9 files changed, 381 insertions, 91 deletions

diff --git a/drivers/infiniband/ulp/ipoib/Makefile b/drivers/infiniband/ulp/ipoib/Makefile
index 98ee38e8c2c4..3090100f0de7 100644
--- a/drivers/infiniband/ulp/ipoib/Makefile
+++ b/drivers/infiniband/ulp/ipoib/Makefile
@@ -4,7 +4,8 @@ ib_ipoib-y					:= ipoib_main.o \
 						   ipoib_ib.o \
 						   ipoib_multicast.o \
 						   ipoib_verbs.o \
-						   ipoib_vlan.o
+						   ipoib_vlan.o \
+						   ipoib_ethtool.o
 ib_ipoib-$(CONFIG_INFINIBAND_IPOIB_CM)		+= ipoib_cm.o
 ib_ipoib-$(CONFIG_INFINIBAND_IPOIB_DEBUG)	+= ipoib_fs.o
 
diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h
index 054fab8e27a0..f1f142dc64b1 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib.h
+++ b/drivers/infiniband/ulp/ipoib/ipoib.h
@@ -56,11 +56,11 @@
 /* constants */
 
 enum {
-	IPOIB_PACKET_SIZE	  = 2048,
-	IPOIB_BUF_SIZE		  = IPOIB_PACKET_SIZE + IB_GRH_BYTES,
-
 	IPOIB_ENCAP_LEN		  = 4,
 
+	IPOIB_UD_HEAD_SIZE	  = IB_GRH_BYTES + IPOIB_ENCAP_LEN,
+	IPOIB_UD_RX_SG		  = 2, /* max buffer needed for 4K mtu */
+
 	IPOIB_CM_MTU		  = 0x10000 - 0x10, /* padding to align header to 16 */
 	IPOIB_CM_BUF_SIZE	  = IPOIB_CM_MTU  + IPOIB_ENCAP_LEN,
 	IPOIB_CM_HEAD_SIZE	  = IPOIB_CM_BUF_SIZE % PAGE_SIZE,
@@ -87,6 +87,7 @@ enum {
 	IPOIB_MCAST_STARTED	  = 8,
 	IPOIB_FLAG_ADMIN_CM	  = 9,
 	IPOIB_FLAG_UMCAST	  = 10,
+	IPOIB_FLAG_CSUM		  = 11,
 
 	IPOIB_MAX_BACKOFF_SECONDS = 16,
 
@@ -138,7 +139,7 @@ struct ipoib_mcast {
 
 struct ipoib_rx_buf {
 	struct sk_buff *skb;
-	u64		mapping;
+	u64		mapping[IPOIB_UD_RX_SG];
 };
 
 struct ipoib_tx_buf {
@@ -241,6 +242,11 @@ struct ipoib_cm_dev_priv {
 	int			num_frags;
 };
 
+struct ipoib_ethtool_st {
+	u16     coalesce_usecs;
+	u16     max_coalesced_frames;
+};
+
 /*
  * Device private locking: tx_lock protects members used in TX fast
  * path (and we use LLTX so upper layers don't do extra locking).
@@ -288,6 +294,7 @@ struct ipoib_dev_priv {
 
 	unsigned int admin_mtu;
 	unsigned int mcast_mtu;
+	unsigned int max_ib_mtu;
 
 	struct ipoib_rx_buf *rx_ring;
 
@@ -299,6 +306,9 @@ struct ipoib_dev_priv {
 	struct ib_send_wr    tx_wr;
 	unsigned	     tx_outstanding;
 
+	struct ib_recv_wr    rx_wr;
+	struct ib_sge	     rx_sge[IPOIB_UD_RX_SG];
+
 	struct ib_wc ibwc[IPOIB_NUM_WC];
 
 	struct list_head dead_ahs;
@@ -318,6 +328,8 @@ struct ipoib_dev_priv {
 	struct dentry *mcg_dentry;
 	struct dentry *path_dentry;
 #endif
+	int	hca_caps;
+	struct ipoib_ethtool_st ethtool;
 };
 
 struct ipoib_ah {
@@ -358,6 +370,14 @@ struct ipoib_neigh {
 	struct list_head    list;
 };
 
+#define IPOIB_UD_MTU(ib_mtu)		(ib_mtu - IPOIB_ENCAP_LEN)
+#define IPOIB_UD_BUF_SIZE(ib_mtu)	(ib_mtu + IB_GRH_BYTES)
+
+static inline int ipoib_ud_need_sg(unsigned int ib_mtu)
+{
+	return IPOIB_UD_BUF_SIZE(ib_mtu) > PAGE_SIZE;
+}
+
 /*
  * We stash a pointer to our private neighbour information after our
  * hardware address in neigh->ha.  The ALIGN() expression here makes
@@ -458,6 +478,8 @@ void ipoib_pkey_poll(struct work_struct *work);
 int ipoib_pkey_dev_delay_open(struct net_device *dev);
 void ipoib_drain_cq(struct net_device *dev);
 
