@@ -46,6 +46,8 @@
#include <asm/vio.h>
#include <asm/iommu.h>
#include <asm/firmware.h>
+#include <net/tcp.h>
+#include <net/ip6_checksum.h>
#include "ibmveth.h"
@@ -808,8 +810,7 @@ static int ibmveth_set_csum_offload(struct net_device *dev, u32 data)
ret = h_illan_attributes(adapter->vdev->unit_address, 0, 0, &ret_attr);
- if (ret == H_SUCCESS && !(ret_attr & IBMVETH_ILLAN_ACTIVE_TRUNK) &&
- !(ret_attr & IBMVETH_ILLAN_TRUNK_PRI_MASK) &&
+ if (ret == H_SUCCESS &&
(ret_attr & IBMVETH_ILLAN_PADDED_PKT_CSUM)) {
ret4 = h_illan_attributes(adapter->vdev->unit_address, clr_attr,
set_attr, &ret_attr);
@@ -1040,6 +1041,15 @@ static netdev_tx_t ibmveth_start_xmit(struct sk_buff *skb,
dma_addr_t dma_addr;
unsigned long mss = 0;
+ /* veth doesn't handle frag_list, so linearize the skb.
+ * When GRO is enabled SKB's can have frag_list.
+ */
+ if (adapter->is_active_trunk &&
+ skb_has_frag_list(skb) && __skb_linearize(skb)) {
+ netdev->stats.tx_dropped++;
+ goto out;
+ }
+
/*
* veth handles a maximum of 6 segments including the header, so
* we have to linearize the skb if there are more than this.
@@ -1064,9 +1074,6 @@ static netdev_tx_t ibmveth_start_xmit(struct sk_buff *skb,
desc_flags = IBMVETH_BUF_VALID;
- if (skb_is_gso(skb) && adapter->fw_large_send_support)
- desc_flags |= IBMVETH_BUF_LRG_SND;
-
if (skb->ip_summed == CHECKSUM_PARTIAL) {
unsigned char *buf = skb_transport_header(skb) +
skb->csum_offset;
@@ -1076,6 +1083,9 @@ static netdev_tx_t ibmveth_start_xmit(struct sk_buff *skb,
/* Need to zero out the checksum */
buf[0] = 0;
buf[1] = 0;
+
+ if (skb_is_gso(skb) && adapter->fw_large_send_support)
+ desc_flags |= IBMVETH_BUF_LRG_SND;
}
retry_bounce:
@@ -1128,7 +1138,7 @@ retry_bounce:
descs[i+1].fields.address = dma_addr;
}
- if (skb_is_gso(skb)) {
+ if (skb->ip_summed == CHECKSUM_PARTIAL && skb_is_gso(skb)) {
if (adapter->fw_large_send_support) {
mss = (unsigned long)skb_shinfo(skb)->gso_size;
adapter->tx_large_packets++;
@@ -1232,6 +1242,71 @@ static void ibmveth_rx_mss_helper(struct sk_buff *skb, u16 mss, int lrg_pkt)
}
}
+static void ibmveth_rx_csum_helper(struct sk_buff *skb,
+ struct ibmveth_adapter *adapter)
+{
+ struct iphdr *iph = NULL;
+ struct ipv6hdr *iph6 = NULL;
+ __be16 skb_proto = 0;
+ u16 iphlen = 0;
+ u16 iph_proto = 0;
+ u16 tcphdrlen = 0;
+
+ skb_proto = be16_to_cpu(skb->protocol);
+
+ if (skb_proto == ETH_P_IP) {
+ iph = (struct iphdr *)skb->data;
+
+ /* If the IP checksum is not offloaded and if the packet
+ * is large send, the checksum must be rebuilt.
+ */
+ if (iph->check == 0xffff) {
+ iph->check = 0;
+ iph->check = ip_fast_csum((unsigned char *)iph,
+ iph->ihl);
+ }
+
+ iphlen = iph->ihl * 4;
+ iph_proto = iph->protocol;
+ } else if (skb_proto == ETH_P_IPV6) {
+ iph6 = (struct ipv6hdr *)skb->data;
+ iphlen = sizeof(struct ipv6hdr);
+ 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
+ * 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.
+ */
+ if (iph_proto == IPPROTO_TCP && adapter->is_active_trunk) {
+ 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,
+ iph->daddr, tcphdrlen, iph_proto, 0);
+ 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);
+ }
+}
+
static int ibmveth_poll(struct napi_struct *napi, int budget)
{
struct ibmveth_adapter *adapter =
@@ -1239,7 +1314,6 @@ static int ibmveth_poll(struct napi_struct *napi, int budget)
struct net_device *netdev = adapter->netdev;
int frames_processed = 0;
unsigned long lpar_rc;
- struct iphdr *iph;
u16 mss = 0;
restart_poll:
@@ -1297,17 +1371,7 @@ restart_poll:
if (csum_good) {
skb->ip_summed = CHECKSUM_UNNECESSARY;
- if (be16_to_cpu(skb->protocol) == ETH_P_IP) {
- iph = (struct iphdr *)skb->data;
-
- /* If the IP checksum is not offloaded and if the packet
- * is large send, the checksum must be rebuilt.
