extent_tree(struct inode *inode)
> {
> struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
>
> - if (!test_opt(sbi, EXTENT_CACHE) ||
> + if (!test_opt(sbi, READ_EXTENT_CACHE) ||
> is_inode_flag_set(inode, FI_NO_EXTENT) ||
> (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
> !f2fs_sb_has_readonly(sbi)))
> diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> index 577f109b4e1d..2c705c60019b 100644
> --- a/fs/f2fs/inode.c
> +++ b/fs/f2fs/inode.c
> @@ -629,7 +629,7 @@ void f2fs_update_inode(struct inode *inode, struct page *node_page)
>
> if (et) {
> read_lock(&et->lock);
> - set_raw_extent(&et->largest, &ri->i_ext);
> + set_raw_read_extent(&et->largest, &ri->i_ext);
> read_unlock(&et->lock);
> } else {
> memset(&ri->i_ext, 0, sizeof(ri->i_ext));
> diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
> index b9ee5a1176a0..84b147966080 100644
> --- a/fs/f2fs/node.c
> +++ b/fs/f2fs/node.c
> @@ -85,7 +85,7 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
> sizeof(struct ino_entry);
> mem_size >>= PAGE_SHIFT;
> res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
> - } else if (type == EXTENT_CACHE) {
> + } else if (type == READ_EXTENT_CACHE) {
> mem_size = (atomic_read(&sbi->total_ext_tree) *
> sizeof(struct extent_tree) +
> atomic_read(&sbi->total_ext_node) *
> diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
> index 3c09cae058b0..0aa48704c77a 100644
> --- a/fs/f2fs/node.h
> +++ b/fs/f2fs/node.h
> @@ -146,7 +146,7 @@ enum mem_type {
> NAT_ENTRIES, /* indicates the cached nat entry */
> DIRTY_DENTS, /* indicates dirty dentry pages */
> INO_ENTRIES, /* indicates inode entries */
> - EXTENT_CACHE, /* indicates extent cache */
> + READ_EXTENT_CACHE, /* indicates read extent cache */
> DISCARD_CACHE, /* indicates memory of cached discard cmds */
> COMPRESS_PAGE, /* indicates memory of cached compressed pages */
> BASE_CHECK, /* check kernel status */
> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
> index 9486ca49ecb1..51de358bc452 100644
> --- a/fs/f2fs/segment.c
> +++ b/fs/f2fs/segment.c
> @@ -449,8 +449,8 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
> return;
>
> /* try to shrink extent cache when there is no enough memory */
> - if (!f2fs_available_free_memory(sbi, EXTENT_CACHE))
> - f2fs_shrink_extent_tree(sbi, EXTENT_CACHE_SHRINK_NUMBER);
> + if (!f2fs_available_free_memory(sbi, READ_EXTENT_CACHE))
> + f2fs_shrink_extent_tree(sbi, READ_EXTENT_CACHE_SHRINK_NUMBER);
>
> /* check the # of cached NAT entries */
> if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
> diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> index 79bf1faf4161..10bd03bbefec 100644
> --- a/fs/f2fs/super.c
> +++ b/fs/f2fs/super.c
> @@ -814,10 +814,10 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
> set_opt(sbi, FASTBOOT);
> break;
> case Opt_extent_cache:
> - set_opt(sbi, EXTENT_CACHE);
> + set_opt(sbi, READ_EXTENT_CACHE);
> break;
> case Opt_noextent_cache:
> - clear_opt(sbi, EXTENT_CACHE);
> + clear_opt(sbi, READ_EXTENT_CACHE);
> break;
> case Opt_noinline_data:
> clear_opt(sbi, INLINE_DATA);
> @@ -1954,10 +1954,10 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
> seq_puts(seq, ",barrier");
> if (test_opt(sbi, FASTBOOT))
> seq_puts(seq, ",fastboot");
> - if (test_opt(sbi, EXTENT_CACHE))
> - seq_puts(seq, ",extent_cache");
> + if (test_opt(sbi, READ_EXTENT_CACHE))
> + seq_puts(seq, ",read_extent_cache");
How about keeping consistent w/ description of f2fs.rst?
> else
> - seq_puts(seq, ",noextent_cache");
> + seq_puts(seq, ",no_read_extent_cache");
Ditto,
Thanks,
> if (test_opt(sbi, DATA_FLUSH))
> seq_puts(seq, ",data_flush");
>
> @@ -2076,7 +2076,7 @@ static void default_options(struct f2fs_sb_info *sbi)
> set_opt(sbi, INLINE_XATTR);
> set_opt(sbi, INLINE_DATA);
> set_opt(sbi, INLINE_DENTRY);
> - set_opt(sbi, EXTENT_CACHE);
> + set_opt(sbi, READ_EXTENT_CACHE);
> set_opt(sbi, NOHEAP);
> clear_opt(sbi, DISABLE_CHECKPOINT);
> set_opt(sbi, MERGE_CHECKPOINT);
> @@ -2218,7 +2218,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
> bool need_restart_ckpt = false, need_stop_ckpt = false;
> bool need_restart_flush = false, need_stop_flush = false;
> bool need_restart_discard = false, need_stop_discard = false;
> - bool no_extent_cache = !test_opt(sbi, EXTENT_CACHE);
> + bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE);
> bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT);
> bool no_io_align = !F2FS_IO_ALIGNED(sbi);
> bool no_atgc = !test_opt(sbi, ATGC);
> @@ -2308,7 +2308,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
> }
>
> /* disallow enable/disable extent_cache dynamically */
> - if (no_extent_cache == !!test_opt(sbi, EXTENT_CACHE)) {
> + if (no_read_extent_cache == !!test_opt(sbi, READ_EXTENT_CACHE)) {
> err = -EINVAL;
> f2fs_warn(sbi, "switch extent_cache option is not allowed");
> goto restore_opts;
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly
2022-12-08 14:43 ` [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly Chao Yu
@ 2022-12-08 17:29 ` Jaegeuk Kim
2022-12-09 1:31 ` Chao Yu
2022-12-09 19:40 ` [f2fs-dev] [PATCH 1/6 v2] " Jaegeuk Kim
1 sibling, 1 reply; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-08 17:29 UTC (permalink / raw)
To: Chao Yu; +Cc: linux-kernel, linux-f2fs-devel
On 12/08, Chao Yu wrote:
> On 2022/12/6 2:54, Jaegeuk Kim wrote:
> > Let's descrbie it's read extent cache.
