linux-kernel - Re: [RFC 1/1] shiftfs: uid/gid shifting bind mount

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 17 Feb 2017 09:24:40 -0800
From:   James Bottomley <James.Bottomley@...senPartnership.com>
To:     Al Viro <viro@...IV.linux.org.uk>
Cc:     Djalal Harouni <tixxdz@...il.com>, Chris Mason <clm@...com>,
        Theodore Tso <tytso@....edu>,
        Josh Triplett <josh@...htriplett.org>,
        "Eric W. Biederman" <ebiederm@...ssion.com>,
        Andy Lutomirski <luto@...nel.org>,
        Seth Forshee <seth.forshee@...onical.com>,
        linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-security-module@...r.kernel.org,
        Dongsu Park <dongsu@...ocode.com>,
        David Herrmann <dh.herrmann@...glemail.com>,
        Miklos Szeredi <mszeredi@...hat.com>,
        Alban Crequy <alban.crequy@...il.com>,
        "Serge E. Hallyn" <serge@...lyn.com>, Phil Estes <estesp@...il.com>
Subject: Re: [RFC 1/1] shiftfs: uid/gid shifting bind mount

On Fri, 2017-02-17 at 02:29 +0000, Al Viro wrote:
> On Sat, Feb 04, 2017 at 11:19:32AM -0800, James Bottomley wrote:
> 
> > +static const struct dentry_operations shiftfs_dentry_ops = {
> > +	.d_release	= shiftfs_d_release,
> > +	.d_real		= shiftfs_d_real,
> > +};
> 
> In other words, those dentries are *never* revalidated.  Nevermind 
> that underlying fs might be mounted elsewhere and be actively 
> modified under you.
> 
> > +static struct dentry *shiftfs_lookup(struct inode *dir, struct
> > dentry *dentry,
> > +				     unsigned int flags)
> > +{
> > +	struct dentry *real = dir->i_private, *new;
> > +	struct inode *reali = real->d_inode, *newi;
> > +	const struct cred *oldcred, *newcred;
> > +
> > +	inode_lock(reali);
> > +	oldcred = shiftfs_new_creds(&newcred, dentry->d_sb);
> > +	new = lookup_one_len(dentry->d_name.name, real, dentry
> > ->d_name.len);
> > +	shiftfs_old_creds(oldcred, &newcred);
> > +	inode_unlock(reali);
> > +
> > +	if (IS_ERR(new))
> > +		return new;
> > +
> > +	dentry->d_fsdata = new;
> > +
> > +	if (!new->d_inode)
> > +		return NULL;
> 
> What happens when somebody comes along and creates the damn thing on 
> the underlying fs?  _Not_ via your code, that is - using the 
> underlying fs mounted elsewhere.

Point taken.  This, I think fixes the dcache revalidation issue.

James

---

diff --git a/fs/shiftfs.c b/fs/shiftfs.c
index a4a1f98..1e71efe 100644
--- a/fs/shiftfs.c
+++ b/fs/shiftfs.c
@@ -118,9 +118,43 @@ static struct dentry *shiftfs_d_real(struct dentry *dentry,
 	return real;
 }
 
+static int shiftfs_d_weak_revalidate(struct dentry *dentry, unsigned int flags)
+{
+	struct dentry *real = dentry->d_fsdata;
+
+	if (d_unhashed(real))
+		return 0;
+
+	if (!(real->d_flags & DCACHE_OP_WEAK_REVALIDATE))
+		return 1;
+
+	return real->d_op->d_weak_revalidate(real, flags);
+}
+
+static int shiftfs_d_revalidate(struct dentry *dentry, unsigned int flags)
+{
+	struct dentry *real = dentry->d_fsdata;
+	int ret;
+
+	if (d_unhashed(real))
+		return 0;
+
+	if (!(real->d_flags & DCACHE_OP_REVALIDATE))
+		return 1;
+
+	ret = real->d_op->d_revalidate(real, flags);
+
+	if (ret == 0 && !(flags & LOOKUP_RCU))
+		d_invalidate(real);
+
+	return ret;
+}
+
 static const struct dentry_operations shiftfs_dentry_ops = {
 	.d_release	= shiftfs_d_release,
 	.d_real		= shiftfs_d_real,
+	.d_revalidate	= shiftfs_d_revalidate,
+	.d_weak_revalidate = shiftfs_d_weak_revalidate,
 };
 
 static int shiftfs_readlink(struct dentry *dentry, char __user *data,
@@ -431,9 +465,7 @@ static struct dentry *shiftfs_lookup(struct inode *dir, struct dentry *dentry,
 		return ERR_PTR(-ENOMEM);
 	}
 
-	d_splice_alias(newi, dentry);
-
-	return NULL;
+	return d_splice_alias(newi, dentry);
 }
 
 static int shiftfs_permission(struct inode *inode, int mask)