[Patch net 0/4] flow_dissector: Fix handling of mixed port and port-range keys

[Patch net 0/4] flow_dissector: Fix handling of mixed port and port-range keys

[Cong Wang]

This patchset contains two fixes for flow_dissector handling of mixed
port and port-range keys, for both tc-flower case and bpf case. Each
of them also comes with a selftest.

---
Cong Wang (4):
  flow_dissector: Fix handling of mixed port and port-range keys
  selftests/net/forwarding: Add a test case for tc-flower of mixed port
    and port-range
  flow_dissector: Fix port range key handling in BPF conversion
  selftests/bpf: Add a specific dst port matching

 net/core/flow_dissector.c                     | 49 +++++++++++--------
 .../flow_dissector_classification.c           |  7 ++-
 .../net/forwarding/tc_flower_port_range.sh    | 46 +++++++++++++++++
 3 files changed, 81 insertions(+), 21 deletions(-)

-- 
2.34.1

[Patch net 1/4] flow_dissector: Fix handling of mixed port and port-range keys

[Cong Wang]

This patch fixes a bug in TC flower filter where rules combining a
specific destination port with a source port range weren't working
correctly.

The specific case was when users tried to configure rules like:

tc filter add dev ens38 ingress protocol ip flower ip_proto udp \
dst_port 5000 src_port 2000-3000 action drop

The root cause was in the flow dissector code. While both
FLOW_DISSECTOR_KEY_PORTS and FLOW_DISSECTOR_KEY_PORTS_RANGE flags
were being set correctly in the classifier, the __skb_flow_dissect_ports()
function was only populating one of them: whichever came first in
the enum check. This meant that when the code needed both a specific
port and a port range, one of them would be left as 0, causing the
filter to not match packets as expected.

Fix it by removing the either/or logic and instead checking and
populating both key types independently when they're in use.

Fixes: 8ffb055beae5 ("cls_flower: Fix the behavior using port ranges with hw-offload")
Reported-by: Qiang Zhang <dtzq01@gmail.com>
Closes: https://lore.kernel.org/netdev/CAPx+-5uvFxkhkz4=j_Xuwkezjn9U6kzKTD5jz4tZ9msSJ0fOJA@mail.gmail.com/
Cc: Yoshiki Komachi <komachi.yoshiki@gmail.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
---
 net/core/flow_dissector.c | 31 +++++++++++++++++++------------
 1 file changed, 19 insertions(+), 12 deletions(-)

diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index 5db41bf2ed93..c33af3ef0b79 100644
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -853,23 +853,30 @@ __skb_flow_dissect_ports(const struct sk_buff *skb,
 			 void *target_container, const void *data,
 			 int nhoff, u8 ip_proto, int hlen)
 {
-	enum flow_dissector_key_id dissector_ports = FLOW_DISSECTOR_KEY_MAX;
-	struct flow_dissector_key_ports *key_ports;
+	struct flow_dissector_key_ports_range *key_ports_range = NULL;
+	struct flow_dissector_key_ports *key_ports = NULL;
+	__be32 ports;
 
 	if (dissector_uses_key(flow_dissector, FLOW_DISSECTOR_KEY_PORTS))
-		dissector_ports = FLOW_DISSECTOR_KEY_PORTS;
-	else if (dissector_uses_key(flow_dissector,
-				    FLOW_DISSECTOR_KEY_PORTS_RANGE))
-		dissector_ports = FLOW_DISSECTOR_KEY_PORTS_RANGE;
+		key_ports = skb_flow_dissector_target(flow_dissector,
+						      FLOW_DISSECTOR_KEY_PORTS,
+						      target_container);
 
-	if (dissector_ports == FLOW_DISSECTOR_KEY_MAX)
+	if (dissector_uses_key(flow_dissector, FLOW_DISSECTOR_KEY_PORTS_RANGE))
+		key_ports_range = skb_flow_dissector_target(flow_dissector,
+							    FLOW_DISSECTOR_KEY_PORTS_RANGE,
+							    target_container);
+
+	if (!key_ports && !key_ports_range)
 		return;
 
-	key_ports = skb_flow_dissector_target(flow_dissector,
-					      dissector_ports,
-					      target_container);
-	key_ports->ports = __skb_flow_get_ports(skb, nhoff, ip_proto,
-						data, hlen);
+	ports = __skb_flow_get_ports(skb, nhoff, ip_proto, data, hlen);
+
+	if (key_ports)
+		key_ports->ports = ports;
+
+	if (key_ports_range)
+		key_ports_range->tp.ports = ports;
 }
 
 static void
-- 
2.34.1

Re: [Patch net 1/4] flow_dissector: Fix handling of mixed port and port-range keys

