The attached patch first checks if an interface is up, before netload goes and
look up the IP-address. If it isn't up, the console used to fill up with ioctl
errors. These aren't there anymore now when the interface is down. 

However, they do appear when the interface is up, but doesn't have an IP -
address. I couldn't find how to detect this, other then someone assuming in his
code that when you get an error on SIOCGIFADDR, this is the result of not having
an IP - address. If that is the only possible error, imho there shouldn't be a
perror-call consuming resources. I suspect that may also be related to the
following topic on Gentoo Forums:
http://forums.gentoo.org/viewtopic-t-353304-highlight-error+ictl.html
Isn't it possible that calling perror causes the high CPU usage?

PS: Fixed a spelling error too.