From: Eric Woudstra <ericwouds@gmail.com>
To: Pablo Neira Ayuso <pablo@netfilter.org>,
Jozsef Kadlecsik <kadlec@netfilter.org>,
"David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
Simon Horman <horms@kernel.org>,
Nikolay Aleksandrov <razor@blackwall.org>
Cc: netfilter-devel@vger.kernel.org, netdev@vger.kernel.org,
Eric Woudstra <ericwouds@gmail.com>
Subject: [PATCH v3 nf-next 2/3] netfilter: nf_flow_table_core: teardown direct xmit when destination changed
Date: Tue, 17 Jun 2025 09:00:06 +0200 [thread overview]
Message-ID: <20250617070007.23812-3-ericwouds@gmail.com> (raw)
In-Reply-To: <20250617070007.23812-1-ericwouds@gmail.com>
In case of a bridge in the forward-fastpath or bridge-fastpath the fdb is
used to create the tuple. In case of roaming at layer 2 level, for example
802.11r, the destination device is changed in the fdb. The destination
device of a direct transmitting tuple is no longer valid and traffic is
send to the wrong destination. Also the hardware offloaded fastpath is not
valid anymore.
In case of roaming, a switchdev notification is send to delete the old fdb
entry. Upon receiving this notification, mark all direct transmitting flows
with the same ifindex, vid and hardware address as the fdb entry to be
teared down. The hardware offloaded fastpath is still in effect, so
minimize the delay of the work queue by setting the delay to zero.
Signed-off-by: Eric Woudstra <ericwouds@gmail.com>
---
net/netfilter/nf_flow_table_core.c | 65 ++++++++++++++++++++++++++++++
1 file changed, 65 insertions(+)
diff --git a/net/netfilter/nf_flow_table_core.c b/net/netfilter/nf_flow_table_core.c
index 992958db4a19..61f03907d103 100644
--- a/net/netfilter/nf_flow_table_core.c
+++ b/net/netfilter/nf_flow_table_core.c
@@ -13,6 +13,7 @@
#include <net/netfilter/nf_conntrack_core.h>
#include <net/netfilter/nf_conntrack_l4proto.h>
#include <net/netfilter/nf_conntrack_tuple.h>
+#include <net/switchdev.h>
static DEFINE_MUTEX(flowtable_lock);
static LIST_HEAD(flowtables);
@@ -745,6 +746,63 @@ void nf_flow_table_cleanup(struct net_device *dev)
}
EXPORT_SYMBOL_GPL(nf_flow_table_cleanup);
+struct flow_cleanup_data {
+ const unsigned char *addr;
+ int ifindex;
+ u16 vid;
+ bool found;
+};
+
+static void nf_flow_table_do_cleanup_addr(struct nf_flowtable *flow_table,
+ struct flow_offload *flow, void *data)
+{
+ struct flow_cleanup_data *cud = data;
+
+ if ((flow->tuplehash[0].tuple.xmit_type == FLOW_OFFLOAD_XMIT_DIRECT &&
+ flow->tuplehash[0].tuple.out.ifidx == cud->ifindex &&
+ flow->tuplehash[0].tuple.out.bridge_vid == cud->vid &&
+ ether_addr_equal(flow->tuplehash[0].tuple.out.h_dest, cud->addr)) ||
+ (flow->tuplehash[1].tuple.xmit_type == FLOW_OFFLOAD_XMIT_DIRECT &&
+ flow->tuplehash[1].tuple.out.ifidx == cud->ifindex &&
+ flow->tuplehash[1].tuple.out.bridge_vid == cud->vid &&
+ ether_addr_equal(flow->tuplehash[1].tuple.out.h_dest, cud->addr))) {
+ flow_offload_teardown(flow);
+ cud->found = true;
+ }
+}
+
+static int nf_flow_table_switchdev_event(struct notifier_block *unused,
+ unsigned long event, void *ptr)
+{
+ struct switchdev_notifier_fdb_info *fdb_info;
+ struct nf_flowtable *flowtable;
+ struct flow_cleanup_data cud;
+
+ if (event != SWITCHDEV_FDB_DEL_TO_DEVICE)
+ return NOTIFY_DONE;
+
+ fdb_info = ptr;
+ cud.addr = fdb_info->addr;
+ cud.vid = fdb_info->vid;
+ cud.ifindex = fdb_info->info.dev->ifindex;
+
+ mutex_lock(&flowtable_lock);
+ list_for_each_entry(flowtable, &flowtables, list) {
+ cud.found = false;
+ nf_flow_table_iterate(flowtable, nf_flow_table_do_cleanup_addr, &cud);
+ if (cud.found)
+ mod_delayed_work(system_power_efficient_wq,
+ &flowtable->gc_work, 0);
+ }
+ mutex_unlock(&flowtable_lock);
+
+ return NOTIFY_DONE;
+}
+
+static struct notifier_block nf_flow_table_switchdev_nb __read_mostly = {
+ .notifier_call = nf_flow_table_switchdev_event,
+};
+
void nf_flow_table_free(struct nf_flowtable *flow_table)
{
mutex_lock(&flowtable_lock);
@@ -818,6 +876,10 @@ static int __init nf_flow_table_module_init(void)
if (ret)
goto out_offload;
+ ret = register_switchdev_notifier(&nf_flow_table_switchdev_nb);
+ if (ret < 0)
+ goto out_sw_noti;
+
ret = nf_flow_register_bpf();
if (ret)
goto out_bpf;
@@ -825,6 +887,8 @@ static int __init nf_flow_table_module_init(void)
return 0;
out_bpf:
+ unregister_switchdev_notifier(&nf_flow_table_switchdev_nb);
+out_sw_noti:
nf_flow_table_offload_exit();
out_offload:
unregister_pernet_subsys(&nf_flow_table_net_ops);
@@ -833,6 +897,7 @@ static int __init nf_flow_table_module_init(void)
static void __exit nf_flow_table_module_exit(void)
{
+ unregister_switchdev_notifier(&nf_flow_table_switchdev_nb);
nf_flow_table_offload_exit();
unregister_pernet_subsys(&nf_flow_table_net_ops);
}
--
2.47.1
next prev parent reply other threads:[~2025-06-17 7:00 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-06-17 7:00 [PATCH v3 nf-next 0/3] flow offload teardown when layer 2 roaming Eric Woudstra
2025-06-17 7:00 ` [PATCH v3 nf-next 1/3] netfilter: flow: Add bridge_vid member Eric Woudstra
2025-06-17 7:00 ` Eric Woudstra [this message]
2025-09-08 21:56 ` [PATCH v3 nf-next 2/3] netfilter: nf_flow_table_core: teardown direct xmit when destination changed Florian Westphal
2025-06-17 7:00 ` [PATCH v3 nf-next 3/3] netfilter: nf_flow_table_ip: don't follow fastpath when marked teardown Eric Woudstra
2025-09-08 21:06 ` Pablo Neira Ayuso
2025-09-02 8:58 ` [PATCH v3 nf-next 0/3] flow offload teardown when layer 2 roaming Eric Woudstra
2025-09-02 12:32 ` Florian Westphal
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250617070007.23812-3-ericwouds@gmail.com \
--to=ericwouds@gmail.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=horms@kernel.org \
--cc=kadlec@netfilter.org \
--cc=kuba@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=netfilter-devel@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=pablo@netfilter.org \
--cc=razor@blackwall.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).