[PATCH v1] ibmveth: Set CHECKSUM_PARTIAL if NULL TCP CSUM.
From: David Wilder <hidden>
Date: 2021-06-22 21:53:00
Subsystem:
ibm power virtual ethernet device driver, linux for powerpc (32-bit and 64-bit), networking drivers, the rest · Maintainers:
Nick Child, Madhavan Srinivasan, Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds
TCP checksums on received packets may be set to NULL by the sender if CSO is enabled. The hypervisor flags these packets as check-sum-ok and the skb is then flagged CHECKSUM_UNNECESSARY. If these packets are then forwarded the sender will not request CSO due to the CHECKSUM_UNNECESSARY flag. The result is a TCP packet sent with a bad checksum. This change sets up CHECKSUM_PARTIAL on these packets causing the sender to correctly request CSUM offload. Signed-off-by: David Wilder <redacted> Reviewed-by: Pradeep Satyanarayana <redacted> Tested-by: Cristobal Forno <redacted> --- drivers/net/ethernet/ibm/ibmveth.c | 51 ++++++++++++++++-------------- 1 file changed, 28 insertions(+), 23 deletions(-)
diff --git a/drivers/net/ethernet/ibm/ibmveth.c b/drivers/net/ethernet/ibm/ibmveth.c
index 7fea9ae60f13..da6c63ff2790 100644
--- a/drivers/net/ethernet/ibm/ibmveth.c
+++ b/drivers/net/ethernet/ibm/ibmveth.c@@ -1285,36 +1285,41 @@ static void ibmveth_rx_csum_helper(struct sk_buff *skb, iph_proto = iph6->nexthdr; } - /* In OVS environment, when a flow is not cached, specifically for a - * new TCP connection, the first packet information is passed up + /* When CSO is enabled the TCP checksum may have be set to NULL by + * the sender given that we zeroed out TCP checksum field in + * transmit path (refer ibmveth_start_xmit routine). In this case set + * up CHECKSUM_PARTIAL. If the packet is forwarded, the checksum will + * then be recalculated by the destination NIC (CSO must be enabled + * on the destination NIC). + * + * In an OVS environment, when a flow is not cached, specifically for a + * new TCP connection, the first packet information is passed up to * the user space for finding a flow. During this process, OVS computes * checksum on the first packet when CHECKSUM_PARTIAL flag is set. * - * Given that we zeroed out TCP checksum field in transmit path - * (refer ibmveth_start_xmit routine) as we set "no checksum bit", - * OVS computed checksum will be incorrect w/o TCP pseudo checksum - * in the packet. This leads to OVS dropping the packet and hence - * TCP retransmissions are seen. - * - * So, re-compute TCP pseudo header checksum. + * So, re-compute TCP pseudo header checksum when configured for + * trunk mode. */ - if (iph_proto == IPPROTO_TCP && adapter->is_active_trunk) { + if (iph_proto == IPPROTO_TCP) { struct tcphdr *tcph = (struct tcphdr *)(skb->data + iphlen); - - tcphdrlen = skb->len - iphlen; - - /* Recompute TCP pseudo header checksum */ - if (skb_proto == ETH_P_IP) - tcph->check = ~csum_tcpudp_magic(iph->saddr, + if (tcph->check == 0x0000) { + /* Recompute TCP pseudo header checksum */ + if (adapter->is_active_trunk) { + tcphdrlen = skb->len - iphlen; + if (skb_proto == ETH_P_IP) + tcph->check = + ~csum_tcpudp_magic(iph->saddr, iph->daddr, tcphdrlen, iph_proto, 0); - else if (skb_proto == ETH_P_IPV6) - tcph->check = ~csum_ipv6_magic(&iph6->saddr, + else if (skb_proto == ETH_P_IPV6) + tcph->check = + ~csum_ipv6_magic(&iph6->saddr, &iph6->daddr, tcphdrlen, iph_proto, 0); - - /* Setup SKB fields for checksum offload */ - skb_partial_csum_set(skb, iphlen, - offsetof(struct tcphdr, check)); - skb_reset_network_header(skb); + } + /* Setup SKB fields for checksum offload */ + skb_partial_csum_set(skb, iphlen, + offsetof(struct tcphdr, check)); + skb_reset_network_header(skb); + } } }
--
2.27.0