Skip to contentSkip to conversation

f2fs: Fix deadlock between f2fs_quota_sync and block_operation

PATCH In review Shachar Raindel 2 participants 2 messages Nov 28, 2020 – Nov 30, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 f2fs: Fix deadlock between f2fs_quota_sync and block_operation

This deadlock is hitting Android users (Pixel 3/3a/4) with Magisk, due to frequent umount/mount operations that trigger quota_sync, hitting the race. See https://github.com/topjohnwu/Magisk/issues/3171 for additional impact discussion.

In commit db6ec53b7e03, we added a semaphore to protect quota flags. As part of this commit, we changed f2fs_quota_sync to call f2fs_lock_op, in an attempt to prevent an AB/BA type deadlock with quota_sem locking in block_operation. However, rwsem in Linux is not recursive. Therefore, the following deadlock can occur:

f2fs_quota_sync
down_read(cp_rwsem) // f2fs_lock_op
filemap_fdatawrite
f2fs_write_data_pages
...
block_opertaion down_write(cp_rwsem) - marks rwsem as
"writer pending"
down_read_trylock(cp_rwsem) - fails as there is
a writer pending.
Code keeps on trying, live-locking the filesystem.

We solve this by creating a new rwsem, used specifically to synchronize this case, instead of attempting to reuse an existing lock.

Fixes db6ec53b7e03 f2fs: add a rw_sem to cover quota flag changes
---
 fs/f2fs/f2fs.h  |  3 +++
 fs/f2fs/super.c | 13 +++++++++++--
 2 files changed, 14 insertions(+), 2 deletions(-)
fs/f2fs/f2fs.h +3 -0
index cb700d797296..b3e55137be7f 100644
@@ -1448,6 +1448,7 @@ struct f2fs_sb_info {
 	struct inode *meta_inode;		/* cache meta blocks */
 	struct mutex cp_mutex;			/* checkpoint procedure lock */
 	struct rw_semaphore cp_rwsem;		/* blocking FS operations */
+	struct rw_semaphore cp_quota_rwsem;    	/* blocking quota sync operations */
 	struct rw_semaphore node_write;		/* locking node writes */
 	struct rw_semaphore node_change;	/* locking node change */
 	wait_queue_head_t cp_wait;
@@ -1961,12 +1962,14 @@ static inline void f2fs_unlock_op(struct f2fs_sb_info *sbi)
 
 static inline void f2fs_lock_all(struct f2fs_sb_info *sbi)
 {
+	down_write(&sbi->cp_quota_rwsem);
 	down_write(&sbi->cp_rwsem);
 }
 
 static inline void f2fs_unlock_all(struct f2fs_sb_info *sbi)
 {
 	up_write(&sbi->cp_rwsem);
+	up_write(&sbi->cp_quota_rwsem);
 }
 
 static inline int __get_cp_reason(struct f2fs_sb_info *sbi)
fs/f2fs/super.c +11 -2
index 00eff2f51807..5ce61147d7e5 100644
@@ -2209,8 +2209,16 @@ int f2fs_quota_sync(struct super_block *sb, int type)
 	 *  f2fs_dquot_commit
 	 *                            block_operation
 	 *                            down_read(quota_sem)
+	 *
+	 * However, we cannot use the cp_rwsem to prevent this
+	 * deadlock, as the cp_rwsem is taken for read inside the
+	 * f2fs_dquot_commit code, and rwsem is not recursive.
+	 *
+	 * We therefore use a special lock to synchronize
+	 * f2fs_quota_sync with block_operations, as this is the only
+	 * place where such recursion occurs.
 	 */
-	f2fs_lock_op(sbi);
+	down_read(&sbi->cp_quota_rwsem);
 
 	down_read(&sbi->quota_sem);
 	ret = dquot_writeback_dquots(sb, type);
@@ -2251,7 +2259,7 @@ int f2fs_quota_sync(struct super_block *sb, int type)
 	if (ret)
 		set_sbi_flag(F2FS_SB(sb), SBI_QUOTA_NEED_REPAIR);
 	up_read(&sbi->quota_sem);
-	f2fs_unlock_op(sbi);
+	up_read(&sbi->cp_quota_rwsem);
 	return ret;
 }
 
@@ -3599,6 +3607,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 
 	init_rwsem(&sbi->cp_rwsem);
 	init_rwsem(&sbi->quota_sem);
+	init_rwsem(&sbi->cp_quota_rwsem);
 	init_waitqueue_head(&sbi->cp_wait);
 	init_sb_info(sbi);
 
Signature · 8 lines
2.29.2



_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
#1 lore raw
Quoted 11 lines · Shachar Raindel

This deadlock is hitting Android users (Pixel 3/3a/4) with Magisk, due to frequent umount/mount operations that trigger quota_sync, hitting the race. See https://github.com/topjohnwu/Magisk/issues/3171 for additional impact discussion.

