Linux: annotate nested xattr setattr znode locks

zfs_setattr() updates both the target znode and its hidden xattr
directory when ownership, mode, or project ID changes. The xattr
directory uses the same z_acl_lock and z_lock classes as the
parent znode, so lockdep reports recursive locking when the
second znode's mutexes are acquired.

This is a lockdep false positive rather than a real deadlock.
attrzp is the target file's hidden xattr directory, and the code
does not acquire these znode mutexes in the reverse order.
Acquire the attrzp mutexes with mutex_enter_nested() so lockdep
treats them as nested.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: ZhengYuan Huang <gality369@gmail.com>
Co-authored-by: gality369 <gality369@example.com>
Closes #18506

This commit is contained in:

Gality

2026-05-09 06:08:21 +08:00

committed by

GitHub

parent c7cfe0805c

commit 9ae9f2e983

1 changed files with 6 additions and 2 deletions

									
										module/os/linux/zfs/zfs_vnops_os.c
									
		+6
		-2
	
												View File
												
				@@ -2434,9 +2434,13 @@ zfs_setattr(znode_t *zp, vattr_t *vap, int flags, cred_t *cr, zidmap_t *mnt_ns)

					    &zp->z_pflags, sizeof (zp->z_pflags));

					if (attrzp) {

						/*

						 * attrzp is zp's hidden xattr directory, so the second

						 * znode lock acquisition is nested rather than recursive.

						 */

						if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE))

							mutex_enter(&attrzp->z_acl_lock);

						mutex_enter(&attrzp->z_lock);

							mutex_enter_nested(&attrzp->z_acl_lock, NESTED_SINGLE);

						mutex_enter_nested(&attrzp->z_lock, NESTED_SINGLE);

						SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,

						    SA_ZPL_FLAGS(zfsvfs), NULL, &attrzp->z_pflags,

						    sizeof (attrzp->z_pflags));