[ofa-general] Re: regression

[Roland Dreier]

 > ethtool ops assume the rtnl_lock is held.  From net/core/dev.c:
 > 
 >                case SIOCETHTOOL:
 >                        dev_load(net, ifr.ifr_name);
 >                        rtnl_lock();
 >                        ret = dev_ethtool(net, &ifr);
 >                        rtnl_unlock();
 > 
 > 
 > 
 > I'm just avoiding the overhead of building the ifr struct and calling
 > the ioctl...

Yes, but cf the argument I'm having with davem on netdev right now about
whether the get_drvinfo ethtool method is safe to call without rtnl...

 - R.