Skip to contentSkip to conversation

Netfilter fixes for net

PATCH Applied Pablo Neira Ayuso 2 participants 10 messages Jan 27, 2022 – Jan 28, 2022

Subsystems (5)
lore
Share & tools
lore

Pablo Neira Ayuso

author maintainer
#0 lore raw

cover Netfilter fixes for net

Hi,

The following patchset contains Netfilter fixes for net:

1) Remove leftovers from flowtable modules, from Geert Uytterhoeven.

2) Missing refcount increment of conntrack template in nft_ct, from Florian Westphal.

3) Reduce nft_zone selftest time, also from Florian.

4) Add selftest to cover stateless NAT on fragments, from Florian Westphal.

5) Do not set net_device when for reject packets from the bridge path, from Phil Sutter.

6) Cancel register tracking info on nft_byteorder operations.

7) Extend nft_concat_range selftest to cover set reload with no elements, from Florian Westphal.

8) Remove useless update of pointer in chain blob builder, reported by kbuild test robot.

Please, pull these changes from:

git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git

Thanks.

----------------------------------------------------------------

The following changes since commit 2f61353cd2f789a4229b6f5c1c24a40a613357bb:

net: hns3: handle empty unknown interrupt for VF (2022-01-25 13:08:05 +0000)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git HEAD

for you to fetch changes up to b07f413732549e5a96e891411fbb5980f2d8e5a1:

netfilter: nf_tables: remove assignment with no effect in chain blob builder (2022-01-27 17:50:56 +0100)

----------------------------------------------------------------
Florian Westphal (4):
netfilter: nft_ct: fix use after free when attaching zone template
selftests: netfilter: reduce zone stress test running time
selftests: netfilter: check stateless nat udp checksum fixup
selftests: nft_concat_range: add test for reload with no element add/del

Geert Uytterhoeven (1):
netfilter: Remove flowtable relics

Pablo Neira Ayuso (2):
netfilter: nft_byteorder: track register operations
netfilter: nf_tables: remove assignment with no effect in chain blob builder

Phil Sutter (1):
netfilter: nft_reject_bridge: Fix for missing reply from prerouting

 net/bridge/netfilter/nft_reject_bridge.c           |   8 +-
 net/ipv4/netfilter/Kconfig                         |   4 -
 net/ipv6/netfilter/Kconfig                         |   4 -
 net/ipv6/netfilter/Makefile                        |   3 -
 net/ipv6/netfilter/nf_flow_table_ipv6.c            |   0
 net/netfilter/nf_tables_api.c                      |   1 -
 net/netfilter/nft_byteorder.c                      |  12 ++
 net/netfilter/nft_ct.c                             |   5 +-
 .../selftests/netfilter/nft_concat_range.sh        |  72 +++++++++-
 tools/testing/selftests/netfilter/nft_nat.sh       | 152 +++++++++++++++++++++
 .../testing/selftests/netfilter/nft_zones_many.sh  |  12 +-
 11 files changed, 249 insertions(+), 24 deletions(-)
 delete mode 100644 net/ipv6/netfilter/nf_flow_table_ipv6.c
author maintainer
#1 lore raw

PATCH 1/8 netfilter: Remove flowtable relics

From: Geert Uytterhoeven <[email protected]>

NF_FLOW_TABLE_IPV4 and NF_FLOW_TABLE_IPV6 are invisble, selected by nothing (so they can no longer be enabled), and their last real users have been removed (nf_flow_table_ipv6.c is empty).

Clean up the leftovers.

