since r3181, nm-netlink-monitor.c uses libnl
remove a now obsolete retry (which seems buggy anyway, I can't figure out how
it could
go out of the recursion if there is a real error).
I didn't tested the change so please review carefully before applying (I just
began to
look at the source and I don't have the environment to test yet).
thanks,
Benoit Boissinot
Index: src/nm-netlink.c
===================================================================
--- src/nm-netlink.c (revision 3375)
+++ src/nm-netlink.c (working copy)
@@ -71,20 +71,8 @@
}
if (nl_connect (def_nl_handle, NETLINK_ROUTE) < 0) {
- /* HACK: try one more time. Because the netlink monitor for
link state
- * inits before we get here, it grabs the port that matches the
PID
- * of the NM process, which also happens to be the PID that
libnl uses
- * the first time too. The real fix is to convert
nm-netlink-monitor.c
- * over to use libnl.
- */
- nl_handle_destroy (def_nl_handle);
- def_nl_handle = NULL;
-
- def_nl_handle = nm_netlink_get_default_handle ();
- if (!def_nl_handle) {
- nm_error ("couldn't connect to netlink: %s",
nl_geterror ());
- return NULL;
- }
+ nm_error ("couldn't connect to netlink: %s", nl_geterror ());
+ return NULL;
}
return def_nl_handle;
--
:wq
_______________________________________________
NetworkManager-list mailing list
[email protected]
http://mail.gnome.org/mailman/listinfo/networkmanager-list
