* [PATCH net 04/16] netfilter: flowtable: fix and simplify IP6IP6 tunnel handling
From: Pablo Neira Ayuso @ 2026-06-19 11:54 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw, horms
In-Reply-To: <20260619115452.93949-1-pablo@netfilter.org>
From: Lorenzo Bianconi <lorenzo@kernel.org>
Fix nf_flow_ip6_tunnel_proto() to use pskb_may_pull() instead of
skb_header_pointer() to ensure the outer IPv6 header is in the skb
headroom, which is required for subsequent packet processing. Move
ctx->offset update inside the IPPROTO_IPV6 conditional block since it
should only be adjusted when an IP6IP6 tunnel is actually detected.
Simplify the rx path by removing ipv6_skip_exthdr() and checking
ip6h->nexthdr directly, as the flowtable fast path only handles simple
IP6IP6 encapsulation without extension headers.
Drop the tunnel encapsulation limit destination option support from the
tx path to match, since the rx path no longer handles extension headers.
Remove the encap_limit parameter from nf_flow_offload_ipv6_forward(),
nf_flow_tunnel_ip6ip6_push() and nf_flow_tunnel_v6_push(), along with
the ipv6_tel_txoption struct and related headroom/MTU adjustments.
Fixes: d98103575dcdd ("netfilter: flowtable: Add IP6IP6 rx sw acceleration")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/ipv6/ip6_tunnel.c | 7 ++
net/netfilter/nf_flow_table_ip.c | 80 +++++--------------
.../selftests/net/netfilter/nft_flowtable.sh | 8 +-
3 files changed, 30 insertions(+), 65 deletions(-)
diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
index d7c90a8533ec..bf8e40af60b0 100644
--- a/net/ipv6/ip6_tunnel.c
+++ b/net/ipv6/ip6_tunnel.c
@@ -1851,6 +1851,13 @@ static int ip6_tnl_fill_forward_path(struct net_device_path_ctx *ctx,
struct dst_entry *dst;
int err;
+ if (!(t->parms.flags & IP6_TNL_F_IGN_ENCAP_LIMIT)) {
+ /* encaplimit option is currently not supported is
+ * sw-acceleration path.
+ */
+ return -EOPNOTSUPP;
+ }
+
dst = ip6_route_output(dev_net(ctx->dev), NULL, &fl6);
if (!dst->error) {
path->type = DEV_PATH_TUN;
diff --git a/net/netfilter/nf_flow_table_ip.c b/net/netfilter/nf_flow_table_ip.c
index 9c05a50d6013..e7a3fb2b2d94 100644
--- a/net/netfilter/nf_flow_table_ip.c
+++ b/net/netfilter/nf_flow_table_ip.c
@@ -347,29 +347,23 @@ static bool nf_flow_ip6_tunnel_proto(struct nf_flowtable_ctx *ctx,
struct sk_buff *skb)
{
#if IS_ENABLED(CONFIG_IPV6)
- struct ipv6hdr *ip6h, _ip6h;
- __be16 frag_off;
- u8 nexthdr;
- int hdrlen;
+ struct ipv6hdr *ip6h;
- ip6h = skb_header_pointer(skb, ctx->offset, sizeof(*ip6h), &_ip6h);
- if (!ip6h)
+ if (!pskb_may_pull(skb, sizeof(*ip6h) + ctx->offset))
return false;
+ ip6h = (struct ipv6hdr *)(skb_network_header(skb) + ctx->offset);
if (ip6h->hop_limit <= 1)
return false;
- nexthdr = ip6h->nexthdr;
- hdrlen = ipv6_skip_exthdr(skb, sizeof(*ip6h) + ctx->offset, &nexthdr,
- &frag_off);
- if (hdrlen < 0)
+ if (ipv6_ext_hdr(ip6h->nexthdr))
return false;
- if (nexthdr == IPPROTO_IPV6) {
- ctx->tun.hdr_size = hdrlen;
- ctx->tun.proto = IPPROTO_IPV6;
+ if (ip6h->nexthdr == IPPROTO_IPV6) {
+ ctx->tun.proto = ip6h->nexthdr;
+ ctx->tun.hdr_size = sizeof(*ip6h);
+ ctx->offset += ctx->tun.hdr_size;
}
- ctx->offset += ctx->tun.hdr_size;
return true;
#else
@@ -648,25 +642,19 @@ static int nf_flow_tunnel_v4_push(struct net *net, struct sk_buff *skb,
return 0;
}
-struct ipv6_tel_txoption {
- struct ipv6_txoptions ops;
- __u8 dst_opt[8];
-};
-
static int nf_flow_tunnel_ip6ip6_push(struct net *net, struct sk_buff *skb,
struct flow_offload_tuple *tuple,
- struct in6_addr **ip6_daddr,
- int encap_limit)
+ struct in6_addr **ip6_daddr)
{
struct ipv6hdr *ip6h = (struct ipv6hdr *)skb_network_header(skb);
- u8 hop_limit = ip6h->hop_limit, proto = IPPROTO_IPV6;
struct rtable *rt = dst_rtable(tuple->dst_cache);
__u8 dsfield = ipv6_get_dsfield(ip6h);
struct flowi6 fl6 = {
.daddr = tuple->tun.src_v6,
.saddr = tuple->tun.dst_v6,
- .flowi6_proto = proto,
+ .flowi6_proto = IPPROTO_IPV6,
};
+ u8 hop_limit = ip6h->hop_limit;
int err, mtu;
u32 headroom;
@@ -674,41 +662,18 @@ static int nf_flow_tunnel_ip6ip6_push(struct net *net, struct sk_buff *skb,
if (err)
return err;
- skb_set_inner_ipproto(skb, proto);
+ skb_set_inner_ipproto(skb, IPPROTO_IPV6);
headroom = sizeof(*ip6h) + LL_RESERVED_SPACE(rt->dst.dev) +
rt->dst.header_len;
- if (encap_limit)
- headroom += 8;
err = skb_cow_head(skb, headroom);
if (err)
return err;
skb_scrub_packet(skb, true);
mtu = dst_mtu(&rt->dst) - sizeof(*ip6h);
- if (encap_limit)
- mtu -= 8;
mtu = max(mtu, IPV6_MIN_MTU);
skb_dst_update_pmtu_no_confirm(skb, mtu);
- if (encap_limit > 0) {
- struct ipv6_tel_txoption opt = {
- .dst_opt[2] = IPV6_TLV_TNL_ENCAP_LIMIT,
- .dst_opt[3] = 1,
- .dst_opt[4] = encap_limit,
- .dst_opt[5] = IPV6_TLV_PADN,
- .dst_opt[6] = 1,
- };
- struct ipv6_opt_hdr *hopt;
-
- opt.ops.dst1opt = (struct ipv6_opt_hdr *)opt.dst_opt;
- opt.ops.opt_nflen = 8;
-
- hopt = skb_push(skb, ipv6_optlen(opt.ops.dst1opt));
- memcpy(hopt, opt.ops.dst1opt, ipv6_optlen(opt.ops.dst1opt));
- hopt->nexthdr = IPPROTO_IPV6;
- proto = NEXTHDR_DEST;
- }
-
skb_push(skb, sizeof(*ip6h));
skb_reset_network_header(skb);
@@ -716,7 +681,7 @@ static int nf_flow_tunnel_ip6ip6_push(struct net *net, struct sk_buff *skb,
ip6_flow_hdr(ip6h, dsfield,
ip6_make_flowlabel(net, skb, fl6.flowlabel, true, &fl6));
ip6h->hop_limit = hop_limit;
- ip6h->nexthdr = proto;
+ ip6h->nexthdr = IPPROTO_IPV6;
ip6h->daddr = tuple->tun.src_v6;
ip6h->saddr = tuple->tun.dst_v6;
ipv6_hdr(skb)->payload_len = htons(skb->len - sizeof(*ip6h));
@@ -729,12 +694,10 @@ static int nf_flow_tunnel_ip6ip6_push(struct net *net, struct sk_buff *skb,
static int nf_flow_tunnel_v6_push(struct net *net, struct sk_buff *skb,
struct flow_offload_tuple *tuple,
- struct in6_addr **ip6_daddr,
- int encap_limit)
+ struct in6_addr **ip6_daddr)
{
if (tuple->tun_num)
- return nf_flow_tunnel_ip6ip6_push(net, skb, tuple, ip6_daddr,
- encap_limit);
+ return nf_flow_tunnel_ip6ip6_push(net, skb, tuple, ip6_daddr);
return 0;
}
@@ -1089,7 +1052,7 @@ static int nf_flow_tuple_ipv6(struct nf_flowtable_ctx *ctx, struct sk_buff *skb,
static int nf_flow_offload_ipv6_forward(struct nf_flowtable_ctx *ctx,
struct nf_flowtable *flow_table,
struct flow_offload_tuple_rhash *tuplehash,
- struct sk_buff *skb, int encap_limit)
+ struct sk_buff *skb)
{
enum flow_offload_tuple_dir dir;
struct flow_offload *flow;
@@ -1100,11 +1063,8 @@ static int nf_flow_offload_ipv6_forward(struct nf_flowtable_ctx *ctx,
flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
mtu = flow->tuplehash[dir].tuple.mtu + ctx->offset;
- if (flow->tuplehash[!dir].tuple.tun_num) {
+ if (flow->tuplehash[!dir].tuple.tun_num)
mtu -= sizeof(*ip6h);
- if (encap_limit > 0)
- mtu -= 8; /* encap limit option */
- }
if (unlikely(nf_flow_exceeds_mtu(skb, mtu)))
return 0;
@@ -1158,7 +1118,6 @@ unsigned int
nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
const struct nf_hook_state *state)
{
- int encap_limit = IPV6_DEFAULT_TNL_ENCAP_LIMIT;
struct flow_offload_tuple_rhash *tuplehash;
struct nf_flowtable *flow_table = priv;
struct flow_offload_tuple *other_tuple;
@@ -1177,8 +1136,7 @@ nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
if (tuplehash == NULL)
return NF_ACCEPT;
- ret = nf_flow_offload_ipv6_forward(&ctx, flow_table, tuplehash, skb,
- encap_limit);
+ ret = nf_flow_offload_ipv6_forward(&ctx, flow_table, tuplehash, skb);
if (ret < 0)
return NF_DROP;
else if (ret == 0)
@@ -1198,7 +1156,7 @@ nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
ip6_daddr = &other_tuple->src_v6;
if (nf_flow_tunnel_v6_push(state->net, skb, other_tuple,
- &ip6_daddr, encap_limit) < 0)
+ &ip6_daddr) < 0)
return NF_DROP;
switch (tuplehash->tuple.xmit_type) {
diff --git a/tools/testing/selftests/net/netfilter/nft_flowtable.sh b/tools/testing/selftests/net/netfilter/nft_flowtable.sh
index 7a34ef468975..08ad07500e8a 100755
--- a/tools/testing/selftests/net/netfilter/nft_flowtable.sh
+++ b/tools/testing/selftests/net/netfilter/nft_flowtable.sh
@@ -592,7 +592,7 @@ ip -net "$nsr1" link set tun0 up
ip -net "$nsr1" addr add 192.168.100.1/24 dev tun0
ip netns exec "$nsr1" sysctl net.ipv4.conf.tun0.forwarding=1 > /dev/null
-ip -net "$nsr1" link add name tun6 type ip6tnl local fee1:2::1 remote fee1:2::2
+ip -net "$nsr1" link add name tun6 type ip6tnl local fee1:2::1 remote fee1:2::2 encaplimit none
ip -net "$nsr1" link set tun6 up
ip -net "$nsr1" addr add fee1:3::1/64 dev tun6 nodad
@@ -601,7 +601,7 @@ ip -net "$nsr2" link set tun0 up
ip -net "$nsr2" addr add 192.168.100.2/24 dev tun0
ip netns exec "$nsr2" sysctl net.ipv4.conf.tun0.forwarding=1 > /dev/null
-ip -net "$nsr2" link add name tun6 type ip6tnl local fee1:2::2 remote fee1:2::1 || ret=1
+ip -net "$nsr2" link add name tun6 type ip6tnl local fee1:2::2 remote fee1:2::1 encaplimit none || ret=1
ip -net "$nsr2" link set tun6 up
ip -net "$nsr2" addr add fee1:3::2/64 dev tun6 nodad
@@ -651,7 +651,7 @@ ip -net "$nsr1" route change default via 192.168.200.2
ip netns exec "$nsr1" sysctl net.ipv4.conf.tun0/10.forwarding=1 > /dev/null
ip netns exec "$nsr1" nft -a insert rule inet filter forward 'meta oif tun0.10 accept'
-ip -net "$nsr1" link add name tun6.10 type ip6tnl local fee1:4::1 remote fee1:4::2
+ip -net "$nsr1" link add name tun6.10 type ip6tnl local fee1:4::1 remote fee1:4::2 encaplimit none