+void ipoib_set_ethtool_ops(struct net_device *dev);
+
 #ifdef CONFIG_INFINIBAND_IPOIB_CM
 
 #define IPOIB_FLAGS_RC		0x80
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_cm.c b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
index 2490b2d79dbb..9db7b0bd9134 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_cm.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
@@ -1007,9 +1007,9 @@ static int ipoib_cm_modify_tx_init(struct net_device *dev,
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct ib_qp_attr qp_attr;
 	int qp_attr_mask, ret;
-	ret = ib_find_cached_pkey(priv->ca, priv->port, priv->pkey, &qp_attr.pkey_index);
+	ret = ib_find_pkey(priv->ca, priv->port, priv->pkey, &qp_attr.pkey_index);
 	if (ret) {
-		ipoib_warn(priv, "pkey 0x%x not in cache: %d\n", priv->pkey, ret);
+		ipoib_warn(priv, "pkey 0x%x not found: %d\n", priv->pkey, ret);
 		return ret;
 	}
 
@@ -1383,6 +1383,10 @@ static ssize_t set_mode(struct device *d, struct device_attribute *attr,
 		set_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags);
 		ipoib_warn(priv, "enabling connected mode "
 			   "will cause multicast packet drops\n");
+
+		dev->features &= ~(NETIF_F_IP_CSUM | NETIF_F_SG | NETIF_F_TSO);
+		priv->tx_wr.send_flags &= ~IB_SEND_IP_CSUM;
+
 		ipoib_flush_paths(dev);
 		return count;
 	}
@@ -1391,6 +1395,13 @@ static ssize_t set_mode(struct device *d, struct device_attribute *attr,
 		clear_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags);
 		dev->mtu = min(priv->mcast_mtu, dev->mtu);
 		ipoib_flush_paths(dev);
+
+		if (test_bit(IPOIB_FLAG_CSUM, &priv->flags)) {
+			dev->features |= NETIF_F_IP_CSUM | NETIF_F_SG;
+			if (priv->hca_caps & IB_DEVICE_UD_TSO)
+				dev->features |= NETIF_F_TSO;
+		}
+
 		return count;
 	}
 
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
new file mode 100644
index 000000000000..9a47428366c9
--- /dev/null
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
@@ -0,0 +1,99 @@
+/*
+ * Copyright (c) 2007 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/ethtool.h>
+#include <linux/netdevice.h>
+
+#include "ipoib.h"
+
+static void ipoib_get_drvinfo(struct net_device *netdev,
+			      struct ethtool_drvinfo *drvinfo)
+{
+	strncpy(drvinfo->driver, "ipoib", sizeof(drvinfo->driver) - 1);
+}
+
+static int ipoib_get_coalesce(struct net_device *dev,
+			      struct ethtool_coalesce *coal)
+{
+	struct ipoib_dev_priv *priv = netdev_priv(dev);
+
+	coal->rx_coalesce_usecs = priv->ethtool.coalesce_usecs;
+	coal->tx_coalesce_usecs = priv->ethtool.coalesce_usecs;
+	coal->rx_max_coalesced_frames = priv->ethtool.max_coalesced_frames;
+	coal->tx_max_coalesced_frames = priv->ethtool.max_coalesced_frames;
+
+	return 0;
+}
+
+static int ipoib_set_coalesce(struct net_device *dev,
+			      struct ethtool_coalesce *coal)
+{
+	struct ipoib_dev_priv *priv = netdev_priv(dev);
+	int ret;
+
+	/*
+	 * Since IPoIB uses a single CQ for both rx and tx, we assume
+	 * that rx params dictate the configuration.  These values are
+	 * saved in the private data and returned when ipoib_get_coalesce()
+	 * is called.
+	 */
+	if (coal->rx_coalesce_usecs       > 0xffff ||
+	    coal->rx_max_coalesced_frames > 0xffff)
+		return -EINVAL;
+
+	ret = ib_modify_cq(priv->cq, coal->rx_max_coalesced_frames,
+			   coal->rx_coalesce_usecs);
+	if (ret && ret != -ENOSYS) {
+		ipoib_warn(priv, "failed modifying CQ (%d)\n", ret);
+		return ret;
+	}
+
+	coal->tx_coalesce_usecs       = coal->rx_coalesce_usecs;
+	coal->tx_max_coalesced_frames = coal->rx_max_coalesced_frames;
+	priv->ethtool.coalesce_usecs       = coal->rx_coalesce_usecs;
+	priv->ethtool.max_coalesced_frames = coal->rx_max_coalesced_frames;
+
+	return 0;
+}
+
+static const struct ethtool_ops ipoib_ethtool_ops = {
+	.get_drvinfo		= ipoib_get_drvinfo,
+	.get_tso		= ethtool_op_get_tso,
+	.get_coalesce		= ipoib_get_coalesce,
+	.set_coalesce		= ipoib_set_coalesce,
+};
+
+void ipoib_set_ethtool_ops(struct net_device *dev)
+{
+	SET_ETHTOOL_OPS(dev, &ipoib_ethtool_ops);
+}
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index 08c4396cf418..7cf1fa7074ab 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -39,6 +39,8 @@
 #include <linux/dma-mapping.h>
 