Fixes c42ba4290b2147aa ("netfilter: flowtable: remove ipv4/ipv6 modules")
---
 net/ipv4/netfilter/Kconfig              | 4 ----
 net/ipv6/netfilter/Kconfig              | 4 ----
 net/ipv6/netfilter/Makefile             | 3 ---
 net/ipv6/netfilter/nf_flow_table_ipv6.c | 0
 4 files changed, 11 deletions(-)
 delete mode 100644 net/ipv6/netfilter/nf_flow_table_ipv6.c
net/ipv4/netfilter/Kconfig +0 -4
index 67087f95579f..aab384126f61 100644
@@ -58,10 +58,6 @@ config NF_TABLES_ARP
 
 endif # NF_TABLES
 
-config NF_FLOW_TABLE_IPV4
-	tristate
-	select NF_FLOW_TABLE_INET
-
 config NF_DUP_IPV4
 	tristate "Netfilter IPv4 packet duplication to alternate destination"
 	depends on !NF_CONNTRACK || NF_CONNTRACK
net/ipv6/netfilter/Kconfig +0 -4
index 97d3d1b36dbc..0ba62f4868f9 100644
@@ -47,10 +47,6 @@ config NFT_FIB_IPV6
 endif # NF_TABLES_IPV6
 endif # NF_TABLES
 
-config NF_FLOW_TABLE_IPV6
-	tristate
-	select NF_FLOW_TABLE_INET
-
 config NF_DUP_IPV6
 	tristate "Netfilter IPv6 packet duplication to alternate destination"
 	depends on !NF_CONNTRACK || NF_CONNTRACK
net/ipv6/netfilter/Makefile +0 -3
index b85383606df7..b8d6dc9aeeb6 100644
@@ -28,9 +28,6 @@ obj-$(CONFIG_NFT_REJECT_IPV6) += nft_reject_ipv6.o
 obj-$(CONFIG_NFT_DUP_IPV6) += nft_dup_ipv6.o
 obj-$(CONFIG_NFT_FIB_IPV6) += nft_fib_ipv6.o
 
-# flow table support
-obj-$(CONFIG_NF_FLOW_TABLE_IPV6) += nf_flow_table_ipv6.o
-
 # matches
 obj-$(CONFIG_IP6_NF_MATCH_AH) += ip6t_ah.o
 obj-$(CONFIG_IP6_NF_MATCH_EUI64) += ip6t_eui64.o
net/ipv6/netfilter/nf_flow_table_ipv6.c deleted +0 -0
deleted file mode 100644
index e69de29bb2d1..000000000000
Signature · 1 line
2.30.2
#2 lore raw

Re: [PATCH net 1/8] netfilter: Remove flowtable relics

This series was applied to netdev/net.git (master) by Pablo Neira Ayuso <[email protected]>:

Hello:

This series was applied to netdev/net.git (master) by Pablo Neira Ayuso <[email protected]>:

Quoted 9 lines · On Fri, 28 Jan 2022 00:52:28 +0100 you wrote:

From: Geert Uytterhoeven <[email protected]>

NF_FLOW_TABLE_IPV4 and NF_FLOW_TABLE_IPV6 are invisble, selected by nothing (so they can no longer be enabled), and their last real users have been removed (nf_flow_table_ipv6.c is empty).

Clean up the leftovers.

[...]

Here is the summary with links:
- [net,1/8] netfilter: Remove flowtable relics

    https://git.kernel.org/netdev/net/c/7355bfe0e0cc
  - [net,2/8] netfilter: nft_ct: fix use after free when attaching zone template
    https://git.kernel.org/netdev/net/c/34243b9ec856
  - [net,3/8] selftests: netfilter: reduce zone stress test running time
    https://git.kernel.org/netdev/net/c/c858620d2ae3
  - [net,4/8] selftests: netfilter: check stateless nat udp checksum fixup
    https://git.kernel.org/netdev/net/c/aad51ca71ad8
  - [net,5/8] netfilter: nft_reject_bridge: Fix for missing reply from prerouting
    https://git.kernel.org/netdev/net/c/aeac4554eb54
  - [net,6/8] netfilter: nft_byteorder: track register operations
    https://git.kernel.org/netdev/net/c/f459bfd4b979
  - [net,7/8] selftests: nft_concat_range: add test for reload with no element add/del
    https://git.kernel.org/netdev/net/c/eda0cf1202ac
  - [net,8/8] netfilter: nf_tables: remove assignment with no effect in chain blob builder
    https://git.kernel.org/netdev/net/c/b07f41373254

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
author maintainer
#3 lore raw

PATCH 2/8 netfilter: nft_ct: fix use after free when attaching zone template

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

The conversion erroneously removed the refcount increment. In case we can use the percpu template, we need to increment the refcount, else it will be released when the skb gets freed.

