Merge branch 'overlayfs-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs

Pull overlayfs updates from Miklos Szeredi:
 "This fixes d_ino correctness in readdir, which brings overlayfs on par
  with normal filesystems regarding inode number semantics, as long as
  all layers are on the same filesystem.

  There are also some bug fixes, one in particular (random ioctl's
  shouldn't be able to modify lower layers) that touches some vfs code,
  but of course no-op for non-overlay fs"

* 'overlayfs-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs:
  ovl: fix false positive ESTALE on lookup
  ovl: don't allow writing ioctl on lower layer
  ovl: fix relatime for directories
  vfs: add flags to d_real()
  ovl: cleanup d_real for negative
  ovl: constant d_ino for non-merge dirs
  ovl: constant d_ino across copy up
  ovl: fix readdir error value
  ovl: check snprintf return

1 files changed, 350 insertions, 33 deletions

diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c
index f0fd3adb1693..62e9b22a2077 100644
--- a/fs/overlayfs/readdir.c
+++ b/fs/overlayfs/readdir.c
@@ -15,11 +15,13 @@
 #include <linux/rbtree.h>
 #include <linux/security.h>
 #include <linux/cred.h>
+#include <linux/ratelimit.h>
 #include "overlayfs.h"
 
 struct ovl_cache_entry {
 	unsigned int len;
 	unsigned int type;
+	u64 real_ino;
 	u64 ino;
 	struct list_head l_node;
 	struct rb_node node;
@@ -32,18 +34,20 @@ struct ovl_dir_cache {
 	long refcount;
 	u64 version;
 	struct list_head entries;
+	struct rb_root root;
 };
 
 struct ovl_readdir_data {
 	struct dir_context ctx;
 	struct dentry *dentry;
 	bool is_lowest;
-	struct rb_root root;
+	struct rb_root *root;
 	struct list_head *list;
 	struct list_head middle;
 	struct ovl_cache_entry *first_maybe_whiteout;
 	int count;
 	int err;
+	bool is_upper;
 	bool d_type_supported;
 };
 
@@ -58,7 +62,33 @@ struct ovl_dir_file {
 
 static struct ovl_cache_entry *ovl_cache_entry_from_node(struct rb_node *n)
 {
-	return container_of(n, struct ovl_cache_entry, node);
+	return rb_entry(n, struct ovl_cache_entry, node);
+}
+
+static bool ovl_cache_entry_find_link(const char *name, int len,
+				      struct rb_node ***link,
+				      struct rb_node **parent)
+{
+	bool found = false;
+	struct rb_node **newp = *link;
+
+	while (!found && *newp) {
+		int cmp;
+		struct ovl_cache_entry *tmp;
+
+		*parent = *newp;
+		tmp = ovl_cache_entry_from_node(*newp);
+		cmp = strncmp(name, tmp->name, len);
+		if (cmp > 0)
+			newp = &tmp->node.rb_right;
+		else if (cmp < 0 || len < tmp->len)
+			newp = &tmp->node.rb_left;
+		else
+			found = true;
+	}
+	*link = newp;
+
+	return found;
 }
 
 static struct ovl_cache_entry *ovl_cache_entry_find(struct rb_root *root,
@@ -82,6 +112,32 @@ static struct ovl_cache_entry *ovl_cache_entry_find(struct rb_root *root,
 	return NULL;
 }
 
+static bool ovl_calc_d_ino(struct ovl_readdir_data *rdd,
+			   struct ovl_cache_entry *p)
+{
+	/* Don't care if not doing ovl_iter() */
+	if (!rdd->dentry)
+		return false;
+
+	/* Always recalc d_ino for parent */
+	if (strcmp(p->name, "..") == 0)
+		return true;
+
+	/* If this is lower, then native d_ino will do */
+	if (!rdd->is_upper)
+		return false;
+
+	/*
+	 * Recalc d_ino for '.' and for all entries if dir is impure (contains
+	 * copied up entries)
+	 */
+	if ((p->name[0] == '.' && p->len == 1) ||
+	    ovl_test_flag(OVL_IMPURE, d_inode(rdd->dentry)))
+		return true;
+
+	return false;
+}
+
 static struct ovl_cache_entry *ovl_cache_entry_new(struct ovl_readdir_data *rdd,
 						   const char *name, int len,
 						   u64 ino, unsigned int d_type)
@@ -97,7 +153,11 @@ static struct ovl_cache_entry *ovl_cache_entry_new(struct ovl_readdir_data *rdd,
 	p->name[len] = '\0';
 	p->len = len;
 	p->type = d_type;
+	p->real_ino = ino;
 	p->ino = ino;
+	/* Defer setting d_ino for upper entry to ovl_iterate() */
+	if (ovl_calc_d_ino(rdd, p))
+		p->ino = 0;
 	p->is_whiteout = false;
 