[Ido Schimmel]

On Mon, Feb 17, 2025 at 08:32:07PM -0800, Cong Wang wrote:
> This patch fixes a bug in TC flower filter where rules combining a
> specific destination port with a source port range weren't working
> correctly.
> 
> The specific case was when users tried to configure rules like:
> 
> tc filter add dev ens38 ingress protocol ip flower ip_proto udp \
> dst_port 5000 src_port 2000-3000 action drop
> 
> The root cause was in the flow dissector code. While both
> FLOW_DISSECTOR_KEY_PORTS and FLOW_DISSECTOR_KEY_PORTS_RANGE flags
> were being set correctly in the classifier, the __skb_flow_dissect_ports()
> function was only populating one of them: whichever came first in
> the enum check. This meant that when the code needed both a specific
> port and a port range, one of them would be left as 0, causing the
> filter to not match packets as expected.
> 
> Fix it by removing the either/or logic and instead checking and
> populating both key types independently when they're in use.
> 
> Fixes: 8ffb055beae5 ("cls_flower: Fix the behavior using port ranges with hw-offload")
> Reported-by: Qiang Zhang <dtzq01@gmail.com>
> Closes: https://lore.kernel.org/netdev/CAPx+-5uvFxkhkz4=j_Xuwkezjn9U6kzKTD5jz4tZ9msSJ0fOJA@mail.gmail.com/
> Cc: Yoshiki Komachi <komachi.yoshiki@gmail.com>
> Cc: Jamal Hadi Salim <jhs@mojatatu.com>
> Cc: Jiri Pirko <jiri@resnulli.us>
> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>

Reviewed-by: Ido Schimmel <idosch@nvidia.com>

[Patch net 2/4] selftests/net/forwarding: Add a test case for tc-flower of mixed port and port-range

[Cong Wang]

After this patch:

 # ./tc_flower_port_range.sh
 TEST: Port range matching - IPv4 UDP                                [ OK ]
 TEST: Port range matching - IPv4 TCP                                [ OK ]
 TEST: Port range matching - IPv6 UDP                                [ OK ]
 TEST: Port range matching - IPv6 TCP                                [ OK ]
 TEST: Port range matching - IPv4 UDP Drop                           [ OK ]

Cc: Qiang Zhang <dtzq01@gmail.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
---
 .../net/forwarding/tc_flower_port_range.sh    | 46 +++++++++++++++++++
 1 file changed, 46 insertions(+)

diff --git a/tools/testing/selftests/net/forwarding/tc_flower_port_range.sh b/tools/testing/selftests/net/forwarding/tc_flower_port_range.sh
index 3885a2a91f7d..baed5e380dae 100755
--- a/tools/testing/selftests/net/forwarding/tc_flower_port_range.sh
+++ b/tools/testing/selftests/net/forwarding/tc_flower_port_range.sh
@@ -20,6 +20,7 @@ ALL_TESTS="
 	test_port_range_ipv4_tcp
 	test_port_range_ipv6_udp
 	test_port_range_ipv6_tcp
+	test_port_range_ipv4_udp_drop
 "
 
 NUM_NETIFS=4
@@ -194,6 +195,51 @@ test_port_range_ipv6_tcp()
 	__test_port_range $proto $ip_proto $sip $dip $mode "$name"
 }
 
