Merge tag 'ovl-fixes-5.2-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs

Pull overlayfs fixes from Miklos Szeredi:
 "Here's one fix for a class of bugs triggered by syzcaller, and one
  that makes xfstests fail less"

* tag 'ovl-fixes-5.2-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs:
  ovl: doc: add non-standard corner cases
  ovl: detect overlapping layers
  ovl: support the FS_IOC_FS[SG]ETXATTR ioctls

8 files changed, 249 insertions, 22 deletions

diff --git a/Documentation/filesystems/overlayfs.txt b/Documentation/filesystems/overlayfs.txt
index eef7d9d259e8..1da2f1668f08 100644
--- a/Documentation/filesystems/overlayfs.txt
+++ b/Documentation/filesystems/overlayfs.txt
@@ -336,8 +336,20 @@ the copied layers will fail the verification of the lower root file handle.
 Non-standard behavior
 ---------------------
 
-Overlayfs can now act as a POSIX compliant filesystem with the following
-features turned on:
+Current version of overlayfs can act as a mostly POSIX compliant
+filesystem.
+
+This is the list of cases that overlayfs doesn't currently handle:
+
+a) POSIX mandates updating st_atime for reads.  This is currently not
+done in the case when the file resides on a lower layer.
+
+b) If a file residing on a lower layer is opened for read-only and then
+memory mapped with MAP_SHARED, then subsequent changes to the file are not
+reflected in the memory mapping.
+
+The following options allow overlayfs to act more like a standards
+compliant filesystem:
 
 1) "redirect_dir"
 
diff --git a/fs/overlayfs/file.c b/fs/overlayfs/file.c
index 540a8b845145..340a6ad45914 100644
--- a/fs/overlayfs/file.c
+++ b/fs/overlayfs/file.c
@@ -426,7 +426,8 @@ static unsigned int ovl_get_inode_flags(struct inode *inode)
 	return ovl_iflags;
 }
 
-static long ovl_ioctl_set_flags(struct file *file, unsigned long arg)
+static long ovl_ioctl_set_flags(struct file *file, unsigned int cmd,
+				unsigned long arg)
 {
 	long ret;
 	struct inode *inode = file_inode(file);
@@ -456,7 +457,7 @@ static long ovl_ioctl_set_flags(struct file *file, unsigned long arg)
 	if (ret)
 		goto unlock;
 
-	ret = ovl_real_ioctl(file, FS_IOC_SETFLAGS, arg);
+	ret = ovl_real_ioctl(file, cmd, arg);
 
 	ovl_copyflags(ovl_inode_real(inode), inode);
 unlock:
@@ -474,11 +475,13 @@ static long ovl_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 
 	switch (cmd) {
 	case FS_IOC_GETFLAGS:
+	case FS_IOC_FSGETXATTR:
 		ret = ovl_real_ioctl(file, cmd, arg);
 		break;
 
 	case FS_IOC_SETFLAGS:
-		ret = ovl_ioctl_set_flags(file, arg);
+	case FS_IOC_FSSETXATTR:
+		ret = ovl_ioctl_set_flags(file, cmd, arg);
 		break;
 
 	default:
diff --git a/fs/overlayfs/inode.c b/fs/overlayfs/inode.c
index b48273e846ad..f7eba21effa5 100644
--- a/fs/overlayfs/inode.c
+++ b/fs/overlayfs/inode.c
@@ -777,6 +777,54 @@ struct inode *ovl_lookup_inode(struct super_block *sb, struct dentry *real,
 	return inode;
 }
 
+bool ovl_lookup_trap_inode(struct super_block *sb, struct dentry *dir)
+{
+	struct inode *key = d_inode(dir);
+	struct inode *trap;
+	bool res;
+
+	trap = ilookup5(sb, (unsigned long) key, ovl_inode_test, key);
+	if (!trap)
+		return false;
+
+	res = IS_DEADDIR(trap) && !ovl_inode_upper(trap) &&
+				  !ovl_inode_lower(trap);
+
+	iput(trap);
+	return res;
+}
+
+/*
+ * Create an inode cache entry for layer root dir, that will intentionally
+ * fail ovl_verify_inode(), so any lookup that will find some layer root
+ * will fail.
+ */
+struct inode *ovl_get_trap_inode(struct super_block *sb, struct dentry *dir)
+{
+	struct inode *key = d_inode(dir);
+	struct inode *trap;
+
+	if (!d_is_dir(dir))
+		return ERR_PTR(-ENOTDIR);
+
+	trap = iget5_locked(sb, (unsigned long) key, ovl_inode_test,
+			    ovl_inode_set, key);
+	if (!trap)
+		return ERR_PTR(-ENOMEM);
+
+	if (!(trap->i_state & I_NEW)) {
+		/* Conflicting layer roots? */
+		iput(trap);
+		return ERR_PTR(-ELOOP);
+	}
+
+	trap->i_mode = S_IFDIR;
+	trap->i_flags = S_DEAD;
+	unlock_new_inode(trap);
+
+	return trap;
+}
+
 /*
  * Does overlay inode need to be hashed by lower inode?
  */
diff --git a/fs/overlayfs/namei.c b/fs/overlayfs/namei.c
index efd372312ef1..badf039267a2 100644
--- a/fs/overlayfs/namei.c
+++ b/fs/overlayfs/namei.c
@@ -18,6 +18,7 @@
 #include "overlayfs.h"
 