 	if (d_type == DT_CHR) {
@@ -111,32 +171,22 @@ static int ovl_cache_entry_add_rb(struct ovl_readdir_data *rdd,
 				  const char *name, int len, u64 ino,
 				  unsigned int d_type)
 {
-	struct rb_node **newp = &rdd->root.rb_node;
+	struct rb_node **newp = &rdd->root->rb_node;
 	struct rb_node *parent = NULL;
 	struct ovl_cache_entry *p;
 
-	while (*newp) {
-		int cmp;
-		struct ovl_cache_entry *tmp;
-
-		parent = *newp;
-		tmp = ovl_cache_entry_from_node(*newp);
-		cmp = strncmp(name, tmp->name, len);
-		if (cmp > 0)
-			newp = &tmp->node.rb_right;
-		else if (cmp < 0 || len < tmp->len)
-			newp = &tmp->node.rb_left;
-		else
-			return 0;
-	}
+	if (ovl_cache_entry_find_link(name, len, &newp, &parent))
+		return 0;
 
 	p = ovl_cache_entry_new(rdd, name, len, ino, d_type);
-	if (p == NULL)
+	if (p == NULL) {
+		rdd->err = -ENOMEM;
 		return -ENOMEM;
+	}
 
 	list_add_tail(&p->l_node, rdd->list);
 	rb_link_node(&p->node, parent, newp);
-	rb_insert_color(&p->node, &rdd->root);
+	rb_insert_color(&p->node, rdd->root);
 
 	return 0;
 }
@@ -147,7 +197,7 @@ static int ovl_fill_lowest(struct ovl_readdir_data *rdd,
 {
 	struct ovl_cache_entry *p;
 
-	p = ovl_cache_entry_find(&rdd->root, name, namelen);
+	p = ovl_cache_entry_find(rdd->root, name, namelen);
 	if (p) {
 		list_move_tail(&p->l_node, &rdd->middle);
 	} else {
@@ -172,6 +222,16 @@ void ovl_cache_free(struct list_head *list)
 	INIT_LIST_HEAD(list);
 }
 
+void ovl_dir_cache_free(struct inode *inode)
+{
+	struct ovl_dir_cache *cache = ovl_dir_cache(inode);
+
+	if (cache) {
+		ovl_cache_free(&cache->entries);
+		kfree(cache);
+	}
+}
+
 static void ovl_cache_put(struct ovl_dir_file *od, struct dentry *dentry)
 {
 	struct ovl_dir_cache *cache = od->cache;
@@ -179,8 +239,8 @@ static void ovl_cache_put(struct ovl_dir_file *od, struct dentry *dentry)
 	WARN_ON(cache->refcount <= 0);
 	cache->refcount--;
 	if (!cache->refcount) {
-		if (ovl_dir_cache(dentry) == cache)
-			ovl_set_dir_cache(dentry, NULL);
+		if (ovl_dir_cache(d_inode(dentry)) == cache)
+			ovl_set_dir_cache(d_inode(dentry), NULL);
 
 		ovl_cache_free(&cache->entries);
 		kfree(cache);
@@ -273,7 +333,8 @@ static void ovl_dir_reset(struct file *file)
 		od->is_real = false;
 }
 
-static int ovl_dir_read_merged(struct dentry *dentry, struct list_head *list)
+static int ovl_dir_read_merged(struct dentry *dentry, struct list_head *list,
+	struct rb_root *root)
 {
 	int err;
 	struct path realpath;
@@ -281,13 +342,14 @@ static int ovl_dir_read_merged(struct dentry *dentry, struct list_head *list)
 		.ctx.actor = ovl_fill_merge,
 		.dentry = dentry,
 		.list = list,
-		.root = RB_ROOT,
+		.root = root,
 		.is_lowest = false,
 	};
 	int idx, next;
 
 	for (idx = 0; idx != -1; idx = next) {
 		next = ovl_path_next(idx, dentry, &realpath);
+		rdd.is_upper = ovl_dentry_upper(dentry) == realpath.dentry;
 
 		if (next != -1) {
 			err = ovl_dir_read(&realpath, &rdd);
@@ -326,12 +388,13 @@ static struct ovl_dir_cache *ovl_cache_get(struct dentry *dentry)
 	int res;
 	struct ovl_dir_cache *cache;
 