> >
> > Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
> > ---
> > fs/f2fs/extent_cache.c | 4 ++--
> > fs/f2fs/f2fs.h | 10 +++++-----
> > fs/f2fs/inode.c | 2 +-
> > fs/f2fs/node.c | 2 +-
> > fs/f2fs/node.h | 2 +-
> > fs/f2fs/segment.c | 4 ++--
> > fs/f2fs/super.c | 16 ++++++++--------
> > 7 files changed, 20 insertions(+), 20 deletions(-)
> >
> > diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
> > index 932c070173b9..8cd87aee0292 100644
> > --- a/fs/f2fs/extent_cache.c
> > +++ b/fs/f2fs/extent_cache.c
> > @@ -383,7 +383,7 @@ static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage)
> > if (!i_ext || !i_ext->len)
> > return;
> > - get_extent_info(&ei, i_ext);
> > + get_read_extent_info(&ei, i_ext);
> > write_lock(&et->lock);
> > if (atomic_read(&et->node_cnt))
> > @@ -710,7 +710,7 @@ unsigned int f2fs_shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
> > unsigned int node_cnt = 0, tree_cnt = 0;
> > int remained;
> > - if (!test_opt(sbi, EXTENT_CACHE))
> > + if (!test_opt(sbi, READ_EXTENT_CACHE))
> > return 0;
> > if (!atomic_read(&sbi->total_zombie_tree))
> > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> > index eb8c27c4e5fc..1c39f8145b61 100644
> > --- a/fs/f2fs/f2fs.h
> > +++ b/fs/f2fs/f2fs.h
> > @@ -92,7 +92,7 @@ extern const char *f2fs_fault_name[FAULT_MAX];
> > #define F2FS_MOUNT_FLUSH_MERGE 0x00000400
> > #define F2FS_MOUNT_NOBARRIER 0x00000800
> > #define F2FS_MOUNT_FASTBOOT 0x00001000
> > -#define F2FS_MOUNT_EXTENT_CACHE 0x00002000
> > +#define F2FS_MOUNT_READ_EXTENT_CACHE 0x00002000
> > #define F2FS_MOUNT_DATA_FLUSH 0x00008000
> > #define F2FS_MOUNT_FAULT_INJECTION 0x00010000
> > #define F2FS_MOUNT_USRQUOTA 0x00080000
> > @@ -600,7 +600,7 @@ enum {
> > #define F2FS_MIN_EXTENT_LEN 64 /* minimum extent length */
> > /* number of extent info in extent cache we try to shrink */
> > -#define EXTENT_CACHE_SHRINK_NUMBER 128
> > +#define READ_EXTENT_CACHE_SHRINK_NUMBER 128
> > #define RECOVERY_MAX_RA_BLOCKS BIO_MAX_VECS
> > #define RECOVERY_MIN_RA_BLOCKS 1
> > @@ -830,7 +830,7 @@ struct f2fs_inode_info {
> > loff_t original_i_size; /* original i_size before atomic write */
> > };
> > -static inline void get_extent_info(struct extent_info *ext,
> > +static inline void get_read_extent_info(struct extent_info *ext,
> > struct f2fs_extent *i_ext)
> > {
> > ext->fofs = le32_to_cpu(i_ext->fofs);
> > @@ -838,7 +838,7 @@ static inline void get_extent_info(struct extent_info *ext,
> > ext->len = le32_to_cpu(i_ext->len);
> > }
> > -static inline void set_raw_extent(struct extent_info *ext,
> > +static inline void set_raw_read_extent(struct extent_info *ext,
> > struct f2fs_extent *i_ext)
> > {
> > i_ext->fofs = cpu_to_le32(ext->fofs);
> > @@ -4407,7 +4407,7 @@ static inline bool f2fs_may_extent_tree(struct inode *inode)
> > {
> > struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
> > - if (!test_opt(sbi, EXTENT_CACHE) ||
> > + if (!test_opt(sbi, READ_EXTENT_CACHE) ||
> > is_inode_flag_set(inode, FI_NO_EXTENT) ||
> > (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
> > !f2fs_sb_has_readonly(sbi)))
> > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> > index 577f109b4e1d..2c705c60019b 100644
> > --- a/fs/f2fs/inode.c
> > +++ b/fs/f2fs/inode.c
> > @@ -629,7 +629,7 @@ void f2fs_update_inode(struct inode *inode, struct page *node_page)
> > if (et) {
> > read_lock(&et->lock);
> > - set_raw_extent(&et->largest, &ri->i_ext);
> > + set_raw_read_extent(&et->largest, &ri->i_ext);
> > read_unlock(&et->lock);
> > } else {
> > memset(&ri->i_ext, 0, sizeof(ri->i_ext));
> > diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
> > index b9ee5a1176a0..84b147966080 100644
> > --- a/fs/f2fs/node.c
> > +++ b/fs/f2fs/node.c
> > @@ -85,7 +85,7 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
> > sizeof(struct ino_entry);
> > mem_size >>= PAGE_SHIFT;
> > res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
> > - } else if (type == EXTENT_CACHE) {
> > + } else if (type == READ_EXTENT_CACHE) {
> > mem_size = (atomic_read(&sbi->total_ext_tree) *
> > sizeof(struct extent_tree) +
> > atomic_read(&sbi->total_ext_node) *
> > diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