 struct ovl_lookup_data {
+	struct super_block *sb;
 	struct qstr name;
 	bool is_dir;
 	bool opaque;
@@ -244,6 +245,12 @@ static int ovl_lookup_single(struct dentry *base, struct ovl_lookup_data *d,
 		if (!d->metacopy || d->last)
 			goto out;
 	} else {
+		if (ovl_lookup_trap_inode(d->sb, this)) {
+			/* Caught in a trap of overlapping layers */
+			err = -ELOOP;
+			goto out_err;
+		}
+
 		if (last_element)
 			d->is_dir = true;
 		if (d->last)
@@ -819,6 +826,7 @@ struct dentry *ovl_lookup(struct inode *dir, struct dentry *dentry,
 	int err;
 	bool metacopy = false;
 	struct ovl_lookup_data d = {
+		.sb = dentry->d_sb,
 		.name = dentry->d_name,
 		.is_dir = false,
 		.opaque = false,
diff --git a/fs/overlayfs/overlayfs.h b/fs/overlayfs/overlayfs.h
index d26efed9f80a..cec40077b522 100644
--- a/fs/overlayfs/overlayfs.h
+++ b/fs/overlayfs/overlayfs.h
@@ -270,6 +270,7 @@ void ovl_clear_flag(unsigned long flag, struct inode *inode);
 bool ovl_test_flag(unsigned long flag, struct inode *inode);
 bool ovl_inuse_trylock(struct dentry *dentry);
 void ovl_inuse_unlock(struct dentry *dentry);
+bool ovl_is_inuse(struct dentry *dentry);
 bool ovl_need_index(struct dentry *dentry);
 int ovl_nlink_start(struct dentry *dentry);
 void ovl_nlink_end(struct dentry *dentry);
@@ -376,6 +377,8 @@ struct ovl_inode_params {
 struct inode *ovl_new_inode(struct super_block *sb, umode_t mode, dev_t rdev);
 struct inode *ovl_lookup_inode(struct super_block *sb, struct dentry *real,
 			       bool is_upper);
+bool ovl_lookup_trap_inode(struct super_block *sb, struct dentry *dir);
+struct inode *ovl_get_trap_inode(struct super_block *sb, struct dentry *dir);
 struct inode *ovl_get_inode(struct super_block *sb,
 			    struct ovl_inode_params *oip);
 static inline void ovl_copyattr(struct inode *from, struct inode *to)
diff --git a/fs/overlayfs/ovl_entry.h b/fs/overlayfs/ovl_entry.h
index ec237035333a..6ed1ace8f8b3 100644
--- a/fs/overlayfs/ovl_entry.h
+++ b/fs/overlayfs/ovl_entry.h
@@ -29,6 +29,8 @@ struct ovl_sb {
 
 struct ovl_layer {
 	struct vfsmount *mnt;
+	/* Trap in ovl inode cache */
+	struct inode *trap;
 	struct ovl_sb *fs;
 	/* Index of this layer in fs root (upper idx == 0) */
 	int idx;
@@ -65,6 +67,10 @@ struct ovl_fs {
 	/* Did we take the inuse lock? */
 	bool upperdir_locked;
 	bool workdir_locked;
+	/* Traps in ovl inode cache */
+	struct inode *upperdir_trap;
+	struct inode *workdir_trap;
+	struct inode *indexdir_trap;
 	/* Inode numbers in all layers do not use the high xino_bits */
 	unsigned int xino_bits;
 };
diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c
index 5ec4fc2f5d7e..746ea36f3171 100644
--- a/fs/overlayfs/super.c
+++ b/fs/overlayfs/super.c
@@ -215,6 +215,9 @@ static void ovl_free_fs(struct ovl_fs *ofs)
 {
 	unsigned i;
 