-	cache = ovl_dir_cache(dentry);
+	cache = ovl_dir_cache(d_inode(dentry));
 	if (cache && ovl_dentry_version_get(dentry) == cache->version) {
+		WARN_ON(!cache->refcount);
 		cache->refcount++;
 		return cache;
 	}
-	ovl_set_dir_cache(dentry, NULL);
+	ovl_set_dir_cache(d_inode(dentry), NULL);
 
 	cache = kzalloc(sizeof(struct ovl_dir_cache), GFP_KERNEL);
 	if (!cache)
@@ -339,8 +402,9 @@ static struct ovl_dir_cache *ovl_cache_get(struct dentry *dentry)
 
 	cache->refcount = 1;
 	INIT_LIST_HEAD(&cache->entries);
+	cache->root = RB_ROOT;
 
-	res = ovl_dir_read_merged(dentry, &cache->entries);
+	res = ovl_dir_read_merged(dentry, &cache->entries, &cache->root);
 	if (res) {
 		ovl_cache_free(&cache->entries);
 		kfree(cache);
@@ -348,22 +412,266 @@ static struct ovl_dir_cache *ovl_cache_get(struct dentry *dentry)
 	}
 
 	cache->version = ovl_dentry_version_get(dentry);
-	ovl_set_dir_cache(dentry, cache);
+	ovl_set_dir_cache(d_inode(dentry), cache);
 
