[PATCH 09/12] HPPFS: add special permission checking from procfs

[Paolo 'Blaisorblade' Giarrusso]

From: Paolo 'Blaisorblade' Giarrusso <blaisorblade@xxxxxxxx>, Al Viro

Procfs uses a special permission method in some cases, so we can't use the
generic one, and must call instead the inode's one if it's not null (only
some inodes use a custom method).

It seems that this method is used to avoid a chroot'ed process to escape
its chroot by accessing another process's root/cwd through procfs (both by
a quick read of the code and by testing on the host), like in "cd
/proc/1/root".

I've not added a ->permission operation on symlink, as there is no procfs
symlink using it, and I wonder if symlink can have permissions set at all
(guess not).

If you disagree, just add it (hppfs_permission() handles null pointers
gracefully).

Note: I've cared to use the proc inode (I missed that in the beginning) -
since the dentry is always created by the underlying ->lookup this should be
safe - or not? I've gone to do some checking, and since we pass to it a
procfs inode, referring to the procfs superblock, it seems that yes, we're
fine.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@xxxxxxxx>
---

 fs/hppfs/hppfs_kern.c |   16 ++++++++++++++++
 1 files changed, 16 insertions(+), 0 deletions(-)

diff --git a/fs/hppfs/hppfs_kern.c b/fs/hppfs/hppfs_kern.c
--- a/fs/hppfs/hppfs_kern.c
+++ b/fs/hppfs/hppfs_kern.c
@@ -214,7 +214,22 @@ static struct dentry *hppfs_lookup(struc
 	return(ERR_PTR(err));
 }
 
+static int hppfs_permission(struct inode *inode, int mask, struct nameidata *nd)
+{
+	struct inode *proc_inode;
+	int (*permission) (struct inode *, int, struct nameidata *);
+
+	proc_inode = HPPFS_I(inode)->proc_dentry->d_inode;
+	permission = proc_inode->i_op->permission;
+
+	if (permission == NULL)
+		return generic_permission(inode, mask, NULL);
+
+	return (*permission)(proc_inode, mask, nd);
+}
+
 static struct inode_operations hppfs_file_iops = {
+	.permission	= hppfs_permission,
 };
 
 /* Pass down ppos when you're being called from userspace.
@@ -779,6 +794,7 @@ static void* hppfs_follow_link(struct de
 }
 
 static struct inode_operations hppfs_dir_iops = {
+	.permission	= hppfs_permission,
 	.lookup		= hppfs_lookup,
 };
 

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/