+test_port_range_ipv4_udp_drop()
+{
+	local proto=ipv4
+	local ip_proto=udp
+	local sip=192.0.2.1
+	local dip=192.0.2.2
+	local mode="-4"
+	local name="IPv4 UDP Drop"
+	local dmac=$(mac_get $h2)
+	local smac=$(mac_get $h1)
+	local sport_min=2000
+	local sport_max=3000
+	local sport_mid=$((sport_min + (sport_max - sport_min) / 2))
+	local dport=5000
+
+	RET=0
+
+	tc filter add dev $swp1 ingress protocol $proto handle 101 pref 1 \
+		flower src_ip $sip dst_ip $dip ip_proto $ip_proto \
+		src_port $sport_min-$sport_max \
+		dst_port $dport \
+		action drop
+
+	# Test ports outside range - should pass
+	$MZ $mode $h1 -c 1 -q -p 100 -a $smac -b $dmac -A $sip -B $dip \
+		-t $ip_proto "sp=$((sport_min - 1)),dp=$dport"
+	$MZ $mode $h1 -c 1 -q -p 100 -a $smac -b $dmac -A $sip -B $dip \
+		-t $ip_proto "sp=$((sport_max + 1)),dp=$dport"
+
+	# Test ports inside range - should be dropped
+	$MZ $mode $h1 -c 1 -q -p 100 -a $smac -b $dmac -A $sip -B $dip \
+		-t $ip_proto "sp=$sport_min,dp=$dport"
+	$MZ $mode $h1 -c 1 -q -p 100 -a $smac -b $dmac -A $sip -B $dip \
+		-t $ip_proto "sp=$sport_mid,dp=$dport"
+	$MZ $mode $h1 -c 1 -q -p 100 -a $smac -b $dmac -A $sip -B $dip \
+		-t $ip_proto "sp=$sport_max,dp=$dport"
+
+	tc_check_packets "dev $swp1 ingress" 101 3
+	check_err $? "Filter did not drop the expected number of packets"
+
+	tc filter del dev $swp1 ingress protocol $proto pref 1 handle 101 flower
+
+	log_test "Port range matching - $name"
+}
+
 setup_prepare()
 {
 	h1=${NETIFS[p1]}
-- 
2.34.1

Re: [Patch net 2/4] selftests/net/forwarding: Add a test case for tc-flower of mixed port and port-range

[Ido Schimmel]

On Mon, Feb 17, 2025 at 08:32:08PM -0800, Cong Wang wrote:
> After this patch:
> 
>  # ./tc_flower_port_range.sh
>  TEST: Port range matching - IPv4 UDP                                [ OK ]
>  TEST: Port range matching - IPv4 TCP                                [ OK ]
>  TEST: Port range matching - IPv6 UDP                                [ OK ]
>  TEST: Port range matching - IPv6 TCP                                [ OK ]
>  TEST: Port range matching - IPv4 UDP Drop                           [ OK ]
> 
> Cc: Qiang Zhang <dtzq01@gmail.com>
> Cc: Jamal Hadi Salim <jhs@mojatatu.com>
> Cc: Jiri Pirko <jiri@resnulli.us>
> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>

Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>

Tested with offload as well.

[Patch net 3/4] flow_dissector: Fix port range key handling in BPF conversion

[Cong Wang]

Fix how port range keys are handled in __skb_flow_bpf_to_target() by:
- Separating PORTS and PORTS_RANGE key handling
- Using correct key_ports_range structure for range keys
- Properly initializing both key types independently

This ensures port range information is correctly stored in its dedicated
structure rather than incorrectly using the regular ports key structure.

Fixes: 59fb9b62fb6c ("flow_dissector: Fix to use new variables for port ranges in bpf hook")
Reported-by: Qiang Zhang <dtzq01@gmail.com>
Closes: https://lore.kernel.org/netdev/CAPx+-5uvFxkhkz4=j_Xuwkezjn9U6kzKTD5jz4tZ9msSJ0fOJA@mail.gmail.com/
Cc: Yoshiki Komachi <komachi.yoshiki@gmail.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
---
 net/core/flow_dissector.c | 18 ++++++++++--------
 1 file changed, 10 insertions(+), 8 deletions(-)

diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index c33af3ef0b79..9cd8de6bebb5 100644
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -931,6 +931,7 @@ static void __skb_flow_bpf_to_target(const struct bpf_flow_keys *flow_keys,
 				     struct flow_dissector *flow_dissector,
 				     void *target_container)
 {
+	struct flow_dissector_key_ports_range *key_ports_range = NULL;
 	struct flow_dissector_key_ports *key_ports = NULL;
 	struct flow_dissector_key_control *key_control;
 	struct flow_dissector_key_basic *key_basic;
@@ -975,20 +976,21 @@ static void __skb_flow_bpf_to_target(const struct bpf_flow_keys *flow_keys,
 		key_control->addr_type = FLOW_DISSECTOR_KEY_IPV6_ADDRS;
 	}
 
