IPoIB: Copy small received SKBs in connected mode

The connected mode implementation in the IPoIB driver has a large
overhead in the way SKBs are handled in the receive flow.  It usually
allocates an SKB with as big as was used in the currently received SKB
and moves unused fragments from the old SKB to the new one. This
involves a loop on all the remaining fragments and incurs overhead on
the CPU.  This patch, for small SKBs, allocates an SKB just large
enough to contain the received data and copies to it the data from the
received SKB.  The newly allocated SKB is passed to the stack and the
old SKB is reposted.

When running netperf, UDP small messages, without this pach I get:

    UDP UNIDIRECTIONAL SEND TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to
    14.4.3.178 (14.4.3.178) port 0 AF_INET
    Socket  Message  Elapsed      Messages
    Size    Size     Time         Okay Errors   Throughput
    bytes   bytes    secs            #      #   10^6bits/sec

    114688     128   10.00     5142034      0     526.31
    114688           10.00     1130489            115.71

With this patch I get both send and receive at ~315 mbps.

The reason that send performance actually slows down is as follows:
When using this patch, the overhead of the CPU for handling RX packets
is dramatically reduced.  As a result, we do not experience RNR NAK
messages from the receiver which cause the connection to be closed and
reopened again; when the patch is not used, the receiver cannot handle
the packets fast enough so there is less time to post new buffers and
hence the mentioned RNR NACKs.  So what happens is that the
application *thinks* it posted a certain number of packets for
transmission but these packets are flushed and do not really get
transmitted.  Since the connection gets opened and closed many times,
each time netperf gets the CPU time that otherwise would have been
given to IPoIB to actually transmit the packets.  This can be verified
when looking at the port counters -- the output of ifconfig and the
oputput of netperf (this is for the case without the patch):

    tx packets
    ==========
    port counter:   1,543,996
    ifconfig:       1,581,426
    netperf:        5,142,034

    rx packets
    ==========
    netperf         1,1304,089

Signed-off-by: Eli Cohen <eli@mellanox.co.il>

diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h
index 0dcbab3203c910cd3860f9d4908048ffb5613505..8754b364f229667280487cd165b7578d389a3d01 100644 (file)
--- a/drivers/infiniband/ulp/ipoib/ipoib.h
+++ b/drivers/infiniband/ulp/ipoib/ipoib.h
@@ -95,6 +95,7 @@ enum {
        IPOIB_MCAST_FLAG_ATTACHED = 3,
 
        MAX_SEND_CQE              = 16,
+       IPOIB_CM_COPYBREAK        = 256,
 };
 
 #define        IPOIB_OP_RECV   (1ul << 31)

diff --git a/drivers/infiniband/ulp/ipoib/ipoib_cm.c b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
index 91c9592999108d2f329f36f19f8f20e3921f3264..6223fc39af7035bacf7e22e0f8222d44c1616128 100644 (file)
--- a/drivers/infiniband/ulp/ipoib/ipoib_cm.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
@@ -523,6 +523,7 @@ void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
        u64 mapping[IPOIB_CM_RX_SG];
        int frags;
        int has_srq;
+       struct sk_buff *small_skb;
 
        ipoib_dbg_data(priv, "cm recv completion: id %d, status: %d\n",
                       wr_id, wc->status);
@@ -577,6 +578,23 @@ void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
                }
        }
 
+       if (wc->byte_len < IPOIB_CM_COPYBREAK) {
+               int dlen = wc->byte_len;
+
+               small_skb = dev_alloc_skb(dlen + 12);
+               if (small_skb) {
+                       skb_reserve(small_skb, 12);
+                       ib_dma_sync_single_for_cpu(priv->ca, rx_ring[wr_id].mapping[0],
+                                                  dlen, DMA_FROM_DEVICE);
+                       skb_copy_from_linear_data(skb, small_skb->data, dlen);
+                       ib_dma_sync_single_for_device(priv->ca, rx_ring[wr_id].mapping[0],
+                                                     dlen, DMA_FROM_DEVICE);
+                       skb_put(small_skb, dlen);
+                       skb = small_skb;
+                       goto copied;
+               }
+       }
+
        frags = PAGE_ALIGN(wc->byte_len - min(wc->byte_len,
                                              (unsigned)IPOIB_CM_HEAD_SIZE)) / PAGE_SIZE;
 
@@ -599,6 +617,7 @@ void ipoib_cm_handle_rx_wc(struct net_device *dev, struct ib_wc *wc)
 
        skb_put_frags(skb, IPOIB_CM_HEAD_SIZE, wc->byte_len, newskb);
 
+copied:
        skb->protocol = ((struct ipoib_header *) skb->data)->proto;
        skb_reset_mac_header(skb);
        skb_pull(skb, IPOIB_ENCAP_LEN);

diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index f217b1edd0acef1600d5fb18f78f9921aad2fab7..bfe1dbf9920775d1a3632bd805d881e858eac99b 100644 (file)
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -1302,6 +1302,12 @@ static int __init ipoib_init_module(void)
        ipoib_max_conn_qp = min(ipoib_max_conn_qp, IPOIB_CM_MAX_CONN_QP);
 #endif
 
+       /*
+        * When copying small received packets, we only copy from the
+        * linear data part of the SKB, so we rely on this condition.
+        */
+       BUILD_BUG_ON(IPOIB_CM_COPYBREAK > IPOIB_CM_HEAD_SIZE);
+
        ret = ipoib_register_debugfs();
        if (ret)
                return ret;