In case the slowpath is taken, the new template already has a refcount of 1.

Fixes 719774377622 ("netfilter: conntrack: convert to refcount_t api")
---
 net/netfilter/nft_ct.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
net/netfilter/nft_ct.c +4 -1
index 518d96c8c247..5adf8bb628a8 100644
@@ -260,9 +260,12 @@ static void nft_ct_set_zone_eval(const struct nft_expr *expr,
 	ct = this_cpu_read(nft_ct_pcpu_template);
 
 	if (likely(refcount_read(&ct->ct_general.use) == 1)) {
+		refcount_inc(&ct->ct_general.use);
 		nf_ct_zone_add(ct, &zone);
 	} else {
-		/* previous skb got queued to userspace */
+		/* previous skb got queued to userspace, allocate temporary
+		 * one until percpu template can be reused.
+		 */
 		ct = nf_ct_tmpl_alloc(nft_net(pkt), &zone, GFP_ATOMIC);
 		if (!ct) {
 			regs->verdict.code = NF_DROP;
Signature · 1 line
2.30.2
author maintainer
#4 lore raw

PATCH 3/8 selftests: netfilter: reduce zone stress test running time

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

This selftests needs almost 3 minutes to complete, reduce the insertes zones to 1000. Test now completes in about 20 seconds.

---
 tools/testing/selftests/netfilter/nft_zones_many.sh | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
tools/testing/selftests/netfilter/nft_zones_many.sh +6 -6
index 04633119b29a..5a8db0b48928 100755
@@ -9,7 +9,7 @@ ns="ns-$sfx"
 # Kselftest framework requirement - SKIP code is 4.
 ksft_skip=4
 
-zones=20000
+zones=2000
 have_ct_tool=0
 ret=0
 
@@ -75,10 +75,10 @@ EOF
 
 	while [ $i -lt $max_zones ]; do
 		local start=$(date +%s%3N)
-		i=$((i + 10000))
+		i=$((i + 1000))
 		j=$((j + 1))
 		# nft rule in output places each packet in a different zone.
-		dd if=/dev/zero of=/dev/stdout bs=8k count=10000 2>/dev/null | ip netns exec "$ns" socat STDIN UDP:127.0.0.1:12345,sourceport=12345
+		dd if=/dev/zero of=/dev/stdout bs=8k count=1000 2>/dev/null | ip netns exec "$ns" socat STDIN UDP:127.0.0.1:12345,sourceport=12345
 		if [ $? -ne 0 ] ;then
 			ret=1
 			break
@@ -86,7 +86,7 @@ EOF
 
 		stop=$(date +%s%3N)
 		local duration=$((stop-start))
-		echo "PASS: added 10000 entries in $duration ms (now $i total, loop $j)"
+		echo "PASS: added 1000 entries in $duration ms (now $i total, loop $j)"
 	done
 
 	if [ $have_ct_tool -eq 1 ]; then
@@ -128,11 +128,11 @@ test_conntrack_tool() {
 			break
 		fi
 
-		if [ $((i%10000)) -eq 0 ];then
+		if [ $((i%1000)) -eq 0 ];then
 			stop=$(date +%s%3N)
 
 			local duration=$((stop-start))
-			echo "PASS: added 10000 entries in $duration ms (now $i total)"
+			echo "PASS: added 1000 entries in $duration ms (now $i total)"
 			start=$stop
 		fi
 	done
Signature · 1 line
2.30.2
author maintainer
#5 lore raw

PATCH 4/8 selftests: netfilter: check stateless nat udp checksum fixup

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

Add a test that sends large udp packet (which is fragmented) via a stateless nft nat rule, i.e. 'ip saddr set 10.2.3.4' and check that the datagram is received by peer.