 #include <rdma/ib_cache.h>
+#include <linux/ip.h>
+#include <linux/tcp.h>
 
 #include "ipoib.h"
 
@@ -87,28 +89,59 @@ void ipoib_free_ah(struct kref *kref)
 	spin_unlock_irqrestore(&priv->lock, flags);
 }
 
+static void ipoib_ud_dma_unmap_rx(struct ipoib_dev_priv *priv,
+				  u64 mapping[IPOIB_UD_RX_SG])
+{
+	if (ipoib_ud_need_sg(priv->max_ib_mtu)) {
+		ib_dma_unmap_single(priv->ca, mapping[0], IPOIB_UD_HEAD_SIZE,
+				    DMA_FROM_DEVICE);
+		ib_dma_unmap_page(priv->ca, mapping[1], PAGE_SIZE,
+				  DMA_FROM_DEVICE);
+	} else
+		ib_dma_unmap_single(priv->ca, mapping[0],
+				    IPOIB_UD_BUF_SIZE(priv->max_ib_mtu),
+				    DMA_FROM_DEVICE);
+}
+
+static void ipoib_ud_skb_put_frags(struct ipoib_dev_priv *priv,
+				   struct sk_buff *skb,
+				   unsigned int length)
+{
+	if (ipoib_ud_need_sg(priv->max_ib_mtu)) {
+		skb_frag_t *frag = &skb_shinfo(skb)->frags[0];
+		unsigned int size;
+		/*
+		 * There is only two buffers needed for max_payload = 4K,
+		 * first buf size is IPOIB_UD_HEAD_SIZE
+		 */
+		skb->tail += IPOIB_UD_HEAD_SIZE;
+		skb->len  += length;
+
+		size = length - IPOIB_UD_HEAD_SIZE;
+
+		frag->size     = size;
+		skb->data_len += size;
+		skb->truesize += size;
+	} else
+		skb_put(skb, length);
+
+}
+
 static int ipoib_ib_post_receive(struct net_device *dev, int id)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
-	struct ib_sge list;
-	struct ib_recv_wr param;
 	struct ib_recv_wr *bad_wr;
 	int ret;
 
-	list.addr     = priv->rx_ring[id].mapping;
-	list.length   = IPOIB_BUF_SIZE;
-	list.lkey     = priv->mr->lkey;
+	priv->rx_wr.wr_id   = id | IPOIB_OP_RECV;
+	priv->rx_sge[0].addr = priv->rx_ring[id].mapping[0];
+	priv->rx_sge[1].addr = priv->rx_ring[id].mapping[1];
 
-	param.next    = NULL;
-	param.wr_id   = id | IPOIB_OP_RECV;
-	param.sg_list = &list;
-	param.num_sge = 1;
 
-	ret = ib_post_recv(priv->qp, &param, &bad_wr);
+	ret = ib_post_recv(priv->qp, &priv->rx_wr, &bad_wr);
 	if (unlikely(ret)) {
 		ipoib_warn(priv, "receive failed for buf %d (%d)\n", id, ret);
-		ib_dma_unmap_single(priv->ca, priv->rx_ring[id].mapping,
-				    IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
+		ipoib_ud_dma_unmap_rx(priv, priv->rx_ring[id].mapping);
 		dev_kfree_skb_any(priv->rx_ring[id].skb);
 		priv->rx_ring[id].skb = NULL;
 	}
@@ -116,15 +149,21 @@ static int ipoib_ib_post_receive(struct net_device *dev, int id)
 	return ret;
 }
 
-static int ipoib_alloc_rx_skb(struct net_device *dev, int id)
+static struct sk_buff *ipoib_alloc_rx_skb(struct net_device *dev, int id)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct sk_buff *skb;
-	u64 addr;
+	int buf_size;
+	u64 *mapping;
 