ip -net "$nsr1" link set tun6.10 up
ip -net "$nsr1" addr add fee1:5::1/64 dev tun6.10 nodad
ip -6 -net "$nsr1" route delete default
@@ -670,7 +670,7 @@ ip -net "$nsr2" addr add 192.168.200.2/24 dev tun0.10
ip -net "$nsr2" route change default via 192.168.200.1
ip netns exec "$nsr2" sysctl net.ipv4.conf.tun0/10.forwarding=1 > /dev/null
-ip -net "$nsr2" link add name tun6.10 type ip6tnl local fee1:4::2 remote fee1:4::1 || ret=1
+ip -net "$nsr2" link add name tun6.10 type ip6tnl local fee1:4::2 remote fee1:4::1 encaplimit none || ret=1
ip -net "$nsr2" link set tun6.10 up
ip -net "$nsr2" addr add fee1:5::2/64 dev tun6.10 nodad
ip -6 -net "$nsr2" route delete default
--
2.47.3
^ permalink raw reply related
* [PATCH net 06/16] netfilter: ipset: Don't use test_bit() in lockless RCU readers in hash types
From: Pablo Neira Ayuso @ 2026-06-19 11:54 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw, horms
In-Reply-To: <20260619115452.93949-1-pablo@netfilter.org>
From: Jozsef Kadlecsik <kadlec@netfilter.org>
Sashiko pointed out that there are a few lockless RCU readers
using test_bit() which is a relaxed atomic operation and
provides no memory barrier guarantees. Use test_bit_acquire()
instead where the operation may run parallel with add/del/gc,
i.e. is not one from the next cases
- protected by region lock
- in a set destroy phase
- in a new/temporary set creation phase
Fixes: 18f84d41d34f ("netfilter: ipset: Introduce RCU locking in hash:* types")
Signed-off-by: Jozsef Kadlecsik <kadlec@netfilter.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/ipset/ip_set_hash_gen.h | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/net/netfilter/ipset/ip_set_hash_gen.h b/net/netfilter/ipset/ip_set_hash_gen.h
index 04e4627ddfc1..00c27b95207f 100644
--- a/net/netfilter/ipset/ip_set_hash_gen.h
+++ b/net/netfilter/ipset/ip_set_hash_gen.h
@@ -689,7 +689,7 @@ mtype_resize(struct ip_set *set, bool retried)
continue;
pos = smp_load_acquire(&n->pos);
for (j = 0; j < pos; j++) {
- if (!test_bit(j, n->used))
+ if (!test_bit_acquire(j, n->used))
continue;
data = ahash_data(n, j, dsize);
if (SET_ELEM_EXPIRED(set, data))
@@ -826,7 +826,7 @@ mtype_ext_size(struct ip_set *set, u32 *elements, size_t *ext_size)
continue;
pos = smp_load_acquire(&n->pos);
for (j = 0; j < pos; j++) {
- if (!test_bit(j, n->used))
+ if (!test_bit_acquire(j, n->used))
continue;
data = ahash_data(n, j, set->dsize);
if (!SET_ELEM_EXPIRED(set, data))
@@ -1201,7 +1201,7 @@ mtype_test_cidrs(struct ip_set *set, struct mtype_elem *d,
continue;
pos = smp_load_acquire(&n->pos);
for (i = 0; i < pos; i++) {
- if (!test_bit(i, n->used))
+ if (!test_bit_acquire(i, n->used))
continue;
data = ahash_data(n, i, set->dsize);
if (!mtype_data_equal(data, d, &multi))
@@ -1259,7 +1259,7 @@ mtype_test(struct ip_set *set, void *value, const struct ip_set_ext *ext,
}
pos = smp_load_acquire(&n->pos);
for (i = 0; i < pos; i++) {
- if (!test_bit(i, n->used))
+ if (!test_bit_acquire(i, n->used))
continue;
data = ahash_data(n, i, set->dsize);
if (!mtype_data_equal(data, d, &multi))
@@ -1396,7 +1396,7 @@ mtype_list(const struct ip_set *set,
continue;
pos = smp_load_acquire(&n->pos);
for (i = 0; i < pos; i++) {
- if (!test_bit(i, n->used))
+ if (!test_bit_acquire(i, n->used))
continue;
e = ahash_data(n, i, set->dsize);
if (SET_ELEM_EXPIRED(set, e))
--
2.47.3
^ permalink raw reply related
* [PATCH net 07/16] netfilter: ipset: Don't use test_bit() in lockless RCU readers in bitmap types
From: Pablo Neira Ayuso @ 2026-06-19 11:54 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw, horms
In-Reply-To: <20260619115452.93949-1-pablo@netfilter.org>
From: Jozsef Kadlecsik <kadlec@netfilter.org>
The pair of the patch "netfilter: ipset: Don't use test_bit() in lockless
RCU readers in hash types" for the bitmap types.
Fixes: 02a3231b6d82 ("netfilter: nf_conntrack_expect: store netns and zone in expectation")
Fixes: b0da3905bb1e ("netfilter: ipset: Bitmap types using the unified code base")
Signed-off-by: Jozsef Kadlecsik <kadlec@netfilter.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/ipset/ip_set_bitmap_gen.h | 4 +++-
net/netfilter/ipset/ip_set_bitmap_ip.c | 2 +-
net/netfilter/ipset/ip_set_bitmap_ipmac.c | 2 +-
net/netfilter/ipset/ip_set_bitmap_port.c | 2 +-
4 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/net/netfilter/ipset/ip_set_bitmap_gen.h b/net/netfilter/ipset/ip_set_bitmap_gen.h
index 798c7993635e..bb9b5bed10e1 100644
--- a/net/netfilter/ipset/ip_set_bitmap_gen.h
+++ b/net/netfilter/ipset/ip_set_bitmap_gen.h
@@ -165,6 +165,7 @@ mtype_add(struct ip_set *set, void *value, const struct ip_set_ext *ext,
ip_set_init_skbinfo(ext_skbinfo(x, set), ext);
/* Activate element */
+ smp_mb__before_atomic();
set_bit(e->id, map->members);
set->elements++;
@@ -219,7 +220,7 @@ mtype_list(const struct ip_set *set,
cond_resched_rcu();
id = cb->args[IPSET_CB_ARG0];
x = get_ext(set, map, id);
- if (!test_bit(id, map->members) ||
+ if (!test_bit_acquire(id, map->members) ||
(SET_WITH_TIMEOUT(set) &&
#ifdef IP_SET_BITMAP_STORED_TIMEOUT
mtype_is_filled(x) &&
@@ -278,6 +279,7 @@ mtype_gc(struct timer_list *t)
x = get_ext(set, map, id);
if (ip_set_timeout_expired(ext_timeout(x, set))) {
clear_bit(id, map->members);
+ smp_mb__after_atomic();
ip_set_ext_destroy(set, x);
set->elements--;
}
diff --git a/net/netfilter/ipset/ip_set_bitmap_ip.c b/net/netfilter/ipset/ip_set_bitmap_ip.c
index 5988b9bb9029..ac7febce074f 100644
--- a/net/netfilter/ipset/ip_set_bitmap_ip.c
+++ b/net/netfilter/ipset/ip_set_bitmap_ip.c
@@ -67,7 +67,7 @@ static int
bitmap_ip_do_test(const struct bitmap_ip_adt_elem *e,
struct bitmap_ip *map, size_t dsize)
{
- return !!test_bit(e->id, map->members);
+ return !!test_bit_acquire(e->id, map->members);
}
static int
diff --git a/net/netfilter/ipset/ip_set_bitmap_ipmac.c b/net/netfilter/ipset/ip_set_bitmap_ipmac.c
index 752f59ef8744..5921fd9d2dca 100644
--- a/net/netfilter/ipset/ip_set_bitmap_ipmac.c
+++ b/net/netfilter/ipset/ip_set_bitmap_ipmac.c
@@ -86,7 +86,7 @@ bitmap_ipmac_do_test(const struct bitmap_ipmac_adt_elem *e,
{
const struct bitmap_ipmac_elem *elem;
- if (!test_bit(e->id, map->members))
+ if (!test_bit_acquire(e->id, map->members))
return 0;
elem = get_const_elem(map->extensions, e->id, dsize);
if (e->add_mac && elem->filled == MAC_FILLED)
diff --git a/net/netfilter/ipset/ip_set_bitmap_port.c b/net/netfilter/ipset/ip_set_bitmap_port.c
index 7138e080def4..ca875c982424 100644
--- a/net/netfilter/ipset/ip_set_bitmap_port.c
+++ b/net/netfilter/ipset/ip_set_bitmap_port.c
@@ -58,7 +58,7 @@ static int
bitmap_port_do_test(const struct bitmap_port_adt_elem *e,
const struct bitmap_port *map, size_t dsize)
{
- return !!test_bit(e->id, map->members);
+ return !!test_bit_acquire(e->id, map->members);
}
static int
--
2.47.3
^ permalink raw reply related
* [PATCH net 05/16] netfilter: nfnetlink: make OOM conditions fatal
From: Pablo Neira Ayuso @ 2026-06-19 11:54 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw, horms
In-Reply-To: <20260619115452.93949-1-pablo@netfilter.org>
From: Florian Westphal <fw@strlen.de>
Batch support design attempts to process the entire batch, even
after a call has returned an error. The advantage is that userspace
gets all errors in one go.
The disadvantages are:
1. ->call() needs to cope with bad-state-due-to-previous-error
2. One error can trigger a cascade of followup errors which
can obfuscate the real problem.
Make -ENOMEM fatal, if we cannot allocate some object there is
a high chance we're going to report followup errors that are
nonsensical from userspace point of view.
Fixes: 0628b123c96d ("netfilter: nfnetlink: add batch support and use it from nf_tables")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nfnetlink.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/net/netfilter/nfnetlink.c b/net/netfilter/nfnetlink.c
index 47f3ed441f64..a1d480e4789c 100644
--- a/net/netfilter/nfnetlink.c
+++ b/net/netfilter/nfnetlink.c
@@ -531,6 +531,13 @@ static void nfnetlink_rcv_batch(struct sk_buff *skb, struct nlmsghdr *nlh,
status |= NFNL_BATCH_REPLAY;
goto done;
}
+
+ /* No point in further processing; followup errors can
+ * be bogus (e.g. -ENOENT because object that next
+ * rule/element wants could not be inserted).
+ */
+ if (err == -ENOMEM)
+ goto ack;
}
ack:
if (nlh->nlmsg_flags & NLM_F_ACK || err) {
--
2.47.3
^ permalink raw reply related
* [PATCH net 03/16] netfilter: xt_cluster: reject template conntracks in hash match
From: Pablo Neira Ayuso @ 2026-06-19 11:54 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw, horms
In-Reply-To: <20260619115452.93949-1-pablo@netfilter.org>
From: Wyatt Feng <bronzed_45_vested@icloud.com>
xt_cluster_mt() treats any non-NULL nf_ct_get() result as a fully
initialized conntrack and passes it to xt_cluster_hash().
This causes a state confusion bug when the raw table CT target attaches
a template conntrack to skb->_nfct before normal conntrack processing.
Templates carry IPS_TEMPLATE status but do not have a valid tuple for
hashing yet, so xt_cluster_hash() can hit its WARN_ON() path on the
zeroed l3num field.
Reject template conntracks before hashing them. This matches existing
netfilter handling for template objects and avoids hashing incomplete
conntrack state.