+	iput(ofs->indexdir_trap);
+	iput(ofs->workdir_trap);
+	iput(ofs->upperdir_trap);
 	dput(ofs->indexdir);
 	dput(ofs->workdir);
 	if (ofs->workdir_locked)
@@ -223,8 +226,10 @@ static void ovl_free_fs(struct ovl_fs *ofs)
 	if (ofs->upperdir_locked)
 		ovl_inuse_unlock(ofs->upper_mnt->mnt_root);
 	mntput(ofs->upper_mnt);
-	for (i = 0; i < ofs->numlower; i++)
+	for (i = 0; i < ofs->numlower; i++) {
+		iput(ofs->lower_layers[i].trap);
 		mntput(ofs->lower_layers[i].mnt);
+	}
 	for (i = 0; i < ofs->numlowerfs; i++)
 		free_anon_bdev(ofs->lower_fs[i].pseudo_dev);
 	kfree(ofs->lower_layers);
@@ -983,7 +988,26 @@ static const struct xattr_handler *ovl_xattr_handlers[] = {
 	NULL
 };
 
-static int ovl_get_upper(struct ovl_fs *ofs, struct path *upperpath)
+static int ovl_setup_trap(struct super_block *sb, struct dentry *dir,
+			  struct inode **ptrap, const char *name)
+{
+	struct inode *trap;
+	int err;
+
+	trap = ovl_get_trap_inode(sb, dir);
+	err = PTR_ERR(trap);
+	if (IS_ERR(trap)) {
+		if (err == -ELOOP)
+			pr_err("overlayfs: conflicting %s path\n", name);
+		return err;
+	}
+
+	*ptrap = trap;
+	return 0;
+}
+
+static int ovl_get_upper(struct super_block *sb, struct ovl_fs *ofs,
+			 struct path *upperpath)
 {
 	struct vfsmount *upper_mnt;
 	int err;
@@ -1003,6 +1027,11 @@ static int ovl_get_upper(struct ovl_fs *ofs, struct path *upperpath)
 	if (err)
 		goto out;
 
+	err = ovl_setup_trap(sb, upperpath->dentry, &ofs->upperdir_trap,
+			     "upperdir");
+	if (err)
+		goto out;
+
 	upper_mnt = clone_private_mount(upperpath);
 	err = PTR_ERR(upper_mnt);
 	if (IS_ERR(upper_mnt)) {
@@ -1029,7 +1058,8 @@ out:
 	return err;
 }
 
-static int ovl_make_workdir(struct ovl_fs *ofs, struct path *workpath)
+static int ovl_make_workdir(struct super_block *sb, struct ovl_fs *ofs,
+			    struct path *workpath)
 {
 	struct vfsmount *mnt = ofs->upper_mnt;
 	struct dentry *temp;
@@ -1044,6 +1074,10 @@ static int ovl_make_workdir(struct ovl_fs *ofs, struct path *workpath)
 	if (!ofs->workdir)
 		goto out;
 
+	err = ovl_setup_trap(sb, ofs->workdir, &ofs->workdir_trap, "workdir");
+	if (err)
+		goto out;
+
 	/*
 	 * Upper should support d_type, else whiteouts are visible.  Given
 	 * workdir and upper are on same fs, we can do iterate_dir() on
@@ -1104,7 +1138,8 @@ out:
 	return err;
 }
 
-static int ovl_get_workdir(struct ovl_fs *ofs, struct path *upperpath)
+static int ovl_get_workdir(struct super_block *sb, struct ovl_fs *ofs,
+			   struct path *upperpath)
 {
 	int err;
 	struct path workpath = { };
@@ -1135,19 +1170,16 @@ static int ovl_get_workdir(struct ovl_fs *ofs, struct path *upperpath)
 		pr_warn("overlayfs: workdir is in-use by another mount, accessing files from both mounts will result in undefined behavior.\n");
 	}
 
-	err = ovl_make_workdir(ofs, &workpath);
-	if (err)
-		goto out;
+	err = ovl_make_workdir(sb, ofs, &workpath);
 