 	return cache;
 }
 
+/*
+ * Set d_ino for upper entries. Non-upper entries should always report
+ * the uppermost real inode ino and should not call this function.
+ *
+ * When not all layer are on same fs, report real ino also for upper.
+ *
+ * When all layers are on the same fs, and upper has a reference to
+ * copy up origin, call vfs_getattr() on the overlay entry to make
+ * sure that d_ino will be consistent with st_ino from stat(2).
+ */
+static int ovl_cache_update_ino(struct path *path, struct ovl_cache_entry *p)
+
+{
+	struct dentry *dir = path->dentry;
+	struct dentry *this = NULL;
+	enum ovl_path_type type;
+	u64 ino = p->real_ino;
+	int err = 0;
+
+	if (!ovl_same_sb(dir->d_sb))
+		goto out;
+
+	if (p->name[0] == '.') {
+		if (p->len == 1) {
+			this = dget(dir);
+			goto get;
+		}
+		if (p->len == 2 && p->name[1] == '.') {
+			/* we shall not be moved */
+			this = dget(dir->d_parent);
+			goto get;
+		}
+	}
+	this = lookup_one_len(p->name, dir, p->len);
+	if (IS_ERR_OR_NULL(this) || !this->d_inode) {
+		if (IS_ERR(this)) {
+			err = PTR_ERR(this);
+			this = NULL;
+			goto fail;
+		}
+		goto out;
+	}
+
+get:
+	type = ovl_path_type(this);
+	if (OVL_TYPE_ORIGIN(type)) {
+		struct kstat stat;
+		struct path statpath = *path;
+
+		statpath.dentry = this;
+		err = vfs_getattr(&statpath, &stat, STATX_INO, 0);
+		if (err)
+			goto fail;
+
+		WARN_ON_ONCE(dir->d_sb->s_dev != stat.dev);
+		ino = stat.ino;
+	}
+
+out:
+	p->ino = ino;
+	dput(this);
+	return err;
+
+fail:
+	pr_warn_ratelimited("overlay: failed to look up (%s) for ino (%i)\n",
+			    p->name, err);
+	goto out;
+}
+
+static int ovl_fill_plain(struct dir_context *ctx, const char *name,
+			  int namelen, loff_t offset, u64 ino,
+			  unsigned int d_type)
+{
+	struct ovl_cache_entry *p;
+	struct ovl_readdir_data *rdd =
+		container_of(ctx, struct ovl_readdir_data, ctx);
+
+	rdd->count++;
+	p = ovl_cache_entry_new(rdd, name, namelen, ino, d_type);
+	if (p == NULL) {
+		rdd->err = -ENOMEM;
+		return -ENOMEM;
+	}
+	list_add_tail(&p->l_node, rdd->list);
+
+	return 0;
+}
+
+static int ovl_dir_read_impure(struct path *path,  struct list_head *list,
+			       struct rb_root *root)
+{
+	int err;
+	struct path realpath;
+	struct ovl_cache_entry *p, *n;
+	struct ovl_readdir_data rdd = {
+		.ctx.actor = ovl_fill_plain,
+		.list = list,
+		.root = root,
+	};
+
+	INIT_LIST_HEAD(list);
+	*root = RB_ROOT;
+	ovl_path_upper(path->dentry, &realpath);
+
+	err = ovl_dir_read(&realpath, &rdd);
+	if (err)
+		return err;
+
+	list_for_each_entry_safe(p, n, list, l_node) {
+		if (strcmp(p->name, ".") != 0 &&
+		    strcmp(p->name, "..") != 0) {
+			err = ovl_cache_update_ino(path, p);
+			if (err)
+				return err;
+		}
+		if (p->ino == p->real_ino) {
+			list_del(&p->l_node);
+			kfree(p);
+		} else {
+			struct rb_node **newp = &root->rb_node;
+			struct rb_node *parent = NULL;
+
+			if (WARN_ON(ovl_cache_entry_find_link(p->name, p->len,
+							      &newp, &parent)))
+				return -EIO;
+
+			rb_link_node(&p->node, parent, newp);
+			rb_insert_color(&p->node, root);
+		}
+	}
+	return 0;
+}
+
+static struct ovl_dir_cache *ovl_cache_get_impure(struct path *path)
+{
+	int res;
+	struct dentry *dentry = path->dentry;
+	struct ovl_dir_cache *cache;
+
+	cache = ovl_dir_cache(d_inode(dentry));
+	if (cache && ovl_dentry_version_get(dentry) == cache->version)
+		return cache;
+
+	/* Impure cache is not refcounted, free it here */
+	ovl_dir_cache_free(d_inode(dentry));
+	ovl_set_dir_cache(d_inode(dentry), NULL);
+
+	cache = kzalloc(sizeof(struct ovl_dir_cache), GFP_KERNEL);
+	if (!cache)
+		return ERR_PTR(-ENOMEM);
+
+	res = ovl_dir_read_impure(path, &cache->entries, &cache->root);
+	if (res) {
+		ovl_cache_free(&cache->entries);
+		kfree(cache);
+		return ERR_PTR(res);
+	}
+	if (list_empty(&cache->entries)) {
+		/* Good oportunity to get rid of an unnecessary "impure" flag */
+		ovl_do_removexattr(ovl_dentry_upper(dentry), OVL_XATTR_IMPURE);
+		ovl_clear_flag(OVL_IMPURE, d_inode(dentry));
+		kfree(cache);
+		return NULL;
+	}
+
+	cache->version = ovl_dentry_version_get(dentry);
+	ovl_set_dir_cache(d_inode(dentry), cache);
+
+	return cache;
+}
+
+struct ovl_readdir_translate {
+	struct dir_context *orig_ctx;
+	struct ovl_dir_cache *cache;
+	struct dir_context ctx;
+	u64 parent_ino;
+};
+
+static int ovl_fill_real(struct dir_context *ctx, const char *name,
+			   int namelen, loff_t offset, u64 ino,
+			   unsigned int d_type)
+{
+	struct ovl_readdir_translate *rdt =
+		container_of(ctx, struct ovl_readdir_translate, ctx);
+	struct dir_context *orig_ctx = rdt->orig_ctx;
+
+	if (rdt->parent_ino && strcmp(name, "..") == 0)
+		ino = rdt->parent_ino;
+	else if (rdt->cache) {
+		struct ovl_cache_entry *p;
+
+		p = ovl_cache_entry_find(&rdt->cache->root, name, namelen);
+		if (p)
+			ino = p->ino;
+	}
+
+	return orig_ctx->actor(orig_ctx, name, namelen, offset, ino, d_type);
+}
+
+static int ovl_iterate_real(struct file *file, struct dir_context *ctx)
+{
+	int err;
+	struct ovl_dir_file *od = file->private_data;
+	struct dentry *dir = file->f_path.dentry;
+	struct ovl_readdir_translate rdt = {
+		.ctx.actor = ovl_fill_real,
+		.orig_ctx = ctx,
+	};
+
+	if (OVL_TYPE_MERGE(ovl_path_type(dir->d_parent))) {
+		struct kstat stat;
+		struct path statpath = file->f_path;
+
+		statpath.dentry = dir->d_parent;
+		err = vfs_getattr(&statpath, &stat, STATX_INO, 0);
+		if (err)
+			return err;
+
+		WARN_ON_ONCE(dir->d_sb->s_dev != stat.dev);
+		rdt.parent_ino = stat.ino;
+	}
+
+	if (ovl_test_flag(OVL_IMPURE, d_inode(dir))) {
+		rdt.cache = ovl_cache_get_impure(&file->f_path);
+		if (IS_ERR(rdt.cache))
+			return PTR_ERR(rdt.cache);
+	}
+
+	return iterate_dir(od->realfile, &rdt.ctx);
+}
+
+
 static int ovl_iterate(struct file *file, struct dir_context *ctx)
 {
 	struct ovl_dir_file *od = file->private_data;
 	struct dentry *dentry = file->f_path.dentry;
 	struct ovl_cache_entry *p;
+	int err;
 