-	skb = dev_alloc_skb(IPOIB_BUF_SIZE + 4);
-	if (!skb)
-		return -ENOMEM;
+	if (ipoib_ud_need_sg(priv->max_ib_mtu))
+		buf_size = IPOIB_UD_HEAD_SIZE;
+	else
+		buf_size = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu);
+
+	skb = dev_alloc_skb(buf_size + 4);
+	if (unlikely(!skb))
+		return NULL;
 
 	/*
 	 * IB will leave a 40 byte gap for a GRH and IPoIB adds a 4 byte
@@ -133,17 +172,32 @@ static int ipoib_alloc_rx_skb(struct net_device *dev, int id)
 	 */
 	skb_reserve(skb, 4);
 
-	addr = ib_dma_map_single(priv->ca, skb->data, IPOIB_BUF_SIZE,
-				 DMA_FROM_DEVICE);
-	if (unlikely(ib_dma_mapping_error(priv->ca, addr))) {
-		dev_kfree_skb_any(skb);
-		return -EIO;
+	mapping = priv->rx_ring[id].mapping;
+	mapping[0] = ib_dma_map_single(priv->ca, skb->data, buf_size,
+				       DMA_FROM_DEVICE);
+	if (unlikely(ib_dma_mapping_error(priv->ca, mapping[0])))
+		goto error;
+
+	if (ipoib_ud_need_sg(priv->max_ib_mtu)) {
+		struct page *page = alloc_page(GFP_ATOMIC);
+		if (!page)
+			goto partial_error;
+		skb_fill_page_desc(skb, 0, page, 0, PAGE_SIZE);
+		mapping[1] =
+			ib_dma_map_page(priv->ca, skb_shinfo(skb)->frags[0].page,
+					0, PAGE_SIZE, DMA_FROM_DEVICE);
+		if (unlikely(ib_dma_mapping_error(priv->ca, mapping[1])))
+			goto partial_error;
 	}
 
-	priv->rx_ring[id].skb     = skb;
-	priv->rx_ring[id].mapping = addr;
+	priv->rx_ring[id].skb = skb;
+	return skb;
 
-	return 0;
+partial_error:
+	ib_dma_unmap_single(priv->ca, mapping[0], buf_size, DMA_FROM_DEVICE);
+error:
+	dev_kfree_skb_any(skb);
+	return NULL;
 }
 
 static int ipoib_ib_post_receives(struct net_device *dev)
@@ -152,7 +206,7 @@ static int ipoib_ib_post_receives(struct net_device *dev)
 	int i;
 
 	for (i = 0; i < ipoib_recvq_size; ++i) {
-		if (ipoib_alloc_rx_skb(dev, i)) {
+		if (!ipoib_alloc_rx_skb(dev, i)) {
 			ipoib_warn(priv, "failed to allocate receive buffer %d\n", i);
 			return -ENOMEM;
 		}
@@ -170,7 +224,7 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	unsigned int wr_id = wc->wr_id & ~IPOIB_OP_RECV;
 	struct sk_buff *skb;
-	u64 addr;
+	u64 mapping[IPOIB_UD_RX_SG];
 
 	ipoib_dbg_data(priv, "recv completion: id %d, status: %d\n",
 		       wr_id, wc->status);
@@ -182,15 +236,13 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
 	}
 
 	skb  = priv->rx_ring[wr_id].skb;
-	addr = priv->rx_ring[wr_id].mapping;
 
 	if (unlikely(wc->status != IB_WC_SUCCESS)) {
 		if (wc->status != IB_WC_WR_FLUSH_ERR)
 			ipoib_warn(priv, "failed recv event "
 				   "(status=%d, wrid=%d vend_err %x)\n",
 				   wc->status, wr_id, wc->vendor_err);
-		ib_dma_unmap_single(priv->ca, addr,
-				    IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
+		ipoib_ud_dma_unmap_rx(priv, priv->rx_ring[wr_id].mapping);
 		dev_kfree_skb_any(skb);
 		priv->rx_ring[wr_id].skb = NULL;
 		return;
@@ -203,11 +255,14 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
 	if (wc->slid == priv->local_lid && wc->src_qp == priv->qp->qp_num)
 		goto repost;
 
+	memcpy(mapping, priv->rx_ring[wr_id].mapping,
+	       IPOIB_UD_RX_SG * sizeof *mapping);
+
 	/*
 	 * If we can't allocate a new RX buffer, dump
 	 * this packet and reuse the old buffer.
 	 */
-	if (unlikely(ipoib_alloc_rx_skb(dev, wr_id))) {
+	if (unlikely(!ipoib_alloc_rx_skb(dev, wr_id))) {
 		++dev->stats.rx_dropped;
 		goto repost;
 	}
@@ -215,9 +270,9 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
 	ipoib_dbg_data(priv, "received %d bytes, SLID 0x%04x\n",
 		       wc->byte_len, wc->slid);
 