On kernels without
commit 4e1860a38637 ("netfilter: nft_payload: do not update layer 4 checksum when mangling fragments")', this will fail with:

cmp: EOF on /tmp/tmp.V1q0iXJyQF which is empty
-rw------- 1 root root 4096 Jan 24 22:03 /tmp/tmp.Aaqnq4rBKS
-rw------- 1 root root 0 Jan 24 22:03 /tmp/tmp.V1q0iXJyQF
ERROR: in and output file mismatch when checking udp with stateless nat
FAIL: nftables v1.0.0 (Fearless Fosdick #2)

On patched kernels, this will show:
PASS: IP statless for ns2-PFp89amx

---
 tools/testing/selftests/netfilter/nft_nat.sh | 152 +++++++++++++++++++
 1 file changed, 152 insertions(+)
tools/testing/selftests/netfilter/nft_nat.sh +152 -0
index 349a319a9e51..79fe627b9e81 100755
@@ -899,6 +899,144 @@ EOF
 	ip netns exec "$ns0" nft delete table $family nat
 }
 
+test_stateless_nat_ip()
+{
+	local lret=0
+
+	ip netns exec "$ns0" sysctl net.ipv4.conf.veth0.forwarding=1 > /dev/null
+	ip netns exec "$ns0" sysctl net.ipv4.conf.veth1.forwarding=1 > /dev/null
+
+	ip netns exec "$ns2" ping -q -c 1 10.0.1.99 > /dev/null # ping ns2->ns1
+	if [ $? -ne 0 ] ; then
+		echo "ERROR: cannot ping $ns1 from $ns2 before loading stateless rules"
+		return 1
+	fi
+
+ip netns exec "$ns0" nft -f /dev/stdin <<EOF
+table ip stateless {
+	map xlate_in {
+		typeof meta iifname . ip saddr . ip daddr : ip daddr
+		elements = {
+			"veth1" . 10.0.2.99 . 10.0.1.99 : 10.0.2.2,
+		}
+	}
+	map xlate_out {
+		typeof meta iifname . ip saddr . ip daddr : ip daddr
+		elements = {
+			"veth0" . 10.0.1.99 . 10.0.2.2 : 10.0.2.99
+		}
+	}
+
+	chain prerouting {
+		type filter hook prerouting priority -400; policy accept;
+		ip saddr set meta iifname . ip saddr . ip daddr map @xlate_in
+		ip daddr set meta iifname . ip saddr . ip daddr map @xlate_out
+	}
+}
+EOF
+	if [ $? -ne 0 ]; then
+		echo "SKIP: Could not add ip statless rules"
+		return $ksft_skip
+	fi
+
+	reset_counters
+
+	ip netns exec "$ns2" ping -q -c 1 10.0.1.99 > /dev/null # ping ns2->ns1
+	if [ $? -ne 0 ] ; then
+		echo "ERROR: cannot ping $ns1 from $ns2 with stateless rules"
+		lret=1
+	fi
+
+	# ns1 should have seen packets from .2.2, due to stateless rewrite.
+	expect="packets 1 bytes 84"
+	cnt=$(ip netns exec "$ns1" nft list counter inet filter ns0insl | grep -q "$expect")
+	if [ $? -ne 0 ]; then
+		bad_counter "$ns1" ns0insl "$expect" "test_stateless 1"
+		lret=1
+	fi
+
+	for dir in "in" "out" ; do
+		cnt=$(ip netns exec "$ns2" nft list counter inet filter ns1${dir} | grep -q "$expect")
+		if [ $? -ne 0 ]; then
+			bad_counter "$ns2" ns1$dir "$expect" "test_stateless 2"
+			lret=1
+		fi
+	done
+
+	# ns1 should not have seen packets from ns2, due to masquerade
+	expect="packets 0 bytes 0"
+	for dir in "in" "out" ; do
+		cnt=$(ip netns exec "$ns1" nft list counter inet filter ns2${dir} | grep -q "$expect")
+		if [ $? -ne 0 ]; then
+			bad_counter "$ns1" ns0$dir "$expect" "test_stateless 3"
+			lret=1
+		fi
+
+		cnt=$(ip netns exec "$ns0" nft list counter inet filter ns1${dir} | grep -q "$expect")
+		if [ $? -ne 0 ]; then
+			bad_counter "$ns0" ns1$dir "$expect" "test_stateless 4"
+			lret=1
+		fi
+	done
+
+	reset_counters
+
+	socat -h > /dev/null 2>&1
+	if [ $? -ne 0 ];then
+		echo "SKIP: Could not run stateless nat frag test without socat tool"
+		if [ $lret -eq 0 ]; then
+			return $ksft_skip
+		fi
+
+		ip netns exec "$ns0" nft delete table ip stateless
+		return $lret
+	fi
+
+	local tmpfile=$(mktemp)
+	dd if=/dev/urandom of=$tmpfile bs=4096 count=1 2>/dev/null
+
+	local outfile=$(mktemp)
+	ip netns exec "$ns1" timeout 3 socat -u UDP4-RECV:4233 OPEN:$outfile < /dev/null &
+	sc_r=$!
+
+	sleep 1
+	# re-do with large ping -> ip fragmentation
+	ip netns exec "$ns2" timeout 3 socat - UDP4-SENDTO:"10.0.1.99:4233" < "$tmpfile" > /dev/null
+	if [ $? -ne 0 ] ; then
+		echo "ERROR: failed to test udp $ns1 to $ns2 with stateless ip nat" 1>&2
+		lret=1
+	fi
+
+	wait
+
+	cmp "$tmpfile" "$outfile"
+	if [ $? -ne 0 ]; then
+		ls -l "$tmpfile" "$outfile"
+		echo "ERROR: in and output file mismatch when checking udp with stateless nat" 1>&2
+		lret=1
+	fi
+
+	rm -f "$tmpfile" "$outfile"
+
+	# ns1 should have seen packets from 2.2, due to stateless rewrite.
+	expect="packets 3 bytes 4164"
+	cnt=$(ip netns exec "$ns1" nft list counter inet filter ns0insl | grep -q "$expect")
+	if [ $? -ne 0 ]; then
+		bad_counter "$ns1" ns0insl "$expect" "test_stateless 5"
+		lret=1
+	fi
+
+	ip netns exec "$ns0" nft delete table ip stateless
+	if [ $? -ne 0 ]; then
+		echo "ERROR: Could not delete table ip stateless" 1>&2
+		lret=1
+	fi
+
+	test $lret -eq 0 && echo "PASS: IP statless for $ns2"
+
+	return $lret
+}
+
 # ip netns exec "$ns0" ping -c 1 -q 10.0.$i.99
 for i in 0 1 2; do
 ip netns exec ns$i-$sfx nft -f /dev/stdin <<EOF
@@ -965,6 +1103,19 @@ table inet filter {
 EOF
 done
 