> > index 3c09cae058b0..0aa48704c77a 100644
> > --- a/fs/f2fs/node.h
> > +++ b/fs/f2fs/node.h
> > @@ -146,7 +146,7 @@ enum mem_type {
> > NAT_ENTRIES, /* indicates the cached nat entry */
> > DIRTY_DENTS, /* indicates dirty dentry pages */
> > INO_ENTRIES, /* indicates inode entries */
> > - EXTENT_CACHE, /* indicates extent cache */
> > + READ_EXTENT_CACHE, /* indicates read extent cache */
> > DISCARD_CACHE, /* indicates memory of cached discard cmds */
> > COMPRESS_PAGE, /* indicates memory of cached compressed pages */
> > BASE_CHECK, /* check kernel status */
> > diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
> > index 9486ca49ecb1..51de358bc452 100644
> > --- a/fs/f2fs/segment.c
> > +++ b/fs/f2fs/segment.c
> > @@ -449,8 +449,8 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
> > return;
> > /* try to shrink extent cache when there is no enough memory */
> > - if (!f2fs_available_free_memory(sbi, EXTENT_CACHE))
> > - f2fs_shrink_extent_tree(sbi, EXTENT_CACHE_SHRINK_NUMBER);
> > + if (!f2fs_available_free_memory(sbi, READ_EXTENT_CACHE))
> > + f2fs_shrink_extent_tree(sbi, READ_EXTENT_CACHE_SHRINK_NUMBER);
> > /* check the # of cached NAT entries */
> > if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
> > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> > index 79bf1faf4161..10bd03bbefec 100644
> > --- a/fs/f2fs/super.c
> > +++ b/fs/f2fs/super.c
> > @@ -814,10 +814,10 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
> > set_opt(sbi, FASTBOOT);
> > break;
> > case Opt_extent_cache:
> > - set_opt(sbi, EXTENT_CACHE);
> > + set_opt(sbi, READ_EXTENT_CACHE);
> > break;
> > case Opt_noextent_cache:
> > - clear_opt(sbi, EXTENT_CACHE);
> > + clear_opt(sbi, READ_EXTENT_CACHE);
> > break;
> > case Opt_noinline_data:
> > clear_opt(sbi, INLINE_DATA);
> > @@ -1954,10 +1954,10 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
> > seq_puts(seq, ",barrier");
> > if (test_opt(sbi, FASTBOOT))
> > seq_puts(seq, ",fastboot");
> > - if (test_opt(sbi, EXTENT_CACHE))
> > - seq_puts(seq, ",extent_cache");
> > + if (test_opt(sbi, READ_EXTENT_CACHE))
> > + seq_puts(seq, ",read_extent_cache");
>
> How about keeping consistent w/ description of f2fs.rst?
This doesn't change the mount options, but do show the exact name which
would be better to say what's going on.
>
> > else
> > - seq_puts(seq, ",noextent_cache");
> > + seq_puts(seq, ",no_read_extent_cache");
>
> Ditto,
>
> Thanks,
>
> > if (test_opt(sbi, DATA_FLUSH))
> > seq_puts(seq, ",data_flush");
> > @@ -2076,7 +2076,7 @@ static void default_options(struct f2fs_sb_info *sbi)
> > set_opt(sbi, INLINE_XATTR);
> > set_opt(sbi, INLINE_DATA);
> > set_opt(sbi, INLINE_DENTRY);
> > - set_opt(sbi, EXTENT_CACHE);
> > + set_opt(sbi, READ_EXTENT_CACHE);
> > set_opt(sbi, NOHEAP);
> > clear_opt(sbi, DISABLE_CHECKPOINT);
> > set_opt(sbi, MERGE_CHECKPOINT);
> > @@ -2218,7 +2218,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
> > bool need_restart_ckpt = false, need_stop_ckpt = false;
> > bool need_restart_flush = false, need_stop_flush = false;
> > bool need_restart_discard = false, need_stop_discard = false;
> > - bool no_extent_cache = !test_opt(sbi, EXTENT_CACHE);
> > + bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE);
> > bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT);
> > bool no_io_align = !F2FS_IO_ALIGNED(sbi);
> > bool no_atgc = !test_opt(sbi, ATGC);
> > @@ -2308,7 +2308,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
> > }
> > /* disallow enable/disable extent_cache dynamically */
> > - if (no_extent_cache == !!test_opt(sbi, EXTENT_CACHE)) {
> > + if (no_read_extent_cache == !!test_opt(sbi, READ_EXTENT_CACHE)) {
> > err = -EINVAL;
> > f2fs_warn(sbi, "switch extent_cache option is not allowed");
> > goto restore_opts;
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly
2022-12-08 17:29 ` Jaegeuk Kim
@ 2022-12-09 1:31 ` Chao Yu
2022-12-09 19:39 ` Jaegeuk Kim
0 siblings, 1 reply; 19+ messages in thread
From: Chao Yu @ 2022-12-09 1:31 UTC (permalink / raw)
To: Jaegeuk Kim; +Cc: linux-kernel, linux-f2fs-devel
On 2022/12/9 1:29, Jaegeuk Kim wrote:
> On 12/08, Chao Yu wrote:
>> On 2022/12/6 2:54, Jaegeuk Kim wrote:
>>> Let's descrbie it's read extent cache.