- */
- if (iph->check == 0xffff) {
- iph->check = 0;
- iph->check = ip_fast_csum((unsigned char *)iph, iph->ihl);
- }
- }
+ ibmveth_rx_csum_helper(skb, adapter);
}
if (length > netdev->mtu + ETH_HLEN) {
@@ -1626,6 +1690,13 @@ static int ibmveth_probe(struct vio_dev *dev, const struct vio_device_id *id)
netdev->hw_features |= NETIF_F_TSO;
}
+ adapter->is_active_trunk = false;
+ if (ret == H_SUCCESS && (ret_attr & IBMVETH_ILLAN_ACTIVE_TRUNK)) {
+ adapter->is_active_trunk = true;
+ netdev->hw_features |= NETIF_F_FRAGLIST;
+ netdev->features |= NETIF_F_FRAGLIST;
+ }
+
netdev->min_mtu = IBMVETH_MIN_MTU;
netdev->max_mtu = ETH_MAX_MTU;
@@ -156,6 +156,7 @@ struct ibmveth_adapter {
int pool_config;
int rx_csum;
int large_send;
+ bool is_active_trunk;
void *bounce_buffer;
dma_addr_t bounce_buffer_dma;
Current largesend and checksum offload feature in ibmveth driver, - Source VM sends the TCP packets with ip_summed field set as CHECKSUM_PARTIAL and TCP pseudo header checksum is placed in checksum field - CHECKSUM_PARTIAL flag in SKB will enable ibmveth driver to mark "no checksum" and "checksum good" bits in transmit buffer descriptor before the packet is delivered to pseries PowerVM Hypervisor - If ibmveth has largesend capability enabled, transmit buffer descriptors are market accordingly before packet is delivered to Hypervisor (along with mss value for packets with length > MSS) - Destination VM's ibmveth driver receives the packet with "checksum good" bit set and so, SKB's ip_summed field is set with CHECKSUM_UNNECESSARY - If "largesend" bit was on, mss value is copied from receive descriptor into SKB's gso_size and other flags are appropriately set for packets > MSS size - The packet is now successfully delivered up the stack in destination VM The offloads described above works fine for TCP communication among VMs in the same pseries server ( VM A <=> PowerVM Hypervisor <=> VM B ) We are now enabling support for OVS in pseries PowerVM environment. One of our requirements is to have ibmveth driver configured in "Trunk" mode, when they are used with OVS. This is because, PowerVM Hypervisor will no more bridge the packets between VMs, instead the packets are delivered to IO Server which hosts OVS to bridge them between VMs or to external networks (flow shown below), VM A <=> PowerVM Hypervisor <=> IO Server(OVS) <=> PowerVM Hypervisor <=> VM B In "IO server" the packet is received by inbound Trunk ibmveth and then delivered to OVS, which is then bridged to outbound Trunk ibmveth (shown below), Inbound Trunk ibmveth <=> OVS <=> Outbound Trunk ibmveth In this model, we hit the following issues which impacted the VM communication performance, - Issue 1: ibmveth doesn't support largesend and checksum offload features when configured as "Trunk". Driver has explicit checks to prevent enabling these offloads. - Issue 2: SYN packet drops seen at destination VM. When the packet originates, it has CHECKSUM_PARTIAL flag set and as it gets delivered to IO server's inbound Trunk ibmveth, on validating "checksum good" bits in ibmveth receive routine, SKB's ip_summed field is set with CHECKSUM_UNNECESSARY flag. This packet is then bridged by OVS (or Linux Bridge) and delivered to outbound Trunk ibmveth. At this point the outbound ibmveth transmit routine will not set "no checksum" and "checksum good" bits in transmit buffer descriptor, as it does so only when the ip_summed field is CHECKSUM_PARTIAL. When this packet gets delivered to destination VM, TCP layer receives the packet with checksum value of 0 and with no checksum related flags in ip_summed field. This leads to packet drops. So, TCP connections never goes through fine. - Issue 3: First packet of a TCP connection will be dropped, if there is no OVS flow cached in datapath. OVS while trying to identify the flow, computes the checksum. The computed checksum will be invalid at the receiving end, as ibmveth transmit routine zeroes out the pseudo checksum value in the packet. This leads to packet drop. - Issue 4: ibmveth driver doesn't have support for SKB's with frag_list. When Physical NIC has GRO enabled and when OVS bridges these packets, OVS vport send code will end up calling dev_queue_xmit, which in turn calls validate_xmit_skb. In validate_xmit_skb routine, the larger packets will get segmented into MSS sized segments, if SKB has a frag_list and if the driver to which they are delivered to doesn't support NETIF_F_FRAGLIST feature. This patch addresses the above four issues, thereby enabling end to end largesend and checksum offload support for better performance. - Fix for Issue 1 : Remove checks which prevent enabling TCP largesend and checksum offloads. - Fix for Issue 2 : When ibmveth receives a packet with "checksum good" bit set and if its configured in Trunk mode, set appropriate SKB fields using skb_partial_csum_set (ip_summed field is set with CHECKSUM_PARTIAL) - Fix for Issue 3: Recompute the pseudo header checksum before sending the SKB up the stack. - Fix for Issue 4: Linearize the SKBs with frag_list. Though we end up allocating buffers and copying data, this fix gives upto 4X throughput increase. Note: All these fixes need to be dropped together as fixing just one of them will lead to other issues immediately (especially for Issues 1,2 & 3). Signed-off-by: Sivakumar Krishnasamy <ksiva@linux.vnet.ibm.com> --- drivers/net/ethernet/ibm/ibmveth.c | 107 ++++++++++++++++++++++++++++++------- drivers/net/ethernet/ibm/ibmveth.h | 1 + 2 files changed, 90 insertions(+), 18 deletions(-)