+# special case for stateless nat check, counter needs to
+# be done before (input) ip defragmentation
+ip netns exec ns1-$sfx nft -f /dev/stdin <<EOF
+table inet filter {
+	counter ns0insl {}
+
+	chain pre {
+		type filter hook prerouting priority -400; policy accept;
+		ip saddr 10.0.2.2 counter name "ns0insl"
+	}
+}
+EOF
+
 sleep 3
 # test basic connectivity
 for i in 1 2; do
@@ -1019,6 +1170,7 @@ $test_inet_nat && test_redirect inet
 $test_inet_nat && test_redirect6 inet
 
 test_port_shadowing
+test_stateless_nat_ip
 
 if [ $ret -ne 0 ];then
 	echo -n "FAIL: "
Signature · 1 line
2.30.2
author maintainer
#6 lore raw

PATCH 5/8 netfilter: nft_reject_bridge: Fix for missing reply from prerouting

In reply to Pablo Neira Ayuso

From: Phil Sutter <[email protected]>

Prior to commit fa538f7cf05aa ("netfilter: nf_reject: add reject skbuff creation helpers"), nft_reject_bridge did not assign to nskb->dev before passing nskb on to br_forward(). The shared skbuff creation helpers introduced in above commit do which seems to confuse br_forward() as reject statements in prerouting hook won't emit a packet anymore.

Fix this by simply passing NULL instead of 'dev' to the helpers - they use the pointer for just that assignment, nothing else.