>>>
>>> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
>>> ---
>>> fs/f2fs/extent_cache.c | 4 ++--
>>> fs/f2fs/f2fs.h | 10 +++++-----
>>> fs/f2fs/inode.c | 2 +-
>>> fs/f2fs/node.c | 2 +-
>>> fs/f2fs/node.h | 2 +-
>>> fs/f2fs/segment.c | 4 ++--
>>> fs/f2fs/super.c | 16 ++++++++--------
>>> 7 files changed, 20 insertions(+), 20 deletions(-)
>>>
>>> diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
>>> index 932c070173b9..8cd87aee0292 100644
>>> --- a/fs/f2fs/extent_cache.c
>>> +++ b/fs/f2fs/extent_cache.c
>>> @@ -383,7 +383,7 @@ static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage)
>>> if (!i_ext || !i_ext->len)
>>> return;
>>> - get_extent_info(&ei, i_ext);
>>> + get_read_extent_info(&ei, i_ext);
>>> write_lock(&et->lock);
>>> if (atomic_read(&et->node_cnt))
>>> @@ -710,7 +710,7 @@ unsigned int f2fs_shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
>>> unsigned int node_cnt = 0, tree_cnt = 0;
>>> int remained;
>>> - if (!test_opt(sbi, EXTENT_CACHE))
>>> + if (!test_opt(sbi, READ_EXTENT_CACHE))
>>> return 0;
>>> if (!atomic_read(&sbi->total_zombie_tree))
>>> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
>>> index eb8c27c4e5fc..1c39f8145b61 100644
>>> --- a/fs/f2fs/f2fs.h
>>> +++ b/fs/f2fs/f2fs.h
>>> @@ -92,7 +92,7 @@ extern const char *f2fs_fault_name[FAULT_MAX];
>>> #define F2FS_MOUNT_FLUSH_MERGE 0x00000400
>>> #define F2FS_MOUNT_NOBARRIER 0x00000800
>>> #define F2FS_MOUNT_FASTBOOT 0x00001000
>>> -#define F2FS_MOUNT_EXTENT_CACHE 0x00002000
>>> +#define F2FS_MOUNT_READ_EXTENT_CACHE 0x00002000
>>> #define F2FS_MOUNT_DATA_FLUSH 0x00008000
>>> #define F2FS_MOUNT_FAULT_INJECTION 0x00010000
>>> #define F2FS_MOUNT_USRQUOTA 0x00080000
>>> @@ -600,7 +600,7 @@ enum {
>>> #define F2FS_MIN_EXTENT_LEN 64 /* minimum extent length */
>>> /* number of extent info in extent cache we try to shrink */
>>> -#define EXTENT_CACHE_SHRINK_NUMBER 128
>>> +#define READ_EXTENT_CACHE_SHRINK_NUMBER 128
>>> #define RECOVERY_MAX_RA_BLOCKS BIO_MAX_VECS
>>> #define RECOVERY_MIN_RA_BLOCKS 1
>>> @@ -830,7 +830,7 @@ struct f2fs_inode_info {
>>> loff_t original_i_size; /* original i_size before atomic write */
>>> };
>>> -static inline void get_extent_info(struct extent_info *ext,
>>> +static inline void get_read_extent_info(struct extent_info *ext,
>>> struct f2fs_extent *i_ext)
>>> {
>>> ext->fofs = le32_to_cpu(i_ext->fofs);
>>> @@ -838,7 +838,7 @@ static inline void get_extent_info(struct extent_info *ext,
>>> ext->len = le32_to_cpu(i_ext->len);
>>> }
>>> -static inline void set_raw_extent(struct extent_info *ext,
>>> +static inline void set_raw_read_extent(struct extent_info *ext,
>>> struct f2fs_extent *i_ext)
>>> {
>>> i_ext->fofs = cpu_to_le32(ext->fofs);
>>> @@ -4407,7 +4407,7 @@ static inline bool f2fs_may_extent_tree(struct inode *inode)
>>> {
>>> struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
>>> - if (!test_opt(sbi, EXTENT_CACHE) ||
>>> + if (!test_opt(sbi, READ_EXTENT_CACHE) ||
>>> is_inode_flag_set(inode, FI_NO_EXTENT) ||
>>> (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
>>> !f2fs_sb_has_readonly(sbi)))
>>> diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
>>> index 577f109b4e1d..2c705c60019b 100644
>>> --- a/fs/f2fs/inode.c
>>> +++ b/fs/f2fs/inode.c
>>> @@ -629,7 +629,7 @@ void f2fs_update_inode(struct inode *inode, struct page *node_page)
>>> if (et) {
>>> read_lock(&et->lock);
>>> - set_raw_extent(&et->largest, &ri->i_ext);
>>> + set_raw_read_extent(&et->largest, &ri->i_ext);
>>> read_unlock(&et->lock);
>>> } else {
>>> memset(&ri->i_ext, 0, sizeof(ri->i_ext));
>>> diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
>>> index b9ee5a1176a0..84b147966080 100644
>>> --- a/fs/f2fs/node.c
>>> +++ b/fs/f2fs/node.c
>>> @@ -85,7 +85,7 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
>>> sizeof(struct ino_entry);
>>> mem_size >>= PAGE_SHIFT;
>>> res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
>>> - } else if (type == EXTENT_CACHE) {
>>> + } else if (type == READ_EXTENT_CACHE) {
>>> mem_size = (atomic_read(&sbi->total_ext_tree) *
>>> sizeof(struct extent_tree) +
>>> atomic_read(&sbi->total_ext_node) *
>>> diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
>>> index 3c09cae058b0..0aa48704c77a 100644
>>> --- a/fs/f2fs/node.h
>>> +++ b/fs/f2fs/node.h
>>> @@ -146,7 +146,7 @@ enum mem_type {
>>> NAT_ENTRIES, /* indicates the cached nat entry */
>>> DIRTY_DENTS, /* indicates dirty dentry pages */
>>> INO_ENTRIES, /* indicates inode entries */
>>> - EXTENT_CACHE, /* indicates extent cache */
>>> + READ_EXTENT_CACHE, /* indicates read extent cache */
>>> DISCARD_CACHE, /* indicates memory of cached discard cmds */
>>> COMPRESS_PAGE, /* indicates memory of cached compressed pages */
>>> BASE_CHECK, /* check kernel status */
>>> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
>>> index 9486ca49ecb1..51de358bc452 100644
>>> --- a/fs/f2fs/segment.c
>>> +++ b/fs/f2fs/segment.c
>>> @@ -449,8 +449,8 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
>>> return;
>>> /* try to shrink extent cache when there is no enough memory */
>>> - if (!f2fs_available_free_memory(sbi, EXTENT_CACHE))
>>> - f2fs_shrink_extent_tree(sbi, EXTENT_CACHE_SHRINK_NUMBER);
>>> + if (!f2fs_available_free_memory(sbi, READ_EXTENT_CACHE))
>>> + f2fs_shrink_extent_tree(sbi, READ_EXTENT_CACHE_SHRINK_NUMBER);
>>> /* check the # of cached NAT entries */
>>> if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
>>> diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
>>> index 79bf1faf4161..10bd03bbefec 100644
>>> --- a/fs/f2fs/super.c
>>> +++ b/fs/f2fs/super.c
>>> @@ -814,10 +814,10 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
>>> set_opt(sbi, FASTBOOT);
>>> break;
>>> case Opt_extent_cache:
>>> - set_opt(sbi, EXTENT_CACHE);
>>> + set_opt(sbi, READ_EXTENT_CACHE);
>>> break;
>>> case Opt_noextent_cache:
>>> - clear_opt(sbi, EXTENT_CACHE);
>>> + clear_opt(sbi, READ_EXTENT_CACHE);
>>> break;
>>> case Opt_noinline_data:
>>> clear_opt(sbi, INLINE_DATA);
>>> @@ -1954,10 +1954,10 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
>>> seq_puts(seq, ",barrier");
>>> if (test_opt(sbi, FASTBOOT))
>>> seq_puts(seq, ",fastboot");
>>> - if (test_opt(sbi, EXTENT_CACHE))
>>> - seq_puts(seq, ",extent_cache");
>>> + if (test_opt(sbi, READ_EXTENT_CACHE))
>>> + seq_puts(seq, ",read_extent_cache");
>>
>> How about keeping consistent w/ description of f2fs.rst?