-	ib_dma_unmap_single(priv->ca, addr, IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
+	ipoib_ud_dma_unmap_rx(priv, mapping);
+	ipoib_ud_skb_put_frags(priv, skb, wc->byte_len);
 
-	skb_put(skb, wc->byte_len);
 	skb_pull(skb, IB_GRH_BYTES);
 
 	skb->protocol = ((struct ipoib_header *) skb->data)->proto;
@@ -231,6 +286,10 @@ static void ipoib_ib_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
 	skb->dev = dev;
 	/* XXX get correct PACKET_ type here */
 	skb->pkt_type = PACKET_HOST;
+
+	if (test_bit(IPOIB_FLAG_CSUM, &priv->flags) && likely(wc->csum_ok))
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
+
 	netif_receive_skb(skb);
 
 repost:
@@ -245,29 +304,37 @@ static int ipoib_dma_map_tx(struct ib_device *ca,
 	struct sk_buff *skb = tx_req->skb;
 	u64 *mapping = tx_req->mapping;
 	int i;
+	int off;
 
-	mapping[0] = ib_dma_map_single(ca, skb->data, skb_headlen(skb),
-				       DMA_TO_DEVICE);
-	if (unlikely(ib_dma_mapping_error(ca, mapping[0])))
-		return -EIO;
+	if (skb_headlen(skb)) {
+		mapping[0] = ib_dma_map_single(ca, skb->data, skb_headlen(skb),
+					       DMA_TO_DEVICE);
+		if (unlikely(ib_dma_mapping_error(ca, mapping[0])))
+			return -EIO;
+
+		off = 1;
+	} else
+		off = 0;
 
 	for (i = 0; i < skb_shinfo(skb)->nr_frags; ++i) {
 		skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
-		mapping[i + 1] = ib_dma_map_page(ca, frag->page,
+		mapping[i + off] = ib_dma_map_page(ca, frag->page,
 						 frag->page_offset, frag->size,
 						 DMA_TO_DEVICE);
-		if (unlikely(ib_dma_mapping_error(ca, mapping[i + 1])))
+		if (unlikely(ib_dma_mapping_error(ca, mapping[i + off])))
 			goto partial_error;
 	}
 	return 0;
 
 partial_error:
-	ib_dma_unmap_single(ca, mapping[0], skb_headlen(skb), DMA_TO_DEVICE);
-
 	for (; i > 0; --i) {
 		skb_frag_t *frag = &skb_shinfo(skb)->frags[i - 1];
-		ib_dma_unmap_page(ca, mapping[i], frag->size, DMA_TO_DEVICE);
+		ib_dma_unmap_page(ca, mapping[i - !off], frag->size, DMA_TO_DEVICE);
 	}
+
+	if (off)
+		ib_dma_unmap_single(ca, mapping[0], skb_headlen(skb), DMA_TO_DEVICE);
+
 	return -EIO;
 }
 
@@ -277,12 +344,17 @@ static void ipoib_dma_unmap_tx(struct ib_device *ca,
 	struct sk_buff *skb = tx_req->skb;
 	u64 *mapping = tx_req->mapping;
 	int i;
+	int off;
 
-	ib_dma_unmap_single(ca, mapping[0], skb_headlen(skb), DMA_TO_DEVICE);
+	if (skb_headlen(skb)) {
+		ib_dma_unmap_single(ca, mapping[0], skb_headlen(skb), DMA_TO_DEVICE);
+		off = 1;
+	} else
+		off = 0;
 
 	for (i = 0; i < skb_shinfo(skb)->nr_frags; ++i) {
 		skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
-		ib_dma_unmap_page(ca, mapping[i + 1], frag->size,
+		ib_dma_unmap_page(ca, mapping[i + off], frag->size,
 				  DMA_TO_DEVICE);
 	}
 }
@@ -388,24 +460,40 @@ void ipoib_ib_completion(struct ib_cq *cq, void *dev_ptr)
 static inline int post_send(struct ipoib_dev_priv *priv,
 			    unsigned int wr_id,
 			    struct ib_ah *address, u32 qpn,
-			    u64 *mapping, int headlen,
-			    skb_frag_t *frags,
-			    int nr_frags)
+			    struct ipoib_tx_buf *tx_req,
+			    void *head, int hlen)
 {
 	struct ib_send_wr *bad_wr;
-	int i;
+	int i, off;
+	struct sk_buff *skb = tx_req->skb;
+	skb_frag_t *frags = skb_shinfo(skb)->frags;
+	int nr_frags = skb_shinfo(skb)->nr_frags;
+	u64 *mapping = tx_req->mapping;
+
+	if (skb_headlen(skb)) {
+		priv->tx_sge[0].addr         = mapping[0];
+		priv->tx_sge[0].length       = skb_headlen(skb);
+		off = 1;
+	} else
+		off = 0;
 
