[ofa-general] [UPDATE] [PATCH 3/3] ib/ipoib: IPoIB-UD RX S/G support

Shirley Ma mashirle at us.ibm.com
Thu Jan 31 12:20:20 PST 2008


This patch enables IPoIB-UD RX to allocate S/G buffer up to payload size
4096. The link IPoIB MTU size is up to 4K - 4.

Signed-off-by: Shirley Ma <xma at us.ibm.com>
---

 drivers/infiniband/ulp/ipoib/ipoib.h    |   14 +----
 drivers/infiniband/ulp/ipoib/ipoib_cm.c |   25 ++++----
 drivers/infiniband/ulp/ipoib/ipoib_ib.c |   95
+++++++++++-------------------
 3 files changed, 50 insertions(+), 84 deletions(-)

diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h
b/drivers/infiniband/ulp/ipoib/ipoib.h
index 004a80b..57d33d5 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib.h
+++ b/drivers/infiniband/ulp/ipoib/ipoib.h
@@ -56,9 +56,6 @@
 /* constants */
 
 enum {
-	IPOIB_PACKET_SIZE	  = 2048,
-	IPOIB_BUF_SIZE		  = IPOIB_PACKET_SIZE + IB_GRH_BYTES,
-
 	IPOIB_ENCAP_LEN		  = 4,
 
 	IPOIB_MAX_IB_MTU	  = 4096,
@@ -142,11 +139,6 @@ struct ipoib_mcast {
 
 struct ipoib_rx_buf {
 	struct sk_buff *skb;
-	u64		mapping;
-};
-
-struct ipoib_cm_rx_buf {
-	struct sk_buff *skb;
 	u64		mapping[IPOIB_CM_RX_SG];
 };
 
@@ -198,7 +190,7 @@ enum ipoib_cm_state {
 struct ipoib_cm_rx {
 	struct ib_cm_id	       *id;
 	struct ib_qp	       *qp;
-	struct ipoib_cm_rx_buf *rx_ring;
+	struct ipoib_rx_buf    *rx_ring;
 	struct list_head	list;
 	struct net_device      *dev;
 	unsigned long		jiffies;
@@ -223,7 +215,7 @@ struct ipoib_cm_tx {
 
 struct ipoib_cm_dev_priv {
 	struct ib_srq	       *srq;
-	struct ipoib_cm_rx_buf *srq_ring;
+	struct ipoib_rx_buf    *srq_ring;
 	struct ib_cm_id	       *id;
 	struct list_head	passive_ids;   /* state: LIVE */
 	struct list_head	rx_error_list; /* state: ERROR */
@@ -473,7 +465,7 @@ int ipoib_pkey_dev_delay_open(struct net_device
*dev);
 void ipoib_drain_cq(struct net_device *dev);
 void skb_put_frags(struct sk_buff *skb, unsigned int hdr_space,
 		   unsigned int length, struct sk_buff *toskb);
-struct sk_buff *ipoib_cm_alloc_rx_skb(struct net_device *dev,
+struct sk_buff *ipoib_alloc_rx_skb(struct net_device *dev,
 				      int id, int frags, int head_size,
 				      int pad, u64 *mapping);
 static void inline ipoib_dma_unmap_rx(struct ipoib_dev_priv *priv, int
frags,
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_cm.c
b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
index 2c2c6b2..b2fe0f8 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_cm.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
@@ -126,7 +126,7 @@ static int ipoib_cm_post_receive_nonsrq(struct
net_device *dev,
 }
 
 static void ipoib_cm_free_rx_ring(struct net_device *dev,
-				  struct ipoib_cm_rx_buf *rx_ring)
+				  struct ipoib_rx_buf *rx_ring)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	int i;
@@ -283,11 +283,11 @@ static int ipoib_cm_nonsrq_init_rx(struct
net_device *dev, struct ib_cm_id *cm_i
 	spin_unlock_irq(&priv->lock);
 