>
> This doesn't change the mount options, but do show the exact name which
> would be better to say what's going on.
mount -t f2fs /dev/vda /mnt/f2fs
mount|grep f2fs
/dev/vda on /mnt/f2fs type f2fs (rw,relatime,lazytime,background_gc=on,discard,no_heap,user_xattr,inline_xattr,acl,inline_data,inline_dentry,flush_merge,barrier,read_extent_cache,mode=adaptive,active_logs=6,alloc_mode=reuse,check)
mount -t f2fs -o remount /mnt/f2fs/
[ 63.142480] F2FS-fs (vda): Unrecognized mount option "read_extent_cache" or missing value
Quoted from manual of mount
"mount -o remount,rw /dir
After this call, mount reads fstab and merges these options with the options
from the command line (-o). If no mountpoint is found in fstab, then a
remount with unspecified source is allowed."
I guess above command tries to use previous mountoptions which were read
via .show_options().
Thanks,
>
>>
>>> else
>>> - seq_puts(seq, ",noextent_cache");
>>> + seq_puts(seq, ",no_read_extent_cache");
>>
>> Ditto,
>>
>> Thanks,
>>
>>> if (test_opt(sbi, DATA_FLUSH))
>>> seq_puts(seq, ",data_flush");
>>> @@ -2076,7 +2076,7 @@ static void default_options(struct f2fs_sb_info *sbi)
>>> set_opt(sbi, INLINE_XATTR);
>>> set_opt(sbi, INLINE_DATA);
>>> set_opt(sbi, INLINE_DENTRY);
>>> - set_opt(sbi, EXTENT_CACHE);
>>> + set_opt(sbi, READ_EXTENT_CACHE);
>>> set_opt(sbi, NOHEAP);
>>> clear_opt(sbi, DISABLE_CHECKPOINT);
>>> set_opt(sbi, MERGE_CHECKPOINT);
>>> @@ -2218,7 +2218,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
>>> bool need_restart_ckpt = false, need_stop_ckpt = false;
>>> bool need_restart_flush = false, need_stop_flush = false;
>>> bool need_restart_discard = false, need_stop_discard = false;
>>> - bool no_extent_cache = !test_opt(sbi, EXTENT_CACHE);
>>> + bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE);
>>> bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT);
>>> bool no_io_align = !F2FS_IO_ALIGNED(sbi);
>>> bool no_atgc = !test_opt(sbi, ATGC);
>>> @@ -2308,7 +2308,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
>>> }
>>> /* disallow enable/disable extent_cache dynamically */
>>> - if (no_extent_cache == !!test_opt(sbi, EXTENT_CACHE)) {
>>> + if (no_read_extent_cache == !!test_opt(sbi, READ_EXTENT_CACHE)) {
>>> err = -EINVAL;
>>> f2fs_warn(sbi, "switch extent_cache option is not allowed");
>>> goto restore_opts;
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly
2022-12-09 1:31 ` Chao Yu
@ 2022-12-09 19:39 ` Jaegeuk Kim
0 siblings, 0 replies; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-09 19:39 UTC (permalink / raw)
To: Chao Yu; +Cc: linux-kernel, linux-f2fs-devel
On 12/09, Chao Yu wrote:
> On 2022/12/9 1:29, Jaegeuk Kim wrote:
> > On 12/08, Chao Yu wrote:
> > > On 2022/12/6 2:54, Jaegeuk Kim wrote:
> > > > Let's descrbie it's read extent cache.
> > > >
> > > > Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
> > > > ---
> > > > fs/f2fs/extent_cache.c | 4 ++--
> > > > fs/f2fs/f2fs.h | 10 +++++-----
> > > > fs/f2fs/inode.c | 2 +-
> > > > fs/f2fs/node.c | 2 +-
> > > > fs/f2fs/node.h | 2 +-
> > > > fs/f2fs/segment.c | 4 ++--
> > > > fs/f2fs/super.c | 16 ++++++++--------
> > > > 7 files changed, 20 insertions(+), 20 deletions(-)
> > > >
> > > > diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
> > > > index 932c070173b9..8cd87aee0292 100644
> > > > --- a/fs/f2fs/extent_cache.c
> > > > +++ b/fs/f2fs/extent_cache.c
> > > > @@ -383,7 +383,7 @@ static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage)
> > > > if (!i_ext || !i_ext->len)
> > > > return;
> > > > - get_extent_info(&ei, i_ext);
> > > > + get_read_extent_info(&ei, i_ext);
> > > > write_lock(&et->lock);
> > > > if (atomic_read(&et->node_cnt))
> > > > @@ -710,7 +710,7 @@ unsigned int f2fs_shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
> > > > unsigned int node_cnt = 0, tree_cnt = 0;
> > > > int remained;
> > > > - if (!test_opt(sbi, EXTENT_CACHE))
> > > > + if (!test_opt(sbi, READ_EXTENT_CACHE))
> > > > return 0;
> > > > if (!atomic_read(&sbi->total_zombie_tree))
> > > > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> > > > index eb8c27c4e5fc..1c39f8145b61 100644
> > > > --- a/fs/f2fs/f2fs.h
> > > > +++ b/fs/f2fs/f2fs.h
> > > > @@ -92,7 +92,7 @@ extern const char *f2fs_fault_name[FAULT_MAX];
> > > > #define F2FS_MOUNT_FLUSH_MERGE 0x00000400
> > > > #define F2FS_MOUNT_NOBARRIER 0x00000800
> > > > #define F2FS_MOUNT_FASTBOOT 0x00001000
> > > > -#define F2FS_MOUNT_EXTENT_CACHE 0x00002000
> > > > +#define F2FS_MOUNT_READ_EXTENT_CACHE 0x00002000
> > > > #define F2FS_MOUNT_DATA_FLUSH 0x00008000
> > > > #define F2FS_MOUNT_FAULT_INJECTION 0x00010000
> > > > #define F2FS_MOUNT_USRQUOTA 0x00080000
> > > > @@ -600,7 +600,7 @@ enum {
> > > > #define F2FS_MIN_EXTENT_LEN 64 /* minimum extent length */
> > > > /* number of extent info in extent cache we try to shrink */
> > > > -#define EXTENT_CACHE_SHRINK_NUMBER 128
