Merge branch 'enforce-global-flow-dissector'

Stanislav Fomichev says:

====================
While having a per-net-ns flow dissector programs is convenient for
testing, security-wise it's better to have only one vetted global
flow dissector implementation.

Let's have a convention that when BPF flow dissector is installed
in the root namespace, child namespaces can't override it.

The intended use-case is to attach global BPF flow dissector
early from the init scripts/systemd. Attaching global dissector
is prohibited if some non-root namespace already has flow dissector
attached. Also, attaching to non-root namespace is prohibited
when there is flow dissector attached to the root namespace.

v3:
* drop extra check and empty line (Andrii Nakryiko)

v2:
* EPERM -> EEXIST (Song Liu)
* Make sure we don't have dissector attached to non-root namespaces
  when attaching the global one (Andrii Nakryiko)
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Documentation/bpf/prog_flow_dissector.rst

@@ -142,3 +142,6 @@ BPF flow dissector doesn't support exporting all the metadata that in-kernel
 C-based implementation can export. Notable example is single VLAN (802.1Q)
 and double VLAN (802.1AD) tags. Please refer to the ``struct bpf_flow_keys``
 for a set of information that's currently can be exported from the BPF context.
+
+When BPF flow dissector is attached to the root network namespace (machine-wide
+policy), users can't override it in their child network namespaces.

net/core/flow_dissector.c

@@ -114,19 +114,46 @@ int skb_flow_dissector_bpf_prog_attach(const union bpf_attr *attr,
 {
 	struct bpf_prog *attached;
 	struct net *net;
+	int ret = 0;
 
 	net = current->nsproxy->net_ns;
 	mutex_lock(&flow_dissector_mutex);
+
+	if (net == &init_net) {
+		/* BPF flow dissector in the root namespace overrides
+		 * any per-net-namespace one. When attaching to root,
+		 * make sure we don't have any BPF program attached
+		 * to the non-root namespaces.
+		 */
+		struct net *ns;
+
+		for_each_net(ns) {
+			if (rcu_access_pointer(ns->flow_dissector_prog)) {
+				ret = -EEXIST;
+				goto out;
+			}
+		}
+	} else {
+		/* Make sure root flow dissector is not attached
+		 * when attaching to the non-root namespace.
+		 */
+		if (rcu_access_pointer(init_net.flow_dissector_prog)) {
+			ret = -EEXIST;
+			goto out;
+		}
+	}
+
 	attached = rcu_dereference_protected(net->flow_dissector_prog,
 					     lockdep_is_held(&flow_dissector_mutex));
 	if (attached) {
 		/* Only one BPF program can be attached at a time */
-		mutex_unlock(&flow_dissector_mutex);
-		return -EEXIST;
+		ret = -EEXIST;
+		goto out;
 	}
 	rcu_assign_pointer(net->flow_dissector_prog, prog);
+out:
 	mutex_unlock(&flow_dissector_mutex);
-	return 0;
+	return ret;
 }
 
 int skb_flow_dissector_bpf_prog_detach(const union bpf_attr *attr)
@@ -910,7 +937,10 @@ bool __skb_flow_dissect(const struct net *net,
 	WARN_ON_ONCE(!net);
 	if (net) {
 		rcu_read_lock();
-		attached = rcu_dereference(net->flow_dissector_prog);
+		attached = rcu_dereference(init_net.flow_dissector_prog);
+
+		if (!attached)
+			attached = rcu_dereference(net->flow_dissector_prog);
 
 		if (attached) {
 			struct bpf_flow_keys flow_keys;

tools/testing/selftests/bpf/test_flow_dissector.sh

@@ -18,19 +18,55 @@ fi
 # this is the case and run it with in_netns.sh if it is being run in the root
 # namespace.
 if [[ -z $(ip netns identify $$) ]]; then
+	err=0
+	if bpftool="$(which bpftool)"; then
+		echo "Testing global flow dissector..."
+
+		$bpftool prog loadall ./bpf_flow.o /sys/fs/bpf/flow \
+			type flow_dissector
+
+		if ! unshare --net $bpftool prog attach pinned \
+			/sys/fs/bpf/flow/flow_dissector flow_dissector; then
+			echo "Unexpected unsuccessful attach in namespace" >&2
+			err=1
+		fi
+
+		$bpftool prog attach pinned /sys/fs/bpf/flow/flow_dissector \
+			flow_dissector
+
+		if unshare --net $bpftool prog attach pinned \
+			/sys/fs/bpf/flow/flow_dissector flow_dissector; then
+			echo "Unexpected successful attach in namespace" >&2
+			err=1
+		fi
+
+		if ! $bpftool prog detach pinned \
+			/sys/fs/bpf/flow/flow_dissector flow_dissector; then
+			echo "Failed to detach flow dissector" >&2
+			err=1
+		fi
+
+		rm -rf /sys/fs/bpf/flow
+	else
+		echo "Skipping root flow dissector test, bpftool not found" >&2
+	fi
+
+	# Run the rest of the tests in a net namespace.
 	../net/in_netns.sh "$0" "$@"
-	exit $?
-fi
+	err=$(( $err + $? ))
 
-# Determine selftest success via shell exit code
-exit_handler()
-{
-	if (( $? == 0 )); then
+	if (( $err == 0 )); then
 		echo "selftests: $TESTNAME [PASS]";
 	else
 		echo "selftests: $TESTNAME [FAILED]";
 	fi
 
+	exit $err
+fi
+
+# Determine selftest success via shell exit code
+exit_handler()
+{
 	set +e
 
 	# Cleanup