[PATCH 1/2] f2fs: avoid down_write on nat_tree_lock during checkpoint

[PATCH 1/2] f2fs: avoid down_write on nat_tree_lock during checkpoint

[Jaegeuk Kim]

Let's cache nat entry if there's no lock contention only.

Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
 fs/f2fs/node.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 556fcd8457f3..b1bc7d76da3b 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -430,6 +430,10 @@ static void cache_nat_entry(struct f2fs_sb_info *sbi, nid_t nid,
 	struct f2fs_nm_info *nm_i = NM_I(sbi);
 	struct nat_entry *new, *e;
 
+	/* Let's mitigate lock contention of nat_tree_lock during checkpoint */
+	if (rwsem_is_locked(&sbi->cp_global_sem))
+		return;
+
 	new = __alloc_nat_entry(sbi, nid, false);
 	if (!new)
 		return;
-- 
2.34.1.173.g76aa8bc2d0-goog

[PATCH 2/2] f2fs: do not bother checkpoint by f2fs_get_node_info

[Jaegeuk Kim]

This patch tries to mitigate lock contention between f2fs_write_checkpoint and
f2fs_get_node_info along with nat_tree_lock.

The idea is, if checkpoint is currently running, other threads that try to grab
nat_tree_lock would be better to wait for checkpoint.

Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
 fs/f2fs/checkpoint.c |  2 +-
 fs/f2fs/compress.c   |  2 +-
 fs/f2fs/data.c       |  8 ++++----
 fs/f2fs/f2fs.h       |  2 +-
 fs/f2fs/file.c       |  2 +-
 fs/f2fs/gc.c         |  6 +++---
 fs/f2fs/inline.c     |  4 ++--
 fs/f2fs/inode.c      |  2 +-
 fs/f2fs/node.c       | 19 ++++++++++---------
 fs/f2fs/recovery.c   |  2 +-
 fs/f2fs/segment.c    |  2 +-
 11 files changed, 26 insertions(+), 25 deletions(-)

diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c
index f1693d45bb78..55e3c54d99c1 100644
--- a/fs/f2fs/checkpoint.c
+++ b/fs/f2fs/checkpoint.c
@@ -664,7 +664,7 @@ static int recover_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino)
 	/* truncate all the data during iput */
 	iput(inode);
 
-	err = f2fs_get_node_info(sbi, ino, &ni);
+	err = f2fs_get_node_info(sbi, ino, &ni, false);
 	if (err)
 		goto err_out;
 
diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c
index 4b49038d150d..d0c3aeba5945 100644
--- a/fs/f2fs/compress.c
+++ b/fs/f2fs/compress.c
@@ -1286,7 +1286,7 @@ static int f2fs_write_compressed_pages(struct compress_ctx *cc,
 
 	psize = (loff_t)(cc->rpages[last_index]->index + 1) << PAGE_SHIFT;
 
-	err = f2fs_get_node_info(fio.sbi, dn.nid, &ni);
+	err = f2fs_get_node_info(fio.sbi, dn.nid, &ni, false);
 	if (err)
 		goto out_put_dnode;
 
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 40c0d1426a11..2737fcc0dbcf 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -1355,7 +1355,7 @@ static int __allocate_data_block(struct dnode_of_data *dn, int seg_type)
 	if (unlikely(is_inode_flag_set(dn->inode, FI_NO_ALLOC)))
 		return -EPERM;
 
-	err = f2fs_get_node_info(sbi, dn->nid, &ni);
+	err = f2fs_get_node_info(sbi, dn->nid, &ni, false);
 	if (err)
 		return err;
 
@@ -1757,7 +1757,7 @@ static int f2fs_xattr_fiemap(struct inode *inode,
 		if (!page)
 			return -ENOMEM;
 
-		err = f2fs_get_node_info(sbi, inode->i_ino, &ni);
+		err = f2fs_get_node_info(sbi, inode->i_ino, &ni, false);
 		if (err) {
 			f2fs_put_page(page, 1);
 			return err;
@@ -1789,7 +1789,7 @@ static int f2fs_xattr_fiemap(struct inode *inode,
 		if (!page)
 			return -ENOMEM;
 