-	priv->tx_sge[0].addr	     = mapping[0];
-	priv->tx_sge[0].length       = headlen;
 	for (i = 0; i < nr_frags; ++i) {
-		priv->tx_sge[i + 1].addr = mapping[i + 1];
-		priv->tx_sge[i + 1].length = frags[i].size;
+		priv->tx_sge[i + off].addr = mapping[i + off];
+		priv->tx_sge[i + off].length = frags[i].size;
 	}
-	priv->tx_wr.num_sge	     = nr_frags + 1;
+	priv->tx_wr.num_sge	     = nr_frags + off;
 	priv->tx_wr.wr_id 	     = wr_id;
 	priv->tx_wr.wr.ud.remote_qpn = qpn;
 	priv->tx_wr.wr.ud.ah 	     = address;
 
+	if (head) {
+		priv->tx_wr.wr.ud.mss	 = skb_shinfo(skb)->gso_size;
+		priv->tx_wr.wr.ud.header = head;
+		priv->tx_wr.wr.ud.hlen	 = hlen;
+		priv->tx_wr.opcode	 = IB_WR_LSO;
+	} else
+		priv->tx_wr.opcode	 = IB_WR_SEND;
+
 	return ib_post_send(priv->qp, &priv->tx_wr, &bad_wr);
 }
 
@@ -414,14 +502,30 @@ void ipoib_send(struct net_device *dev, struct sk_buff *skb,
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct ipoib_tx_buf *tx_req;
-
-	if (unlikely(skb->len > priv->mcast_mtu + IPOIB_ENCAP_LEN)) {
-		ipoib_warn(priv, "packet len %d (> %d) too long to send, dropping\n",
-			   skb->len, priv->mcast_mtu + IPOIB_ENCAP_LEN);
-		++dev->stats.tx_dropped;
-		++dev->stats.tx_errors;
-		ipoib_cm_skb_too_long(dev, skb, priv->mcast_mtu);
-		return;
+	int hlen;
+	void *phead;
+
+	if (skb_is_gso(skb)) {
+		hlen = skb_transport_offset(skb) + tcp_hdrlen(skb);
+		phead = skb->data;
+		if (unlikely(!skb_pull(skb, hlen))) {
+			ipoib_warn(priv, "linear data too small\n");
+			++dev->stats.tx_dropped;
+			++dev->stats.tx_errors;
+			dev_kfree_skb_any(skb);
+			return;
+		}
+	} else {
+		if (unlikely(skb->len > priv->mcast_mtu + IPOIB_ENCAP_LEN)) {
+			ipoib_warn(priv, "packet len %d (> %d) too long to send, dropping\n",
+				   skb->len, priv->mcast_mtu + IPOIB_ENCAP_LEN);
+			++dev->stats.tx_dropped;
+			++dev->stats.tx_errors;
+			ipoib_cm_skb_too_long(dev, skb, priv->mcast_mtu);
+			return;
+		}
+		phead = NULL;
+		hlen  = 0;
 	}
 
 	ipoib_dbg_data(priv, "sending packet, length=%d address=%p qpn=0x%06x\n",
@@ -442,10 +546,13 @@ void ipoib_send(struct net_device *dev, struct sk_buff *skb,
 		return;
 	}
 
+	if (skb->ip_summed == CHECKSUM_PARTIAL)
+		priv->tx_wr.send_flags |= IB_SEND_IP_CSUM;
+	else
+		priv->tx_wr.send_flags &= ~IB_SEND_IP_CSUM;
+
 	if (unlikely(post_send(priv, priv->tx_head & (ipoib_sendq_size - 1),
-			       address->ah, qpn,
-			       tx_req->mapping, skb_headlen(skb),
-			       skb_shinfo(skb)->frags, skb_shinfo(skb)->nr_frags))) {
+			       address->ah, qpn, tx_req, phead, hlen))) {
 		ipoib_warn(priv, "post_send failed\n");
 		++dev->stats.tx_errors;
 		ipoib_dma_unmap_tx(priv->ca, tx_req);
@@ -540,7 +647,7 @@ static void ipoib_pkey_dev_check_presence(struct net_device *dev)
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	u16 pkey_index = 0;
 