-	err = 0;
 out:
 	path_put(&workpath);
 
 	return err;
 }
 
-static int ovl_get_indexdir(struct ovl_fs *ofs, struct ovl_entry *oe,
-			    struct path *upperpath)
+static int ovl_get_indexdir(struct super_block *sb, struct ovl_fs *ofs,
+			    struct ovl_entry *oe, struct path *upperpath)
 {
 	struct vfsmount *mnt = ofs->upper_mnt;
 	int err;
@@ -1166,6 +1198,11 @@ static int ovl_get_indexdir(struct ovl_fs *ofs, struct ovl_entry *oe,
 
 	ofs->indexdir = ovl_workdir_create(ofs, OVL_INDEXDIR_NAME, true);
 	if (ofs->indexdir) {
+		err = ovl_setup_trap(sb, ofs->indexdir, &ofs->indexdir_trap,
+				     "indexdir");
+		if (err)
+			goto out;
+
 		/*
 		 * Verify upper root is exclusively associated with index dir.
 		 * Older kernels stored upper fh in "trusted.overlay.origin"
@@ -1253,8 +1290,8 @@ static int ovl_get_fsid(struct ovl_fs *ofs, const struct path *path)
 	return ofs->numlowerfs;
 }
 
-static int ovl_get_lower_layers(struct ovl_fs *ofs, struct path *stack,
-				unsigned int numlower)
+static int ovl_get_lower_layers(struct super_block *sb, struct ovl_fs *ofs,
+				struct path *stack, unsigned int numlower)
 {
 	int err;
 	unsigned int i;
@@ -1272,16 +1309,28 @@ static int ovl_get_lower_layers(struct ovl_fs *ofs, struct path *stack,
 
 	for (i = 0; i < numlower; i++) {
 		struct vfsmount *mnt;
+		struct inode *trap;
 		int fsid;
 
 		err = fsid = ovl_get_fsid(ofs, &stack[i]);
 		if (err < 0)
 			goto out;
 
+		err = -EBUSY;
+		if (ovl_is_inuse(stack[i].dentry)) {
+			pr_err("overlayfs: lowerdir is in-use as upperdir/workdir\n");
+			goto out;
+		}
+
+		err = ovl_setup_trap(sb, stack[i].dentry, &trap, "lowerdir");
+		if (err)
+			goto out;
+
 		mnt = clone_private_mount(&stack[i]);
 		err = PTR_ERR(mnt);
 		if (IS_ERR(mnt)) {
 			pr_err("overlayfs: failed to clone lowerpath\n");
+			iput(trap);
 			goto out;
 		}
 
@@ -1291,6 +1340,7 @@ static int ovl_get_lower_layers(struct ovl_fs *ofs, struct path *stack,
 		 */
 		mnt->mnt_flags |= MNT_READONLY | MNT_NOATIME;
 
+		ofs->lower_layers[ofs->numlower].trap = trap;
 		ofs->lower_layers[ofs->numlower].mnt = mnt;
 		ofs->lower_layers[ofs->numlower].idx = i + 1;
 		ofs->lower_layers[ofs->numlower].fsid = fsid;
@@ -1385,7 +1435,7 @@ static struct ovl_entry *ovl_get_lowerstack(struct super_block *sb,
 		goto out_err;
 	}
 
-	err = ovl_get_lower_layers(ofs, stack, numlower);
+	err = ovl_get_lower_layers(sb, ofs, stack, numlower);
 	if (err)
 		goto out_err;
 
