Skip to contentSkip to conversation

fs/affs: Replace binary semaphores with mutexes

PATCH In review Davidlohr Bueso 2 participants 2 messages Oct 27, 2019 – Oct 30, 2019 Affs file system

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 fs/affs: Replace binary semaphores with mutexes

At a slight footprint cost (24 vs 32 bytes), mutexes are more optimal than semaphores; it's also a nicer interface for mutual exclusion, which is why they are encouraged over binary semaphores, when possible.

For both i_link_lock and i_ext_lock (and hence i_hash_lock which I annotated for the hash lock mapping hackery for lockdep), their semantics imply traditional lock ownership; that is, the lock owner is the same for both lock/unlock operations and does not run in irq context. Therefore it is safe to convert.

---
Compile tested only.

 fs/affs/affs.h  | 16 ++++++++--------
 fs/affs/super.c |  4 ++--
 2 files changed, 10 insertions(+), 10 deletions(-)
fs/affs/affs.h +8 -8
index a92eb6ae2ae2..a755bef7c4c7 100644
@@ -43,8 +43,8 @@ struct affs_ext_key {
  */
 struct affs_inode_info {
 	atomic_t i_opencnt;
-	struct semaphore i_link_lock;		/* Protects internal inode access. */
-	struct semaphore i_ext_lock;		/* Protects internal inode access. */
+	struct mutex i_link_lock;		/* Protects internal inode access. */
+	struct mutex i_ext_lock;		/* Protects internal inode access. */
 #define i_hash_lock i_ext_lock
 	u32	 i_blkcnt;			/* block count */
 	u32	 i_extcnt;			/* extended block count */
@@ -293,30 +293,30 @@ affs_adjust_bitmapchecksum(struct buffer_head *bh, u32 val)
 static inline void
 affs_lock_link(struct inode *inode)
 {
-	down(&AFFS_I(inode)->i_link_lock);
+	mutex_lock(&AFFS_I(inode)->i_link_lock);
 }
 static inline void
 affs_unlock_link(struct inode *inode)
 {
-	up(&AFFS_I(inode)->i_link_lock);
+	mutex_unlock(&AFFS_I(inode)->i_link_lock);
 }
 static inline void
 affs_lock_dir(struct inode *inode)
 {
-	down(&AFFS_I(inode)->i_hash_lock);
+	mutex_lock_nested(&AFFS_I(inode)->i_hash_lock, SINGLE_DEPTH_NESTING);
 }
 static inline void
 affs_unlock_dir(struct inode *inode)
 {
-	up(&AFFS_I(inode)->i_hash_lock);
+	mutex_unlock(&AFFS_I(inode)->i_hash_lock);
 }
 static inline void
 affs_lock_ext(struct inode *inode)
 {
-	down(&AFFS_I(inode)->i_ext_lock);
+	mutex_lock(&AFFS_I(inode)->i_ext_lock);
 }
 static inline void
 affs_unlock_ext(struct inode *inode)
 {
-	up(&AFFS_I(inode)->i_ext_lock);
+	mutex_unlock(&AFFS_I(inode)->i_ext_lock);
 }
fs/affs/super.c +2 -2
index cc463ae47c12..9fcff55c65af 100644
@@ -121,8 +121,8 @@ static void init_once(void *foo)
 {
 	struct affs_inode_info *ei = (struct affs_inode_info *) foo;
 
-	sema_init(&ei->i_link_lock, 1);
-	sema_init(&ei->i_ext_lock, 1);
+	mutex_init(&ei->i_link_lock);
+	mutex_init(&ei->i_ext_lock);
 	inode_init_once(&ei->vfs_inode);
 }
 
Signature · 1 line
2.16.4
#1 lore raw
Quoted 11 lines · Davidlohr Bueso

At a slight footprint cost (24 vs 32 bytes), mutexes are more optimal than semaphores; it's also a nicer interface for mutual exclusion, which is why they are encouraged over binary semaphores, when possible.

For both i_link_lock and i_ext_lock (and hence i_hash_lock which I annotated for the hash lock mapping hackery for lockdep), their semantics imply traditional lock ownership; that is, the lock owner is the same for both lock/unlock operations and does not run in irq context. Therefore it is safe to convert.

Thanks, I'll add it to affs queue.