Fixes fa538f7cf05aa ("netfilter: nf_reject: add reject skbuff creation helpers")
---
 net/bridge/netfilter/nft_reject_bridge.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
net/bridge/netfilter/nft_reject_bridge.c +4 -4
index eba0efe64d05..fbf858ddec35 100644
@@ -49,7 +49,7 @@ static void nft_reject_br_send_v4_tcp_reset(struct net *net,
 {
 	struct sk_buff *nskb;
 
-	nskb = nf_reject_skb_v4_tcp_reset(net, oldskb, dev, hook);
+	nskb = nf_reject_skb_v4_tcp_reset(net, oldskb, NULL, hook);
 	if (!nskb)
 		return;
 
@@ -65,7 +65,7 @@ static void nft_reject_br_send_v4_unreach(struct net *net,
 {
 	struct sk_buff *nskb;
 
-	nskb = nf_reject_skb_v4_unreach(net, oldskb, dev, hook, code);
+	nskb = nf_reject_skb_v4_unreach(net, oldskb, NULL, hook, code);
 	if (!nskb)
 		return;
 
@@ -81,7 +81,7 @@ static void nft_reject_br_send_v6_tcp_reset(struct net *net,
 {
 	struct sk_buff *nskb;
 
-	nskb = nf_reject_skb_v6_tcp_reset(net, oldskb, dev, hook);
+	nskb = nf_reject_skb_v6_tcp_reset(net, oldskb, NULL, hook);
 	if (!nskb)
 		return;
 
@@ -98,7 +98,7 @@ static void nft_reject_br_send_v6_unreach(struct net *net,
 {
 	struct sk_buff *nskb;
 
-	nskb = nf_reject_skb_v6_unreach(net, oldskb, dev, hook, code);
+	nskb = nf_reject_skb_v6_unreach(net, oldskb, NULL, hook, code);
 	if (!nskb)
 		return;
 
Signature · 1 line
2.30.2
author maintainer
#7 lore raw

PATCH 6/8 netfilter: nft_byteorder: track register operations

In reply to Pablo Neira Ayuso

Cancel tracking for byteorder operation, otherwise selector + byteorder operation is incorrectly reduced if source and destination registers are the same.

---
 net/netfilter/nft_byteorder.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
net/netfilter/nft_byteorder.c +12 -0
index 9d5947ab8d4e..e646e9ee4a98 100644
@@ -167,12 +167,24 @@ static int nft_byteorder_dump(struct sk_buff *skb, const struct nft_expr *expr)
 	return -1;
 }
 
+static bool nft_byteorder_reduce(struct nft_regs_track *track,
+				 const struct nft_expr *expr)
+{
+	struct nft_byteorder *priv = nft_expr_priv(expr);
+
+	track->regs[priv->dreg].selector = NULL;
+	track->regs[priv->dreg].bitwise = NULL;
+
+	return false;
+}
+
 static const struct nft_expr_ops nft_byteorder_ops = {
 	.type		= &nft_byteorder_type,
 	.size		= NFT_EXPR_SIZE(sizeof(struct nft_byteorder)),
 	.eval		= nft_byteorder_eval,
 	.init		= nft_byteorder_init,
 	.dump		= nft_byteorder_dump,
+	.reduce		= nft_byteorder_reduce,
 };
 
 struct nft_expr_type nft_byteorder_type __read_mostly = {
Signature · 1 line
2.30.2
author maintainer
#8 lore raw

PATCH 7/8 selftests: nft_concat_range: add test for reload with no element add/del

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

Add a specific test for the reload issue fixed with
commit 23c54263efd7cb ("netfilter: nft_set_pipapo: allocate pcpu scratch maps on clone").

Add to set, then flush set content + restore without other add/remove in the transaction.

