packet: fix reserve calculation
authorWillem de Bruijn <willemb@google.com>
Thu, 24 May 2018 22:10:30 +0000 (18:10 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Mon, 11 Jun 2018 20:49:20 +0000 (22:49 +0200)
commit0d751192afdef9f40746f36a47e802c35bc1075a
treeaab5b892202ad0d0112ff438857c05bbf23eb896
parent1c2c7767e876c547d5c74cacfbe5b6665cf56c5e
packet: fix reserve calculation

[ Upstream commit 9aad13b087ab0a588cd68259de618f100053360e ]

Commit b84bbaf7a6c8 ("packet: in packet_snd start writing at link
layer allocation") ensures that packet_snd always starts writing
the link layer header in reserved headroom allocated for this
purpose.

This is needed because packets may be shorter than hard_header_len,
in which case the space up to hard_header_len may be zeroed. But
that necessary padding is not accounted for in skb->len.

The fix, however, is buggy. It calls skb_push, which grows skb->len
when moving skb->data back. But in this case packet length should not
change.

Instead, call skb_reserve, which moves both skb->data and skb->tail
back, without changing length.

Fixes: b84bbaf7a6c8 ("packet: in packet_snd start writing at link layer allocation")
Reported-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
net/packet/af_packet.c