-	if (ib_find_cached_pkey(priv->ca, priv->port, priv->pkey, &pkey_index))
+	if (ib_find_pkey(priv->ca, priv->port, priv->pkey, &pkey_index))
 		clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
 	else
 		set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
@@ -679,10 +786,8 @@ int ipoib_ib_dev_stop(struct net_device *dev, int flush)
 				rx_req = &priv->rx_ring[i];
 				if (!rx_req->skb)
 					continue;
-				ib_dma_unmap_single(priv->ca,
-						    rx_req->mapping,
-						    IPOIB_BUF_SIZE,
-						    DMA_FROM_DEVICE);
+				ipoib_ud_dma_unmap_rx(priv,
+						      priv->rx_ring[i].mapping);
 				dev_kfree_skb_any(rx_req->skb);
 				rx_req->skb = NULL;
 			}
@@ -781,13 +886,13 @@ static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv, int pkey_event)
 			clear_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
 			ipoib_ib_dev_down(dev, 0);
 			ipoib_ib_dev_stop(dev, 0);
-			ipoib_pkey_dev_delay_open(dev);
-			return;
+			if (ipoib_pkey_dev_delay_open(dev))
+				return;
 		}
-		set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags);
 
 		/* restart QP only if P_Key index is changed */
-		if (new_index == priv->pkey_index) {
+		if (test_and_set_bit(IPOIB_PKEY_ASSIGNED, &priv->flags) &&
+		    new_index == priv->pkey_index) {
 			ipoib_dbg(priv, "Not flushing - P_Key index not changed.\n");
 			return;
 		}
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 57282048865c..7a4ed9d3d844 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -195,7 +195,7 @@ static int ipoib_change_mtu(struct net_device *dev, int new_mtu)
 		return 0;
 	}
 
-	if (new_mtu > IPOIB_PACKET_SIZE - IPOIB_ENCAP_LEN)
+	if (new_mtu > IPOIB_UD_MTU(priv->max_ib_mtu))
 		return -EINVAL;
 
 	priv->admin_mtu = new_mtu;
@@ -359,8 +359,7 @@ void ipoib_flush_paths(struct net_device *dev)
 	spin_lock_irq(&priv->tx_lock);
 	spin_lock(&priv->lock);
 
-	list_splice(&priv->path_list, &remove_list);
-	INIT_LIST_HEAD(&priv->path_list);
+	list_splice_init(&priv->path_list, &remove_list);
 
 	list_for_each_entry(path, &remove_list, list)
 		rb_erase(&path->rb_node, &priv->path_tree);
@@ -952,6 +951,8 @@ static void ipoib_setup(struct net_device *dev)
 	dev->set_multicast_list	 = ipoib_set_mcast_list;
 	dev->neigh_setup	 = ipoib_neigh_setup_dev;
 
+	ipoib_set_ethtool_ops(dev);
+
 	netif_napi_add(dev, &priv->napi, ipoib_poll, 100);
 
 	dev->watchdog_timeo	 = HZ;
@@ -970,10 +971,6 @@ static void ipoib_setup(struct net_device *dev)
 				    NETIF_F_LLTX		|
 				    NETIF_F_HIGHDMA);
 
-	/* MTU will be reset when mcast join happens */
-	dev->mtu		 = IPOIB_PACKET_SIZE - IPOIB_ENCAP_LEN;
-	priv->mcast_mtu		 = priv->admin_mtu = dev->mtu;
-
 	memcpy(dev->broadcast, ipv4_bcast_addr, INFINIBAND_ALEN);
 
 	netif_carrier_off(dev);
@@ -1105,6 +1102,8 @@ static struct net_device *ipoib_add_port(const char *format,
 					 struct ib_device *hca, u8 port)
 {
 	struct ipoib_dev_priv *priv;
+	struct ib_device_attr *device_attr;
+	struct ib_port_attr attr;
 	int result = -ENOMEM;
 
 	priv = ipoib_intf_alloc(format);
@@ -1113,6 +1112,18 @@ static struct net_device *ipoib_add_port(const char *format,
 
 	SET_NETDEV_DEV(priv->dev, hca->dma_device);
 
+	if (!ib_query_port(hca, port, &attr))
+		priv->max_ib_mtu = ib_mtu_enum_to_int(attr.max_mtu);
+	else {
+		printk(KERN_WARNING "%s: ib_query_port %d failed\n",
+		       hca->name, port);
+		goto device_init_failed;
+	}
+
+	/* MTU will be reset when mcast join happens */
+	priv->dev->mtu  = IPOIB_UD_MTU(priv->max_ib_mtu);
+	priv->mcast_mtu  = priv->admin_mtu = priv->dev->mtu;
+
 	result = ib_query_pkey(hca, port, 0, &priv->pkey);
 	if (result) {
 		printk(KERN_WARNING "%s: ib_query_pkey port %d failed (ret = %d)\n",
@@ -1120,6 +1131,29 @@ static struct net_device *ipoib_add_port(const char *format,
 		goto device_init_failed;
 	}
 
