Partially revert 5a6ac4c

Reinstate the zpl_revalidate() functionality to resolve a regression
where dentries for open files during a rollback are not invalidated.

The unrelated functionality for automatically unmounting .zfs/snapshots
was not reverted.  Nor was the addition of shrink_dcache_sb() to the
zfs_resume_fs() function.

This issue was not immediately caught by the CI because the test case
intended to catch it was included in the list of ZTS tests which may
occasionally fail for unrelated reasons.  Remove all of the rollback
tests from this list to help identify the frequency of any spurious
failures.

The rollback_003_pos.ksh test case exposes a real issue with the
long standing code which needs to be investigated.  Regardless,
it has been enable with a small workaround in the test case itself.

Reviewed-by: Matt Ahrens <matt@delphix.com>
Reviewed-by: Pavel Snajdr <snajpa@snajpa.net>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #9587
Closes #9592

config/kernel-dentry-operations.m4

@@ -147,6 +147,28 @@ AC_DEFUN([ZFS_AC_KERNEL_CONST_DENTRY_OPERATIONS], [
 	])
 ])
 
+dnl #
+dnl # 2.6.38 API change
+dnl # Added sb->s_d_op default dentry_operations member
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_S_D_OP], [
+	ZFS_LINUX_TEST_SRC([super_block_s_d_op], [
+		#include <linux/fs.h>
+	],[
+		struct super_block sb __attribute__ ((unused));
+		sb.s_d_op = NULL;
+	])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_S_D_OP], [
+	AC_MSG_CHECKING([whether super_block has s_d_op])
+	ZFS_LINUX_TEST_RESULT([super_block_s_d_op], [
+		AC_MSG_RESULT(yes)
+	], [
+		ZFS_LINUX_TEST_ERROR([super_block s_d_op])
+	])
+])
+
 AC_DEFUN([ZFS_AC_KERNEL_SRC_DENTRY], [
         ZFS_AC_KERNEL_SRC_D_MAKE_ROOT
         ZFS_AC_KERNEL_SRC_D_OBTAIN_ALIAS
@@ -154,6 +176,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_DENTRY], [
         ZFS_AC_KERNEL_SRC_D_SET_D_OP
         ZFS_AC_KERNEL_SRC_D_REVALIDATE_NAMEIDATA
         ZFS_AC_KERNEL_SRC_CONST_DENTRY_OPERATIONS
+        ZFS_AC_KERNEL_SRC_S_D_OP
 ])
 
 AC_DEFUN([ZFS_AC_KERNEL_DENTRY], [
@@ -163,4 +186,5 @@ AC_DEFUN([ZFS_AC_KERNEL_DENTRY], [
         ZFS_AC_KERNEL_D_SET_D_OP
         ZFS_AC_KERNEL_D_REVALIDATE_NAMEIDATA
         ZFS_AC_KERNEL_CONST_DENTRY_OPERATIONS
+        ZFS_AC_KERNEL_S_D_OP
 ])

include/os/linux/zfs/sys/zpl.h

@@ -45,6 +45,7 @@ extern const struct inode_operations zpl_inode_operations;
 extern const struct inode_operations zpl_dir_inode_operations;
 extern const struct inode_operations zpl_symlink_inode_operations;
 extern const struct inode_operations zpl_special_inode_operations;
+extern dentry_operations_t zpl_dentry_operations;
 
 /* zpl_file.c */
 extern ssize_t zpl_read_common(struct inode *ip, const char *buf,

module/os/linux/zfs/zfs_vfsops.c

@@ -1926,6 +1926,7 @@ zfs_domount(struct super_block *sb, zfs_mnt_t *zm, int silent)
 	sb->s_op = &zpl_super_operations;
 	sb->s_xattr = zpl_xattr_handlers;
 	sb->s_export_op = &zpl_export_operations;
+	sb->s_d_op = &zpl_dentry_operations;
 
 	/* Set features for file system. */
 	zfs_set_fuid_feature(zfsvfs);

module/os/linux/zfs/zpl_inode.c

@@ -602,6 +602,46 @@ zpl_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry)
 	return (error);
 }
 
