On Thu, Oct 20, 2016 at 09:19:29PM +0000, Asbjoern Sloth Toennesen wrote: > On Thu, 20 Oct 2016 20:08:26 +0000 (UTC), a...@asbjorn.biz wrote: > > diff --git a/net/l2tp/l2tp_eth.c b/net/l2tp/l2tp_eth.c > > index 965f7e3..ba82dcc 100644 > > --- a/net/l2tp/l2tp_eth.c > > +++ b/net/l2tp/l2tp_eth.c > > @@ -259,6 +259,7 @@ static int l2tp_eth_create(struct net *net, u32 > > tunnel_id, u32 session_id, u32 p > > session->mtu = dev->mtu - session->hdr_len; > > dev->mtu = session->mtu; > > dev->needed_headroom += session->hdr_len; > > + dev->max_mtu = ETH_MAX_MTU - dev->needed_headroom; > > I forgot to add that subtracting dev->needed_headroom doesn't > give the exact maximum MTU, but one that is a bit too high. > It could also just be set to ETH_MAX_MTU, to indicate that > the encapsulation is not included. > > Calculations like in R. Parameswaran's patch is needed for the > exact one. > https://lkml.org/lkml/2016/10/17/3
Argh. So there are a tiny handful of drivers that call ether_setup, but have no ndo_change_mtu at all, which thus previously, had no mtu range checks other than mtu < 0, and now have a range check of 68 to 1500... Since there was no range at all before, feature-parity here should be min_mtu = 0, max_mtu = ETH_MAX_MTU. I think I see five other cases that are in a similar situation: net/atm/br2684.c net/bluetooth/bnep/netdev.c net/dsa/slave.c net/irda/irda_device.c net/irda/irlan/irlan_eth.c Honestly, I think dsa is the only one where it likely actually matters, but I'll send a patch addressing all of those plus l2tp_eth in half a sec. -- Jarod Wilson ja...@redhat.com
