Merge tag 'vfs-6.15-rc4.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs

Pull vfs fixes from Christian Brauner:

 - For some reason we went from zero to three maintainers for HFS/HFS+
   in a matter of days. The lesson to learn from this might just be that
   we need to threaten code removal more often!?

 - Fix a regression introduced by enabling large folios for lage logical
   block sizes. This has caused issues for noref migration with large
   folios due to sleeping while in an atomic context.

   New sleeping variants of pagecache lookup helpers are introduced.
   These helpers take the folio lock instead of the mapping's private
   spinlock. The problematic users are converted to the sleeping
   variants and serialize against noref migration. Atomic users will
   bail on seeing the new BH_Migrate flag.

   This also shrinks the critical region of the mapping's private lock
   and the new blocking callers reduce contention on the spinlock for
   bdev mappings.

 - Fix two bugs in do_move_mount() when with MOVE_MOUNT_BENEATH. The
   first bug is using a mountpoint that is located on a mount we're not
   holding a reference to. The second bug is putting the mountpoint
   after we've called namespace_unlock() as it's no longer guaranteed
   that it does stay a mountpoint.

 - Remove a pointless call to vfs_getattr_nosec() in the devtmpfs code
   just to query i_mode instead of simply querying the inode directly.
   This also avoids lifetime issues for the dm code by an earlier bugfix
   this cycle that moved bdev_statx() handling into vfs_getattr_nosec().

 - Fix AT_FDCWD handling with getname_maybe_null() in the xattr code.

 - Fix a performance regression for files when multiple callers issue a
   close when it's not the last reference.

 - Remove a duplicate noinline annotation from pipe_clear_nowait().

* tag 'vfs-6.15-rc4.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
  fs/xattr: Fix handling of AT_FDCWD in setxattrat(2) and getxattrat(2)
  MAINTAINERS: hfs/hfsplus: add myself as maintainer
  splice: remove duplicate noinline from pipe_clear_nowait
  devtmpfs: don't use vfs_getattr_nosec to query i_mode
  fix a couple of races in MNT_TREE_BENEATH handling by do_move_mount()
  fs: fall back to file_ref_put() for non-last reference
  mm/migrate: fix sleep in atomic for large folios and buffer heads
  fs/ext4: use sleeping version of sb_find_get_block()
  fs/jbd2: use sleeping version of __find_get_block()
  fs/ocfs2: use sleeping version of __find_get_block()
  fs/buffer: use sleeping version of __find_get_block()
  fs/buffer: introduce sleeping flavors for pagecache lookups
  MAINTAINERS: add HFS/HFS+ maintainers
  fs/buffer: split locking for pagecache lookups

2 files changed, 15 insertions, 13 deletions

diff --git a/include/linux/buffer_head.h b/include/linux/buffer_head.h
index f0a4ad7839b6..0029ff880e27 100644
--- a/include/linux/buffer_head.h
+++ b/include/linux/buffer_head.h
@@ -34,6 +34,7 @@ enum bh_state_bits {
 	BH_Meta,	/* Buffer contains metadata */
 	BH_Prio,	/* Buffer should be submitted with REQ_PRIO */
 	BH_Defer_Completion, /* Defer AIO completion to workqueue */
+	BH_Migrate,     /* Buffer is being migrated (norefs) */
 
 	BH_PrivateStart,/* not a state bit, but the first bit available
 			 * for private allocation by other entities
@@ -222,6 +223,8 @@ void __wait_on_buffer(struct buffer_head *);
 wait_queue_head_t *bh_waitq_head(struct buffer_head *bh);
 struct buffer_head *__find_get_block(struct block_device *bdev, sector_t block,
 			unsigned size);
+struct buffer_head *__find_get_block_nonatomic(struct block_device *bdev,
+			sector_t block, unsigned size);
 struct buffer_head *bdev_getblk(struct block_device *bdev, sector_t block,
 		unsigned size, gfp_t gfp);
 void __brelse(struct buffer_head *);
@@ -397,6 +400,12 @@ sb_find_get_block(struct super_block *sb, sector_t block)
 	return __find_get_block(sb->s_bdev, block, sb->s_blocksize);
 }
 
+static inline struct buffer_head *
+sb_find_get_block_nonatomic(struct super_block *sb, sector_t block)
+{
+	return __find_get_block_nonatomic(sb->s_bdev, block, sb->s_blocksize);
+}
+
 static inline void
 map_bh(struct buffer_head *bh, struct super_block *sb, sector_t block)
 {
diff --git a/include/linux/file_ref.h b/include/linux/file_ref.h
index 7db62fbc0500..31551e4cb8f3 100644
--- a/include/linux/file_ref.h
+++ b/include/linux/file_ref.h
@@ -61,7 +61,6 @@ static inline void file_ref_init(file_ref_t *ref, unsigned long cnt)
 	atomic_long_set(&ref->refcnt, cnt - 1);
 }
 
-bool __file_ref_put_badval(file_ref_t *ref, unsigned long cnt);
 bool __file_ref_put(file_ref_t *ref, unsigned long cnt);
 
 /**
@@ -178,20 +177,14 @@ static __always_inline __must_check bool file_ref_put(file_ref_t *ref)
  */
 static __always_inline __must_check bool file_ref_put_close(file_ref_t *ref)
 {
-	long old, new;
+	long old;
 
 	old = atomic_long_read(&ref->refcnt);
-	do {
-		if (unlikely(old < 0))
-			return __file_ref_put_badval(ref, old);
-
-		if (old == FILE_REF_ONEREF)
-			new = FILE_REF_DEAD;
-		else
-			new = old - 1;
-	} while (!atomic_long_try_cmpxchg(&ref->refcnt, &old, new));
-
-	return new == FILE_REF_DEAD;
+	if (likely(old == FILE_REF_ONEREF)) {
+		if (likely(atomic_long_try_cmpxchg(&ref->refcnt, &old, FILE_REF_DEAD)))
+			return true;
+	}
+	return file_ref_put(ref);
 }
 
 /**