+	device_attr = kmalloc(sizeof *device_attr, GFP_KERNEL);
+	if (!device_attr) {
+		printk(KERN_WARNING "%s: allocation of %zu bytes failed\n",
+		       hca->name, sizeof *device_attr);
+		goto device_init_failed;
+	}
+
+	result = ib_query_device(hca, device_attr);
+	if (result) {
+		printk(KERN_WARNING "%s: ib_query_device failed (ret = %d)\n",
+		       hca->name, result);
+		kfree(device_attr);
+		goto device_init_failed;
+	}
+	priv->hca_caps = device_attr->device_cap_flags;
+
+	kfree(device_attr);
+
+	if (priv->hca_caps & IB_DEVICE_UD_IP_CSUM) {
+		set_bit(IPOIB_FLAG_CSUM, &priv->flags);
+		priv->dev->features |= NETIF_F_SG | NETIF_F_IP_CSUM;
+	}
+
 	/*
 	 * Set the full membership bit, so that we join the right
 	 * broadcast group, etc.
@@ -1137,7 +1171,6 @@ static struct net_device *ipoib_add_port(const char *format,
 	} else
 		memcpy(priv->dev->dev_addr + 4, priv->local_gid.raw, sizeof (union ib_gid));
 
-
 	result = ipoib_dev_init(priv->dev, hca, port);
 	if (result < 0) {
 		printk(KERN_WARNING "%s: failed to initialize port %d (ret = %d)\n",
@@ -1155,6 +1188,9 @@ static struct net_device *ipoib_add_port(const char *format,
 		goto event_failed;
 	}
 
+	if (priv->dev->features & NETIF_F_SG && priv->hca_caps & IB_DEVICE_UD_TSO)
+		priv->dev->features |= NETIF_F_TSO;
+
 	result = register_netdev(priv->dev);
 	if (result) {
 		printk(KERN_WARNING "%s: couldn't register ipoib port %d; error %d\n",
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
index 31a53c5bcb13..d00a2c174aee 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
@@ -567,8 +567,7 @@ void ipoib_mcast_join_task(struct work_struct *work)
 		return;
 	}
 
-	priv->mcast_mtu = ib_mtu_enum_to_int(priv->broadcast->mcmember.mtu) -
-		IPOIB_ENCAP_LEN;
+	priv->mcast_mtu = IPOIB_UD_MTU(ib_mtu_enum_to_int(priv->broadcast->mcmember.mtu));
 
 	if (!ipoib_cm_admin_enabled(dev))
 		dev->mtu = min(priv->mcast_mtu, priv->admin_mtu);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
index a3aeb911f024..07c03f178a49 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
@@ -150,7 +150,7 @@ int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca)
 			.max_send_wr  = ipoib_sendq_size,
 			.max_recv_wr  = ipoib_recvq_size,
 			.max_send_sge = 1,
-			.max_recv_sge = 1
+			.max_recv_sge = IPOIB_UD_RX_SG
 		},
 		.sq_sig_type = IB_SIGNAL_ALL_WR,
 		.qp_type     = IB_QPT_UD
@@ -192,6 +192,9 @@ int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca)
 	init_attr.send_cq = priv->cq;
 	init_attr.recv_cq = priv->cq;
 
+	if (priv->hca_caps & IB_DEVICE_UD_TSO)
+		init_attr.create_flags = IB_QP_CREATE_IPOIB_UD_LSO;
+
 	if (dev->features & NETIF_F_SG)
 		init_attr.cap.max_send_sge = MAX_SKB_FRAGS + 1;
 
@@ -212,6 +215,19 @@ int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca)
 	priv->tx_wr.sg_list	= priv->tx_sge;
 	priv->tx_wr.send_flags	= IB_SEND_SIGNALED;
 
+	priv->rx_sge[0].lkey = priv->mr->lkey;
+	if (ipoib_ud_need_sg(priv->max_ib_mtu)) {
+		priv->rx_sge[0].length = IPOIB_UD_HEAD_SIZE;
+		priv->rx_sge[1].length = PAGE_SIZE;
+		priv->rx_sge[1].lkey = priv->mr->lkey;
+		priv->rx_wr.num_sge = IPOIB_UD_RX_SG;
+	} else {
+		priv->rx_sge[0].length = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu);
+		priv->rx_wr.num_sge = 1;
+	}
+	priv->rx_wr.next = NULL;
+	priv->rx_wr.sg_list = priv->rx_sge;
+
 	return 0;
 
 out_free_cq:
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
index 293f5b892e3f..431fdeaa2dc4 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
@@ -89,6 +89,7 @@ int ipoib_vlan_add(struct net_device *pdev, unsigned short pkey)
 		goto err;
 	}
 
+	priv->max_ib_mtu = ppriv->max_ib_mtu;
 	set_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags);
 
 	priv->pkey = pkey;