+static int
+#ifdef HAVE_D_REVALIDATE_NAMEIDATA
+zpl_revalidate(struct dentry *dentry, struct nameidata *nd)
+{
+	unsigned int flags = (nd ? nd->flags : 0);
+#else
+zpl_revalidate(struct dentry *dentry, unsigned int flags)
+{
+#endif /* HAVE_D_REVALIDATE_NAMEIDATA */
+	/* CSTYLED */
+	zfsvfs_t *zfsvfs = dentry->d_sb->s_fs_info;
+	int error;
+
+	if (flags & LOOKUP_RCU)
+		return (-ECHILD);
+
+	/*
+	 * After a rollback negative dentries created before the rollback
+	 * time must be invalidated.  Otherwise they can obscure files which
+	 * are only present in the rolled back dataset.
+	 */
+	if (dentry->d_inode == NULL) {
+		spin_lock(&dentry->d_lock);
+		error = time_before(dentry->d_time, zfsvfs->z_rollback_time);
+		spin_unlock(&dentry->d_lock);
+
+		if (error)
+			return (0);
+	}
+
+	/*
+	 * The dentry may reference a stale inode if a mounted file system
+	 * was rolled back to a point in time where the object didn't exist.
+	 */
+	if (dentry->d_inode && ITOZ(dentry->d_inode)->z_is_stale)
+		return (0);
+
+	return (1);
+}
+
 const struct inode_operations zpl_inode_operations = {
 	.setattr	= zpl_setattr,
 	.getattr	= zpl_getattr,
@@ -690,3 +730,7 @@ const struct inode_operations zpl_special_inode_operations = {
 	.get_acl	= zpl_get_acl,
 #endif /* CONFIG_FS_POSIX_ACL */
 };
+
+dentry_operations_t zpl_dentry_operations = {
+	.d_revalidate	= zpl_revalidate,
+};

tests/test-runner/bin/zts-report.py

@@ -189,7 +189,6 @@
     'removal/removal_with_zdb': ['SKIP', known_reason],
     'rootpool/setup': ['SKIP', na_reason],
     'rsend/rsend_008_pos': ['SKIP', '6066'],
-    'snapshot/rollback_003_pos': ['SKIP', '6143'],
     'vdev_zaps/vdev_zaps_007_pos': ['FAIL', known_reason],
     'xattr/xattr_008_pos': ['SKIP', na_reason],
     'xattr/xattr_009_neg': ['SKIP', na_reason],
@@ -222,8 +221,6 @@
     'cli_root/zdb/zdb_006_pos': ['FAIL', known_reason],
     'cli_root/zfs_get/zfs_get_004_pos': ['FAIL', known_reason],
     'cli_root/zfs_get/zfs_get_009_pos': ['SKIP', '5479'],
-    'cli_root/zfs_rollback/zfs_rollback_001_pos': ['FAIL', '6415'],
-    'cli_root/zfs_rollback/zfs_rollback_002_pos': ['FAIL', '6416'],
     'cli_root/zfs_share/setup': ['SKIP', share_reason],
     'cli_root/zfs_snapshot/zfs_snapshot_002_neg': ['FAIL', known_reason],
     'cli_root/zfs_unshare/setup': ['SKIP', share_reason],

tests/zfs-tests/tests/functional/snapshot/rollback_003_pos.ksh

@@ -48,10 +48,6 @@
 
 verify_runnable "both"
 
-if is_linux; then
-	log_unsupported "Test case is known to fail on Linux"
-fi
-
 function cleanup
 {
 	typeset snap=""
@@ -61,18 +57,16 @@ function cleanup
 	log_must zfs mount -a
 	unset __ZFS_POOL_RESTRICT
 
-	for snap in "$SNAPPOOL.1" "$SNAPPOOL"
-	do
-		snapexists $snap
-		[[ $? -eq 0 ]] && \
-			log_must zfs destroy $snap
+	for snap in "$SNAPPOOL.1" "$SNAPPOOL"; do
+		if snapexists $snap; then
+			destroy_snapshot $snap
+		fi
 	done
 
-	for fs in "$TESTPOOL/$TESTFILE/$TESTFILE.1" "$TESTPOOL/$TESTFILE"
-	do
-		datasetexists $fs
-		[[ $? -eq 0 ]] && \
-			log_must zfs destroy -r $fs
+	for fs in "$TESTPOOL/$TESTFILE/$TESTFILE.1" "$TESTPOOL/$TESTFILE"; do
+		if datasetexists $fs; then
+			destroy_dataset $fs -r
+		fi
 	done
 
 	[[ -e /$TESTPOOL ]] && \
@@ -107,4 +101,11 @@ log_must touch /$TESTPOOL/$TESTFILE/$TESTFILE.1
 
 log_must zfs rollback $SNAPPOOL.1
 
+#
+# Workaround for issue #6143.  Issuing a `df` seems to properly force any
+# negative dcache entries to be invalidated preventing subsequent failures
+# when accessing the mount point.  Additional investigation required.
+#
+log_must df
+
 log_pass "Rollbacks succeed when nested file systems are present."