From mboxrd@z Thu Jan 1 00:00:00 1970 From: Helmut Buchsbaum Subject: [PATCH v2 1/1] net: macb: initialize checksum when using checksum offloading Date: Mon, 29 Aug 2016 15:57:25 +0200 Message-ID: <1472479045-8814-2-git-send-email-helmut.buchsbaum@gmail.com> References: <1471611898-3852-2-git-send-email-helmut.buchsbaum@gmail.com> <1472479045-8814-1-git-send-email-helmut.buchsbaum@gmail.com> Cc: Michal Simek , netdev@vger.kernel.org, Cyrille Pitchen , Helmut Buchsbaum To: David Miller , tom@herbertland.com, Nicolas Ferre Return-path: Received: from mail-wm0-f65.google.com ([74.125.82.65]:35939 "EHLO mail-wm0-f65.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757252AbcH2N6I (ORCPT ); Mon, 29 Aug 2016 09:58:08 -0400 Received: by mail-wm0-f65.google.com with SMTP id i138so9669337wmf.3 for ; Mon, 29 Aug 2016 06:58:07 -0700 (PDT) In-Reply-To: <1472479045-8814-1-git-send-email-helmut.buchsbaum@gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: MACB/GEM needs the checksum field initialized to 0 to get correct results on transmit in all cases, e.g. on Zynq, UDP packets with payload <= 2 otherwise contain a wrong checksums. Signed-off-by: Helmut Buchsbaum --- drivers/net/ethernet/cadence/macb.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/drivers/net/ethernet/cadence/macb.c b/drivers/net/ethernet/cadence/macb.c index 89c0cfa..de2f791 100644 --- a/drivers/net/ethernet/cadence/macb.c +++ b/drivers/net/ethernet/cadence/macb.c @@ -1323,6 +1323,19 @@ dma_error: return 0; } +static inline void macb_clear_csum(struct sk_buff *skb) +{ + /* no change for packets without checksum offloading */ + if (skb->ip_summed != CHECKSUM_PARTIAL) + return; + + /* initialize checksum field + * This is required - at least for Zynq, which otherwise calculates + * wrong UDP header checksums for UDP packets with UDP data len <=2 + */ + *(__sum16 *)(skb->head + skb->csum_start + skb->csum_offset) = 0; +} + static int macb_start_xmit(struct sk_buff *skb, struct net_device *dev) { u16 queue_index = skb_get_queue_mapping(skb); @@ -1362,6 +1375,8 @@ static int macb_start_xmit(struct sk_buff *skb, struct net_device *dev) return NETDEV_TX_BUSY; } + macb_clear_csum(skb); + /* Map socket buffer for DMA transfer */ if (!macb_tx_map(bp, queue, skb)) { dev_kfree_skb_any(skb); -- 2.1.4