diff options
| author | David S. Miller <davem@davemloft.net> | 2019-08-05 21:17:40 +0300 |
|---|---|---|
| committer | David S. Miller <davem@davemloft.net> | 2019-08-05 21:17:40 +0300 |
| commit | 4130741736d4bc27add2ab4f7e53bc401846ce27 (patch) | |
| tree | 8646d4782ad35e0fce6e8ebcecc3896fd694154d /net | |
| parent | 60d60c8fbd8d1acf25b041ecd72ae4fa16e9405b (diff) | |
| parent | 065af355470519bd184019a93ac579f22b036045 (diff) | |
| download | linux-4130741736d4bc27add2ab4f7e53bc401846ce27.tar.xz | |
Merge branch 'net-fix-regressions-for-generic-XDP'
Jesper Dangaard Brouer says:
====================
net: fix regressions for generic-XDP
Thanks to Brandon Cazander, who wrote a very detailed bug report that
even used perf probe's on xdp-newbies mailing list, we discovered that
generic-XDP contains some regressions when using bpf_xdp_adjust_head().
First issue were that my selftests script, that use bpf_xdp_adjust_head(),
by mistake didn't use generic-XDP any-longer. That selftest should have
caught the real regression introduced in commit 458bf2f224f0 ("net: core:
support XDP generic on stacked devices.").
To verify this patchset fix the regressions, you can invoked manually via:
cd tools/testing/selftests/bpf/
sudo ./test_xdp_vlan_mode_generic.sh
sudo ./test_xdp_vlan_mode_native.sh
====================
Link: https://www.spinics.net/lists/xdp-newbies/msg01231.html
Fixes: 458bf2f224f0 ("net: core: support XDP generic on stacked devices.")
Reported by: Brandon Cazander <brandon.cazander@multapplied.net>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net')
| -rw-r--r-- | net/core/dev.c | 15 |
1 files changed, 10 insertions, 5 deletions
diff --git a/net/core/dev.c b/net/core/dev.c index 2f341b850845..0891f499c1bb 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4374,12 +4374,17 @@ static u32 netif_receive_generic_xdp(struct sk_buff *skb, act = bpf_prog_run_xdp(xdp_prog, xdp); + /* check if bpf_xdp_adjust_head was used */ off = xdp->data - orig_data; - if (off > 0) - __skb_pull(skb, off); - else if (off < 0) - __skb_push(skb, -off); - skb->mac_header += off; + if (off) { + if (off > 0) + __skb_pull(skb, off); + else if (off < 0) + __skb_push(skb, -off); + + skb->mac_header += off; + skb_reset_network_header(skb); + } /* check if bpf_xdp_adjust_tail was used. it can only "shrink" * pckt. |