When rule policy is changed, ipv6 socket cache is not refreshed.
The sock's skb still uses a outdated route cache and was sent to
a wrong interface.
To avoid this error we should update fib node's version when
rule is changed. Then skb's route will be reroute checked as
route cache version is already different with fib node version.
The route cache is refreshed to match the latest rule.
Fixes: 101367c2f8c4 ("[IPV6]: Policy Routing Rules")
Signed-off-by: Shiming Cheng <[email protected]>
Signed-off-by: Lena Wang <[email protected]>
---
v5:
- rebase on the top of latest net/main branch.
v4:
- add "Fixes:" tag.
- update subject as requested.
v3:
- update patch description and name format in commit message.
v2:
- modify flush function same way as ipv4 flush cache.
- use tabs to aligh with existing code.
---
---
net/ipv6/fib6_rules.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/net/ipv6/fib6_rules.c b/net/ipv6/fib6_rules.c
index 7523c4baef35..52c04f0ac498 100644
--- a/net/ipv6/fib6_rules.c
+++ b/net/ipv6/fib6_rules.c
@@ -449,6 +449,11 @@ 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)
+{
+ rt_genid_bump_ipv6(ops->fro_net);
+}
+
static const struct fib_rules_ops __net_initconst fib6_rules_ops_template = {
.family = AF_INET6,
.rule_size = sizeof(struct fib6_rule),
@@ -461,6 +466,7 @@ 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
On 3/7/24 3:01 AM, Shiming Cheng wrote:
> When rule policy is changed, ipv6 socket cache is not refreshed.
> The sock's skb still uses a outdated route cache and was sent to
> a wrong interface.
>
> To avoid this error we should update fib node's version when
> rule is changed. Then skb's route will be reroute checked as
> route cache version is already different with fib node version.
> The route cache is refreshed to match the latest rule.
>
> Fixes: 101367c2f8c4 ("[IPV6]: Policy Routing Rules")
> Signed-off-by: Shiming Cheng <[email protected]>
> Signed-off-by: Lena Wang <[email protected]>
> ---
> v5:
> - rebase on the top of latest net/main branch.
> v4:
> - add "Fixes:" tag.
> - update subject as requested.
> v3:
> - update patch description and name format in commit message.
> v2:
> - modify flush function same way as ipv4 flush cache.
> - use tabs to aligh with existing code.
> ---
> ---
> net/ipv6/fib6_rules.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
Reviewed-by: David Ahern <[email protected]>
Hello:
This patch was applied to netdev/net.git (main)
by David S. Miller <[email protected]>:
On Thu, 7 Mar 2024 18:01:57 +0800 you wrote:
> When rule policy is changed, ipv6 socket cache is not refreshed.
> The sock's skb still uses a outdated route cache and was sent to
> a wrong interface.
>
> To avoid this error we should update fib node's version when
> rule is changed. Then skb's route will be reroute checked as
> route cache version is already different with fib node version.
> The route cache is refreshed to match the latest rule.
>
> [...]
Here is the summary with links:
- [net,v5] ipv6: fib6_rules: flush route cache when rule is changed
https://git.kernel.org/netdev/net/c/c4386ab4f6c6
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html