Fixes: 0269ea493734 ("netfilter: xtables: add cluster match")
Cc: stable@vger.kernel.org
Reported-by: Yuan Tan <yuantan098@gmail.com>
Reported-by: Yifan Wu <yifanwucs@gmail.com>
Reported-by: Juefei Pu <tomapufckgml@gmail.com>
Reported-by: Zhengchuan Liang <zcliangcn@gmail.com>
Reported-by: Xin Liu <bird@lzu.edu.cn>
Assisted-by: Codex:GPT-5.4
Signed-off-by: Wyatt Feng <bronzed_45_vested@icloud.com>
Signed-off-by: Ren Wei <n05ec@lzu.edu.cn>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/xt_cluster.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/netfilter/xt_cluster.c b/net/netfilter/xt_cluster.c
index 908fd5f2c3c8..eaf2511d63f0 100644
--- a/net/netfilter/xt_cluster.c
+++ b/net/netfilter/xt_cluster.c
@@ -107,7 +107,7 @@ xt_cluster_mt(const struct sk_buff *skb, struct xt_action_param *par)
}
ct = nf_ct_get(skb, &ctinfo);
- if (ct == NULL)
+ if (!ct || nf_ct_is_template(ct))
return false;
if (ct->master)
--
2.47.3
^ permalink raw reply related
* [PATCH net 02/16] netfilter: nf_queue: pin bridge device while NFQUEUE holds fake dst
From: Pablo Neira Ayuso @ 2026-06-19 11:54 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw, horms
In-Reply-To: <20260619115452.93949-1-pablo@netfilter.org>
From: Haoze Xie <royenheart@gmail.com>
The br_netfilter fake rtable is embedded in struct net_bridge and is
attached to bridged packets with skb_dst_set_noref(). If such a packet is
queued to NFQUEUE, __nf_queue() upgrades that fake dst with
skb_dst_force().
At that point the queued skb can hold a real dst reference after bridge
teardown has started. The problem is not that every bridged packet needs
its own dst reference. The problem is that NFQUEUE can keep the bridge
private fake dst alive after unregister begins.
Fix this by keeping the bridge fake dst model unchanged and pinning the
bridge master device only while the packet sits in NFQUEUE. Record the
bridge device in nf_queue_entry when the queued skb carries a bridge fake
dst, take a device reference for the queue lifetime, and drop it when the
queue entry is freed.
Also make sure queued entries are reaped when that bridge device goes
down, and drop the redundant nf_bridge_info_exists() test from the fake
dst detection.
This keeps netdev_priv(br->dev) alive until verdict completion, so the
embedded fake rtable and its metrics backing storage cannot be freed out
from under dst_release(). It also avoids the constant refcount bump and
avoids using ipv4-specific dst helpers for IPv6 bridge traffic.
Fixes: 34666d467cbf ("netfilter: bridge: move br_netfilter out of the core")
Cc: stable@kernel.org
Reported-by: Yuan Tan <yuantan098@gmail.com>
Reported-by: Yifan Wu <yifanwucs@gmail.com>
Reported-by: Juefei Pu <tomapufckgml@gmail.com>
Reported-by: Xin Liu <bird@lzu.edu.cn>
Signed-off-by: Haoze Xie <royenheart@gmail.com>
Signed-off-by: Ren Wei <n05ec@lzu.edu.cn>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
include/net/netfilter/nf_queue.h | 1 +
net/netfilter/nf_queue.c | 14 ++++++++++++++
net/netfilter/nfnetlink_queue.c | 3 +++
3 files changed, 18 insertions(+)
diff --git a/include/net/netfilter/nf_queue.h b/include/net/netfilter/nf_queue.h
index 3978c3174cdb..fc3e81c07364 100644
--- a/include/net/netfilter/nf_queue.h
+++ b/include/net/netfilter/nf_queue.h
@@ -18,6 +18,7 @@ struct nf_queue_entry {
unsigned int id;
unsigned int hook_index; /* index in hook_entries->hook[] */
#if IS_ENABLED(CONFIG_BRIDGE_NETFILTER)
+ struct net_device *bridge_dev;
struct net_device *physin;
struct net_device *physout;
#endif
diff --git a/net/netfilter/nf_queue.c b/net/netfilter/nf_queue.c
index 57b450024a99..73363ceedebe 100644
--- a/net/netfilter/nf_queue.c
+++ b/net/netfilter/nf_queue.c
@@ -68,6 +68,7 @@ static void nf_queue_entry_release_refs(struct nf_queue_entry *entry)
nf_queue_sock_put(state->sk);
#if IS_ENABLED(CONFIG_BRIDGE_NETFILTER)
+ dev_put(entry->bridge_dev);
dev_put(entry->physin);
dev_put(entry->physout);
#endif
@@ -84,6 +85,8 @@ static void __nf_queue_entry_init_physdevs(struct nf_queue_entry *entry)
{
#if IS_ENABLED(CONFIG_BRIDGE_NETFILTER)
const struct sk_buff *skb = entry->skb;
+ struct dst_entry *dst = skb_dst(skb);
+ struct net_device *dev = NULL;
if (nf_bridge_info_exists(skb)) {
entry->physin = nf_bridge_get_physindev(skb, entry->state.net);
@@ -92,6 +95,16 @@ static void __nf_queue_entry_init_physdevs(struct nf_queue_entry *entry)
entry->physin = NULL;
entry->physout = NULL;
}
+
+ if (entry->state.pf == NFPROTO_BRIDGE &&
+ dst && (dst->flags & DST_FAKE_RTABLE))
+ dev = dst_dev_rcu(dst);
+
+ /* Must hold a reference on the bridge device: dst_hold() protects
+ * the dst itself, but the fake rtable is embedded in bridge-private
+ * storage that netdevice teardown can free independently.
+ */
+ entry->bridge_dev = dev;
#endif
}
@@ -108,6 +121,7 @@ bool nf_queue_entry_get_refs(struct nf_queue_entry *entry)
dev_hold(state->out);
#if IS_ENABLED(CONFIG_BRIDGE_NETFILTER)
+ dev_hold(entry->bridge_dev);
dev_hold(entry->physin);
dev_hold(entry->physout);
#endif
diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
index c5e29fec419b..80ca077b81bd 100644
--- a/net/netfilter/nfnetlink_queue.c
+++ b/net/netfilter/nfnetlink_queue.c
@@ -1262,6 +1262,9 @@ dev_cmp(struct nf_queue_entry *entry, unsigned long ifindex)
if (physinif == ifindex || physoutif == ifindex)
return 1;
+
+ if (entry->bridge_dev && entry->bridge_dev->ifindex == ifindex)
+ return 1;
#endif
if (entry->skb_dev && entry->skb_dev->ifindex == ifindex)
return 1;
--
2.47.3
^ permalink raw reply related
* [PATCH net 01/16] netfilter: flowtable: fix offloaded ct timeout never being extended
From: Pablo Neira Ayuso @ 2026-06-19 11:54 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw, horms
In-Reply-To: <20260619115452.93949-1-pablo@netfilter.org>
From: Adrian Bente <adibente@gmail.com>
OpenWrt has recently migrated many platforms to kernel 6.18. On the
MediaTek platform, which supports hardware network offloading, WiFi
connections accelerated via the WED path were observed to drop after
roughly 300 seconds.
After several debugging sessions, assisted by the Claude LLM, the
problem was narrowed down as follows:
nf_flow_table_extend_ct_timeout() extends ct->timeout for offloaded
flows using:
cmpxchg(&ct->timeout, expires, new_timeout);
'expires' comes from nf_ct_expires(ct) and is a relative value, while
ct->timeout holds an absolute timestamp. The two are never equal, so
the cmpxchg always fails and the timeout is never extended.
This goes unnoticed for most flows, but a long-lived hardware (WED)
offloaded flow on MediaTek MT7986 eventually has ct->timeout decay to
zero, the conntrack entry is reaped and the connection breaks.
Open-code the relative value from a single READ_ONCE(ct->timeout)
snapshot and compare against that same absolute snapshot in the
cmpxchg, so the timeout extension actually takes effect while the
datapath remains authoritative if it updates ct->timeout concurrently.
Fixes: 03428ca5cee9 ("netfilter: conntrack: rework offload nf_conn timeout extension logic")
Cc: stable@vger.kernel.org
Suggested-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Adrian Bente <adibente@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nf_flow_table_core.c | 13 +++++++++----
1 file changed, 9 insertions(+), 4 deletions(-)
diff --git a/net/netfilter/nf_flow_table_core.c b/net/netfilter/nf_flow_table_core.c
index 785d8c244a77..99c5b9d671a0 100644
--- a/net/netfilter/nf_flow_table_core.c
+++ b/net/netfilter/nf_flow_table_core.c
@@ -505,8 +505,13 @@ static u32 nf_flow_table_tcp_timeout(const struct nf_conn *ct)
*/
static void nf_flow_table_extend_ct_timeout(struct nf_conn *ct)
{
- static const u32 min_timeout = 5 * 60 * HZ;
- u32 expires = nf_ct_expires(ct);
+ static const s32 min_timeout = 5 * 60 * HZ;
+ u32 ct_timeout = READ_ONCE(ct->timeout);
+ s32 expires;
+
+ expires = ct_timeout - nfct_time_stamp;
+ if (expires <= 0) /* already expired */
+ return;
/* normal case: large enough timeout, nothing to do. */
if (likely(expires >= min_timeout))
@@ -524,7 +529,7 @@ static void nf_flow_table_extend_ct_timeout(struct nf_conn *ct)
if (nf_ct_is_confirmed(ct) &&
test_bit(IPS_OFFLOAD_BIT, &ct->status)) {
u8 l4proto = nf_ct_protonum(ct);
- u32 new_timeout = true;
+ u32 new_timeout = 1;
switch (l4proto) {
case IPPROTO_UDP:
@@ -549,7 +554,7 @@ static void nf_flow_table_extend_ct_timeout(struct nf_conn *ct)
*/
if (new_timeout) {
new_timeout += nfct_time_stamp;
- cmpxchg(&ct->timeout, expires, new_timeout);
+ cmpxchg(&ct->timeout, ct_timeout, new_timeout);
}
}
--
2.47.3
^ permalink raw reply related
* [PATCH net 00/16] Netfilter fixes for net
From: Pablo Neira Ayuso @ 2026-06-19 11:54 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw, horms
Hi,
The following patchset contains Netfilter fixes for net, this contains
fixes for a few crash, but many of the patches are trivial/correctness
fixes. There is too one rework of the conntrack expectation timeout
strategy to deal with a possible race when removing an expectation.
1) Fix the incorrect flowtable timeout extension for entries in
hw offload, from Adrian Bente. This is correcting a defect in
the functionality, no crash.
2) Hold reference to device under the fake dst in br_netfilter,
from Haoze Xie. This is fixing a possible UaF if the device
is removed while packet is sitting in nfqueue.
3) Reject template conntrack in xt_cluster, otherwise access to
uninitialize conntrack fields are possible leading to WARN_ON
due to unset layer 3 protocol. From Wyatt Feng.
4) Make sure the IPv6 tunnel header is in the linear skb data
area before pulling. While at it remove incomplete NEXTHDR_DEST
support. From Lorenzo Bianconi. This possibly leading to crash
if IPv4 header is not linear, but GRO already guarantees this,
unlikely but still possible.
5) Bail out immediately if ENOMEM is seen in a nfnetlink batch,
no further processing since this will accumulate more bogus
errors. From Florian Westphal. Functionally improvements
under memory stress, no crash.
6) Use test_bit_acquire in ipset hash set to avoid reordering
of subsequent memory access. This is addressing a LLM related
report, no crash has been observed. From Jozsef Kadlecsik.
7) Use test_bit_acquire in ipset bitmap set too, for the same
reason as in the previous patch, from Jozsef Kadlecsik.
8) Call kfree_rcu() after rcu_assign_pointer() to address a
possible UaF, very hard to trigger. Never observed in practise,
reported by LLM. Also from Jozsef Kadlecsik.
9) Use disable_delayed_work_sync() instead cancel_delayed_work_sync()
to avoid that ipset GC handler re-queues work as reported by LLM.
From Jozsef Kadlecsik. This is for correctness.
10) Restore the check in nft_payload for exceeding payloda offset
over 2^16. From Florian Westphal. This fixes a silent truncation,
not a big deal, but better be assertive and reject it.
11) Validate NFT_META_BRI_IIFHWADDR can only run from bridge
prerouting. From Florian Westphal. Harmless but it could allow
to read bytes from skb->cb.
12) Zero out destination hardware address during the flowtable
path setup, also from Florian. This is a correctness fix, LLM
points that possible infoleak can happen but topology to achieve
it is not clear.
13) Skip IPv4 options if present when building the IPV4 reject reply.
Otherwise bytes in the IPv4 options header can be sent back to
origin where the ICMP header is being expected. Again from
Florian Westphal.
14) Replace timer API for expectation by GC worker approach. This
is implicitly fixing a race between nf_ct_remove_expectations()
which might fail to remove the expectation due to timer_del()
returning false because timer has expired and callback is
being run concurrently. This fix is addressing a crash that has
been already reported with a reproducer.
