ovl: Add an alternative type of whiteout

An xattr whiteout (called "xwhiteout" in the code) is a reguar file of
zero size with the "overlay.whiteout" xattr set. A file like this in a
directory with the "overlay.whiteouts" xattrs set will be treated the
same way as a regular whiteout.

The "overlay.whiteouts" directory xattr is used in order to
efficiently handle overlay checks in readdir(), as we only need to
checks xattrs in affected directories.

The advantage of this kind of whiteout is that they can be escaped
using the standard overlay xattr escaping mechanism. So, a file with a
"overlay.overlay.whiteout" xattr would be unescaped to
"overlay.whiteout", which could then be consumed by another overlayfs
as a whiteout.

Overlayfs itself doesn't create whiteouts like this, but a userspace
mechanism could use this alternative mechanism to convert images that
may contain whiteouts to be used with overlayfs.

To work as a whiteout for both regular overlayfs mounts as well as
userxattr mounts both the "user.overlay.whiteout*" and the
"trusted.overlay.whiteout*" xattrs will need to be created.

Signed-off-by: Alexander Larsson <alexl@redhat.com>
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Amir Goldstein <amir73il@gmail.com>

1 files changed, 20 insertions, 7 deletions

diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c
index de39e067ae65..a490fc47c3e7 100644
--- a/fs/overlayfs/readdir.c
+++ b/fs/overlayfs/readdir.c
@@ -25,6 +25,7 @@ struct ovl_cache_entry {
 	struct ovl_cache_entry *next_maybe_whiteout;
 	bool is_upper;
 	bool is_whiteout;
+	bool check_xwhiteout;
 	char name[];
 };
 
@@ -47,6 +48,7 @@ struct ovl_readdir_data {
 	int err;
 	bool is_upper;
 	bool d_type_supported;
+	bool in_xwhiteouts_dir;
 };
 
 struct ovl_dir_file {
@@ -162,6 +164,8 @@ static struct ovl_cache_entry *ovl_cache_entry_new(struct ovl_readdir_data *rdd,
 		p->ino = 0;
 	p->is_upper = rdd->is_upper;
 	p->is_whiteout = false;
+	/* Defer check for overlay.whiteout to ovl_iterate() */
+	p->check_xwhiteout = rdd->in_xwhiteouts_dir && d_type == DT_REG;
 
 	if (d_type == DT_CHR) {
 		p->next_maybe_whiteout = rdd->first_maybe_whiteout;
@@ -301,6 +305,8 @@ static inline int ovl_dir_read(const struct path *realpath,
 	if (IS_ERR(realfile))
 		return PTR_ERR(realfile);
 
+	rdd->in_xwhiteouts_dir = rdd->dentry &&
+		ovl_path_check_xwhiteouts_xattr(OVL_FS(rdd->dentry->d_sb), realpath);
 	rdd->first_maybe_whiteout = NULL;
 	rdd->ctx.pos = 0;
 	do {
@@ -447,7 +453,7 @@ static u64 ovl_remap_lower_ino(u64 ino, int xinobits, int fsid,
 }
 
 /*
- * Set d_ino for upper entries. Non-upper entries should always report
+ * Set d_ino for upper entries if needed. 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.
@@ -455,8 +461,11 @@ static u64 ovl_remap_lower_ino(u64 ino, int xinobits, int fsid,
  * 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).
+ *
+ * Also checks the overlay.whiteout xattr by doing a full lookup which will return
+ * negative in this case.
  */
-static int ovl_cache_update_ino(const struct path *path, struct ovl_cache_entry *p)
+static int ovl_cache_update(const struct path *path, struct ovl_cache_entry *p, bool update_ino)
 
 {
 	struct dentry *dir = path->dentry;
@@ -467,7 +476,7 @@ static int ovl_cache_update_ino(const struct path *path, struct ovl_cache_entry
 	int xinobits = ovl_xino_bits(ofs);
 	int err = 0;
 
-	if (!ovl_same_dev(ofs))
+	if (!ovl_same_dev(ofs) && !p->check_xwhiteout)
 		goto out;
 
 	if (p->name[0] == '.') {
@@ -481,6 +490,7 @@ static int ovl_cache_update_ino(const struct path *path, struct ovl_cache_entry
 			goto get;
 		}
 	}
+	/* This checks also for xwhiteouts */
 	this = lookup_one(mnt_idmap(path->mnt), p->name, dir, p->len);
 	if (IS_ERR_OR_NULL(this) || !this->d_inode) {
 		/* Mark a stale entry */
@@ -494,6 +504,9 @@ static int ovl_cache_update_ino(const struct path *path, struct ovl_cache_entry
 	}
 
 get:
+	if (!ovl_same_dev(ofs) || !update_ino)
+		goto out;
+
 	type = ovl_path_type(this);
 	if (OVL_TYPE_ORIGIN(type)) {
 		struct kstat stat;
@@ -572,7 +585,7 @@ static int ovl_dir_read_impure(const struct path *path,  struct list_head *list,
 	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);
+			err = ovl_cache_update(path, p, true);
 			if (err)
 				return err;
 		}
@@ -778,13 +791,13 @@ 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->ino) {
-				err = ovl_cache_update_ino(&file->f_path, p);
+			if (!p->ino || p->check_xwhiteout) {
+				err = ovl_cache_update(&file->f_path, p, !p->ino);
 				if (err)
 					goto out;
 			}
 		}
-		/* ovl_cache_update_ino() sets is_whiteout on stale entry */
+		/* ovl_cache_update() sets is_whiteout on stale entry */
 		if (!p->is_whiteout) {
 			if (!dir_emit(ctx, p->name, p->len, p->ino, p->type))
 				break;