On 2026-03-04 09:20:27 [+0100], Daniel Borkmann wrote: > Arguably it's a corner case, but could we not just do sth like this to > actually check if the device is up and if not drop? > > diff --git a/net/core/filter.c b/net/core/filter.c > index ba019ded773d..c447fd989a27 100644 > --- a/net/core/filter.c > +++ b/net/core/filter.c > @@ -4387,6 +4387,9 @@ u32 xdp_master_redirect(struct xdp_buff *xdp) > struct net_device *master, *slave; > master = netdev_master_upper_dev_get_rcu(xdp->rxq->dev); > + if (unlikely(!(master->flags & IFF_UP))) > + return XDP_ABORTED;
preemption (vcpu scheduled), while paused the other CPU could put the device down, or is it too much of a corner case? Also, there is also ndo_get_xmit_slave which at this time is only used by Infiniband and supports only active-backup mode [0]. [0] https://lore.kernel.org/netdev/999129.1772247707@famine/ > + > slave = master->netdev_ops->ndo_xdp_get_xmit_slave(master, xdp); > if (slave && slave != xdp->rxq->dev) { > /* The target device is different from the receiving device, > so Sebastian