15) Store the master tuple in the expectation, since SLAB_TYPESAFE_BY_RCU
does not guarantee that accessing exp->master under rcu read lock
refer to the right master conntrack. Found by initial round of
fixes for expectation by LLM also found this.
16) Check if br_vlan_get_pvid_rcu() fails to address a possible stack
infoleak of 4-bytes. From Florian Westphal.
This is slightly over the 15 patch limit in batches, please, allow this
round to exceed it by one.
Please, pull these changes from:
git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf.git nf-26-06-19
Thanks.
----------------------------------------------------------------
The following changes since commit 96e7f9122aae0ed000ee321f324b812a447906d9:
eth: fbnic: take netif_addr_lock_bh() around rx mode address programming (2026-06-18 18:36:26 -0700)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf.git tags/nf-26-06-19
for you to fetch changes up to 05477f7a037c127854b58441f60b34210668f5c3:
netfilter: nft_meta_bridge: fix NFT_META_BRI_IIFPVID stack leak (2026-06-19 12:27:08 +0200)
----------------------------------------------------------------
netfilter pull request 26-06-19
----------------------------------------------------------------
Adrian Bente (1):
netfilter: flowtable: fix offloaded ct timeout never being extended
Florian Westphal (6):
netfilter: nfnetlink: make OOM conditions fatal
netfilter: nft_payload: reject offsets exceeding 65535 bytes
netfilter: nft_meta_bridge: add validate callback for get operations
netfilter: nft_flow_offload: zero device address for non-ether case
netfilter: nf_reject: skip iphdr options when looking for icmp header
netfilter: nft_meta_bridge: fix NFT_META_BRI_IIFPVID stack leak
Haoze Xie (1):
netfilter: nf_queue: pin bridge device while NFQUEUE holds fake dst
Jozsef Kadlecsik (4):
netfilter: ipset: Don't use test_bit() in lockless RCU readers in hash types
netfilter: ipset: Don't use test_bit() in lockless RCU readers in bitmap types
netfilter: ipset: fix order of kfree_rcu() and rcu_assign_pointer()
netfilter: ipset: make sure gc is properly stopped
Lorenzo Bianconi (1):
netfilter: flowtable: fix and simplify IP6IP6 tunnel handling
Pablo Neira Ayuso (2):
netfilter: nf_conntrack_expect: use conntrack GC to reap expectations
netfilter: nf_conntrack_expect: store master_tuple in expectation
Wyatt Feng (1):
netfilter: xt_cluster: reject template conntracks in hash match
include/net/netfilter/nf_conntrack_expect.h | 17 ++-
include/net/netfilter/nf_queue.h | 1 +
include/net/netfilter/nft_meta.h | 2 +
include/uapi/linux/netfilter/nf_conntrack_common.h | 1 +
net/bridge/netfilter/nft_meta_bridge.c | 23 +++-
net/ipv4/netfilter/nf_reject_ipv4.c | 2 +-
net/ipv6/ip6_tunnel.c | 7 +
net/netfilter/ipset/ip_set_bitmap_gen.h | 4 +-
net/netfilter/ipset/ip_set_bitmap_ip.c | 2 +-
net/netfilter/ipset/ip_set_bitmap_ipmac.c | 2 +-
net/netfilter/ipset/ip_set_bitmap_port.c | 2 +-
net/netfilter/ipset/ip_set_core.c | 4 +-
net/netfilter/ipset/ip_set_hash_gen.h | 12 +-
net/netfilter/nf_conntrack_broadcast.c | 1 +
net/netfilter/nf_conntrack_core.c | 33 ++++-
net/netfilter/nf_conntrack_expect.c | 147 +++++++++++----------
net/netfilter/nf_conntrack_h323_main.c | 4 +-
net/netfilter/nf_conntrack_helper.c | 10 +-
net/netfilter/nf_conntrack_netlink.c | 31 ++---
net/netfilter/nf_conntrack_sip.c | 13 +-
net/netfilter/nf_flow_table_core.c | 13 +-
net/netfilter/nf_flow_table_ip.c | 80 +++--------
net/netfilter/nf_flow_table_path.c | 4 +-
net/netfilter/nf_queue.c | 14 ++
net/netfilter/nfnetlink.c | 7 +
net/netfilter/nfnetlink_queue.c | 3 +
net/netfilter/nft_ct.c | 3 +-
net/netfilter/nft_meta.c | 5 +-
net/netfilter/nft_payload.c | 16 ++-
net/netfilter/xt_cluster.c | 2 +-
.../selftests/net/netfilter/nft_flowtable.sh | 8 +-
31 files changed, 268 insertions(+), 205 deletions(-)
^ permalink raw reply
* Re: [PATCH net v2] net: airoha: Fix TX scheduler queue mask loop upper bound
From: Lorenzo Bianconi @ 2026-06-19 11:39 UTC (permalink / raw)
To: Wayen Yan
Cc: netdev, horms, pabeni, kuba, edumazet, andrew+netdev,
angelogioacchino.delregno, matthias.bgg, linux-arm-kernel,
linux-mediatek
In-Reply-To: <178185574223.2378148.13454900445528174929@gmail.com>
[-- Attachment #1: Type: text/plain, Size: 2499 bytes --]
> In airoha_qdma_set_chan_tx_sched(), the loop clearing queue mask was
> using AIROHA_NUM_TX_RING (32) instead of AIROHA_NUM_QOS_QUEUES (8).
>
> Each channel has 8 queues, and TXQ_DISABLE_CHAN_QUEUE_MASK(channel, i)
> computes BIT(i + (channel * 8)). With i ranging 0..31, this causes:
> - channel 0: clears bit 0..31 (all 4 channels) instead of 0..7
> - channel 1: clears bit 8..31 (channels 1-3) instead of 8..15
> - channel 2: clears bit 16..31 (channels 2-3) instead of 16..23
> - channel 3: clears bit 24..31 (channel 3 only) - correct by accident
>
> While BIT(32+) on arm64 produces 64-bit values truncated to 0 in u32
> mask parameter, the loop still incorrectly clears queues within the
> same channel beyond queue 7.
>
> Even though this is functionally harmless (the register resets to 0
> and is only ever cleared, never set — so clearing extra bits is a
> no-op), the loop bound is semantically wrong and should be fixed for
> correctness and clarity.
>
> Fix by using AIROHA_NUM_QOS_QUEUES (8) as the loop upper bound.
>
> Fixes: ef1ca9271313 ("net: airoha: Add sched HTB offload support")
> Acked-by: Lorenzo Bianconi <lorenzo@kernel.org>
> Signed-off-by: Wayen Yan <win847@gmail.com>
> ---
> Changes in v2:
> - Add Lorenzo's Acked-by tag.
> - Clarify in commit message that this is semantically wrong but
> functionally harmless (register resets to 0, only cleared), as
> Lorenzo pointed out in review.
> - Rebase on current net tree.
>
> Link: https://lore.kernel.org/netdev/ajJIWMs4dVbfkHZ5@lore-desk/
> Link: https://lore.kernel.org/netdev/CAL_ptrs6J3Ryw_4mVTq5VgzkB4RreF5S0huHyLvd9YwWr1m6jAA@mail.gmail.com/
>
> drivers/net/ethernet/airoha/airoha_eth.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
> index d0c0c0ec8a..ca77747b44 100644
> --- a/drivers/net/ethernet/airoha/airoha_eth.c
> +++ b/drivers/net/ethernet/airoha/airoha_eth.c
> @@ -2212,7 +2212,7 @@ static int airoha_qdma_set_chan_tx_sched(struct net_device *dev,
> struct airoha_gdm_port *port = netdev_priv(dev);
it seems you have not rebased on top of net tree.
Regards,
Lorenzo
> int i;
>
> - for (i = 0; i < AIROHA_NUM_TX_RING; i++)
> + for (i = 0; i < AIROHA_NUM_QOS_QUEUES; i++)
> airoha_qdma_clear(port->qdma, REG_QUEUE_CLOSE_CFG(channel),
> TXQ_DISABLE_CHAN_QUEUE_MASK(channel, i));
>
> --
> 2.51.0
>
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]
^ permalink raw reply
* [PATCH net v2 2/2] net: airoha: fix netif_set_real_num_tx_queues for sparse QoS channels
From: Lorenzo Bianconi @ 2026-06-19 11:37 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Lorenzo Bianconi
Cc: Simon Horman, Wayen Yan, linux-arm-kernel, linux-mediatek, netdev
In-Reply-To: <20260619-airoha-qos-fixes-v2-0-5c43485038f9@kernel.org>
airoha_tc_htb_alloc_leaf_queue() assigns queue IDs based on the channel
index (opt->qid = AIROHA_NUM_TX_RING + channel), but updates
real_num_tx_queues with a simple increment (num_tx_queues + 1). When QoS
channels are allocated sparsely (e.g., channels 0 and 3 without 1 and
2), the returned qid can exceed real_num_tx_queues, causing out-of-bounds
accesses in the networking stack.
For example, allocating channel 0 then channel 3 results in
real_num_tx_queues = 34 but qid = 35, which is out of range [0, 34).
Fix this by computing real_num_tx_queues based on the highest active
channel index rather than using a simple counter, in both the allocation
and deletion paths.
Fixes: ef1ca9271313b ("net: airoha: Add sched HTB offload support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
---
drivers/net/ethernet/airoha/airoha_eth.c | 25 ++++++++++++++++---------
1 file changed, 16 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
index aa98d1823ab6..aa2ddfd3af9f 100644
--- a/drivers/net/ethernet/airoha/airoha_eth.c
+++ b/drivers/net/ethernet/airoha/airoha_eth.c
@@ -2789,7 +2789,7 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
struct tc_htb_qopt_offload *opt)
{
u32 channel = TC_H_MIN(opt->classid) % AIROHA_NUM_QOS_CHANNELS;
- int err, num_tx_queues = netdev->real_num_tx_queues;
+ int err, num_tx_queues = AIROHA_NUM_TX_RING + channel + 1;
struct airoha_gdm_dev *dev = netdev_priv(netdev);
struct airoha_qdma *qdma = dev->qdma;
@@ -2806,13 +2806,15 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
if (err)
goto error;
- err = netif_set_real_num_tx_queues(netdev, num_tx_queues + 1);
- if (err) {
- airoha_qdma_set_tx_rate_limit(netdev, channel, 0,
- opt->quantum);
- NL_SET_ERR_MSG_MOD(opt->extack,
- "failed setting real_num_tx_queues");
- goto error;
+ if (num_tx_queues > netdev->real_num_tx_queues) {
+ err = netif_set_real_num_tx_queues(netdev, num_tx_queues);
+ if (err) {
+ airoha_qdma_set_tx_rate_limit(netdev, channel, 0,
+ opt->quantum);
+ NL_SET_ERR_MSG_MOD(opt->extack,
+ "failed setting real_num_tx_queues");
+ goto error;
+ }
}
set_bit(channel, dev->qos_sq_bmap);
@@ -3003,13 +3005,18 @@ static int airoha_dev_setup_tc_block(struct net_device *dev,
static void airoha_tc_remove_htb_queue(struct net_device *netdev, int queue)
{
struct airoha_gdm_dev *dev = netdev_priv(netdev);
+ int num_tx_queues = AIROHA_NUM_TX_RING;
struct airoha_qdma *qdma = dev->qdma;
- netif_set_real_num_tx_queues(netdev, netdev->real_num_tx_queues - 1);
airoha_qdma_set_tx_rate_limit(netdev, queue, 0, 0);
clear_bit(queue, qdma->qos_channel_map);
clear_bit(queue, dev->qos_sq_bmap);
+
+ if (!bitmap_empty(dev->qos_sq_bmap, AIROHA_NUM_QOS_CHANNELS))
+ num_tx_queues += find_last_bit(dev->qos_sq_bmap,
+ AIROHA_NUM_QOS_CHANNELS) + 1;
+ netif_set_real_num_tx_queues(netdev, num_tx_queues);
}
static int airoha_tc_htb_delete_leaf_queue(struct net_device *netdev,
--
2.54.0
^ permalink raw reply related
* [PATCH net v2 1/2] net: airoha: Fix off-by-one in airoha_tc_remove_htb_queue()
From: Lorenzo Bianconi @ 2026-06-19 11:37 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Lorenzo Bianconi
Cc: Simon Horman, Wayen Yan, linux-arm-kernel, linux-mediatek, netdev
In-Reply-To: <20260619-airoha-qos-fixes-v2-0-5c43485038f9@kernel.org>
airoha_tc_htb_alloc_leaf_queue() computes the HTB QoS channel index
as opt->classid % AIROHA_NUM_QOS_CHANNELS and stores it in qos_sq_bmap.
