Skip to contentSkip to conversation

netpoll: fix IPv6 local-address corruption

PATCH Applied Breno Leitao 2 participants 2 messages Apr 24 – Apr 28

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 netpoll: fix IPv6 local-address corruption

netpoll_setup() decides whether to auto-populate the local source address by testing np->local_ip.ip, which only inspects the first 4 bytes of the union inet_addr storage.

For an IPv6 netpoll whose caller-supplied local address has a zero high-32 bits (::1, ::<suffix>, IPv4-mapped ::ffff:a.b.c.d, etc.), this misdetects the address as unset (which they are not, but the first 4 bytes are empty), calls netpoll_take_ipv6() and overwrites it with whatever matching link-local/global address the device happens to expose first.

Introduce a helper netpoll_local_ip_unset() that picks the correct family-aware test (ipv6_addr_any() for IPv6, !.ip for IPv4) and use it from netpoll_setup().

Reproducer is something like:

echo "::2" > local_ip
echo 1 > enabled
cat local_ip

  # before this fix: 2001:db8::1   (caller-supplied ::2 was clobbered)
  # after  this fix: ::2
Fixes b7394d2429c1 ("netpoll: prepare for ipv6")

---
I've found this problem while using Chris' new AI toy (kres - Kernel code RESearch agent), against netpoll code.

https://github.com/masoncl/kres

---
 net/core/netpoll.c | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)
net/core/netpoll.c +18 -1
index cd74beffd209c..4381e0fc25bf4 100644
@@ -704,6 +704,23 @@ static int netpoll_take_ipv4(struct netpoll *np, struct net_device *ndev)
 	return 0;
 }
 
+/*
+ * Test whether the caller left np->local_ip unset, so that
+ * netpoll_setup() should auto-populate it from the egress device.
+ *
+ * np->local_ip is a union of __be32 (IPv4) and struct in6_addr (IPv6),
+ * so an IPv6 address whose first 4 bytes are zero (e.g. ::1, ::2,
+ * IPv4-mapped ::ffff:a.b.c.d) must not be tested via the IPv4 arm —
+ * doing so would misclassify a caller-supplied address as unset and
+ * silently overwrite it with whatever address the device exposes.
+ */
+static bool netpoll_local_ip_unset(const struct netpoll *np)
+{
+	if (np->ipv6)
+		return ipv6_addr_any(&np->local_ip.in6);
+	return !np->local_ip.ip;
+}
+
 int netpoll_setup(struct netpoll *np)
 {
 	struct net *net = current->nsproxy->net_ns;
@@ -747,7 +764,7 @@ int netpoll_setup(struct netpoll *np)
 		rtnl_lock();
 	}
 
-	if (!np->local_ip.ip) {
+	if (netpoll_local_ip_unset(np)) {
 		if (!np->ipv6) {
 			err = netpoll_take_ipv4(np, ndev);
 			if (err)

---
base-commit: e728258debd553c95d2e70f9cd97c9fde27c7130
change-id: 20260424-netpoll_fix-88842a798663

Best regards,

Signature · 1 line
Breno Leitao <[email protected]>
#1 lore raw
This patch was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:

Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:

Quoted 4 lines · On Fri, 24 Apr 2026 08:31:16 -0700 you wrote:

netpoll_setup() decides whether to auto-populate the local source address by testing np->local_ip.ip, which only inspects the first 4 bytes of the union inet_addr storage.

Quoted 8 lines

For an IPv6 netpoll whose caller-supplied local address has a zero high-32 bits (::1, ::<suffix>, IPv4-mapped ::ffff:a.b.c.d, etc.), this misdetects the address as unset (which they are not, but the first 4 bytes are empty), calls netpoll_take_ipv6() and overwrites it with whatever matching link-local/global address the device happens to expose first.

[...]

Here is the summary with links:
- netpoll: fix IPv6 local-address corruption

    https://git.kernel.org/netdev/net/c/3bc179bc7146

You are awesome, thank you!

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