-		err = f2fs_get_node_info(sbi, xnid, &ni);
+		err = f2fs_get_node_info(sbi, xnid, &ni, false);
 		if (err) {
 			f2fs_put_page(page, 1);
 			return err;
@@ -2649,7 +2649,7 @@ int f2fs_do_write_data_page(struct f2fs_io_info *fio)
 		fio->need_lock = LOCK_REQ;
 	}
 
-	err = f2fs_get_node_info(fio->sbi, dn.nid, &ni);
+	err = f2fs_get_node_info(fio->sbi, dn.nid, &ni, false);
 	if (err)
 		goto out_writepage;
 
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 5da592286721..a9ed2fd3fffb 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -3412,7 +3412,7 @@ int f2fs_need_dentry_mark(struct f2fs_sb_info *sbi, nid_t nid);
 bool f2fs_is_checkpointed_node(struct f2fs_sb_info *sbi, nid_t nid);
 bool f2fs_need_inode_block_update(struct f2fs_sb_info *sbi, nid_t ino);
 int f2fs_get_node_info(struct f2fs_sb_info *sbi, nid_t nid,
-						struct node_info *ni);
+				struct node_info *ni, bool checkpoint_context);
 pgoff_t f2fs_get_next_page_offset(struct dnode_of_data *dn, pgoff_t pgofs);
 int f2fs_get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode);
 int f2fs_truncate_inode_blocks(struct inode *inode, pgoff_t from);
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index 5ec6bef3937f..f540c1cbddca 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -1233,7 +1233,7 @@ static int __clone_blkaddrs(struct inode *src_inode, struct inode *dst_inode,
 			if (ret)
 				return ret;
 
-			ret = f2fs_get_node_info(sbi, dn.nid, &ni);
+			ret = f2fs_get_node_info(sbi, dn.nid, &ni, false);
 			if (ret) {
 				f2fs_put_dnode(&dn);
 				return ret;
diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
index 7fbe46477a5a..a6accec60d04 100644
--- a/fs/f2fs/gc.c
+++ b/fs/f2fs/gc.c
@@ -959,7 +959,7 @@ static int gc_node_segment(struct f2fs_sb_info *sbi,
 			continue;
 		}
 
-		if (f2fs_get_node_info(sbi, nid, &ni)) {
+		if (f2fs_get_node_info(sbi, nid, &ni, false)) {
 			f2fs_put_page(node_page, 1);
 			continue;
 		}
@@ -1027,7 +1027,7 @@ static bool is_alive(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
 	if (IS_ERR(node_page))
 		return false;
 
-	if (f2fs_get_node_info(sbi, nid, dni)) {
+	if (f2fs_get_node_info(sbi, nid, dni, false)) {
 		f2fs_put_page(node_page, 1);
 		return false;
 	}
@@ -1221,7 +1221,7 @@ static int move_data_block(struct inode *inode, block_t bidx,
 
 	f2fs_wait_on_block_writeback(inode, dn.data_blkaddr);
 
-	err = f2fs_get_node_info(fio.sbi, dn.nid, &ni);
+	err = f2fs_get_node_info(fio.sbi, dn.nid, &ni, false);
 	if (err)
 		goto put_out;
 
diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c
index ea08f0dfa1bd..4b5cefa3f90c 100644
--- a/fs/f2fs/inline.c
+++ b/fs/f2fs/inline.c
@@ -131,7 +131,7 @@ int f2fs_convert_inline_page(struct dnode_of_data *dn, struct page *page)
 	if (err)
 		return err;
 
-	err = f2fs_get_node_info(fio.sbi, dn->nid, &ni);
+	err = f2fs_get_node_info(fio.sbi, dn->nid, &ni, false);
 	if (err) {
 		f2fs_truncate_data_blocks_range(dn, 1);
 		f2fs_put_dnode(dn);
@@ -786,7 +786,7 @@ int f2fs_inline_data_fiemap(struct inode *inode,
 		ilen = start + len;
 	ilen -= start;
 
-	err = f2fs_get_node_info(F2FS_I_SB(inode), inode->i_ino, &ni);
+	err = f2fs_get_node_info(F2FS_I_SB(inode), inode->i_ino, &ni, false);
 	if (err)
 		goto out;
 