However, airoha_tc_remove_htb_queue() clears the HTB configuration
using queue + 1 as the channel index, causing an off-by-one error.
Use queue directly as the QoS channel index to match the allocation
logic.
Fixes: ef1ca9271313b ("net: airoha: Add sched HTB offload support")
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
---
drivers/net/ethernet/airoha/airoha_eth.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
index 64dde6464f3f..aa98d1823ab6 100644
--- a/drivers/net/ethernet/airoha/airoha_eth.c
+++ b/drivers/net/ethernet/airoha/airoha_eth.c
@@ -3006,7 +3006,7 @@ static void airoha_tc_remove_htb_queue(struct net_device *netdev, int queue)
struct airoha_qdma *qdma = dev->qdma;
netif_set_real_num_tx_queues(netdev, netdev->real_num_tx_queues - 1);
- airoha_qdma_set_tx_rate_limit(netdev, queue + 1, 0, 0);
+ airoha_qdma_set_tx_rate_limit(netdev, queue, 0, 0);
clear_bit(queue, qdma->qos_channel_map);
clear_bit(queue, dev->qos_sq_bmap);
--
2.54.0
^ permalink raw reply related
* [PATCH net v2 0/2] airoha: fixes for sched HTB offload support
From: Lorenzo Bianconi @ 2026-06-19 11:37 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Lorenzo Bianconi
Cc: Simon Horman, Wayen Yan, linux-arm-kernel, linux-mediatek, netdev
---
Changes in v2:
- cosmetics
- Link to v1: https://lore.kernel.org/r/20260618-airoha-qos-fixes-v1-0-37192652157f@kernel.org
---
Lorenzo Bianconi (2):
net: airoha: Fix off-by-one in airoha_tc_remove_htb_queue()
net: airoha: fix netif_set_real_num_tx_queues for sparse QoS channels
drivers/net/ethernet/airoha/airoha_eth.c | 27 +++++++++++++++++----------
1 file changed, 17 insertions(+), 10 deletions(-)
---
base-commit: 96e7f9122aae0ed000ee321f324b812a447906d9
change-id: 20260618-airoha-qos-fixes-b6460b085680
Best regards,
--
Lorenzo Bianconi <lorenzo@kernel.org>
^ permalink raw reply
* Re: [PATCH net 2/2] net: airoha: fix netif_set_real_num_tx_queues for sparse QoS channels
From: Lorenzo Bianconi @ 2026-06-19 11:34 UTC (permalink / raw)
To: Simon Horman
Cc: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Wayen Yan, linux-arm-kernel, linux-mediatek, netdev
In-Reply-To: <20260619093529.GV827683@horms.kernel.org>
[-- Attachment #1: Type: text/plain, Size: 2052 bytes --]
> On Thu, Jun 18, 2026 at 08:00:30AM +0200, Lorenzo Bianconi wrote:
> > airoha_tc_htb_alloc_leaf_queue() assigns queue IDs based on the channel
> > index (opt->qid = AIROHA_NUM_TX_RING + channel), but updates
> > real_num_tx_queues with a simple increment (num_tx_queues + 1). When QoS
> > channels are allocated sparsely (e.g., channels 0 and 3 without 1 and
> > 2), the returned qid can exceed real_num_tx_queues, causing out-of-bounds
> > accesses in the networking stack.
> > For example, allocating channel 0 then channel 3 results in
> > real_num_tx_queues = 34 but qid = 35, which is out of range [0, 34).
> > Fix this by computing real_num_tx_queues based on the highest active
> > channel index rather than using a simple counter, in both the allocation
> > and deletion paths.
> >
> > Fixes: ef1ca9271313b ("net: airoha: Add sched HTB offload support")
> > Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> > ---
> > drivers/net/ethernet/airoha/airoha_eth.c | 15 ++++++++++++---
> > 1 file changed, 12 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
>
> ...
>
> > @@ -2806,7 +2806,10 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
> > if (err)
> > goto error;
> >
> > - err = netif_set_real_num_tx_queues(netdev, num_tx_queues + 1);
> > + if (num_tx_queues <= netdev->real_num_tx_queues)
> > + goto set_qos_sq_bmap;
> > +
> > + err = netif_set_real_num_tx_queues(netdev, num_tx_queues);
> > if (err) {
> > airoha_qdma_set_tx_rate_limit(netdev, channel, 0,
> > opt->quantum);
> > @@ -2815,6 +2818,7 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
> > goto error;
> > }
> >
> > +set_qos_sq_bmap:
>
> I would prefer if this could be achieved without a goto.
ack, I will fix it in v2.
Regards,
Lorenzo
>
> > set_bit(channel, dev->qos_sq_bmap);
> > opt->qid = AIROHA_NUM_TX_RING + channel;
> >
>
> ...
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]
^ permalink raw reply
* airoha_eth: PPE flow entries and EGRESS TRTCM shaping
From: Wayen Yan @ 2026-06-19 11:32 UTC (permalink / raw)
To: Lorenzo Bianconi; +Cc: netdev, linux-kernel
Hi Lorenzo,
While reviewing the airoha HTB offload code and your recent fix series,
I noticed that EGRESS TRTCM rate-limit buckets configured by
airoha_qdma_set_tx_rate_limit() do not appear to be referenced by
either the CPU xmit path or PPE-accelerated flows.
Specifically:
1) CPU xmit path always disables the meter:
In airoha_dev_xmit(), TXMSG.METER is hardcoded to 0x7f:
msg1 = FIELD_PREP(QDMA_ETH_TXMSG_METER_MASK, 0x7f);
The register comment notes "0x7f no meters", so even though
airoha_tc_htb_alloc_leaf_queue() configures TRTCM bucket[channel]
with the requested rate, CPU-path packets never hit those buckets.
2) PPE flow entries never bind to a TRTCM bucket:
In airoha_ppe_foe_entry_prepare(), the FOE 'data' field is
initialized with:
qdata = FIELD_PREP(AIROHA_FOE_SHAPER_ID, 0x7f);
And neither CHANNEL[15:11] nor QID[10:8] are ever set — they
remain zero in all PPE entry creation paths (tc offload,
HW autolearn, L2 subflow commit). This means PPE-accelerated
flows bypass TRTCM shaping entirely.
The only QoS-related bit set is IB2.PSE_QOS for wired LAN ports,
which is a flag rather than an index.
3) airoha_ppe_foe_flow_stats_update() does read CHANNEL|QID from
the data field and moves them to ACTDP, but since they were
never populated, this is effectively a no-op in practice.
I have a few questions:
- Does AIROHA_FOE_SHAPER_ID map to the same QDMA TRTCM meter index
space used by airoha_qdma_set_tx_rate_limit()? If so, setting
SHAPER_ID = channel for flows belonging to an HTB class would
enable per-flow egress shaping via PPE.
- Is the current 0x7f / disabled behavior intentional — i.e., PPE
flow shaping is simply not yet implemented, or is there a hardware
constraint I'm missing?
- For the CPU xmit path, is there a reason TXMSG.METER cannot be
set to the channel derived from skb_get_queue_mapping(), so that
CPU-path packets also respect the TRTCM rate limits?
Thanks,
Wayen
^ permalink raw reply
* Re: [PATCH] net: add sock_open() for unified socket creation
From: Alex Goltsev @ 2026-06-19 10:35 UTC (permalink / raw)
To: Al Viro; +Cc: davem, netdev, linux-kernel
In-Reply-To: <20260618211231.GB2636677@ZenIV>
> What's the point (and why not make it inline, while we are at it)?
> Are there really callers that would pass a non-constant value as the last argument,
> and if so, what are they doing next?
As for `inline`: in this case, it would have no practical significance.
The compiler already treats a simple inline function as a regular
symbol within the `EXPORT_SYMBOL` context, whereas a static inline
function (the standard
kernel template for helper functions) would completely break the
export to the LKM.
This function solves the problem of actual API fragmentation:
currently, there are
three nearly identical functions (sock_create, sock_create_kern,
sock_create_lite) with slightly different signatures. If new variants
are added in the future, this will turn into a “zoo,” similar to what
happened with `kmalloc` before it was unified. I propose unifying this
now, while maintaining backward compatibility (all three existing
functions will remain unchanged).
As for the last argument, yes, today it is usually a constant,
but that’s not the point. The purpose of the enumeration is to provide
a unified, explicit control interface. It’s important that if, in the future,
someone adds a new type of socket creation, existing calling programs won’t
panic or throw a compilation error, but will smoothly fall back to
the default case and return -EINVAL, which is a safe failure mode.
I’m also aware that in `sock_open`, the first argument passed to the
`sock_create_kern` branch is not safe; I’ve placed it there as a
placeholder and am considering how to elegantly pass the `struct net`
to the `sock_create_kern` branch.
^ permalink raw reply
* [PATCH net v2] net: airoha: fix BQL underflow and UAF in shared QDMA TX ring
From: Lorenzo Bianconi @ 2026-06-19 10:30 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Lorenzo Bianconi
Cc: Wayen Yan, linux-arm-kernel, linux-mediatek, netdev
When multiple netdevs share a QDMA TX ring and one device is stopped,
netdev_tx_reset_subqueue() zeroes that device's BQL counters while its
pending skbs remain in the shared HW TX ring. When NAPI later completes
those skbs via netdev_tx_completed_queue(), the already-zeroed
dql->num_queued counter underflows.
Moreover, in the airoha_remove() path, netdevs are unregistered
sequentially while skbs from previously unregistered netdevs may still
reference freed net_device memory via skb->dev, causing a use-after-free
during BQL accounting.
Fix both issues:
- Remove netdev_tx_reset_subqueue() from airoha_dev_stop() so pending
skbs are completed naturally by NAPI with proper BQL accounting.
- Introduce airoha_qdma_tx_flush() to stop NAPI and flush BQL counters
for all pending skbs while skb->dev references are still valid.
- Guard airoha_dev_xmit() with DEV_STATE_FLUSH to drop packets during
teardown.
- Move DMA engine start into probe and stop into airoha_qdma_cleanup().
Fixes: a9c2ca61fec7 ("net: airoha: Support multiple net_devices for a single FE GDM port")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
---
Changes in v2:
- Introduce airoha_qdma_tx_flush() to account BQL in airoha_remove() or
airoha_probe() error path.
- Fix possible NULL pointer dereference in airoha_qdma_cleanup().
- Introduce DEV_STATE_FLUSH().
- Move back airoha_hw_cleanup().
- Set proper Fixes tag.
- Link to v1: https://lore.kernel.org/r/20260618-airoha-bql-fixes-v1-1-ffd2c2089518@kernel.org
---
drivers/net/ethernet/airoha/airoha_eth.c | 87 +++++++++++++++++++++++---------
drivers/net/ethernet/airoha/airoha_eth.h | 1 +
2 files changed, 63 insertions(+), 25 deletions(-)
diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
index 64dde6464f3f..e81cd806b57b 100644
--- a/drivers/net/ethernet/airoha/airoha_eth.c
+++ b/drivers/net/ethernet/airoha/airoha_eth.c
@@ -1004,6 +1004,7 @@ static int airoha_qdma_tx_napi_poll(struct napi_struct *napi, int budget)
e = &q->entry[index];
skb = e->skb;
+ e->skb = NULL;
dma_unmap_single(eth->dev, e->dma_addr, e->dma_len,
DMA_TO_DEVICE);
@@ -1523,10 +1524,26 @@ static int airoha_qdma_init(struct platform_device *pdev,
return airoha_qdma_hw_init(qdma);
}
-static void airoha_qdma_cleanup(struct airoha_qdma *qdma)
+static void airoha_qdma_cleanup(struct airoha_eth *eth,
+ struct airoha_qdma *qdma)
{
int i;
+ if (test_bit(DEV_STATE_INITIALIZED, ð->state)) {
+ u32 status;
+
+ airoha_qdma_clear(qdma, REG_QDMA_GLOBAL_CFG,
+ GLOBAL_CFG_TX_DMA_EN_MASK |
+ GLOBAL_CFG_RX_DMA_EN_MASK);
+ if (read_poll_timeout(airoha_qdma_rr, status,
+ !(status & (GLOBAL_CFG_TX_DMA_BUSY_MASK |
+ GLOBAL_CFG_RX_DMA_BUSY_MASK)),
+ USEC_PER_MSEC, 50 * USEC_PER_MSEC, true,
+ qdma, REG_QDMA_GLOBAL_CFG))
+ dev_warn(eth->dev,
+ "QDMA DMA engine busy timeout\n");
+ }
+
for (i = 0; i < ARRAY_SIZE(qdma->q_rx); i++) {
if (!qdma->q_rx[i].ndesc)
continue;
@@ -1593,7 +1610,7 @@ static int airoha_hw_init(struct platform_device *pdev,
return 0;
error:
for (i = 0; i < ARRAY_SIZE(eth->qdma); i++)
- airoha_qdma_cleanup(ð->qdma[i]);
+ airoha_qdma_cleanup(eth, ð->qdma[i]);
return err;
}
@@ -1603,7 +1620,7 @@ static void airoha_hw_cleanup(struct airoha_eth *eth)
int i;
for (i = 0; i < ARRAY_SIZE(eth->qdma); i++)
- airoha_qdma_cleanup(ð->qdma[i]);
+ airoha_qdma_cleanup(eth, ð->qdma[i]);
airoha_ppe_deinit(eth);
}
@@ -1637,6 +1654,35 @@ static void airoha_qdma_stop_napi(struct airoha_qdma *qdma)
}
}
+static void airoha_qdma_tx_flush(struct airoha_qdma *qdma)
+{
+ int i;
+
+ airoha_qdma_stop_napi(qdma);
+
+ for (i = 0; i < ARRAY_SIZE(qdma->q_tx); i++) {
+ struct airoha_queue *q = &qdma->q_tx[i];
+ int j;
+
+ if (!q->ndesc)
+ continue;
+
+ spin_lock_bh(&q->lock);
+ for (j = 0; j < q->ndesc; j++) {
+ struct airoha_queue_entry *e = &q->entry[j];
+ struct sk_buff *skb = e->skb;
+ struct netdev_queue *txq;
+
+ if (!skb)
+ continue;
+
+ txq = skb_get_tx_queue(skb->dev, skb);
+ netdev_tx_completed_queue(txq, 1, skb->len);
+ }
+ spin_unlock_bh(&q->lock);
+ }
+}
+
static void airoha_dev_get_hw_stats(struct airoha_gdm_dev *dev)
{
struct airoha_gdm_port *port = dev->port;
@@ -1837,9 +1883,6 @@ static int airoha_dev_open(struct net_device *netdev)
}
port->users++;
- airoha_qdma_set(qdma, REG_QDMA_GLOBAL_CFG,
- GLOBAL_CFG_TX_DMA_EN_MASK |
- GLOBAL_CFG_RX_DMA_EN_MASK);
qdma->users++;
if (!airoha_is_lan_gdm_dev(dev) &&
@@ -1880,12 +1923,9 @@ static int airoha_dev_stop(struct net_device *netdev)
struct airoha_gdm_dev *dev = netdev_priv(netdev);
struct airoha_gdm_port *port = dev->port;
struct airoha_qdma *qdma = dev->qdma;
- int i;
netif_tx_disable(netdev);
airoha_set_vip_for_gdm_port(dev, false);
- for (i = 0; i < netdev->num_tx_queues; i++)
- netdev_tx_reset_subqueue(netdev, i);
if (--port->users)
airoha_set_port_mtu(dev->eth, port);
@@ -1893,19 +1933,7 @@ static int airoha_dev_stop(struct net_device *netdev)
airoha_set_gdm_port_fwd_cfg(qdma->eth,
REG_GDM_FWD_CFG(port->id),
FE_PSE_PORT_DROP);
-
- if (!--qdma->users) {
- airoha_qdma_clear(qdma, REG_QDMA_GLOBAL_CFG,
- GLOBAL_CFG_TX_DMA_EN_MASK |
- GLOBAL_CFG_RX_DMA_EN_MASK);
-
- for (i = 0; i < ARRAY_SIZE(qdma->q_tx); i++) {
- if (!qdma->q_tx[i].ndesc)
- continue;
-
- airoha_qdma_cleanup_tx_queue(&qdma->q_tx[i]);
- }
- }
+ qdma->users--;
return 0;
}
@@ -2191,6 +2219,9 @@ static netdev_tx_t airoha_dev_xmit(struct sk_buff *skb,
u16 index;
u8 fport;
+ if (test_bit(DEV_STATE_FLUSH, &dev->eth->state))
+ goto error;
+
qid = airoha_qdma_get_txq(qdma, skb_get_queue_mapping(skb));
tag = airoha_get_dsa_tag(skb, netdev);
@@ -3413,8 +3444,12 @@ static int airoha_probe(struct platform_device *pdev)
if (err)
goto error_netdev_free;
- for (i = 0; i < ARRAY_SIZE(eth->qdma); i++)
+ for (i = 0; i < ARRAY_SIZE(eth->qdma); i++) {
airoha_qdma_start_napi(ð->qdma[i]);
+ airoha_qdma_set(ð->qdma[i], REG_QDMA_GLOBAL_CFG,
+ GLOBAL_CFG_TX_DMA_EN_MASK |
+ GLOBAL_CFG_RX_DMA_EN_MASK);
+ }
for_each_child_of_node(pdev->dev.of_node, np) {
if (!of_device_is_compatible(np, "airoha,eth-mac"))
@@ -3437,8 +3472,9 @@ static int airoha_probe(struct platform_device *pdev)
return 0;
error_napi_stop:
+ set_bit(DEV_STATE_FLUSH, ð->state);
for (i = 0; i < ARRAY_SIZE(eth->qdma); i++)
- airoha_qdma_stop_napi(ð->qdma[i]);
+ airoha_qdma_tx_flush(ð->qdma[i]);
for (i = 0; i < ARRAY_SIZE(eth->ports); i++) {
struct airoha_gdm_port *port = eth->ports[i];
@@ -3474,8 +3510,9 @@ static void airoha_remove(struct platform_device *pdev)
struct airoha_eth *eth = platform_get_drvdata(pdev);
int i;
+ set_bit(DEV_STATE_FLUSH, ð->state);
for (i = 0; i < ARRAY_SIZE(eth->qdma); i++)
- airoha_qdma_stop_napi(ð->qdma[i]);
+ airoha_qdma_tx_flush(ð->qdma[i]);
for (i = 0; i < ARRAY_SIZE(eth->ports); i++) {
struct airoha_gdm_port *port = eth->ports[i];
diff --git a/drivers/net/ethernet/airoha/airoha_eth.h b/drivers/net/ethernet/airoha/airoha_eth.h
index 41d2e7a1f9fb..f6dce5e74e02 100644
--- a/drivers/net/ethernet/airoha/airoha_eth.h
+++ b/drivers/net/ethernet/airoha/airoha_eth.h
@@ -92,6 +92,7 @@ enum {
enum {
DEV_STATE_INITIALIZED,
DEV_STATE_REGISTERED,
+ DEV_STATE_FLUSH,
};
enum {
---
base-commit: a887f2c7da66a805a55fd8706d45faec85f646db
change-id: 20260618-airoha-bql-fixes-f57b2d108573
Best regards,
--
Lorenzo Bianconi <lorenzo@kernel.org>
^ permalink raw reply related
* Re: [PATCH net 3/6] ipv6: fix error handling in forwarding sysctl
From: Fernando Fernandez Mancera @ 2026-06-19 10:28 UTC (permalink / raw)
To: nicolas.dichtel, netdev
Cc: shemminger, dforster, gospo, ddutt, brian.haley, horms, pabeni,
kuba, edumazet, davem, idosch, dsahern
In-Reply-To: <fa7cbedb-33f3-4a79-acff-2b23cf3f57d8@6wind.com>
On 6/19/26 11:34 AM, Nicolas Dichtel wrote:
> Le 18/06/2026 à 18:22, Fernando Fernandez Mancera a écrit :
>> When writing to the forwarding sysctl, if proc_dointvec() fails to parse
>> the input, it returns a negative error code. The current implementation
>> is overwriting that error for write operations.
>>
>> This results in a silent failure, it returns a successful write although
>> the configuration was not modified at all. When modifying the "all"
>> variant it can also modify the configuration of existing interfaces to
>> the wrong value.
>>
>> Fix this by checking the return value of proc_dointvec() and returning
>> early on failure.
>>
>> Fixes: b325fddb7f86 ("ipv6: Fix sysctl unregistration deadlock")
> The bug existed before the git era.
> Maybe
> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
>
>
Hm, not really, AFAICS b325fddb7f86 is the first commit overwriting the
return value from proc_dointvec(). See:
@@ -3983,7 +3986,7 @@ int addrconf_sysctl_forward(ctl_table *ctl, int
write, struct file * filp,
ret = proc_dointvec(ctl, write, filp, buffer, lenp, ppos);
if (write)
- addrconf_fixup_forwarding(ctl, valp, val);
+ ret = addrconf_fixup_forwarding(ctl, valp, val);
return ret;
}
^ permalink raw reply
* Re: [PATCHv2 3/4] mmc: sdhci-esdhc-mcf: do not use readl()/writel() on ColdFire
From: Angelo Dureghello @ 2026-06-19 10:24 UTC (permalink / raw)
To: Greg Ungerer
Cc: linux-m68k, linux-kernel, arnd, wei.fang, frank.li, shenwei.wang,
imx, netdev, nico, adureghello, ulfh, linux-mmc, linux-can,
linux-spi, olteanv
In-Reply-To: <20260609142139.1563360-5-gerg@linux-m68k.org>
Hi Greg,
the driver behaves as before, and it is not using "edma",
but his internal dma interface, no new issues are introduced.
Over testing btw i noticed a possible issue, totally unrelated
with your changes. Will fix it eventually.
Tested-by: Angelo Dureghello <adureghello@baylibre.com>
Acked-by: Angelo Dureghello <adureghello@baylibre.com>
On Wed, Jun 10, 2026 at 12:13:00AM +1000, Greg Ungerer wrote:
> The implementation of the readX() and writeX() family of IO access
> functions is non-standard on ColdFire platforms. They check the supplied
> IO address and will return either big or little endian results based on
> that check. This is non-standard, they are expected to always return
> little-endian byte ordered data. Unfortunately this behavior also means
> that ioreadX()/iowroteX() and their big-endian counter parts
> ioreadXbe()/iowriteXbe() are wrong. This is now in the process of being
> cleaned up and fixed.
>
> Change the use of the readX() and writeX() access functions in this driver
> to use the recently defined specific ColdFire internal SoC hardware IO
> access functions mcf_read8()/mcf_read16()/mcf_read32() and
> mcf_write8()/mcf_write16()/mcf_write32().
>
> There is no functional change to the driver. Though it does have the
> effect of making the IO access slightly more efficient, since there is
> no longer a need to do the address check at every register access.
>
> Acked-by: Angelo Dureghello <adureghello@baylibre.com>
> Tested-by: Angelo Dureghello <adureghello@baylibre.com>
> Signed-off-by: Greg Ungerer <gerg@linux-m68k.org>
> ---
> v2: moved from RFC to PATCH
>
> drivers/mmc/host/sdhci-esdhc-mcf.c | 24 ++++++++++++------------
> 1 file changed, 12 insertions(+), 12 deletions(-)
>
> diff --git a/drivers/mmc/host/sdhci-esdhc-mcf.c b/drivers/mmc/host/sdhci-esdhc-mcf.c
> index 375fce5639d7..6853521e8b2c 100644
> --- a/drivers/mmc/host/sdhci-esdhc-mcf.c
> +++ b/drivers/mmc/host/sdhci-esdhc-mcf.c
> @@ -55,7 +55,7 @@ static inline void esdhc_clrset_be(struct sdhci_host *host,
> if (reg == SDHCI_HOST_CONTROL)
> val |= ESDHC_PROCTL_D3CD;
>
> - writel((readl(base) & ~mask) | val, base);
> + mcf_write32((mcf_read32(base) & ~mask) | val, base);
> }
>
> /*
> @@ -71,7 +71,7 @@ static void esdhc_mcf_writeb_be(struct sdhci_host *host, u8 val, int reg)
> if (reg == SDHCI_HOST_CONTROL) {
> u32 host_ctrl = ESDHC_DEFAULT_HOST_CONTROL;
> u8 dma_bits = (val & SDHCI_CTRL_DMA_MASK) >> 3;
> - u8 tmp = readb(host->ioaddr + SDHCI_HOST_CONTROL + 1);
> + u8 tmp = mcf_read8(host->ioaddr + SDHCI_HOST_CONTROL + 1);
>
> tmp &= ~0x03;
> tmp |= dma_bits;
> @@ -82,12 +82,12 @@ static void esdhc_mcf_writeb_be(struct sdhci_host *host, u8 val, int reg)
> */
> host_ctrl |= val;
> host_ctrl |= (dma_bits << 8);
> - writel(host_ctrl, host->ioaddr + SDHCI_HOST_CONTROL);
> + mcf_write32(host_ctrl, host->ioaddr + SDHCI_HOST_CONTROL);
>
> return;
> }
>
> - writel((readl(base) & mask) | (val << shift), base);
> + mcf_write32((mcf_read32(base) & mask) | (val << shift), base);
> }
>
> static void esdhc_mcf_writew_be(struct sdhci_host *host, u16 val, int reg)
> @@ -110,24 +110,24 @@ static void esdhc_mcf_writew_be(struct sdhci_host *host, u16 val, int reg)
> * As for the fsl driver,
> * we have to set the mode in a single write here.
