Merge branch 'ovs-drop-reasons'

Adrian Moreno says:

====================
openvswitch: add drop reasons

There is currently a gap in drop visibility in the openvswitch module.
This series tries to improve this by adding a new drop reason subsystem
for OVS.

Apart from adding a new drop reasson subsystem and some common drop
reasons, this series takes Eric's preliminary work [1] on adding an
explicit drop action and integrates it into the same subsystem.

A limitation of this series is that it does not report upcall errors.
The reason is that there could be many sources of upcall drops and the
most common one, which is the netlink buffer overflow, cannot be
reported via kfree_skb() because the skb is freed in the netlink layer
(see [2]). Therefore, using a reason for the rare events and not the
common one would be even more misleading. I'd propose we add (in a
follow up patch) a tracepoint to better report upcall errors.

[1] https://lore.kernel.org/netdev/202306300609.tdRdZscy-lkp@intel.com/T/
[2] commit 1100248a5c5c ("openvswitch: Fix double reporting of drops in dropwatch")

---
v4 -> v5:
- Rebased
- Added a helper function to explicitly convert drop reason enum types

v3 -> v4:
- Changed names of errors following Ilya's suggestions
- Moved the ovs-dpctl.py changes from patch 7/7 to 3/7
- Added a test to ensure actions following a drop are rejected

rfc2 -> v3:
- Rebased on top of latest net-next

rfc1 -> rfc2:
- Fail when an explicit drop is not the last
- Added a drop reason for action errors
- Added braces around macros
- Dropped patch that added support for masks in ovs-dpctl.py as it's now
  included in Aaron's series [2].
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

2 files changed, 119 insertions, 5 deletions

diff --git a/tools/testing/selftests/net/openvswitch/openvswitch.sh b/tools/testing/selftests/net/openvswitch/openvswitch.sh
index dced4f612a78..9c2012d70b08 100755
--- a/tools/testing/selftests/net/openvswitch/openvswitch.sh
+++ b/tools/testing/selftests/net/openvswitch/openvswitch.sh
@@ -16,7 +16,8 @@ tests="
 	connect_v4				ip4-xon: Basic ipv4 ping between two NS
 	nat_connect_v4				ip4-nat-xon: Basic ipv4 tcp connection via NAT
 	netlink_checks				ovsnl: validate netlink attrs and settings
-	upcall_interfaces			ovs: test the upcall interfaces"
+	upcall_interfaces			ovs: test the upcall interfaces
+	drop_reason				drop: test drop reasons are emitted"
 
 info() {
     [ $VERBOSE = 0 ] || echo $*
@@ -141,6 +142,25 @@ ovs_add_flow () {
 	return 0
 }
 
+ovs_drop_record_and_run () {
+	local sbx=$1
+	shift
+
+	perf record -a -q -e skb:kfree_skb -o ${ovs_dir}/perf.data $* \
+		>> ${ovs_dir}/stdout 2>> ${ovs_dir}/stderr
+	return $?
+}
+
+ovs_drop_reason_count()
+{
+	local reason=$1
+
+	local perf_output=`perf script -i ${ovs_dir}/perf.data -F trace:event,trace`
+	local pattern="skb:kfree_skb:.*reason: $reason"
+
+	return `echo "$perf_output" | grep "$pattern" | wc -l`
+}
+
 usage() {
 	echo
 	echo "$0 [OPTIONS] [TEST]..."
@@ -155,6 +175,76 @@ usage() {
 	exit 1
 }
 