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index 935016e56010..2ab3b424735a 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -881,7 +881,7 @@ void f2fs_handle_failed_inode(struct inode *inode)
 	 * so we can prevent losing this orphan when encoutering checkpoint
 	 * and following suddenly power-off.
 	 */
-	err = f2fs_get_node_info(sbi, inode->i_ino, &ni);
+	err = f2fs_get_node_info(sbi, inode->i_ino, &ni, false);
 	if (err) {
 		set_sbi_flag(sbi, SBI_NEED_FSCK);
 		f2fs_warn(sbi, "May loss orphan inode, run fsck to fix.");
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index b1bc7d76da3b..e0b5eb28d383 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -543,7 +543,7 @@ int f2fs_try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink)
 }
 
 int f2fs_get_node_info(struct f2fs_sb_info *sbi, nid_t nid,
-						struct node_info *ni)
+				struct node_info *ni, bool checkpoint_context)
 {
 	struct f2fs_nm_info *nm_i = NM_I(sbi);
 	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
@@ -576,9 +576,10 @@ int f2fs_get_node_info(struct f2fs_sb_info *sbi, nid_t nid,
 	 * nat_tree_lock. Therefore, we should retry, if we failed to grab here
 	 * while not bothering checkpoint.
 	 */
-	if (!rwsem_is_locked(&sbi->cp_global_sem)) {
+	if (!rwsem_is_locked(&sbi->cp_global_sem) || checkpoint_context) {
 		down_read(&curseg->journal_rwsem);
-	} else if (!down_read_trylock(&curseg->journal_rwsem)) {
+	} else if (rwsem_is_contended(&nm_i->nat_tree_lock) ||
+				!down_read_trylock(&curseg->journal_rwsem)) {
 		up_read(&nm_i->nat_tree_lock);
 		goto retry;
 	}
@@ -891,7 +892,7 @@ static int truncate_node(struct dnode_of_data *dn)
 	int err;
 	pgoff_t index;
 
-	err = f2fs_get_node_info(sbi, dn->nid, &ni);
+	err = f2fs_get_node_info(sbi, dn->nid, &ni, false);
 	if (err)
 		return err;
 
@@ -1290,7 +1291,7 @@ struct page *f2fs_new_node_page(struct dnode_of_data *dn, unsigned int ofs)
 		goto fail;
 
 #ifdef CONFIG_F2FS_CHECK_FS
-	err = f2fs_get_node_info(sbi, dn->nid, &new_ni);
+	err = f2fs_get_node_info(sbi, dn->nid, &new_ni, false);
 	if (err) {
 		dec_valid_node_count(sbi, dn->inode, !ofs);
 		goto fail;
@@ -1352,7 +1353,7 @@ static int read_node_page(struct page *page, int op_flags)
 		return LOCKED_PAGE;
 	}
 
-	err = f2fs_get_node_info(sbi, page->index, &ni);
+	err = f2fs_get_node_info(sbi, page->index, &ni, false);
 	if (err)
 		return err;
 
@@ -1604,7 +1605,7 @@ static int __write_node_page(struct page *page, bool atomic, bool *submitted,
 	nid = nid_of_node(page);
 	f2fs_bug_on(sbi, page->index != nid);
 
-	if (f2fs_get_node_info(sbi, nid, &ni))
+	if (f2fs_get_node_info(sbi, nid, &ni, !do_balance))
 		goto redirty_out;
 
 	if (wbc->for_reclaim) {
@@ -2705,7 +2706,7 @@ int f2fs_recover_xattr_data(struct inode *inode, struct page *page)
 		goto recover_xnid;
 
 	/* 1: invalidate the previous xattr nid */
-	err = f2fs_get_node_info(sbi, prev_xnid, &ni);
+	err = f2fs_get_node_info(sbi, prev_xnid, &ni, false);
 	if (err)
 		return err;
 
@@ -2745,7 +2746,7 @@ int f2fs_recover_inode_page(struct f2fs_sb_info *sbi, struct page *page)
 	struct page *ipage;
 	int err;
 
-	err = f2fs_get_node_info(sbi, ino, &old_ni);
+	err = f2fs_get_node_info(sbi, ino, &old_ni, false);
 	if (err)
 		return err;
 
diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c
index 6a1b4668d933..e65c73c4411d 100644
--- a/fs/f2fs/recovery.c
+++ b/fs/f2fs/recovery.c
@@ -595,7 +595,7 @@ static int do_recover_data(struct f2fs_sb_info *sbi, struct inode *inode,
 
 	f2fs_wait_on_page_writeback(dn.node_page, NODE, true, true);
 