> > > > +#define READ_EXTENT_CACHE_SHRINK_NUMBER 128
> > > > #define RECOVERY_MAX_RA_BLOCKS BIO_MAX_VECS
> > > > #define RECOVERY_MIN_RA_BLOCKS 1
> > > > @@ -830,7 +830,7 @@ struct f2fs_inode_info {
> > > > loff_t original_i_size; /* original i_size before atomic write */
> > > > };
> > > > -static inline void get_extent_info(struct extent_info *ext,
> > > > +static inline void get_read_extent_info(struct extent_info *ext,
> > > > struct f2fs_extent *i_ext)
> > > > {
> > > > ext->fofs = le32_to_cpu(i_ext->fofs);
> > > > @@ -838,7 +838,7 @@ static inline void get_extent_info(struct extent_info *ext,
> > > > ext->len = le32_to_cpu(i_ext->len);
> > > > }
> > > > -static inline void set_raw_extent(struct extent_info *ext,
> > > > +static inline void set_raw_read_extent(struct extent_info *ext,
> > > > struct f2fs_extent *i_ext)
> > > > {
> > > > i_ext->fofs = cpu_to_le32(ext->fofs);
> > > > @@ -4407,7 +4407,7 @@ static inline bool f2fs_may_extent_tree(struct inode *inode)
> > > > {
> > > > struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
> > > > - if (!test_opt(sbi, EXTENT_CACHE) ||
> > > > + if (!test_opt(sbi, READ_EXTENT_CACHE) ||
> > > > is_inode_flag_set(inode, FI_NO_EXTENT) ||
> > > > (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
> > > > !f2fs_sb_has_readonly(sbi)))
> > > > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> > > > index 577f109b4e1d..2c705c60019b 100644
> > > > --- a/fs/f2fs/inode.c
> > > > +++ b/fs/f2fs/inode.c
> > > > @@ -629,7 +629,7 @@ void f2fs_update_inode(struct inode *inode, struct page *node_page)
> > > > if (et) {
> > > > read_lock(&et->lock);
> > > > - set_raw_extent(&et->largest, &ri->i_ext);
> > > > + set_raw_read_extent(&et->largest, &ri->i_ext);
> > > > read_unlock(&et->lock);
> > > > } else {
> > > > memset(&ri->i_ext, 0, sizeof(ri->i_ext));
> > > > diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
> > > > index b9ee5a1176a0..84b147966080 100644
> > > > --- a/fs/f2fs/node.c
> > > > +++ b/fs/f2fs/node.c
> > > > @@ -85,7 +85,7 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
> > > > sizeof(struct ino_entry);
> > > > mem_size >>= PAGE_SHIFT;
> > > > res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
> > > > - } else if (type == EXTENT_CACHE) {
> > > > + } else if (type == READ_EXTENT_CACHE) {
> > > > mem_size = (atomic_read(&sbi->total_ext_tree) *
> > > > sizeof(struct extent_tree) +
> > > > atomic_read(&sbi->total_ext_node) *
> > > > diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
> > > > index 3c09cae058b0..0aa48704c77a 100644
> > > > --- a/fs/f2fs/node.h
> > > > +++ b/fs/f2fs/node.h
> > > > @@ -146,7 +146,7 @@ enum mem_type {
> > > > NAT_ENTRIES, /* indicates the cached nat entry */
> > > > DIRTY_DENTS, /* indicates dirty dentry pages */
> > > > INO_ENTRIES, /* indicates inode entries */
> > > > - EXTENT_CACHE, /* indicates extent cache */
> > > > + READ_EXTENT_CACHE, /* indicates read extent cache */
> > > > DISCARD_CACHE, /* indicates memory of cached discard cmds */
> > > > COMPRESS_PAGE, /* indicates memory of cached compressed pages */
> > > > BASE_CHECK, /* check kernel status */
> > > > diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
> > > > index 9486ca49ecb1..51de358bc452 100644
> > > > --- a/fs/f2fs/segment.c
> > > > +++ b/fs/f2fs/segment.c
> > > > @@ -449,8 +449,8 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
> > > > return;
> > > > /* try to shrink extent cache when there is no enough memory */
> > > > - if (!f2fs_available_free_memory(sbi, EXTENT_CACHE))
> > > > - f2fs_shrink_extent_tree(sbi, EXTENT_CACHE_SHRINK_NUMBER);
> > > > + if (!f2fs_available_free_memory(sbi, READ_EXTENT_CACHE))
> > > > + f2fs_shrink_extent_tree(sbi, READ_EXTENT_CACHE_SHRINK_NUMBER);
> > > > /* check the # of cached NAT entries */
> > > > if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
> > > > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> > > > index 79bf1faf4161..10bd03bbefec 100644
> > > > --- a/fs/f2fs/super.c
> > > > +++ b/fs/f2fs/super.c
> > > > @@ -814,10 +814,10 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
> > > > set_opt(sbi, FASTBOOT);
> > > > break;
> > > > case Opt_extent_cache:
> > > > - set_opt(sbi, EXTENT_CACHE);
> > > > + set_opt(sbi, READ_EXTENT_CACHE);
> > > > break;
> > > > case Opt_noextent_cache:
> > > > - clear_opt(sbi, EXTENT_CACHE);
> > > > + clear_opt(sbi, READ_EXTENT_CACHE);
> > > > break;
> > > > case Opt_noinline_data:
> > > > clear_opt(sbi, INLINE_DATA);
> > > > @@ -1954,10 +1954,10 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
> > > > seq_puts(seq, ",barrier");
> > > > if (test_opt(sbi, FASTBOOT))
> > > > seq_puts(seq, ",fastboot");
> > > > - if (test_opt(sbi, EXTENT_CACHE))
> > > > - seq_puts(seq, ",extent_cache");
> > > > + if (test_opt(sbi, READ_EXTENT_CACHE))
> > > > + seq_puts(seq, ",read_extent_cache");
> > >
> > > How about keeping consistent w/ description of f2fs.rst?