@@ -1417,6 +1467,85 @@ out_err:
 	goto out;
 }
 
+/*
+ * Check if this layer root is a descendant of:
+ * - another layer of this overlayfs instance
+ * - upper/work dir of any overlayfs instance
+ * - a disconnected dentry (detached root)
+ */
+static int ovl_check_layer(struct super_block *sb, struct dentry *dentry,
+			   const char *name)
+{
+	struct dentry *next, *parent;
+	bool is_root = false;
+	int err = 0;
+
+	if (!dentry || dentry == dentry->d_sb->s_root)
+		return 0;
+
+	next = dget(dentry);
+	/* Walk back ancestors to fs root (inclusive) looking for traps */
+	do {
+		parent = dget_parent(next);
+		is_root = (parent == next);
+		if (ovl_is_inuse(parent)) {
+			err = -EBUSY;
+			pr_err("overlayfs: %s path overlapping in-use upperdir/workdir\n",
+			       name);
+		} else if (ovl_lookup_trap_inode(sb, parent)) {
+			err = -ELOOP;
+			pr_err("overlayfs: overlapping %s path\n", name);
+		}
+		dput(next);
+		next = parent;
+	} while (!err && !is_root);
+
+	/* Did we really walk to fs root or found a detached root? */
+	if (!err && next != dentry->d_sb->s_root) {
+		err = -ESTALE;
+		pr_err("overlayfs: disconnected %s path\n", name);
+	}
+
+	dput(next);
+
+	return err;
+}
+
+/*
+ * Check if any of the layers or work dirs overlap.
+ */
+static int ovl_check_overlapping_layers(struct super_block *sb,
+					struct ovl_fs *ofs)
+{
+	int i, err;
+
+	if (ofs->upper_mnt) {
+		err = ovl_check_layer(sb, ofs->upper_mnt->mnt_root, "upperdir");
+		if (err)
+			return err;
+
+		/*
+		 * Checking workbasedir avoids hitting ovl_is_inuse(parent) of
+		 * this instance and covers overlapping work and index dirs,
+		 * unless work or index dir have been moved since created inside
+		 * workbasedir.  In that case, we already have their traps in
+		 * inode cache and we will catch that case on lookup.
+		 */
+		err = ovl_check_layer(sb, ofs->workbasedir, "workdir");
+		if (err)
+			return err;
+	}
+
+	for (i = 0; i < ofs->numlower; i++) {
+		err = ovl_check_layer(sb, ofs->lower_layers[i].mnt->mnt_root,
+				      "lowerdir");
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
 static int ovl_fill_super(struct super_block *sb, void *data, int silent)
 {
 	struct path upperpath = { };
@@ -1456,17 +1585,20 @@ static int ovl_fill_super(struct super_block *sb, void *data, int silent)
 	if (ofs->config.xino != OVL_XINO_OFF)
 		ofs->xino_bits = BITS_PER_LONG - 32;
 
+	/* alloc/destroy_inode needed for setting up traps in inode cache */
+	sb->s_op = &ovl_super_operations;
+
 	if (ofs->config.upperdir) {
 		if (!ofs->config.workdir) {
 			pr_err("overlayfs: missing 'workdir'\n");
 			goto out_err;
 		}
 
-		err = ovl_get_upper(ofs, &upperpath);
+		err = ovl_get_upper(sb, ofs, &upperpath);
 		if (err)
 			goto out_err;
 
-		err = ovl_get_workdir(ofs, &upperpath);
+		err = ovl_get_workdir(sb, ofs, &upperpath);
 		if (err)
 			goto out_err;
 
@@ -1487,7 +1619,7 @@ static int ovl_fill_super(struct super_block *sb, void *data, int silent)
 		sb->s_flags |= SB_RDONLY;
 
 	if (!(ovl_force_readonly(ofs)) && ofs->config.index) {
-		err = ovl_get_indexdir(ofs, oe, &upperpath);
+		err = ovl_get_indexdir(sb, ofs, oe, &upperpath);
 		if (err)
 			goto out_free_oe;
 
@@ -1500,6 +1632,10 @@ static int ovl_fill_super(struct super_block *sb, void *data, int silent)
 
 	}
 
+	err = ovl_check_overlapping_layers(sb, ofs);
+	if (err)
+		goto out_free_oe;
+
 	/* Show index=off in /proc/mounts for forced r/o mount */
 	if (!ofs->indexdir) {
 		ofs->config.index = false;
@@ -1521,7 +1657,6 @@ static int ovl_fill_super(struct super_block *sb, void *data, int silent)
 	cap_lower(cred->cap_effective, CAP_SYS_RESOURCE);
 
 	sb->s_magic = OVERLAYFS_SUPER_MAGIC;
-	sb->s_op = &ovl_super_operations;
 	sb->s_xattr = ovl_xattr_handlers;
 	sb->s_fs_info = ofs;
 	sb->s_flags |= SB_POSIXACL;
diff --git a/fs/overlayfs/util.c b/fs/overlayfs/util.c
index 4035e640f402..e135064e87ad 100644
--- a/fs/overlayfs/util.c
+++ b/fs/overlayfs/util.c
@@ -652,6 +652,18 @@ void ovl_inuse_unlock(struct dentry *dentry)
 	}
 }
 
+bool ovl_is_inuse(struct dentry *dentry)
+{
+	struct inode *inode = d_inode(dentry);
+	bool inuse;
+
+	spin_lock(&inode->i_lock);
+	inuse = (inode->i_state & I_OVL_INUSE);
+	spin_unlock(&inode->i_lock);
+
+	return inuse;
+}
+
 /*
  * Does this overlay dentry need to be indexed on copy up?
  */