+# drop_reason test
+# - drop packets and verify the right drop reason is reported
+test_drop_reason() {
+	which perf >/dev/null 2>&1 || return $ksft_skip
+
+	sbx_add "test_drop_reason" || return $?
+
+	ovs_add_dp "test_drop_reason" dropreason || return 1
+
+	info "create namespaces"
+	for ns in client server; do
+		ovs_add_netns_and_veths "test_drop_reason" "dropreason" "$ns" \
+			"${ns:0:1}0" "${ns:0:1}1" || return 1
+	done
+
+	# Setup client namespace
+	ip netns exec client ip addr add 172.31.110.10/24 dev c1
+	ip netns exec client ip link set c1 up
+
+	# Setup server namespace
+	ip netns exec server ip addr add 172.31.110.20/24 dev s1
+	ip netns exec server ip link set s1 up
+
+	# Allow ARP
+	ovs_add_flow "test_drop_reason" dropreason \
+		'in_port(1),eth(),eth_type(0x0806),arp()' '2' || return 1
+	ovs_add_flow "test_drop_reason" dropreason \
+		'in_port(2),eth(),eth_type(0x0806),arp()' '1' || return 1
+
+	# Allow client ICMP traffic but drop return path
+	ovs_add_flow "test_drop_reason" dropreason \
+		"in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10,proto=1),icmp()" '2'
+	ovs_add_flow "test_drop_reason" dropreason \
+		"in_port(2),eth(),eth_type(0x0800),ipv4(src=172.31.110.20,proto=1),icmp()" 'drop'
+
+	ovs_drop_record_and_run "test_drop_reason" ip netns exec client ping -c 2 172.31.110.20
+	ovs_drop_reason_count 0x30001 # OVS_DROP_FLOW_ACTION
+	if [[ "$?" -ne "2" ]]; then
+		info "Did not detect expected drops: $?"
+		return 1
+	fi
+
+	# Drop UDP 6000 traffic with an explicit action and an error code.
+	ovs_add_flow "test_drop_reason" dropreason \
+		"in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10,proto=17),udp(dst=6000)" \
+                'drop(42)'
+	# Drop UDP 7000 traffic with an explicit action with no error code.
+	ovs_add_flow "test_drop_reason" dropreason \
+		"in_port(1),eth(),eth_type(0x0800),ipv4(src=172.31.110.10,proto=17),udp(dst=7000)" \
+                'drop(0)'
+
+	ovs_drop_record_and_run \
+            "test_drop_reason" ip netns exec client nc -i 1 -zuv 172.31.110.20 6000
+	ovs_drop_reason_count 0x30004 # OVS_DROP_EXPLICIT_ACTION_ERROR
+	if [[ "$?" -ne "1" ]]; then
+		info "Did not detect expected explicit error drops: $?"
+		return 1
+	fi
+
+	ovs_drop_record_and_run \
+            "test_drop_reason" ip netns exec client nc -i 1 -zuv 172.31.110.20 7000
+	ovs_drop_reason_count 0x30003 # OVS_DROP_EXPLICIT_ACTION
+	if [[ "$?" -ne "1" ]]; then
+		info "Did not detect expected explicit drops: $?"
+		return 1
+	fi
+
+	return 0
+}
+
 # arp_ping test
 # - client has 1500 byte MTU
 # - server has 1500 byte MTU
@@ -393,6 +483,16 @@ test_netlink_checks () {
 	    wc -l) == 2 ] || \
 	      return 1
 
+	ERR_MSG="Flow actions may not be safe on all matching packets"
+	PRE_TEST=$(dmesg | grep -c "${ERR_MSG}")
+	ovs_add_flow "test_netlink_checks" nv0 \
+		'in_port(1),eth(),eth_type(0x0806),arp()' 'drop(0),2' \
+		&> /dev/null && return 1
+	POST_TEST=$(dmesg | grep -c "${ERR_MSG}")
+	if [ "$PRE_TEST" == "$POST_TEST" ]; then
+		info "failed - error not generated"
+		return 1
+	fi
 	return 0
 }
 
diff --git a/tools/testing/selftests/net/openvswitch/ovs-dpctl.py b/tools/testing/selftests/net/openvswitch/ovs-dpctl.py
index fbdac15e3134..912dc8c49085 100644
--- a/tools/testing/selftests/net/openvswitch/ovs-dpctl.py
+++ b/tools/testing/selftests/net/openvswitch/ovs-dpctl.py
@@ -301,6 +301,7 @@ class ovsactions(nla):
         ("OVS_ACTION_ATTR_CHECK_PKT_LEN", "none"),
         ("OVS_ACTION_ATTR_ADD_MPLS", "none"),
         ("OVS_ACTION_ATTR_DEC_TTL", "none"),
+        ("OVS_ACTION_ATTR_DROP", "uint32"),
     )
 
     class ctact(nla):
@@ -447,6 +448,8 @@ class ovsactions(nla):
                     print_str += "recirc(0x%x)" % int(self.get_attr(field[0]))
                 elif field[0] == "OVS_ACTION_ATTR_TRUNC":
                     print_str += "trunc(%d)" % int(self.get_attr(field[0]))
+                elif field[0] == "OVS_ACTION_ATTR_DROP":
+                    print_str += "drop(%d)" % int(self.get_attr(field[0]))
             elif field[1] == "flag":
                 if field[0] == "OVS_ACTION_ATTR_CT_CLEAR":
                     print_str += "ct_clear"
@@ -468,10 +471,21 @@ class ovsactions(nla):
         while len(actstr) != 0:
             parsed = False
             if actstr.startswith("drop"):
-                # for now, drops have no explicit action, so we
-                # don't need to set any attributes.  The final
-                # act of the processing chain will just drop the packet
-                return
+                # If no reason is provided, the implicit drop is used (i.e no
+                # action). If some reason is given, an explicit action is used.
+                actstr, reason = parse_extract_field(
+                    actstr,
+                    "drop(",
+                    "([0-9]+)",
+                    lambda x: int(x, 0),
+                    False,
+                    None,
+                )
+                if reason is not None:
+                    self["attrs"].append(["OVS_ACTION_ATTR_DROP", reason])
+                    parsed = True
+                else:
+                    return
 
             elif parse_starts_block(actstr, "^(\d+)", False, True):
                 actstr, output = parse_extract_field(