Fix FreeBSD/OpenBSD/NetBSD compiler warnings in get_default_gateway()

On these platforms (including DragonFly), get_default_gateway() would in some cases return false. As get_default_gateway() is defined as a void function, and none of the callers expect a return value -> just return without any value.

---
 route.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)

index d632b45..be23a89 100644
@@ -2581,7 +2581,7 @@ get_default_gateway (struct route_gateway_info *rgi)
       warn("writing to routing socket");
       gc_free (&gc);
       close(s);
-      return false;
+      return;
     }
 
   do {

@@ -2605,7 +2605,7 @@ get_default_gateway (struct route_gateway_info *rgi)
   else
     {
       gc_free (&gc);
-      return false;
+      return;
     }
 
 

@@ -2947,7 +2947,7 @@ get_default_gateway (struct route_gateway_info *rgi)
       warn("writing to routing socket");
       gc_free (&gc);
       close(s);
-      return false;
+      return;
     }
 
   do {

@@ -2971,7 +2971,7 @@ get_default_gateway (struct route_gateway_info *rgi)
   else
     {
       gc_free (&gc);
-      return false;
+      return;
     }
 
 

1.7.4.4

As the body of the check_subnet_conflict() function was disabled with an #if 0 clause, remove this function and clean up all places this function was called.

This function also expected get_default_gateway() to return data, which it does not do any more. Hence this function is "out-of-date" and no longer usefull.

---
 route.c |    1 -
 tun.c   |   42 ------------------------------------------
 tun.h   |    4 ----
 3 files changed, 0 insertions(+), 47 deletions(-)

index be23a89..9357cb6 100644
@@ -1001,7 +1001,6 @@ add_routes (struct route_list *rl, struct route_ipv6_list *rl6, const struct tun
       for (i = 0; i < rl->n; ++i)
 	{
 	  struct route *r = &rl->routes[i];
-	  check_subnet_conflict (r->network, r->netmask, "route");
 	  if (flags & ROUTE_DELETE_FIRST)
 	    delete_route (r, tt, flags, &rl->rgi, es);
 	  add_route (r, tt, flags, &rl->rgi, es);

index f28b8d8..de330b6 100644
@@ -248,10 +248,6 @@ const char *ifconfig_options_string (const struct tuntap* tt, bool remote, bool
 
 bool is_tun_p2p (const struct tuntap *tt);
 
-void check_subnet_conflict (const in_addr_t ip,
-			    const in_addr_t netmask,
-			    const char *prefix);
-
 void warn_on_use_of_common_subnets (void);
 
 /*

1.7.4.4

hi,

I agree with the assessment, but come to a different conclusion - I think the functionality is worthwile to have, so maybe we should try to repair it, instead of throwing it out completely.

Meeting topic?

gert

USENET is *not* the non-clickable part of WWW!
                                                           //www.muc.de/~gert/
Gert Doering - Munich, Germany                             [email protected]...
fax: +49-89-35655025                        [email protected]...

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Agreed, lets put this on the agenda. I've Cc'ed James, so that he can have a look at it too.

This warning feature got disabled in
commit 7fb0e07ec3f7c5f6514523085dbe02ea6b8933e2 [1], which refactored a lot of the depending code of check_subnet_conflict(). We would need to fix get_default_gateway() again. And further figure out what the reason for all those false positives.

This commit doesn't change any behaviour, as this code is already removed via the #if 0. Hopefully most compilers won't bring in calls to an empty function (which is the current state of check_subnet_conflict() now)

kind regards,

David Sommerseth

[1]
<http://openvpn.git.sourceforge.net/git/gitweb.cgi?p=openvpn/openvpn-testing.git;a=commitdiff;h=7fb0e07ec3f7c5f6514523085dbe02ea6b8933e2>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk7KQbkACgkQDC186MBRfrotHQCfWSA/TXLueXP5+fp3pjlDXBcp f2YAnjRuuc7WbPtT1Mgn+GtZbUC+Ewwp
=k46c
-----END PGP SIGNATURE-----

Hi,

ACK. Fallout of the recent Linux / route refactoring, as far as I can see - it's on my to-be-cleaned-up-List for FreeBSD, but having it out of the way already means less patches to send in for me :-)

gert

USENET is *not* the non-clickable part of WWW!
                                                           //www.muc.de/~gert/
Gert Doering - Munich, Germany                             [email protected]...
fax: +49-89-35655025                        [email protected]...

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Applied to -testing and -stable master branches

commit 93626f2cf71b4803aa1248a928bfc1ac8c9da18b

    Signed-off-by: David Sommerseth <[email protected]...>
    Acked-by: Gert Doering <[email protected]...>

kind regards,

David Sommerseth
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk7KRKUACgkQDC186MBRfroGogCggUMfWYvTxmiiOBRyp+/FcGSL lJAAn3Z79b24KQazk/Rw+7fKKKMe0xmu
=MdL+
-----END PGP SIGNATURE-----