Re: [PATCH net-next v23 18/23] ovpn: implement peer add/get/dump/delete via netlink

On 17/03/2025 13:47, Sabrina Dubroca wrote:

Hello,

A few comments since it seems you'll have to send one more version
(otherwise they could be fixed later).

2025-03-12, 21:54:27 +0100, Antonio Quartulli wrote:

quoted

diff --git a/drivers/net/ovpn/netlink.c b/drivers/net/ovpn/netlink.c
index 8d267d4c82283d9b5f989478102086ce385195d5..407b5b908be431625a3c258e7887211ef0f4b197 100644
--- a/drivers/net/ovpn/netlink.c
+++ b/drivers/net/ovpn/netlink.c

[...]

quoted

+static int ovpn_nl_peer_modify(struct ovpn_peer *peer, struct genl_info *info,
+			       struct nlattr **attrs)
+{
+	struct sockaddr_storage ss = {};
+	void *local_ip = NULL;
+	u32 interv, timeout;
+	bool rehash = false;
+	int ret;
+
+	spin_lock_bh(&peer->lock);
+
+	if (ovpn_nl_attr_sockaddr_remote(attrs, &ss)) {
+		/* we carry the local IP in a generic container.
+		 * ovpn_peer_reset_sockaddr() will properly interpret it
+		 * based on ss.ss_family
+		 */
+		local_ip = ovpn_nl_attr_local_ip(attrs);
+
+		/* set peer sockaddr */
+		ret = ovpn_peer_reset_sockaddr(peer, &ss, local_ip);
+		if (ret < 0) {
+			NL_SET_ERR_MSG_FMT_MOD(info->extack,
+					       "cannot set peer sockaddr: %d",
+					       ret);
+			goto err_unlock;
+		}

Similar to the floating case, this should reset the peer dst_cache? In
both cases we're updating the remote peer's address and our local
address.

makes sense.
Then I should probably move the call to dst_cache_reset() to ovpn_peer_reset_sockaddr() and slightly adjust ovpn_peer_endpoints_update() to avoid a double invocation.

...

quoted

diff --git a/drivers/net/ovpn/peer.c b/drivers/net/ovpn/peer.c
index 0d8b12fd5de4cd6fe15455b435c7d6807203a825..aead1d75400a604a320c886aed5308fb20475da3 100644
--- a/drivers/net/ovpn/peer.c
+++ b/drivers/net/ovpn/peer.c

...

quoted

@@ -1335,8 +1356,11 @@ void ovpn_peer_keepalive_work(struct work_struct *work)
  		netdev_dbg(ovpn->dev,
  			   "scheduling keepalive work: now=%llu next_run=%llu delta=%llu\n",
  			   next_run, now, next_run - now);
+		/* due to the waiting above, the next_run deadline may have
+		 * passed: in this case we reschedule the worker immediately
+		 */
  		schedule_delayed_work(&ovpn->keepalive_work,
-				      (next_run - now) * HZ);
+				      (now - next_run) * HZ);

This whole hunk should be dropped, no? The comment is outdated, and
the sign swap is confusing me.

ouch right. fix-during-rebase gone slightly wrong.

Thanks!

-- 
Antonio Quartulli
OpenVPN Inc.