[08/12] vfs: allow unprivileged whiteout creation
diff mbox series

Message ID 20191128155940.17530-9-mszeredi@redhat.com
State New
Headers show
Series
  • various vfs patches
Related show

Commit Message

Miklos Szeredi Nov. 28, 2019, 3:59 p.m. UTC
Whiteouts are special, but unlike real device nodes they should not require
privileges to create.

The 0 char device number should already be reserved, but make this explicit
in cdev_add() to be on the safe side.

Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
---
 fs/char_dev.c                 |  3 +++
 fs/namei.c                    | 17 ++++-------------
 include/linux/device_cgroup.h |  3 +++
 3 files changed, 10 insertions(+), 13 deletions(-)

Comments

Al Viro Dec. 17, 2019, 3:51 a.m. UTC | #1
On Thu, Nov 28, 2019 at 04:59:36PM +0100, Miklos Szeredi wrote:
> Whiteouts are special, but unlike real device nodes they should not require
> privileges to create.

More detailed analysis, please.  Maybe I'm missing something obvious,
but I don't see off-hand why it's safe.
Miklos Szeredi Dec. 17, 2019, 4:22 a.m. UTC | #2
On Tue, Dec 17, 2019 at 4:51 AM Al Viro <viro@zeniv.linux.org.uk> wrote:
>
> On Thu, Nov 28, 2019 at 04:59:36PM +0100, Miklos Szeredi wrote:
> > Whiteouts are special, but unlike real device nodes they should not require
> > privileges to create.
>
> More detailed analysis, please.  Maybe I'm missing something obvious,
> but I don't see off-hand why it's safe.

The general concern with device nodes is that opening them can have
side effects.  The kernel already avoids zero major, so opening will
result in ENODEV, but the patch also makes sure that registering a
device with 0/0 number is forbidden (fs/char_dev.c).

Thanks,
Miklos

Patch
diff mbox series

diff --git a/fs/char_dev.c b/fs/char_dev.c
index 00dfe17871ac..8bf66f40e5e0 100644
--- a/fs/char_dev.c
+++ b/fs/char_dev.c
@@ -483,6 +483,9 @@  int cdev_add(struct cdev *p, dev_t dev, unsigned count)
 	p->dev = dev;
 	p->count = count;
 
+	if (WARN_ON(dev == WHITEOUT_DEV))
+		return -EBUSY;
+
 	error = kobj_map(cdev_map, dev, count, NULL,
 			 exact_match, exact_lock, p);
 	if (error)
diff --git a/fs/namei.c b/fs/namei.c
index 671c3c1a3425..05ca98595b62 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -3687,12 +3687,14 @@  EXPORT_SYMBOL(user_path_create);
 
 int vfs_mknod(struct inode *dir, struct dentry *dentry, umode_t mode, dev_t dev)
 {
+	bool is_whiteout = S_ISCHR(mode) && dev == WHITEOUT_DEV;
 	int error = may_create(dir, dentry);
 
 	if (error)
 		return error;
 
-	if ((S_ISCHR(mode) || S_ISBLK(mode)) && !capable(CAP_MKNOD))
+	if ((S_ISCHR(mode) || S_ISBLK(mode)) && !capable(CAP_MKNOD) &&
+	    !is_whiteout)
 		return -EPERM;
 
 	if (!dir->i_op->mknod)
@@ -4527,9 +4529,6 @@  static int do_renameat2(int olddfd, const char __user *oldname, int newdfd,
 	    (flags & RENAME_EXCHANGE))
 		return -EINVAL;
 
-	if ((flags & RENAME_WHITEOUT) && !capable(CAP_MKNOD))
-		return -EPERM;
-
 	if (flags & RENAME_EXCHANGE)
 		target_flags = 0;
 
@@ -4667,15 +4666,7 @@  SYSCALL_DEFINE2(rename, const char __user *, oldname, const char __user *, newna
 
 int vfs_whiteout(struct inode *dir, struct dentry *dentry)
 {
-	int error = may_create(dir, dentry);
-	if (error)
-		return error;
-
-	if (!dir->i_op->mknod)
-		return -EPERM;
-
-	return dir->i_op->mknod(dir, dentry,
-				S_IFCHR | WHITEOUT_MODE, WHITEOUT_DEV);
+	return vfs_mknod(dir, dentry, S_IFCHR | WHITEOUT_MODE, WHITEOUT_DEV);
 }
 EXPORT_SYMBOL(vfs_whiteout);
 
diff --git a/include/linux/device_cgroup.h b/include/linux/device_cgroup.h
index 8557efe096dc..fc989487c273 100644
--- a/include/linux/device_cgroup.h
+++ b/include/linux/device_cgroup.h
@@ -62,6 +62,9 @@  static inline int devcgroup_inode_mknod(int mode, dev_t dev)
 	if (!S_ISBLK(mode) && !S_ISCHR(mode))
 		return 0;
 
+	if (S_ISCHR(mode) && dev == WHITEOUT_DEV)
+		return 0;
+
 	if (S_ISBLK(mode))
 		type = DEVCG_DEV_BLOCK;
 	else