> >
> > This doesn't change the mount options, but do show the exact name which
> > would be better to say what's going on.
>
> mount -t f2fs /dev/vda /mnt/f2fs
> mount|grep f2fs
> /dev/vda on /mnt/f2fs type f2fs (rw,relatime,lazytime,background_gc=on,discard,no_heap,user_xattr,inline_xattr,acl,inline_data,inline_dentry,flush_merge,barrier,read_extent_cache,mode=adaptive,active_logs=6,alloc_mode=reuse,check)
>
> mount -t f2fs -o remount /mnt/f2fs/
> [ 63.142480] F2FS-fs (vda): Unrecognized mount option "read_extent_cache" or missing value
>
> Quoted from manual of mount
>
> "mount -o remount,rw /dir
>
> After this call, mount reads fstab and merges these options with the options
> from the command line (-o). If no mountpoint is found in fstab, then a
> remount with unspecified source is allowed."
>
> I guess above command tries to use previous mountoptions which were read
> via .show_options().
Oh, great. Let me fix this.
>
> Thanks,
>
> >
> > >
> > > > else
> > > > - seq_puts(seq, ",noextent_cache");
> > > > + seq_puts(seq, ",no_read_extent_cache");
> > >
> > > Ditto,
> > >
> > > Thanks,
> > >
> > > > if (test_opt(sbi, DATA_FLUSH))
> > > > seq_puts(seq, ",data_flush");
> > > > @@ -2076,7 +2076,7 @@ static void default_options(struct f2fs_sb_info *sbi)
> > > > set_opt(sbi, INLINE_XATTR);
> > > > set_opt(sbi, INLINE_DATA);
> > > > set_opt(sbi, INLINE_DENTRY);
> > > > - set_opt(sbi, EXTENT_CACHE);
> > > > + set_opt(sbi, READ_EXTENT_CACHE);
> > > > set_opt(sbi, NOHEAP);
> > > > clear_opt(sbi, DISABLE_CHECKPOINT);
> > > > set_opt(sbi, MERGE_CHECKPOINT);
> > > > @@ -2218,7 +2218,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
> > > > bool need_restart_ckpt = false, need_stop_ckpt = false;
> > > > bool need_restart_flush = false, need_stop_flush = false;
> > > > bool need_restart_discard = false, need_stop_discard = false;
> > > > - bool no_extent_cache = !test_opt(sbi, EXTENT_CACHE);
> > > > + bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE);
> > > > bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT);
> > > > bool no_io_align = !F2FS_IO_ALIGNED(sbi);
> > > > bool no_atgc = !test_opt(sbi, ATGC);
> > > > @@ -2308,7 +2308,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
> > > > }
> > > > /* disallow enable/disable extent_cache dynamically */
> > > > - if (no_extent_cache == !!test_opt(sbi, EXTENT_CACHE)) {
> > > > + if (no_read_extent_cache == !!test_opt(sbi, READ_EXTENT_CACHE)) {
> > > > err = -EINVAL;
> > > > f2fs_warn(sbi, "switch extent_cache option is not allowed");
> > > > goto restore_opts;
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [f2fs-dev] [PATCH 1/6 v2] f2fs: specify extent cache for read explicitly
2022-12-08 14:43 ` [f2fs-dev] [PATCH 1/6] f2fs: specify extent cache for read explicitly Chao Yu
2022-12-08 17:29 ` Jaegeuk Kim
@ 2022-12-09 19:40 ` Jaegeuk Kim
2022-12-11 1:59 ` Chao Yu
1 sibling, 1 reply; 19+ messages in thread
From: Jaegeuk Kim @ 2022-12-09 19:40 UTC (permalink / raw)
To: Chao Yu; +Cc: linux-kernel, linux-f2fs-devel
Let's descrbie it's read extent cache.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
Change log from v1:
- keep showing the mount option
fs/f2fs/extent_cache.c | 4 ++--
fs/f2fs/f2fs.h | 10 +++++-----
fs/f2fs/inode.c | 2 +-
fs/f2fs/node.c | 2 +-
fs/f2fs/node.h | 2 +-
fs/f2fs/segment.c | 4 ++--
fs/f2fs/super.c | 12 ++++++------
7 files changed, 18 insertions(+), 18 deletions(-)
diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
index 932c070173b9..8cd87aee0292 100644
--- a/fs/f2fs/extent_cache.c
+++ b/fs/f2fs/extent_cache.c
@@ -383,7 +383,7 @@ static void __f2fs_init_extent_tree(struct inode *inode, struct page *ipage)
if (!i_ext || !i_ext->len)
return;
- get_extent_info(&ei, i_ext);
+ get_read_extent_info(&ei, i_ext);
write_lock(&et->lock);
if (atomic_read(&et->node_cnt))
@@ -710,7 +710,7 @@ unsigned int f2fs_shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
unsigned int node_cnt = 0, tree_cnt = 0;
int remained;
- if (!test_opt(sbi, EXTENT_CACHE))
+ if (!test_opt(sbi, READ_EXTENT_CACHE))
return 0;
if (!atomic_read(&sbi->total_zombie_tree))
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index eb8c27c4e5fc..1c39f8145b61 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -92,7 +92,7 @@ extern const char *f2fs_fault_name[FAULT_MAX];
#define F2FS_MOUNT_FLUSH_MERGE 0x00000400
#define F2FS_MOUNT_NOBARRIER 0x00000800
#define F2FS_MOUNT_FASTBOOT 0x00001000
-#define F2FS_MOUNT_EXTENT_CACHE 0x00002000
+#define F2FS_MOUNT_READ_EXTENT_CACHE 0x00002000
#define F2FS_MOUNT_DATA_FLUSH 0x00008000
#define F2FS_MOUNT_FAULT_INJECTION 0x00010000
#define F2FS_MOUNT_USRQUOTA 0x00080000
@@ -600,7 +600,7 @@ enum {
#define F2FS_MIN_EXTENT_LEN 64 /* minimum extent length */
/* number of extent info in extent cache we try to shrink */
-#define EXTENT_CACHE_SHRINK_NUMBER 128
+#define READ_EXTENT_CACHE_SHRINK_NUMBER 128
#define RECOVERY_MAX_RA_BLOCKS BIO_MAX_VECS
#define RECOVERY_MIN_RA_BLOCKS 1
@@ -830,7 +830,7 @@ struct f2fs_inode_info {
loff_t original_i_size; /* original i_size before atomic write */
};