On kernels before the fix, this test case fails:
net,mac with reload [FAIL]

---
 .../selftests/netfilter/nft_concat_range.sh   | 72 ++++++++++++++++++-
 1 file changed, 71 insertions(+), 1 deletion(-)
tools/testing/selftests/netfilter/nft_concat_range.sh +71 -1
index ed61f6cab60f..df322e47a54f 100755
@@ -27,7 +27,7 @@ TYPES="net_port port_net net6_port port_proto net6_port_mac net6_port_mac_proto
        net6_port_net6_port net_port_mac_proto_net"
 
 # Reported bugs, also described by TYPE_ variables below
-BUGS="flush_remove_add"
+BUGS="flush_remove_add reload"
 
 # List of possible paths to pktgen script from kernel tree for performance tests
 PKTGEN_SCRIPT_PATHS="
@@ -354,6 +354,23 @@ TYPE_flush_remove_add="
 display		Add two elements, flush, re-add
 "
 
+TYPE_reload="
+display		net,mac with reload
+type_spec	ipv4_addr . ether_addr
+chain_spec	ip daddr . ether saddr
+dst		addr4
+src		mac
+start		1
+count		1
+src_delta	2000
+tools		sendip nc bash
+proto		udp
+
+race_repeat	0
+
+perf_duration	0
+"
+
 # Set template for all tests, types and rules are filled in depending on test
 set_template='
 flush ruleset
@@ -1473,6 +1490,59 @@ test_bug_flush_remove_add() {
 	nft flush ruleset
 }
 
+# - add ranged element, check that packets match it
+# - reload the set, check packets still match
+test_bug_reload() {
+	setup veth send_"${proto}" set || return ${KSELFTEST_SKIP}
+	rstart=${start}
+
+	range_size=1
+	for i in $(seq "${start}" $((start + count))); do
+		end=$((start + range_size))
+
+		# Avoid negative or zero-sized port ranges
+		if [ $((end / 65534)) -gt $((start / 65534)) ]; then
+			start=${end}
+			end=$((end + 1))
+		fi
+		srcstart=$((start + src_delta))
+		srcend=$((end + src_delta))
+
+		add "$(format)" || return 1
+		range_size=$((range_size + 1))
+		start=$((end + range_size))
+	done
+
+	# check kernel does allocate pcpu sctrach map
+	# for reload with no elemet add/delete
+	( echo flush set inet filter test ;
+	  nft list set inet filter test ) | nft -f -
+
+	start=${rstart}
+	range_size=1
+
+	for i in $(seq "${start}" $((start + count))); do
+		end=$((start + range_size))
+
+		# Avoid negative or zero-sized port ranges
+		if [ $((end / 65534)) -gt $((start / 65534)) ]; then
+			start=${end}
+			end=$((end + 1))
+		fi
+		srcstart=$((start + src_delta))
+		srcend=$((end + src_delta))
+
+		for j in $(seq ${start} $((range_size / 2 + 1)) ${end}); do
+			send_match "${j}" $((j + src_delta)) || return 1
+		done
+
+		range_size=$((range_size + 1))
+		start=$((end + range_size))
+	done
+
+	nft flush ruleset
+}
+
 test_reported_issues() {
 	eval test_bug_"${subtest}"
 }
Signature · 1 line
2.30.2
author maintainer
#9 lore raw

PATCH 8/8 netfilter: nf_tables: remove assignment with no effect in chain blob builder

In reply to Pablo Neira Ayuso

cppcheck possible warnings:

Quoted 1 line
Quoted 1 line

net/netfilter/nf_tables_api.c:2014:2: warning: Assignment of function parameter has no effect outside the function. Did you forget dereferencing it? [uselessAssignmentPtrArg]

    ptr += offsetof(struct nft_rule_dp, data);
    ^
---
 net/netfilter/nf_tables_api.c | 1 -
 1 file changed, 1 deletion(-)
net/netfilter/nf_tables_api.c +0 -1
index cf454f8ca2b0..5fa16990da95 100644
@@ -2011,7 +2011,6 @@ static void nft_last_rule(struct nft_rule_blob *blob, const void *ptr)
 
 	prule = (struct nft_rule_dp *)ptr;
 	prule->is_last = 1;
-	ptr += offsetof(struct nft_rule_dp, data);
 	/* blob size does not include the trailer rule */
 }
 
Signature · 1 line
2.30.2