 	if (!ctx->pos)
 		ovl_dir_reset(file);
 
-	if (od->is_real)
+	if (od->is_real) {
+		/*
+		 * If parent is merge, then need to adjust d_ino for '..', if
+		 * dir is impure then need to adjust d_ino for copied up
+		 * entries.
+		 */
+		if (ovl_same_sb(dentry->d_sb) &&
+		    (ovl_test_flag(OVL_IMPURE, d_inode(dentry)) ||
+		     OVL_TYPE_MERGE(ovl_path_type(dentry->d_parent)))) {
+			return ovl_iterate_real(file, ctx);
+		}
 		return iterate_dir(od->realfile, ctx);
+	}
 
 	if (!od->cache) {
 		struct ovl_dir_cache *cache;
@@ -378,9 +686,15 @@ static int ovl_iterate(struct file *file, struct dir_context *ctx)
 
 	while (od->cursor != &od->cache->entries) {
 		p = list_entry(od->cursor, struct ovl_cache_entry, l_node);
-		if (!p->is_whiteout)
+		if (!p->is_whiteout) {
+			if (!p->ino) {
+				err = ovl_cache_update_ino(&file->f_path, p);
+				if (err)
+					return err;
+			}
 			if (!dir_emit(ctx, p->name, p->len, p->ino, p->type))
 				break;
+		}
 		od->cursor = p->l_node.next;
 		ctx->pos++;
 	}
@@ -522,8 +836,9 @@ int ovl_check_empty_dir(struct dentry *dentry, struct list_head *list)
 {
 	int err;
 	struct ovl_cache_entry *p;
+	struct rb_root root = RB_ROOT;
 
-	err = ovl_dir_read_merged(dentry, list);
+	err = ovl_dir_read_merged(dentry, list, &root);
 	if (err)
 		return err;
 
@@ -612,12 +927,13 @@ static void ovl_workdir_cleanup_recurse(struct path *path, int level)
 	int err;
 	struct inode *dir = path->dentry->d_inode;
 	LIST_HEAD(list);
+	struct rb_root root = RB_ROOT;
 	struct ovl_cache_entry *p;
 	struct ovl_readdir_data rdd = {
 		.ctx.actor = ovl_fill_merge,
 		.dentry = NULL,
 		.list = &list,
-		.root = RB_ROOT,
+		.root = &root,
 		.is_lowest = false,
 	};
 
@@ -675,12 +991,13 @@ int ovl_indexdir_cleanup(struct dentry *dentry, struct vfsmount *mnt,
 	struct inode *dir = dentry->d_inode;
 	struct path path = { .mnt = mnt, .dentry = dentry };
 	LIST_HEAD(list);
+	struct rb_root root = RB_ROOT;
 	struct ovl_cache_entry *p;
 	struct ovl_readdir_data rdd = {
 		.ctx.actor = ovl_fill_merge,
 		.dentry = NULL,
 		.list = &list,
-		.root = RB_ROOT,
+		.root = &root,
 		.is_lowest = false,
 	};