In commit db6ec53b7e03, we added a semaphore to protect quota flags. As part of this commit, we changed f2fs_quota_sync to call f2fs_lock_op, in an attempt to prevent an AB/BA type deadlock with quota_sem locking in block_operation. However, rwsem in Linux is not recursive. Therefore, the following deadlock can occur:

Quoted 12 lines

f2fs_quota_sync
down_read(cp_rwsem) // f2fs_lock_op
filemap_fdatawrite
f2fs_write_data_pages
...
block_opertaion
down_write(cp_rwsem) - marks rwsem as
"writer pending"
down_read_trylock(cp_rwsem) - fails as there is
a writer pending.
Code keeps on trying, live-locking the filesystem.

f2fs_write_single_data_page() will not grab read lock of cp_rwsem now, could you please check f2fs code in mainline?

	/* Dentry/quota blocks are controlled by checkpoint */
	if (S_ISDIR(inode->i_mode) || IS_NOQUOTA(inode)) {
		/*
		 * We need to wait for node_write to avoid block allocation during
		 * checkpoint. This can only happen to quota writes which can cause
		 * the below discard race condition.
		 */
		if (IS_NOQUOTA(inode))
			down_read(&sbi->node_write);

		fio.need_lock = LOCK_DONE;
		err = f2fs_do_write_data_page(&fio);

		if (IS_NOQUOTA(inode))
			up_read(&sbi->node_write);

		goto done;
	}

Thanks,

Quoted 73 lines

We solve this by creating a new rwsem, used specifically to synchronize this case, instead of attempting to reuse an existing lock.

Fixes db6ec53b7e03 f2fs: add a rw_sem to cover quota flag changes
---
  fs/f2fs/f2fs.h  |  3 +++
  fs/f2fs/super.c | 13 +++++++++++--
  2 files changed, 14 insertions(+), 2 deletions(-)
fs/f2fs/f2fs.h +3 -0
index cb700d797296..b3e55137be7f 100644
@@ -1448,6 +1448,7 @@ struct f2fs_sb_info {
  	struct inode *meta_inode;		/* cache meta blocks */
  	struct mutex cp_mutex;			/* checkpoint procedure lock */
  	struct rw_semaphore cp_rwsem;		/* blocking FS operations */
+	struct rw_semaphore cp_quota_rwsem;    	/* blocking quota sync operations */
  	struct rw_semaphore node_write;		/* locking node writes */
  	struct rw_semaphore node_change;	/* locking node change */
  	wait_queue_head_t cp_wait;
@@ -1961,12 +1962,14 @@ static inline void f2fs_unlock_op(struct f2fs_sb_info *sbi)
 
  static inline void f2fs_lock_all(struct f2fs_sb_info *sbi)
  {
+	down_write(&sbi->cp_quota_rwsem);
  	down_write(&sbi->cp_rwsem);
  }
 
  static inline void f2fs_unlock_all(struct f2fs_sb_info *sbi)
  {
  	up_write(&sbi->cp_rwsem);
+	up_write(&sbi->cp_quota_rwsem);
  }
 
  static inline int __get_cp_reason(struct f2fs_sb_info *sbi)
fs/f2fs/super.c +10 -2
index 00eff2f51807..5ce61147d7e5 100644
@@ -2209,8 +2209,16 @@ int f2fs_quota_sync(struct super_block *sb, int type)
  	 *  f2fs_dquot_commit
  	 *                            block_operation
  	 *                            down_read(quota_sem)
+	 *
+	 * However, we cannot use the cp_rwsem to prevent this
+	 * deadlock, as the cp_rwsem is taken for read inside the
+	 * f2fs_dquot_commit code, and rwsem is not recursive.
+	 *
+	 * We therefore use a special lock to synchronize
+	 * f2fs_quota_sync with block_operations, as this is the only
+	 * place where such recursion occurs.
  	 */
-	f2fs_lock_op(sbi);
+	down_read(&sbi->cp_quota_rwsem);
 
  	down_read(&sbi->quota_sem);
  	ret = dquot_writeback_dquots(sb, type);
@@ -2251,7 +2259,7 @@ int f2fs_quota_sync(struct super_block *sb, int type)
  	if (ret)
  		set_sbi_flag(F2FS_SB(sb), SBI_QUOTA_NEED_REPAIR);
  	up_read(&sbi->quota_sem);
-	f2fs_unlock_op(sbi);
+	up_read(&sbi->cp_quota_rwsem);
  	return ret;
  }
 

fs/f2fs/super.c static int f2fs_fill_super(struct super_block *sb, void *data, int silent) · line 3613

Quoted 7 lines
fs/f2fs/super.c +1 -0
@@ -3599,6 +3607,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 
  	init_rwsem(&sbi->cp_rwsem);
  	init_rwsem(&sbi->quota_sem);
+	init_rwsem(&sbi->cp_quota_rwsem);
  	init_waitqueue_head(&sbi->cp_wait);
  	init_sb_info(sbi);
 

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel