[PATCH] Net:cache didn't flush when ipv6 rule changed

From: Lena Wang (王娜)
Date: Mon Feb 26 2024 - 08:11:26 EST


From a415abaf4fc90a27dd86357ea8be62d32956f7d8 Mon Sep 17 00:00:00 2001
From: shiming cheng <shiming.cheng@xxxxxxxxxxxx>
Date: Mon, 26 Feb 2024 20:17:58 +0800
Subject: [PATCH] Net:cache didn't flush when ipv6 rule changed

When changed from old rule&route configure to new one as below,
ipv6 cache dst_entry did not change to new route table as no
cache flush callback function, then forward to wrong out interface.
When fib6_check dst_entry, the fib6_node version[fn_sernm] is
always the same with socket dst_cookie, old cache dst_entry is
always used and no chance to update.

So we need to update fib6_node version when rule changed and
flush cache to avoid dispatching a wrong interface.

Signed-off-by: shiming cheng <shiming.cheng@xxxxxxxxxxxx>
---
net/ipv6/fib6_rules.c | 10 ++++++++++
1 files changed, 10 insertions(+)

diff --git a/net/ipv6/fib6_rules.c b/net/ipv6/fib6_rules.c
index 7523c4baef35..bec2cf4436e1 100644
--- a/net/ipv6/fib6_rules.c
+++ b/net/ipv6/fib6_rules.c
@@ -449,6 +449,15 @@ static size_t fib6_rule_nlmsg_payload(struct
fib_rule *rule)
+ nla_total_size(16); /* src */
}

+static void fib6_rule_flush_cache(struct fib_rules_ops *ops)
+{
+ struct net *net = ops->fro_net;
+ if (!net)
+ return;
+ rt_genid_bump_ipv6(net);
+ return;
+}
+
static const struct fib_rules_ops __net_initconst
fib6_rules_ops_template = {
.family = AF_INET6,
.rule_size = sizeof(struct fib6_rule),
@@ -461,6 +470,7 @@ static const struct fib_rules_ops __net_initconst
fib6_rules_ops_template = {
.compare = fib6_rule_compare,
.fill = fib6_rule_fill,
.nlmsg_payload = fib6_rule_nlmsg_payload,
+ .flush_cache = fib6_rule_flush_cache,
.nlgroup = RTNLGRP_IPV6_RULE,
.owner = THIS_MODULE,
.fro_net = &init_net,
--
2.18.0