Merge tag 'pull-dcache' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

Pull dcache updates from Al Viro:
 "Change of locking rules for __dentry_kill(), regularized refcounting
  rules in that area, assorted cleanups and removal of weird corner
  cases (e.g. now ->d_iput() on child is always called before the parent
  might hit __dentry_kill(), etc)"

* tag 'pull-dcache' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (40 commits)
  dcache: remove unnecessary NULL check in dget_dlock()
  kill DCACHE_MAY_FREE
  __d_unalias() doesn't use inode argument
  d_alloc_parallel(): in-lookup hash insertion doesn't need an RCU variant
  get rid of DCACHE_GENOCIDE
  d_genocide(): move the extern into fs/internal.h
  simple_fill_super(): don't bother with d_genocide() on failure
  nsfs: use d_make_root()
  d_alloc_pseudo(): move setting ->d_op there from the (sole) caller
  kill d_instantate_anon(), fold __d_instantiate_anon() into remaining caller
  retain_dentry(): introduce a trimmed-down lockless variant
  __dentry_kill(): new locking scheme
  d_prune_aliases(): use a shrink list
  switch select_collect{,2}() to use of to_shrink_list()
  to_shrink_list(): call only if refcount is 0
  fold dentry_kill() into dput()
  don't try to cut corners in shrink_lock_dentry()
  fold the call of retain_dentry() into fast_dput()
  Call retain_dentry() with refcount 0
  dentry_kill(): don't bother with retain_dentry() on slow path
  ...

Author: torvalds

Documentation/filesystems/porting.rst

@@ -1091,6 +1091,40 @@ would need to do so.
 
 ---
 
+**mandatory**
+
+The list of children anchored in parent dentry got turned into hlist now.
+Field names got changed (->d_children/->d_sib instead of ->d_subdirs/->d_child
+for anchor/entries resp.), so any affected places will be immediately caught
+by compiler.
+
+---
+
+**mandatory**
+
+->d_delete() instances are now called for dentries with ->d_lock held
+and refcount equal to 0.  They are not permitted to drop/regain ->d_lock.
+None of in-tree instances did anything of that sort.  Make sure yours do not...
+
+---
+
+**mandatory**
+
+->d_prune() instances are now called without ->d_lock held on the parent.
+->d_lock on dentry itself is still held; if you need per-parent exclusions (none
+of the in-tree instances did), use your own spinlock.
+
+->d_iput() and ->d_release() are called with victim dentry still in the
+list of parent's children.  It is still unhashed, marked killed, etc., just not
+removed from parent's ->d_children yet.
+
+Anyone iterating through the list of children needs to be aware of the
+half-killed dentries that might be seen there; taking ->d_lock on those will
+see them negative, unhashed and with negative refcount, which means that most
+of the in-kernel users would've done the right thing anyway without any adjustment.
+
+---
+
 **recommended**
 
 Block device freezing and thawing have been moved to holder operations.

arch/powerpc/platforms/cell/spufs/inode.c

@@ -145,10 +145,11 @@ spufs_evict_inode(struct inode *inode)
 
 static void spufs_prune_dir(struct dentry *dir)
 {
-	struct dentry *dentry, *tmp;
+	struct dentry *dentry;
+	struct hlist_node *n;
 
 	inode_lock(d_inode(dir));
-	list_for_each_entry_safe(dentry, tmp, &dir->d_subdirs, d_child) {
+	hlist_for_each_entry_safe(dentry, n, &dir->d_children, d_sib) {
 		spin_lock(&dentry->d_lock);
 		if (simple_positive(dentry)) {
 			dget_dlock(dentry);

fs/afs/dynroot.c

@@ -373,7 +373,7 @@ int afs_dynroot_populate(struct super_block *sb)
 void afs_dynroot_depopulate(struct super_block *sb)
 {
 	struct afs_net *net = afs_sb2net(sb);
-	struct dentry *root = sb->s_root, *subdir, *tmp;
+	struct dentry *root = sb->s_root, *subdir;
 
 	/* Prevent more subdirs from being created */
 	mutex_lock(&net->proc_cells_lock);
@@ -382,10 +382,11 @@ void afs_dynroot_depopulate(struct super_block *sb)
 	mutex_unlock(&net->proc_cells_lock);
 
 	if (root) {
+		struct hlist_node *n;
 		inode_lock(root->d_inode);
 
 		/* Remove all the pins for dirs created for manually added cells */
-		list_for_each_entry_safe(subdir, tmp, &root->d_subdirs, d_child) {
+		hlist_for_each_entry_safe(subdir, n, &root->d_children, d_sib) {
 			if (subdir->d_fsdata) {
 				subdir->d_fsdata = NULL;
 				dput(subdir);

fs/autofs/expire.c

@@ -73,12 +73,9 @@ static int autofs_mount_busy(struct vfsmount *mnt,
 /* p->d_lock held */
 static struct dentry *positive_after(struct dentry *p, struct dentry *child)
 {
-	if (child)
-		child = list_next_entry(child, d_child);
-	else
-		child = list_first_entry(&p->d_subdirs, struct dentry, d_child);
+	child = child ? d_next_sibling(child) : d_first_child(p);
 
-	list_for_each_entry_from(child, &p->d_subdirs, d_child) {
+	hlist_for_each_entry_from(child, d_sib) {
 		spin_lock_nested(&child->d_lock, DENTRY_D_LOCK_NESTED);
 		if (simple_positive(child)) {
 			dget_dlock(child);

fs/ceph/dir.c

@@ -174,7 +174,7 @@ __dcache_find_get_entry(struct dentry *parent, u64 idx,
 /*
  * When possible, we try to satisfy a readdir by peeking at the
  * dcache.  We make this work by carefully ordering dentries on
- * d_child when we initially get results back from the MDS, and
+ * d_children when we initially get results back from the MDS, and
  * falling back to a "normal" sync readdir if any dentries in the dir
  * are dropped.
  *

fs/ceph/mds_client.c

@@ -2128,7 +2128,7 @@ static bool drop_negative_children(struct dentry *dentry)
 		goto out;
 
 	spin_lock(&dentry->d_lock);
-	list_for_each_entry(child, &dentry->d_subdirs, d_child) {
+	hlist_for_each_entry(child, &dentry->d_children, d_sib) {
 		if (d_really_is_positive(child)) {
 			all_negative = false;
 			break;

fs/coda/cache.c

@@ -93,13 +93,13 @@ static void coda_flag_children(struct dentry *parent, int flag)
 	struct dentry *de;
 
 	spin_lock(&parent->d_lock);
-	list_for_each_entry(de, &parent->d_subdirs, d_child) {
+	hlist_for_each_entry(de, &parent->d_children, d_sib) {
+		struct inode *inode = d_inode_rcu(de);
 		/* don't know what to do with negative dentries */
-		if (d_inode(de) ) 
-			coda_flag_inode(d_inode(de), flag);
+		if (inode)
+			coda_flag_inode(inode, flag);
 	}
 	spin_unlock(&parent->d_lock);
-	return; 
 }
 
 void coda_flag_inode_children(struct inode *inode, int flag)