-	if (dissector_uses_key(flow_dissector, FLOW_DISSECTOR_KEY_PORTS))
+	if (dissector_uses_key(flow_dissector, FLOW_DISSECTOR_KEY_PORTS)) {
 		key_ports = skb_flow_dissector_target(flow_dissector,
 						      FLOW_DISSECTOR_KEY_PORTS,
 						      target_container);
-	else if (dissector_uses_key(flow_dissector,
-				    FLOW_DISSECTOR_KEY_PORTS_RANGE))
-		key_ports = skb_flow_dissector_target(flow_dissector,
-						      FLOW_DISSECTOR_KEY_PORTS_RANGE,
-						      target_container);
-
-	if (key_ports) {
 		key_ports->src = flow_keys->sport;
 		key_ports->dst = flow_keys->dport;
 	}
+	if (dissector_uses_key(flow_dissector,
+			       FLOW_DISSECTOR_KEY_PORTS_RANGE)) {
+		key_ports_range = skb_flow_dissector_target(flow_dissector,
+							    FLOW_DISSECTOR_KEY_PORTS_RANGE,
+							    target_container);
+		key_ports_range->tp.src = flow_keys->sport;
+		key_ports_range->tp.dst = flow_keys->dport;
+	}
 
 	if (dissector_uses_key(flow_dissector,
 			       FLOW_DISSECTOR_KEY_FLOW_LABEL)) {
-- 
2.34.1

[Patch net 4/4] selftests/bpf: Add a specific dst port matching

[Cong Wang]

After this patch:

 #102/1   flow_dissector_classification/ipv4:OK
 #102/2   flow_dissector_classification/ipv4_continue_dissect:OK
 #102/3   flow_dissector_classification/ipip:OK
 #102/4   flow_dissector_classification/gre:OK
 #102/5   flow_dissector_classification/port_range:OK
 #102/6   flow_dissector_classification/ipv6:OK
 #102     flow_dissector_classification:OK
 Summary: 1/6 PASSED, 0 SKIPPED, 0 FAILED

Cc: bpf@vger.kernel.org
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
---
 .../bpf/prog_tests/flow_dissector_classification.c         | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/tools/testing/selftests/bpf/prog_tests/flow_dissector_classification.c b/tools/testing/selftests/bpf/prog_tests/flow_dissector_classification.c
index 3729fbfd3084..80b153d3ddec 100644
--- a/tools/testing/selftests/bpf/prog_tests/flow_dissector_classification.c
+++ b/tools/testing/selftests/bpf/prog_tests/flow_dissector_classification.c
@@ -542,8 +542,12 @@ static void detach_program(struct bpf_flow *skel, int prog_fd)
 
 static int set_port_drop(int pf, bool multi_port)
 {
+	char dst_port[16];
+
+	snprintf(dst_port, sizeof(dst_port), "%d", CFG_PORT_INNER);
+
 	SYS(fail, "tc qdisc add dev lo ingress");
-	SYS(fail_delete_qdisc, "tc filter add %s %s %s %s %s %s %s %s %s %s",
+	SYS(fail_delete_qdisc, "tc filter add %s %s %s %s %s %s %s %s %s %s %s %s",
 	    "dev lo",
 	    "parent FFFF:",
 	    "protocol", pf == PF_INET6 ? "ipv6" : "ip",
@@ -551,6 +555,7 @@ static int set_port_drop(int pf, bool multi_port)
 	    "flower",
 	    "ip_proto udp",
 	    "src_port", multi_port ? "8-10" : "9",
+	    "dst_port", dst_port,
 	    "action drop");
 	return 0;
 
-- 
2.34.1

Re: [Patch net 0/4] flow_dissector: Fix handling of mixed port and port-range keys

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Mon, 17 Feb 2025 20:32:06 -0800 you wrote:
> This patchset contains two fixes for flow_dissector handling of mixed
> port and port-range keys, for both tc-flower case and bpf case. Each
> of them also comes with a selftest.
> 
> ---
> Cong Wang (4):
>   flow_dissector: Fix handling of mixed port and port-range keys
>   selftests/net/forwarding: Add a test case for tc-flower of mixed port
>     and port-range
>   flow_dissector: Fix port range key handling in BPF conversion
>   selftests/bpf: Add a specific dst port matching
> 
> [...]

Here is the summary with links:
  - [net,1/4] flow_dissector: Fix handling of mixed port and port-range keys
    https://git.kernel.org/netdev/net/c/3e5796862c69
  - [net,2/4] selftests/net/forwarding: Add a test case for tc-flower of mixed port and port-range
    https://git.kernel.org/netdev/net/c/dfc1580f960b
  - [net,3/4] flow_dissector: Fix port range key handling in BPF conversion
    https://git.kernel.org/netdev/net/c/69ab34f705fb
  - [net,4/4] selftests/bpf: Add a specific dst port matching
    https://git.kernel.org/netdev/net/c/15de6ba95dbe

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html