-	err = f2fs_get_node_info(sbi, dn.nid, &ni);
+	err = f2fs_get_node_info(sbi, dn.nid, &ni, false);
 	if (err)
 		goto err;
 
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index df9ed75f0b7a..b4a2f8c36149 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -254,7 +254,7 @@ static int __revoke_inmem_pages(struct inode *inode,
 				goto next;
 			}
 
-			err = f2fs_get_node_info(sbi, dn.nid, &ni);
+			err = f2fs_get_node_info(sbi, dn.nid, &ni, false);
 			if (err) {
 				f2fs_put_dnode(&dn);
 				return err;
-- 
2.34.1.173.g76aa8bc2d0-goog

Re: [f2fs-dev] [PATCH 2/2] f2fs: do not bother checkpoint by f2fs_get_node_info

[Chao Yu]

On 2021/12/15 2:24, Jaegeuk Kim wrote:
> This patch tries to mitigate lock contention between f2fs_write_checkpoint and
> f2fs_get_node_info along with nat_tree_lock.
> 
> The idea is, if checkpoint is currently running, other threads that try to grab
> nat_tree_lock would be better to wait for checkpoint.
> 
> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>

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

Thanks,

Re: [f2fs-dev] [PATCH 1/2] f2fs: avoid down_write on nat_tree_lock during checkpoint

[Chao Yu]

On 2021/12/15 2:24, Jaegeuk Kim wrote:
> Let's cache nat entry if there's no lock contention only.
> 
> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
> ---
>   fs/f2fs/node.c | 4 ++++
>   1 file changed, 4 insertions(+)
> 
> diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
> index 556fcd8457f3..b1bc7d76da3b 100644
> --- a/fs/f2fs/node.c
> +++ b/fs/f2fs/node.c
> @@ -430,6 +430,10 @@ static void cache_nat_entry(struct f2fs_sb_info *sbi, nid_t nid,
>   	struct f2fs_nm_info *nm_i = NM_I(sbi);
>   	struct nat_entry *new, *e;
>   
> +	/* Let's mitigate lock contention of nat_tree_lock during checkpoint */
> +	if (rwsem_is_locked(&sbi->cp_global_sem))

Why not down_write_trylock(nat_tree_lock)? cp_global_sem lock coverage is larger than
nat_tree_lock's in f2fs_write_checkpoint().

Thanks,

> +		return;
> +
>   	new = __alloc_nat_entry(sbi, nid, false);
>   	if (!new)
>   		return;

Re: [f2fs-dev] [PATCH 1/2] f2fs: avoid down_write on nat_tree_lock during checkpoint

[Jaegeuk Kim]

On 12/19, Chao Yu wrote:
> On 2021/12/15 2:24, Jaegeuk Kim wrote:
> > Let's cache nat entry if there's no lock contention only.
> > 
> > Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
> > ---
> >   fs/f2fs/node.c | 4 ++++
> >   1 file changed, 4 insertions(+)
> > 
> > diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
> > index 556fcd8457f3..b1bc7d76da3b 100644
> > --- a/fs/f2fs/node.c
> > +++ b/fs/f2fs/node.c
> > @@ -430,6 +430,10 @@ static void cache_nat_entry(struct f2fs_sb_info *sbi, nid_t nid,
> >   	struct f2fs_nm_info *nm_i = NM_I(sbi);
> >   	struct nat_entry *new, *e;
> > +	/* Let's mitigate lock contention of nat_tree_lock during checkpoint */
> > +	if (rwsem_is_locked(&sbi->cp_global_sem))
> 
> Why not down_write_trylock(nat_tree_lock)? cp_global_sem lock coverage is larger than
> nat_tree_lock's in f2fs_write_checkpoint().

I'm focusing on faster checkpoint by minimizing the nat_tree_lock holding time.

> 
> Thanks,
> 
> > +		return;
> > +
> >   	new = __alloc_nat_entry(sbi, nid, false);
> >   	if (!new)
> >   		return;

Re: [f2fs-dev] [PATCH 1/2] f2fs: avoid down_write on nat_tree_lock during checkpoint

[Chao Yu]

On 2021/12/15 2:24, Jaegeuk Kim wrote:
> Let's cache nat entry if there's no lock contention only.
> 
> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>

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

Thanks,