-static inline void get_extent_info(struct extent_info *ext,
+static inline void get_read_extent_info(struct extent_info *ext,
struct f2fs_extent *i_ext)
{
ext->fofs = le32_to_cpu(i_ext->fofs);
@@ -838,7 +838,7 @@ static inline void get_extent_info(struct extent_info *ext,
ext->len = le32_to_cpu(i_ext->len);
}
-static inline void set_raw_extent(struct extent_info *ext,
+static inline void set_raw_read_extent(struct extent_info *ext,
struct f2fs_extent *i_ext)
{
i_ext->fofs = cpu_to_le32(ext->fofs);
@@ -4407,7 +4407,7 @@ static inline bool f2fs_may_extent_tree(struct inode *inode)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
- if (!test_opt(sbi, EXTENT_CACHE) ||
+ if (!test_opt(sbi, READ_EXTENT_CACHE) ||
is_inode_flag_set(inode, FI_NO_EXTENT) ||
(is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
!f2fs_sb_has_readonly(sbi)))
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index 577f109b4e1d..2c705c60019b 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -629,7 +629,7 @@ void f2fs_update_inode(struct inode *inode, struct page *node_page)
if (et) {
read_lock(&et->lock);
- set_raw_extent(&et->largest, &ri->i_ext);
+ set_raw_read_extent(&et->largest, &ri->i_ext);
read_unlock(&et->lock);
} else {
memset(&ri->i_ext, 0, sizeof(ri->i_ext));
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index b9ee5a1176a0..84b147966080 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -85,7 +85,7 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
sizeof(struct ino_entry);
mem_size >>= PAGE_SHIFT;
res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
- } else if (type == EXTENT_CACHE) {
+ } else if (type == READ_EXTENT_CACHE) {
mem_size = (atomic_read(&sbi->total_ext_tree) *
sizeof(struct extent_tree) +
atomic_read(&sbi->total_ext_node) *
diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
index 3c09cae058b0..0aa48704c77a 100644
--- a/fs/f2fs/node.h
+++ b/fs/f2fs/node.h
@@ -146,7 +146,7 @@ enum mem_type {
NAT_ENTRIES, /* indicates the cached nat entry */
DIRTY_DENTS, /* indicates dirty dentry pages */
INO_ENTRIES, /* indicates inode entries */
- EXTENT_CACHE, /* indicates extent cache */
+ READ_EXTENT_CACHE, /* indicates read extent cache */
DISCARD_CACHE, /* indicates memory of cached discard cmds */
COMPRESS_PAGE, /* indicates memory of cached compressed pages */
BASE_CHECK, /* check kernel status */
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 9486ca49ecb1..51de358bc452 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -449,8 +449,8 @@ void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg)
return;
/* try to shrink extent cache when there is no enough memory */
- if (!f2fs_available_free_memory(sbi, EXTENT_CACHE))
- f2fs_shrink_extent_tree(sbi, EXTENT_CACHE_SHRINK_NUMBER);
+ if (!f2fs_available_free_memory(sbi, READ_EXTENT_CACHE))
+ f2fs_shrink_extent_tree(sbi, READ_EXTENT_CACHE_SHRINK_NUMBER);
/* check the # of cached NAT entries */
if (!f2fs_available_free_memory(sbi, NAT_ENTRIES))
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 79bf1faf4161..412c2e7352c0 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -814,10 +814,10 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount)
set_opt(sbi, FASTBOOT);
break;
case Opt_extent_cache:
- set_opt(sbi, EXTENT_CACHE);
+ set_opt(sbi, READ_EXTENT_CACHE);
break;
case Opt_noextent_cache:
- clear_opt(sbi, EXTENT_CACHE);
+ clear_opt(sbi, READ_EXTENT_CACHE);
break;
case Opt_noinline_data:
clear_opt(sbi, INLINE_DATA);
@@ -1954,7 +1954,7 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
seq_puts(seq, ",barrier");
if (test_opt(sbi, FASTBOOT))
seq_puts(seq, ",fastboot");
- if (test_opt(sbi, EXTENT_CACHE))
+ if (test_opt(sbi, READ_EXTENT_CACHE))
seq_puts(seq, ",extent_cache");
else
seq_puts(seq, ",noextent_cache");
@@ -2076,7 +2076,7 @@ static void default_options(struct f2fs_sb_info *sbi)
set_opt(sbi, INLINE_XATTR);
set_opt(sbi, INLINE_DATA);
set_opt(sbi, INLINE_DENTRY);
- set_opt(sbi, EXTENT_CACHE);
+ set_opt(sbi, READ_EXTENT_CACHE);
set_opt(sbi, NOHEAP);
clear_opt(sbi, DISABLE_CHECKPOINT);
set_opt(sbi, MERGE_CHECKPOINT);
@@ -2218,7 +2218,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
bool need_restart_ckpt = false, need_stop_ckpt = false;
bool need_restart_flush = false, need_stop_flush = false;
bool need_restart_discard = false, need_stop_discard = false;
- bool no_extent_cache = !test_opt(sbi, EXTENT_CACHE);
+ bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE);
bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT);
bool no_io_align = !F2FS_IO_ALIGNED(sbi);
bool no_atgc = !test_opt(sbi, ATGC);
@@ -2308,7 +2308,7 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
}
/* disallow enable/disable extent_cache dynamically */
- if (no_extent_cache == !!test_opt(sbi, EXTENT_CACHE)) {
+ if (no_read_extent_cache == !!test_opt(sbi, READ_EXTENT_CACHE)) {
err = -EINVAL;
f2fs_warn(sbi, "switch extent_cache option is not allowed");
goto restore_opts;
--
2.39.0.rc1.256.g54fd8350bd-goog
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
^ permalink raw reply related [flat|nested] 19+ messages in thread