Skip to contentSkip to conversation

multi: skip IPv4 logic in multi_select_virtual_addr() if no pool is configured

PATCH In review Antonio Quartulli 2 participants 2 messages Jun 10, 2020

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 multi: skip IPv4 logic in multi_select_virtual_addr() if no pool is configured

When no IPv4 pool is configured (but we have an IPv6 pool only), the multi_select_virtual_addr() function will spit a warning when allocating an address for a new client. This happens because the code will check for some IPv4 bits and will see that they are missing.

However, these bits are not really important, because in this use case we don't want to configure any IPv4 address at all.

For this reason it is safe to wrap this entire logic in an if-block that just does not execute when no IPv4 pool is configured.

This avoids the warning and will also avoid any other hidden side effect.

---
 src/openvpn/multi.c | 50 ++++++++++++++++++++++++---------------------
 1 file changed, 27 insertions(+), 23 deletions(-)
src/openvpn/multi.c +27 -23
index 2fbbe9ec..99472f14 100644
@@ -1504,36 +1504,40 @@ multi_select_virtual_addr(struct multi_context *m, struct multi_instance *mi)
                   ? print_in6_addr( remote_ipv6, 0, &gc )
                   : "(Not enabled)") );
 
-            /* set push_ifconfig_remote_netmask from pool ifconfig address(es) */
-            mi->context.c2.push_ifconfig_local = remote;
-            if (tunnel_type == DEV_TYPE_TAP || (tunnel_type == DEV_TYPE_TUN && tunnel_topology == TOP_SUBNET))
+            if (mi->context.options.ifconfig_pool_defined)
             {
-                mi->context.c2.push_ifconfig_remote_netmask = mi->context.options.ifconfig_pool_netmask;
-                if (!mi->context.c2.push_ifconfig_remote_netmask)
+                /* set push_ifconfig_remote_netmask from pool ifconfig address(es) */
+                mi->context.c2.push_ifconfig_local = remote;
+                if (tunnel_type == DEV_TYPE_TAP || (tunnel_type == DEV_TYPE_TUN && tunnel_topology == TOP_SUBNET))
                 {
-                    mi->context.c2.push_ifconfig_remote_netmask = mi->context.c1.tuntap->remote_netmask;
+                    mi->context.c2.push_ifconfig_remote_netmask = mi->context.options.ifconfig_pool_netmask;
+                    if (!mi->context.c2.push_ifconfig_remote_netmask)
+                    {
+                        mi->context.c2.push_ifconfig_remote_netmask = mi->context.c1.tuntap->remote_netmask;
+                    }
                 }
-            }
-            else if (tunnel_type == DEV_TYPE_TUN)
-            {
-                if (tunnel_topology == TOP_P2P)
+                else if (tunnel_type == DEV_TYPE_TUN)
                 {
-                    mi->context.c2.push_ifconfig_remote_netmask = mi->context.c1.tuntap->local;
+                    if (tunnel_topology == TOP_P2P)
+                    {
+                        mi->context.c2.push_ifconfig_remote_netmask = mi->context.c1.tuntap->local;
+                    }
+                    else if (tunnel_topology == TOP_NET30)
+                    {
+                        mi->context.c2.push_ifconfig_remote_netmask = local;
+                    }
                 }
-                else if (tunnel_topology == TOP_NET30)
+
+                if (mi->context.c2.push_ifconfig_remote_netmask)
                 {
-                    mi->context.c2.push_ifconfig_remote_netmask = local;
+                    mi->context.c2.push_ifconfig_defined = true;
+                }
+                else
+                {
+                    msg(D_MULTI_ERRORS,
+                        "MULTI: no --ifconfig-pool netmask parameter is available to push to %s",
+                        multi_instance_string(mi, false, &gc));
                 }
-            }
-
-            if (mi->context.c2.push_ifconfig_remote_netmask)
-            {
-                mi->context.c2.push_ifconfig_defined = true;
-            }
-            else
-            {
-                msg(D_MULTI_ERRORS, "MULTI: no --ifconfig-pool netmask parameter is available to push to %s",
-                    multi_instance_string(mi, false, &gc));
             }
 
             if (mi->context.options.ifconfig_ipv6_pool_defined)
Signature · 1 line
2.27.0
Acked-by on v1 1/1
#1 lore raw

The patch is quite trivial - if seen with "git show -w" :-) - it basically just encapsulates the whole "the pool has returned something, how does it need to be wrapped for pushing out?" into "is there an IPv4 pool at all?".

Tested on the t_server setup - warnings are gone, and ccd/ifconfig-push still works (which it should, on no-v4-pool or no-pool-at-all setups).

Your patch has been applied to the master branch.

commit 9002885bd8a10d9375dc4f0baf2df05395c86f1a

Author: Antonio Quartulli
Date:   Wed Jun 10 10:45:49 2020 +0200
     multi: skip IPv4 logic in multi_select_virtual_addr() if no pool is configured

     Signed-off-by: Antonio Quartulli <[email protected]...>
     Acked-by: Gert Doering <[email protected]...>
     Message-Id: <[email protected]...>
     URL: https://www.mail-archive.com/[email protected]/msg20012.html
     Signed-off-by: Gert Doering <[email protected]...>

--
kind regards,

Gert Doering