>> and with the reproducer, got: >> [ 54.255469] Bogus gso_type: 7 >> [ 54.258801] Bogus gso_size: 63464 >> [ 54.262532] ------------[ cut here ]------------ >> [ 54.267703] syz0: caps=(0x00000800000058c1, 0x0000000000000000) len=32 >> data_len=0 gso_size=63464 gso_type=7 ip_summed0 >> [ 54.279777] WARNING: CPU: 1 PID: 13005 at /root/linux/net/core/dev.c:2600 >> skb_warn_bad_offload+0xd6/0xec > I couldn't reproduce this call trace on net-next, maybe it's been fixed by: > commit 8d74e9f88d65af8bb2e095aff506aa6eac755ada > Author: Willem de Bruijn <will...@google.com> > Date: Tue Dec 12 11:39:04 2017 -0500 > > net: avoid skb_warn_bad_offload on IS_ERR
Yes, I forgot to mention that that has been fixed in net-next. It does not address the crash, but does suppress the gratuitous warning once we make segmentation return with error for bad packets like these.