> */
> - writel(val << 16 | mcf_data->aside,
> + mcf_write32(val << 16 | mcf_data->aside,
> host->ioaddr + SDHCI_TRANSFER_MODE);
> return;
> }
>
> - writel((readl(base) & mask) | (val << shift), base);
> + mcf_write32((mcf_read32(base) & mask) | (val << shift), base);
> }
>
> static void esdhc_mcf_writel_be(struct sdhci_host *host, u32 val, int reg)
> {
> - writel(val, host->ioaddr + reg);
> + mcf_write32(val, host->ioaddr + reg);
> }
>
> static u8 esdhc_mcf_readb_be(struct sdhci_host *host, int reg)
> {
> if (reg == SDHCI_HOST_CONTROL) {
> u8 __iomem *base = host->ioaddr + (reg & ~3);
> - u16 val = readw(base + 2);
> + u16 val = mcf_read16(base + 2);
> u8 dma_bits = (val >> 5) & SDHCI_CTRL_DMA_MASK;
> u8 host_ctrl = val & 0xff;
>
> @@ -137,7 +137,7 @@ static u8 esdhc_mcf_readb_be(struct sdhci_host *host, int reg)
> return host_ctrl;
> }
>
> - return readb(host->ioaddr + (reg ^ 0x3));
> + return mcf_read8(host->ioaddr + (reg ^ 0x3));
> }
>
> static u16 esdhc_mcf_readw_be(struct sdhci_host *host, int reg)
> @@ -149,14 +149,14 @@ static u16 esdhc_mcf_readw_be(struct sdhci_host *host, int reg)
> if (reg == SDHCI_HOST_VERSION)
> reg -= 2;
>
> - return readw(host->ioaddr + (reg ^ 0x2));
> + return mcf_read16(host->ioaddr + (reg ^ 0x2));
> }
>
> static u32 esdhc_mcf_readl_be(struct sdhci_host *host, int reg)
> {
> u32 val;
>
> - val = readl(host->ioaddr + reg);
> + val = mcf_read32(host->ioaddr + reg);
>
> /*
> * RM (25.3.9) sd pin clock must never exceed 25Mhz.
> @@ -245,7 +245,7 @@ static void esdhc_mcf_pltfm_set_clock(struct sdhci_host *host,
> * fvco = fsys * outdvi1 + 1
> * fshdc = fvco / outdiv3 + 1
> */
> - temp = readl(pll_dr);
> + temp = mcf_read32(pll_dr);
> fsys = pltfm_host->clock;
> fvco = fsys * ((temp & 0x1f) + 1);
> fesdhc = fvco / (((temp >> 10) & 0x1f) + 1);
> --
> 2.54.0
>
^ permalink raw reply
* Re: [PATCH net 5/6] ipv6: reset value and position for proxy_ndp sysctl restart
From: Fernando Fernandez Mancera @ 2026-06-19 10:09 UTC (permalink / raw)
To: nicolas.dichtel, netdev
Cc: shemminger, dforster, gospo, ddutt, brian.haley, horms, pabeni,
kuba, edumazet, davem, idosch, dsahern
In-Reply-To: <491465c3-0d1a-42f6-86fe-6c31812e23c9@6wind.com>
On 6/19/26 11:58 AM, Nicolas Dichtel wrote:
> Le 18/06/2026 à 18:22, Fernando Fernandez Mancera a écrit :
>> When handling proxy_ndp, if rtnl_net_trylock() fails, the operation is
>> retried but as the value was already modified by the initial
>> proc_dointvec() call, the restarted syscall will read the newly modified
>> value as the 'old' state.
>>
>> Fix this by restoring the original value and position pointer before
>> restarting the syscall.
> Is it not better to call rtnl_net_trylock() at the beginning of the function?
> It avoids flapping the sysctl value.
>
IMHO it is not better if we want to reduce the time we are holding RTNL
lock. I think the idea is that if the user introduces a invalid value,
we don't need to take the lock at all.
That is the general pattern I see around the sysctl code (IPv4 and
IPv6). Given the current efforts to reduce the usage of RTNL I think
this approach would be better.
In any case, it is not a blocker for me so if we all agree that your
suggestion is better I don't mind taking that path.
Thanks for all the reviews!
>>
>> Fixes: c92d5491a6d9 ("netconf: add support for IPv6 proxy_ndp")
>> Signed-off-by: Fernando Fernandez Mancera <fmancera@suse.de>
>> ---
>> net/ipv6/addrconf.c | 9 +++++++--
>> 1 file changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
>> index 8ff015975e27..1cfb223476bd 100644
>> --- a/net/ipv6/addrconf.c
>> +++ b/net/ipv6/addrconf.c
>> @@ -6483,8 +6483,9 @@ static int addrconf_sysctl_proxy_ndp(const struct ctl_table *ctl, int write,
>> void *buffer, size_t *lenp, loff_t *ppos)
>> {
>> int *valp = ctl->data;
>> - int ret;
>> + loff_t pos = *ppos;
>> int old, new;
>> + int ret;
>>
>> old = *valp;
>> ret = proc_dointvec(ctl, write, buffer, lenp, ppos);
>> @@ -6493,8 +6494,12 @@ static int addrconf_sysctl_proxy_ndp(const struct ctl_table *ctl, int write,
>> if (write && old != new) {
>> struct net *net = ctl->extra2;
>>
>> - if (!rtnl_net_trylock(net))
>> + if (!rtnl_net_trylock(net)) {
>> + /* Restore the original values before restarting */
>> + *valp = old;
>> + *ppos = pos;
>> return restart_syscall();
>> + }
>>
>> if (valp == &net->ipv6.devconf_dflt->proxy_ndp) {
>> inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
>
>
^ permalink raw reply
* Re: [PATCH net 6/6] ipv6: fix missing notification for ignore_routes_with_linkdown
From: Nicolas Dichtel @ 2026-06-19 10:02 UTC (permalink / raw)
To: Fernando Fernandez Mancera, netdev
Cc: shemminger, dforster, gospo, ddutt, brian.haley, horms, pabeni,
kuba, edumazet, davem, idosch, dsahern
In-Reply-To: <20260618162225.4588-7-fmancera@suse.de>
Le 18/06/2026 à 18:22, Fernando Fernandez Mancera a écrit :
> When changing the ignore_routes_with_linkdown sysctl for a specific
> interface, the RTM_NEWNETCONF netlink notification was not being emitted
> to userspace. Fix this by emitting the notification when needed.
>
> In addition, fix bogus return value for successful "all" and specific
> interface write operation leading to a wrong reset of the position
> pointer.
>
> Fixes: 35103d11173b ("net: ipv6 sysctl option to ignore routes when nexthop link is down")
> Signed-off-by: Fernando Fernandez Mancera <fmancera@suse.de>
Reviewed-by: Nicolas Dichtel <nicolas.dichtel@6wind.com
^ permalink raw reply
* Re: [PATCH net 5/6] ipv6: reset value and position for proxy_ndp sysctl restart
From: Nicolas Dichtel @ 2026-06-19 9:58 UTC (permalink / raw)
To: Fernando Fernandez Mancera, netdev
Cc: shemminger, dforster, gospo, ddutt, brian.haley, horms, pabeni,
kuba, edumazet, davem, idosch, dsahern
In-Reply-To: <20260618162225.4588-6-fmancera@suse.de>
Le 18/06/2026 à 18:22, Fernando Fernandez Mancera a écrit :
> When handling proxy_ndp, if rtnl_net_trylock() fails, the operation is
> retried but as the value was already modified by the initial
> proc_dointvec() call, the restarted syscall will read the newly modified
> value as the 'old' state.
>
> Fix this by restoring the original value and position pointer before
> restarting the syscall.
Is it not better to call rtnl_net_trylock() at the beginning of the function?
It avoids flapping the sysctl value.
>
> Fixes: c92d5491a6d9 ("netconf: add support for IPv6 proxy_ndp")
> Signed-off-by: Fernando Fernandez Mancera <fmancera@suse.de>
> ---
> net/ipv6/addrconf.c | 9 +++++++--
> 1 file changed, 7 insertions(+), 2 deletions(-)
>
> diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
> index 8ff015975e27..1cfb223476bd 100644
> --- a/net/ipv6/addrconf.c
> +++ b/net/ipv6/addrconf.c
> @@ -6483,8 +6483,9 @@ static int addrconf_sysctl_proxy_ndp(const struct ctl_table *ctl, int write,
> void *buffer, size_t *lenp, loff_t *ppos)
> {
> int *valp = ctl->data;
> - int ret;
> + loff_t pos = *ppos;
> int old, new;
> + int ret;
>
> old = *valp;
> ret = proc_dointvec(ctl, write, buffer, lenp, ppos);
> @@ -6493,8 +6494,12 @@ static int addrconf_sysctl_proxy_ndp(const struct ctl_table *ctl, int write,
> if (write && old != new) {
> struct net *net = ctl->extra2;
>
> - if (!rtnl_net_trylock(net))
> + if (!rtnl_net_trylock(net)) {
> + /* Restore the original values before restarting */
> + *valp = old;
> + *ppos = pos;
> return restart_syscall();
> + }
>
> if (valp == &net->ipv6.devconf_dflt->proxy_ndp) {
> inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
^ permalink raw reply
* [PATCH v6.1 3/3] netfilter: nf_tables: unconditionally bump set->nelems before insertion
From: Shivani Agarwal @ 2026-06-19 9:28 UTC (permalink / raw)
To: stable, gregkh
Cc: pablo, fw, phil, davem, edumazet, kuba, pabeni, horms,
netfilter-devel, coreteam, netdev, linux-kernel, ajay.kaher,
alexey.makhalov, vamsi-krishna.brahmajosyula, yin.ding,
tapas.kundu, Inseo An, Li hongliang, Sasha Levin, Shivani Agarwal
In-Reply-To: <20260619092850.1274076-1-shivani.agarwal@broadcom.com>
From: Pablo Neira Ayuso <pablo@netfilter.org>
[ Upstream commit def602e498a4f951da95c95b1b8ce8ae68aa733a ]
In case that the set is full, a new element gets published then removed
without waiting for the RCU grace period, while RCU reader can be
walking over it already.
To address this issue, add the element transaction even if set is full,
but toggle the set_full flag to report -ENFILE so the abort path safely
unwinds the set to its previous state.
As for element updates, decrement set->nelems to restore it.
A simpler fix is to call synchronize_rcu() in the error path.
However, with a large batch adding elements to already maxed-out set,
this could cause noticeable slowdown of such batches.