 	for (i = 0; i < ipoib_recvq_size; ++i) {
-		rx->rx_ring[i].skb = ipoib_cm_alloc_rx_skb(dev, i,
-							   IPOIB_CM_RX_SG - 1,
-							   IPOIB_CM_HEAD_SIZE, 
-							   12,
-							   rx->rx_ring[i].mapping);
+		rx->rx_ring[i].skb = ipoib_alloc_rx_skb(dev, i,
+							IPOIB_CM_RX_SG - 1,
+							IPOIB_CM_HEAD_SIZE, 
+							12,
+							rx->rx_ring[i].mapping);
 		if (!rx->rx_ring[i].skb) {
 			ipoib_warn(priv, "failed to allocate receive buffer %d\n", i);
 				ret = -ENOMEM;
@@ -426,7 +426,7 @@ static int ipoib_cm_rx_handler(struct ib_cm_id
*cm_id,
 void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
-	struct ipoib_cm_rx_buf *rx_ring;
+	struct ipoib_rx_buf *rx_ring;
 	unsigned int wr_id = wc->wr_id & ~(IPOIB_OP_CM | IPOIB_OP_RECV);
 	struct sk_buff *skb, *newskb;
 	struct ipoib_cm_rx *p;
@@ -491,8 +491,7 @@ void ipoib_cm_handle_rx_wc(struct net_device *dev,
struct ib_wc *wc)
 	frags = PAGE_ALIGN(wc->byte_len - min(wc->byte_len,
 					      (unsigned)IPOIB_CM_HEAD_SIZE)) / PAGE_SIZE;
 
-	newskb = ipoib_cm_alloc_rx_skb(dev, wr_id, frags, IPOIB_CM_HEAD_SIZE, 
-				       12, mapping);
+	newskb = ipoib_alloc_rx_skb(dev, wr_id, frags, IPOIB_CM_HEAD_SIZE, 12,
mapping);
 	if (unlikely(!newskb)) {
 		/*
 		 * If we can't allocate a new RX buffer, dump
@@ -1396,10 +1395,10 @@ int ipoib_cm_dev_init(struct net_device *dev)
 	if (ipoib_cm_has_srq(dev)) {
 		for (i = 0; i < ipoib_recvq_size; ++i) {
 			priv->cm.srq_ring[i].skb = 
-			      ipoib_cm_alloc_rx_skb(dev, i, 
-						    priv->cm.num_frags - 1, 
-						    IPOIB_CM_HEAD_SIZE, 12, 
-						    priv->cm.srq_ring[i].mapping);
+			      ipoib_alloc_rx_skb(dev, i, 
+						 priv->cm.num_frags - 1, 
+						 IPOIB_CM_HEAD_SIZE, 12, 
+						 priv->cm.srq_ring[i].mapping);
 			if (!priv->cm.srq_ring[i].skb) {	
 				ipoib_warn(priv, "failed to allocate "
 					   "receive buffer %d\n", i);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index c40329f..e6540a4 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -90,25 +90,16 @@ void ipoib_free_ah(struct kref *kref)
 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;
-
-	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);
+	priv->rx_wr.wr_id = id | IPOIB_OP_RECV;
+	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_dma_unmap_rx(priv, IPOIB_UD_RX_SG(priv->max_ib_mtu) - 1,
+				   IPOIB_UD_HEAD_SIZE(priv->max_ib_mtu),
+				   priv->rx_ring[id].mapping);
 		dev_kfree_skb_any(priv->rx_ring[id].skb);
 		priv->rx_ring[id].skb = NULL;
 	}
@@ -116,9 +107,9 @@ static int ipoib_ib_post_receive(struct net_device
*dev, int id)
 	return ret;
 }
 
-struct sk_buff *ipoib_cm_alloc_rx_skb(struct net_device *dev,
-				      int id, int frags, int head_size,
-				      int pad, u64 *mapping)
+struct sk_buff *ipoib_alloc_rx_skb(struct net_device *dev,
+				   int id, int frags, int head_size,
+				   int pad, u64 *mapping)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct sk_buff *skb;
@@ -201,43 +192,17 @@ void skb_put_frags(struct sk_buff *skb, unsigned
int hdr_space,
 	}
 }
 
