[PATCH] f2fs: Avoid double lock for cp_rwsem during checkpoint

From: Sayali Lokhande
Date: Thu Apr 30 2020 - 06:58:29 EST


There could be a scenario where f2fs_sync_node_pages gets
called during checkpoint, which in turn tries to flush
inline data and calls iput(). This results in deadlock as
iput() tries to hold cp_rwsem, which is already held at the
beginning by checkpoint->block_operations().

Call stack :

Thread A Thread B
f2fs_write_checkpoint()
- block_operations(sbi)
- f2fs_lock_all(sbi);
- down_write(&sbi->cp_rwsem);

- open()
- igrab()
- write() write inline data
- unlink()
- f2fs_sync_node_pages()
- if (is_inline_node(page))
- flush_inline_data()
- ilookup()
page = f2fs_pagecache_get_page()
if (!page)
goto iput_out;
iput_out:
-close()
-iput()
iput(inode);
- f2fs_evict_inode()
- f2fs_truncate_blocks()
- f2fs_lock_op()
- down_read(&sbi->cp_rwsem);

Fixes: 2049d4fcb057 ("f2fs: avoid multiple node page writes due to inline_data")
Signed-off-by: Sayali Lokhande <sayalil@xxxxxxxxxxxxxx>
Signed-off-by: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
---
fs/f2fs/checkpoint.c | 9 ++++++++-
fs/f2fs/f2fs.h | 4 ++--
fs/f2fs/node.c | 10 +++++-----
3 files changed, 15 insertions(+), 8 deletions(-)

diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c
index d49f7a01d8a26..928aea4ff663d 100644
--- a/fs/f2fs/checkpoint.c
+++ b/fs/f2fs/checkpoint.c
@@ -1168,6 +1168,12 @@ static int block_operations(struct f2fs_sb_info *sbi)
};
int err = 0, cnt = 0;

+ /*
+ * Let's flush node pages first to flush inline_data.
+ * We'll actually guarantee everything below under f2fs_lock_all.
+ */
+ f2fs_sync_node_pages(sbi, &wbc, false, false, FS_CP_NODE_IO);
+
retry_flush_quotas:
f2fs_lock_all(sbi);
if (__need_flush_quota(sbi)) {
@@ -1222,7 +1228,8 @@ static int block_operations(struct f2fs_sb_info *sbi)
if (get_pages(sbi, F2FS_DIRTY_NODES)) {
up_write(&sbi->node_write);
atomic_inc(&sbi->wb_sync_req[NODE]);
- err = f2fs_sync_node_pages(sbi, &wbc, false, FS_CP_NODE_IO);
+ err = f2fs_sync_node_pages(sbi, &wbc, false,
+ true, FS_CP_NODE_IO);
atomic_dec(&sbi->wb_sync_req[NODE]);
if (err) {
up_write(&sbi->node_change);
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index d916540f12813..ac6ae42b9dd4e 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -3286,8 +3286,8 @@ int f2fs_fsync_node_pages(struct f2fs_sb_info *sbi, struct inode *inode,
struct writeback_control *wbc, bool atomic,
unsigned int *seq_id);
int f2fs_sync_node_pages(struct f2fs_sb_info *sbi,
- struct writeback_control *wbc,
- bool do_balance, enum iostat_type io_type);
+ struct writeback_control *wbc,
+ bool do_balance, bool sync, enum iostat_type io_type);
int f2fs_build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount);
bool f2fs_alloc_nid(struct f2fs_sb_info *sbi, nid_t *nid);
void f2fs_alloc_nid_done(struct f2fs_sb_info *sbi, nid_t nid);
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 1db8cabf727ef..fd00a8c119088 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -1808,8 +1808,8 @@ static bool flush_dirty_inode(struct page *page)
}

int f2fs_sync_node_pages(struct f2fs_sb_info *sbi,
- struct writeback_control *wbc,
- bool do_balance, enum iostat_type io_type)
+ struct writeback_control *wbc,
+ bool do_balance, bool sync, enum iostat_type io_type)
{
pgoff_t index;
struct pagevec pvec;
@@ -1870,8 +1870,8 @@ int f2fs_sync_node_pages(struct f2fs_sb_info *sbi,
goto continue_unlock;
}

- /* flush inline_data */
- if (is_inline_node(page)) {
+ /* flush inline_data, if it's async context. */
+ if (!sync && is_inline_node(page)) {
clear_inline_node(page);
unlock_page(page);
flush_inline_data(sbi, ino_of_node(page));
@@ -1999,7 +1999,7 @@ static int f2fs_write_node_pages(struct address_space *mapping,

diff = nr_pages_to_write(sbi, NODE, wbc);
blk_start_plug(&plug);
- f2fs_sync_node_pages(sbi, wbc, true, FS_NODE_IO);
+ f2fs_sync_node_pages(sbi, wbc, true, false, FS_NODE_IO);
blk_finish_plug(&plug);
wbc->nr_to_write = max((long)0, wbc->nr_to_write - diff);

--
2.26.2.645.ge9eca65c58-goog