Fixes: 35d0ac9070ef ("netfilter: nf_tables: fix set->nelems counting with no NLM_F_EXCL")
Reported-by: Inseo An <y0un9sa@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
[ Minor conflict resolved. ]
Signed-off-by: Li hongliang <1468888505@139.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Shivani: Modified to apply on 6.1.y ]
Signed-off-by: Shivani Agarwal <shivani.agarwal@broadcom.com>
---
net/netfilter/nf_tables_api.c | 28 +++++++++++++++-------------
1 file changed, 15 insertions(+), 13 deletions(-)
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 15bfdf07c..196ac4e76 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -6388,6 +6388,7 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
struct nft_data_desc desc;
enum nft_registers dreg;
struct nft_trans *trans;
+ bool set_full = false;
u64 timeout;
u64 expiration;
int err, i;
@@ -6680,10 +6681,18 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
if (err < 0)
goto err_elem_free;
+ if (!(flags & NFT_SET_ELEM_CATCHALL)) {
+ unsigned int max = nft_set_maxsize(set), nelems;
+
+ nelems = atomic_inc_return(&set->nelems);
+ if (nelems > max)
+ set_full = true;
+ }
+
trans = nft_trans_elem_alloc(ctx, NFT_MSG_NEWSETELEM, set);
if (trans == NULL) {
err = -ENOMEM;
- goto err_elem_free;
+ goto err_set_size;
}
ext->genmask = nft_genmask_cur(ctx->net);
@@ -6715,23 +6724,16 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
goto err_element_clash;
}
- if (!(flags & NFT_SET_ELEM_CATCHALL)) {
- unsigned int max = nft_set_maxsize(set);
-
- if (!atomic_add_unless(&set->nelems, 1, max)) {
- err = -ENFILE;
- goto err_set_full;
- }
- }
-
nft_trans_elem(trans) = elem;
nft_trans_commit_list_add_tail(ctx->net, trans);
- return 0;
-err_set_full:
- nft_setelem_remove(ctx->net, set, &elem);
+ return set_full ? -ENFILE : 0;
+
err_element_clash:
kfree(trans);
+err_set_size:
+ if (!(flags & NFT_SET_ELEM_CATCHALL))
+ atomic_dec(&set->nelems);
err_elem_free:
nf_tables_set_elem_destroy(ctx, set, elem.priv);
err_parse_data:
--
2.53.0
^ permalink raw reply related
* [PATCH v6.1 2/3] netfilter: nf_tables: fix set size with rbtree backend
From: Shivani Agarwal @ 2026-06-19 9:28 UTC (permalink / raw)
To: stable, gregkh
Cc: pablo, fw, phil, davem, edumazet, kuba, pabeni, horms,
netfilter-devel, coreteam, netdev, linux-kernel, ajay.kaher,
alexey.makhalov, vamsi-krishna.brahmajosyula, yin.ding,
tapas.kundu, Sasha Levin, Shivani Agarwal
In-Reply-To: <20260619092850.1274076-1-shivani.agarwal@broadcom.com>
From: Pablo Neira Ayuso <pablo@netfilter.org>
[ Upstream commit 8d738c1869f611955d91d8d0fd0012d9ef207201 ]
The existing rbtree implementation uses singleton elements to represent
ranges, however, userspace provides a set size according to the number
of ranges in the set.
Adjust provided userspace set size to the number of singleton elements
in the kernel by multiplying the range by two.
Check if the no-match all-zero element is already in the set, in such
case release one slot in the set size.
Fixes: 0ed6389c483d ("netfilter: nf_tables: rename set implementations")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Shivani: Modified to apply on 6.1.y ]
Signed-off-by: Shivani Agarwal <shivani.agarwal@broadcom.com>
---
include/net/netfilter/nf_tables.h | 6 ++++
net/netfilter/nf_tables_api.c | 49 +++++++++++++++++++++++++++++--
net/netfilter/nft_set_rbtree.c | 43 +++++++++++++++++++++++++++
3 files changed, 96 insertions(+), 2 deletions(-)
diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index dafa0a32e..3329c2eae 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -422,6 +422,9 @@ struct nft_set_ext;
* @remove: remove element from set
* @walk: iterate over all set elements
* @get: get set elements
+ * @ksize: kernel set size
+ * @usize: userspace set size
+ * @adjust_maxsize: delta to adjust maximum set size
* @privsize: function to return size of set private data
* @init: initialize private data of new set instance
* @destroy: destroy private data of set instance
@@ -470,6 +473,9 @@ struct nft_set_ops {
const struct nft_set *set,
const struct nft_set_elem *elem,
unsigned int flags);
+ u32 (*ksize)(u32 size);
+ u32 (*usize)(u32 size);
+ u32 (*adjust_maxsize)(const struct nft_set *set);
void (*commit)(struct nft_set *set);
void (*abort)(const struct nft_set *set);
u64 (*privsize)(const struct nlattr * const nla[],
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index ec4bfe53b..15bfdf07c 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -4264,6 +4264,14 @@ static int nf_tables_fill_set_concat(struct sk_buff *skb,
return 0;
}
+static u32 nft_set_userspace_size(const struct nft_set_ops *ops, u32 size)
+{
+ if (ops->usize)
+ return ops->usize(size);
+
+ return size;
+}
+
static int nf_tables_fill_set(struct sk_buff *skb, const struct nft_ctx *ctx,
const struct nft_set *set, u16 event, u16 flags)
{
@@ -4328,7 +4336,8 @@ static int nf_tables_fill_set(struct sk_buff *skb, const struct nft_ctx *ctx,
if (!nest)
goto nla_put_failure;
if (set->size &&
- nla_put_be32(skb, NFTA_SET_DESC_SIZE, htonl(set->size)))
+ nla_put_be32(skb, NFTA_SET_DESC_SIZE,
+ htonl(nft_set_userspace_size(set->ops, set->size))))
goto nla_put_failure;
if (set->field_count > 1 &&
@@ -4698,6 +4707,15 @@ static bool nft_set_is_same(const struct nft_set *set,
return true;
}
+static u32 nft_set_kernel_size(const struct nft_set_ops *ops,
+ const struct nft_set_desc *desc)
+{
+ if (ops->ksize)
+ return ops->ksize(desc->size);
+
+ return desc->size;
+}
+
static int nf_tables_newset(struct sk_buff *skb, const struct nfnl_info *info,
const struct nlattr * const nla[])
{
@@ -4880,6 +4898,9 @@ static int nf_tables_newset(struct sk_buff *skb, const struct nfnl_info *info,
if (err < 0)
return err;
+ if (desc.size)
+ desc.size = nft_set_kernel_size(set->ops, &desc);
+
err = 0;
if (!nft_set_is_same(set, &desc, exprs, num_exprs, flags)) {
NL_SET_BAD_ATTR(extack, nla[NFTA_SET_NAME]);
@@ -4902,6 +4923,9 @@ static int nf_tables_newset(struct sk_buff *skb, const struct nfnl_info *info,
if (IS_ERR(ops))
return PTR_ERR(ops);
+ if (desc.size)
+ desc.size = nft_set_kernel_size(ops, &desc);
+
udlen = 0;
if (nla[NFTA_SET_USERDATA])
udlen = nla_len(nla[NFTA_SET_USERDATA]);
@@ -6327,6 +6351,27 @@ static bool nft_setelem_valid_key_end(const struct nft_set *set,
return true;
}
+static u32 nft_set_maxsize(const struct nft_set *set)
+{
+ u32 maxsize, delta;
+
+ if (!set->size)
+ return UINT_MAX;
+
+ if (set->ops->adjust_maxsize)
+ delta = set->ops->adjust_maxsize(set);
+ else
+ delta = 0;
+
+ if (check_add_overflow(set->size, set->ndeact, &maxsize))
+ return UINT_MAX;
+
+ if (check_add_overflow(maxsize, delta, &maxsize))
+ return UINT_MAX;
+
+ return maxsize;
+}
+
static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
const struct nlattr *attr, u32 nlmsg_flags)
{
@@ -6671,7 +6716,7 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
}
if (!(flags & NFT_SET_ELEM_CATCHALL)) {
- unsigned int max = set->size ? set->size + set->ndeact : UINT_MAX;
+ unsigned int max = nft_set_maxsize(set);
if (!atomic_add_unless(&set->nelems, 1, max)) {
err = -ENFILE;
diff --git a/net/netfilter/nft_set_rbtree.c b/net/netfilter/nft_set_rbtree.c
index 426becaad..26e1d994f 100644
--- a/net/netfilter/nft_set_rbtree.c
+++ b/net/netfilter/nft_set_rbtree.c
@@ -775,6 +775,46 @@ static bool nft_rbtree_estimate(const struct nft_set_desc *desc, u32 features,
return true;
}
+/* rbtree stores ranges as singleton elements, each range is composed of two
+ * elements ...
+ */
+static u32 nft_rbtree_ksize(u32 size)
+{
+ return size * 2;
+}
+
+/* ... hide this detail to userspace. */
+static u32 nft_rbtree_usize(u32 size)
+{
+ if (!size)
+ return 0;
+
+ return size / 2;
+}
+
+static u32 nft_rbtree_adjust_maxsize(const struct nft_set *set)
+{
+ struct nft_rbtree *priv = nft_set_priv(set);
+ struct nft_rbtree_elem *rbe;
+ struct rb_node *node;
+ const void *key;
+
+ node = rb_last(&priv->root);
+ if (!node)
+ return 0;
+
+ rbe = rb_entry(node, struct nft_rbtree_elem, node);
+ if (!nft_rbtree_interval_end(rbe))
+ return 0;
+
+ key = nft_set_ext_key(&rbe->ext);
+ if (memchr(key, 1, set->klen))
+ return 0;
+
+ /* this is the all-zero no-match element. */
+ return 1;
+}
+
const struct nft_set_type nft_set_rbtree_type = {
.features = NFT_SET_INTERVAL | NFT_SET_MAP | NFT_SET_OBJECT | NFT_SET_TIMEOUT,
.ops = {
@@ -791,5 +831,8 @@ const struct nft_set_type nft_set_rbtree_type = {
.lookup = nft_rbtree_lookup,
.walk = nft_rbtree_walk,
.get = nft_rbtree_get,
+ .ksize = nft_rbtree_ksize,
+ .usize = nft_rbtree_usize,
+ .adjust_maxsize = nft_rbtree_adjust_maxsize,
},
};
--
2.53.0
^ permalink raw reply related
* [PATCH v6.1 1/3] netfilter: nf_tables: always increment set element count
From: Shivani Agarwal @ 2026-06-19 9:28 UTC (permalink / raw)
To: stable, gregkh
Cc: pablo, fw, phil, davem, edumazet, kuba, pabeni, horms,
netfilter-devel, coreteam, netdev, linux-kernel, ajay.kaher,
alexey.makhalov, vamsi-krishna.brahmajosyula, yin.ding,
tapas.kundu, Shivani Agarwal
In-Reply-To: <20260619092850.1274076-1-shivani.agarwal@broadcom.com>
From: Florian Westphal <fw@strlen.de>
[ Upstream commit d4b7f29eb85c93893bc27388b37709efbc3c9a0e ]
At this time, set->nelems counter only increments when the set has
a maximum size.
All set elements decrement the counter unconditionally, this is
confusing.
Increment the counter unconditionally to make this symmetrical.
This would also allow changing the set maximum size after set creation
in a later patch.
Signed-off-by: Florian Westphal <fw@strlen.de>
[ Shivani: Modified to apply on 6.1.y ]
Signed-off-by: Shivani Agarwal <shivani.agarwal@broadcom.com>
---
net/netfilter/nf_tables_api.c | 11 +++++++----
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 0c4224282..ec4bfe53b 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -6670,10 +6670,13 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
goto err_element_clash;
}
- if (!(flags & NFT_SET_ELEM_CATCHALL) && set->size &&
- !atomic_add_unless(&set->nelems, 1, set->size + set->ndeact)) {
- err = -ENFILE;
- goto err_set_full;
+ if (!(flags & NFT_SET_ELEM_CATCHALL)) {
+ unsigned int max = set->size ? set->size + set->ndeact : UINT_MAX;
+
+ if (!atomic_add_unless(&set->nelems, 1, max)) {
+ err = -ENFILE;
+ goto err_set_full;
+ }
}
nft_trans_elem(trans) = elem;
--
2.53.0
^ permalink raw reply related
* [PATCH v6.1 0/3] Fix CVE-2026-23272
From: Shivani Agarwal @ 2026-06-19 9:28 UTC (permalink / raw)
To: stable, gregkh
Cc: pablo, fw, phil, davem, edumazet, kuba, pabeni, horms,
netfilter-devel, coreteam, netdev, linux-kernel, ajay.kaher,
alexey.makhalov, vamsi-krishna.brahmajosyula, yin.ding,
tapas.kundu, Shivani Agarwal
To fix CVE-2026-23272, commit def602e498a4 is required; however,
it depends on commit d4b7f29eb85c and 8d738c1869f6. Therefore,
both patches have been backported to v6.1.
Florian Westphal (1):
netfilter: nf_tables: always increment set element count
Pablo Neira Ayuso (2):
netfilter: nf_tables: fix set size with rbtree backend
netfilter: nf_tables: unconditionally bump set->nelems before
insertion
include/net/netfilter/nf_tables.h | 6 +++
net/netfilter/nf_tables_api.c | 72 ++++++++++++++++++++++++++-----
net/netfilter/nft_set_rbtree.c | 43 ++++++++++++++++++
3 files changed, 110 insertions(+), 11 deletions(-)
--
2.53.0
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox