[PATCH] overlayfs: copy up i_uid/i_gid from the underlying inode

[PATCH] overlayfs: copy up i_uid/i_gid from the underlying inode

[Andy Whitcroft]

YAMA et al rely on on i_uid/i_gid to be populated in order to perform
their checks.  While these really cannot be guarenteed as the underlying
filesystem may not even have the concept, they are expected to be filled
when possible.  To quote Al Viro:

    "Ideally, yes, we'd want to have ->i_uid used only by fs-specific
     code and helpers used by that fs (including those that are
     implicit defaults). [...]   In practice we have enough places
     where uid/gid is used directly to make setting them practically
     a requirement - places like /proc/<pid>/ can get away with
     not doing that, but only because shitloads of syscalls are
     not allowed on those anyway, permissions or no permissions.
     In anything general-purpose you really need to set it."

Copy up the underlying filesystem information into the overlayfs inode
when we create it.

BugLink: http://bugs.launchpad.net/bugs/944386
Signed-off-by: Andy Whitcroft <apw@canonical.com>
---
 fs/overlayfs/dir.c       |    2 ++
 fs/overlayfs/overlayfs.h |    6 ++++++
 fs/overlayfs/super.c     |    1 +
 3 files changed, 9 insertions(+)

	After a long hiatus I have had time to look into the issues
	highlighted by the i_uid/i_gid requirements from the VFS.
	I have identified a number of places which definatly did need
	the ids copying up and those are reflected in the patch below.
	I am not 100% convinced I have hit all of the places this might
	be needed but it cirtainly helps with the issues I was seeing
	with link and YAMA (which given YAMA is now gaining the link
	constraints in mainline in v3.6 we will see more issues here).
	were seeing and identify the places where

	Please consider for overlayfs.

	-apw

diff --git a/fs/overlayfs/dir.c b/fs/overlayfs/dir.c
index c914c97..084e527 100644
--- a/fs/overlayfs/dir.c
+++ b/fs/overlayfs/dir.c
@@ -304,6 +304,7 @@ static int ovl_create_object(struct dentry *dentry, int mode, dev_t rdev,
 		}
 	}
 	ovl_dentry_update(dentry, newdentry);
+	ovl_copyattr(newdentry->d_inode, inode);
 	d_instantiate(dentry, inode);
 	inode = NULL;
 	newdentry = NULL;
@@ -446,6 +447,7 @@ static int ovl_link(struct dentry *old, struct inode *newdir,
 				new->d_fsdata);
 		if (!newinode)
 			goto link_fail;
+		ovl_copyattr(upperdir->d_inode, newinode);
 
 		ovl_dentry_version_inc(new->d_parent);
 		ovl_dentry_update(new, newdentry);
diff --git a/fs/overlayfs/overlayfs.h b/fs/overlayfs/overlayfs.h
index 1dd05f7..3495a55 100644
--- a/fs/overlayfs/overlayfs.h
+++ b/fs/overlayfs/overlayfs.h
@@ -56,6 +56,12 @@ int ovl_removexattr(struct dentry *dentry, const char *name);
 
 struct inode *ovl_new_inode(struct super_block *sb, umode_t mode,
 			    struct ovl_entry *oe);
+static inline void ovl_copyattr(struct inode *from, struct inode *to)
+{
+	to->i_uid = from->i_uid;
+	to->i_gid = from->i_gid;
+}
+
 /* dir.c */
 extern const struct inode_operations ovl_dir_inode_operations;
 
diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c
index 1d2d1e2..23cac54 100644
--- a/fs/overlayfs/super.c
+++ b/fs/overlayfs/super.c
@@ -343,6 +343,7 @@ static int ovl_do_lookup(struct dentry *dentry)
 				      oe);
 		if (!inode)
 			goto out_dput;
+		ovl_copyattr(realdentry->d_inode, inode);
 	}
 
 	if (upperdentry)
-- 
1.7.9.5

Re: [PATCH] overlayfs: copy up i_uid/i_gid from the underlying inode

[Miklos Szeredi]

Andy Whitcroft <apw@canonical.com> writes:
> 	After a long hiatus I have had time to look into the issues
> 	highlighted by the i_uid/i_gid requirements from the VFS.
> 	I have identified a number of places which definatly did need
> 	the ids copying up and those are reflected in the patch below.
> 	I am not 100% convinced I have hit all of the places this might
> 	be needed but it cirtainly helps with the issues I was seeing
> 	with link and YAMA (which given YAMA is now gaining the link
> 	constraints in mainline in v3.6 we will see more issues here).
> 	were seeing and identify the places where
>
> 	Please consider for overlayfs.

ovl_setattr() also needs this, I think.

Updated patch below.  Also pushed overlayfs.v14 with this patch to:

  git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs.git overlayfs.v14


Thanks,
Miklos
----


>From 1182ebc7718994c66f3a966a90ed861aea2a5d01 Mon Sep 17 00:00:00 2001
From: Andy Whitcroft <apw@canonical.com>
Date: Thu, 9 Aug 2012 16:47:21 +0100
Subject: [PATCH] overlayfs: copy up i_uid/i_gid from the underlying inode

YAMA et al rely on on i_uid/i_gid to be populated in order to perform
their checks.  While these really cannot be guarenteed as the underlying
filesystem may not even have the concept, they are expected to be filled
when possible.  To quote Al Viro:

    "Ideally, yes, we'd want to have ->i_uid used only by fs-specific
     code and helpers used by that fs (including those that are
     implicit defaults). [...]   In practice we have enough places
     where uid/gid is used directly to make setting them practically
     a requirement - places like /proc/<pid>/ can get away with
     not doing that, but only because shitloads of syscalls are
     not allowed on those anyway, permissions or no permissions.
     In anything general-purpose you really need to set it."

Copy up the underlying filesystem information into the overlayfs inode
when we create it.

BugLink: http://bugs.launchpad.net/bugs/944386
Signed-off-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
---
 fs/overlayfs/dir.c       |    2 ++
 fs/overlayfs/inode.c     |    2 ++
 fs/overlayfs/overlayfs.h |    6 ++++++
 fs/overlayfs/super.c     |    1 +
 4 files changed, 11 insertions(+), 0 deletions(-)

diff --git a/fs/overlayfs/dir.c b/fs/overlayfs/dir.c
index 40650c4..c4446c4 100644
--- a/fs/overlayfs/dir.c
+++ b/fs/overlayfs/dir.c
@@ -304,6 +304,7 @@ static int ovl_create_object(struct dentry *dentry, int mode, dev_t rdev,
 		}
 	}
 	ovl_dentry_update(dentry, newdentry);
+	ovl_copyattr(newdentry->d_inode, inode);
 	d_instantiate(dentry, inode);
 	inode = NULL;
 	newdentry = NULL;
@@ -446,6 +447,7 @@ static int ovl_link(struct dentry *old, struct inode *newdir,
 				new->d_fsdata);
 		if (!newinode)
 			goto link_fail;
+		ovl_copyattr(upperdir->d_inode, newinode);
 
 		ovl_dentry_version_inc(new->d_parent);
 		ovl_dentry_update(new, newdentry);
diff --git a/fs/overlayfs/inode.c b/fs/overlayfs/inode.c
index f3a534f..e7ab09b 100644
--- a/fs/overlayfs/inode.c
+++ b/fs/overlayfs/inode.c
@@ -31,6 +31,8 @@ int ovl_setattr(struct dentry *dentry, struct iattr *attr)
 
 	mutex_lock(&upperdentry->d_inode->i_mutex);
 	err = notify_change(upperdentry, attr);
+	if (!err)
+		ovl_copyattr(upperdentry->d_inode, dentry->d_inode);
 	mutex_unlock(&upperdentry->d_inode->i_mutex);
 
 	return err;
diff --git a/fs/overlayfs/overlayfs.h b/fs/overlayfs/overlayfs.h
index fe1241d..1cba38f 100644
--- a/fs/overlayfs/overlayfs.h
+++ b/fs/overlayfs/overlayfs.h
@@ -56,6 +56,12 @@ int ovl_removexattr(struct dentry *dentry, const char *name);
 
 struct inode *ovl_new_inode(struct super_block *sb, umode_t mode,
 			    struct ovl_entry *oe);
+static inline void ovl_copyattr(struct inode *from, struct inode *to)
+{
+	to->i_uid = from->i_uid;
+	to->i_gid = from->i_gid;
+}
+
 /* dir.c */
 extern const struct inode_operations ovl_dir_inode_operations;
 
diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c
index 64d2695..9808408 100644
--- a/fs/overlayfs/super.c
+++ b/fs/overlayfs/super.c
@@ -343,6 +343,7 @@ static int ovl_do_lookup(struct dentry *dentry)
 				      oe);
 		if (!inode)
 			goto out_dput;
+		ovl_copyattr(realdentry->d_inode, inode);
 	}
 
 	if (upperdentry)
-- 
1.7.7