-static int ipoib_alloc_rx_skb(struct net_device *dev, int id)
-{
-	struct ipoib_dev_priv *priv = netdev_priv(dev);
-	struct sk_buff *skb;
-	u64 addr;
-
-	skb = dev_alloc_skb(IPOIB_BUF_SIZE + 4);
-	if (!skb)
-		return -ENOMEM;
-
-	/*
-	 * IB will leave a 40 byte gap for a GRH and IPoIB adds a 4 byte
-	 * header.  So we need 4 more bytes to get to 48 and align the
-	 * IP header to a multiple of 16.
-	 */
-	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;
-	}
-
-	priv->rx_ring[id].skb     = skb;
-	priv->rx_ring[id].mapping = addr;
-
-	return 0;
-}
-
 static int ipoib_ib_post_receives(struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	int i;
 
 	for (i = 0; i < ipoib_recvq_size; ++i) {
-		if (ipoib_alloc_rx_skb(dev, i)) {
+		priv->rx_ring[i].skb = ipoib_alloc_rx_skb(dev, i,
+					IPOIB_UD_RX_SG(priv->max_ib_mtu) - 1,
+					IPOIB_UD_HEAD_SIZE(priv->max_ib_mtu), 4,
+					priv->rx_ring[i].mapping);
+		if (!priv->rx_ring[i].skb) {
 			ipoib_warn(priv, "failed to allocate receive buffer %d\n", i);
 			return -ENOMEM;
 		}
@@ -254,8 +219,9 @@ 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;
+	struct sk_buff *skb, *newskb;
+	u64 mapping[IPOIB_UD_RX_SG(priv->max_ib_mtu)];
+	int frags;
 
 	ipoib_dbg_data(priv, "recv completion: id %d, status: %d\n",
 		       wr_id, wc->status);
@@ -267,15 +233,15 @@ 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_dma_unmap_rx(priv, IPOIB_UD_RX_SG(priv->max_ib_mtu) - 1,
+				   IPOIB_UD_HEAD_SIZE(priv->max_ib_mtu),
+				   priv->rx_ring[wr_id].mapping);
 		dev_kfree_skb_any(skb);
 		priv->rx_ring[wr_id].skb = NULL;
 		return;
@@ -288,11 +254,17 @@ 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;
 
+	frags = PAGE_ALIGN(wc->byte_len - min(wc->byte_len,
+					      (unsigned)(IPOIB_UD_HEAD_SIZE(priv->max_ib_mtu)))) /
PAGE_SIZE;
+	newskb = ipoib_alloc_rx_skb(dev, wr_id, frags,
+				    IPOIB_UD_HEAD_SIZE(priv->max_ib_mtu),
+				    4, 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(newskb)) {
+		ipoib_dbg(priv, "failed to allocate receive buffer %d\n", wr_id);	
 		++dev->stats.rx_dropped;
 		goto repost;
 	}
@@ -300,9 +272,12 @@ 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_dma_unmap_rx(priv, frags, IPOIB_UD_HEAD_SIZE(priv->max_ib_mtu),
+			   priv->rx_ring[wr_id].mapping);
+	memcpy(priv->rx_ring[wr_id].mapping, mapping,
+	       (frags + 1) * sizeof *mapping);
 
-	skb_put(skb, wc->byte_len);
+	skb_put_frags(skb, IPOIB_UD_HEAD_SIZE(priv->max_ib_mtu), wc->byte_len,
newskb);
 	skb_pull(skb, IB_GRH_BYTES);
 
 	skb->protocol = ((struct ipoib_header *) skb->data)->proto;
@@ -715,10 +690,10 @@ 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_dma_unmap_rx(priv,
+						   IPOIB_UD_RX_SG(priv->max_ib_mtu) - 1,
+						   IPOIB_UD_HEAD_SIZE(priv->max_ib_mtu),
+						   priv->rx_ring[i].mapping);
 				dev_kfree_skb_any(rx_req->skb);
 				rx_req->skb = NULL;
 			}





More information about the general mailing list