l2: input performance

Type: improvement

- cache the values form the BD on the input config to avoid loading
- avoid the short write long read on the sequence number
- use vlib_buffer_enqueue_to_next

Signed-off-by: Neale Ranns <nranns@cisco.com>
Change-Id: I33442b9104b457e4c638d26e9ad3bc965687a0bc

1 files changed, 401 insertions, 0 deletions

diff --git a/src/vnet/l2/l2_input_node.c b/src/vnet/l2/l2_input_node.c
new file mode 100644
index 00000000000..898e16ab347
--- /dev/null
+++ b/src/vnet/l2/l2_input_node.c
@@ -0,0 +1,401 @@
+/*
+ * l2_input.c : layer 2 input packet processing
+ *
+ * Copyright (c) 2013 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <vlib/vlib.h>
+#include <vnet/vnet.h>
+#include <vnet/pg/pg.h>
+#include <vnet/ethernet/ethernet.h>
+#include <vnet/ethernet/packet.h>
+#include <vnet/ip/ip_packet.h>
+#include <vnet/ip/ip4_packet.h>
+#include <vnet/ip/ip6_packet.h>
+#include <vnet/fib/fib_node.h>
+#include <vnet/ethernet/arp_packet.h>
+#include <vlib/cli.h>
+#include <vnet/l2/l2_input.h>
+#include <vnet/l2/l2_output.h>
+#include <vnet/l2/feat_bitmap.h>
+#include <vnet/l2/l2_bvi.h>
+#include <vnet/l2/l2_fib.h>
+#include <vnet/l2/l2_bd.h>
+
+#include <vppinfra/error.h>
+#include <vppinfra/hash.h>
+#include <vppinfra/cache.h>
+
+/**
+ * @file
+ * @brief Interface Input Mode (Layer 2 Cross-Connect or Bridge / Layer 3).
+ *
+ * This file contains the CLI Commands that modify the input mode of an
+ * interface. For interfaces in a Layer 2 cross-connect, all packets
+ * received on one interface will be transmitted to the other. For
+ * interfaces in a bridge-domain, packets will be forwarded to other
+ * interfaces in the same bridge-domain based on destination mac address.
+ * For interfaces in Layer 3 mode, the packets will be routed.
+ */
+
+typedef struct
+{
+  /* per-pkt trace data */
+  u8 dst_and_src[12];
+  u32 next_index;
+  u32 sw_if_index;
+  u32 feat_mask;
+} l2input_trace_t;
+
+/* packet trace format function */
+static u8 *
+format_l2input_trace (u8 * s, va_list * args)
+{
+  CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
+  CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
+  l2input_trace_t *t = va_arg (*args, l2input_trace_t *);
+
+  s = format (s, "l2-input: sw_if_index %d dst %U src %U [%U]",
+	      t->sw_if_index,
+	      format_ethernet_address, t->dst_and_src,
+	      format_ethernet_address, t->dst_and_src + 6,
+	      format_l2_input_feature_bitmap, t->feat_mask, 0);
+  return s;
+}
+
+extern l2input_main_t l2input_main;
+
+#ifndef CLIB_MARCH_VARIANT
+l2input_main_t l2input_main;
+#endif /* CLIB_MARCH_VARIANT */
+
+#define foreach_l2input_error			\
+_(L2INPUT,     "L2 input packets")		\
+_(DROP,        "L2 input drops")
+
+typedef enum
+{
+#define _(sym,str) L2INPUT_ERROR_##sym,
+  foreach_l2input_error
+#undef _
+    L2INPUT_N_ERROR,
+} l2input_error_t;
+
+static char *l2input_error_strings[] = {
+#define _(sym,string) string,
+  foreach_l2input_error
+#undef _
+};
+
+typedef enum
+{				/*  */
+  L2INPUT_NEXT_LEARN,
+  L2INPUT_NEXT_FWD,
+  L2INPUT_NEXT_DROP,
+  L2INPUT_N_NEXT,
+} l2input_next_t;
+
+static_always_inline void
+classify_and_dispatch (l2input_main_t * msm, vlib_buffer_t * b0, u16 * next0)
+{
+  /*
+   * Load L2 input feature struct
+   * Load bridge domain struct
+   * Parse ethernet header to determine unicast/mcast/broadcast
+   * take L2 input stat
+   * classify packet as IP/UDP/TCP, control, other
+   * mask feature bitmap
+   * go to first node in bitmap
+   * Later: optimize VTM
+   *
+   * For L2XC,
+   *   set tx sw-if-handle
+   */
+
+  u32 feat_mask = ~0;
+  u32 sw_if_index0 = vnet_buffer (b0)->sw_if_index[VLIB_RX];
+  ethernet_header_t *h0 = vlib_buffer_get_current (b0);
+
+  /* Get config for the input interface */
+  l2_input_config_t *config = vec_elt_at_index (msm->configs, sw_if_index0);
+
+  /* Save split horizon group */
+  vnet_buffer (b0)->l2.shg = config->shg;
+
+  /* determine layer2 kind for stat and mask */
+  if (PREDICT_FALSE (ethernet_address_cast (h0->dst_address)))
+    {
+      u8 *l3h0 = (u8 *) h0 + vnet_buffer (b0)->l2.l2_len;
+
+#define get_u16(addr) ( *((u16 *)(addr)) )
+      u16 ethertype = clib_net_to_host_u16 (get_u16 (l3h0 - 2));
+      u8 protocol = ((ip6_header_t *) l3h0)->protocol;
+
+      /* Disable bridge forwarding (flooding will execute instead if not xconnect) */
+      feat_mask &= ~(L2INPUT_FEAT_FWD |
+		     L2INPUT_FEAT_UU_FLOOD |
+		     L2INPUT_FEAT_UU_FWD | L2INPUT_FEAT_GBP_FWD);
+
+      if (ethertype != ETHERNET_TYPE_ARP)
+	feat_mask &= ~(L2INPUT_FEAT_ARP_UFWD);
+
+      /* Disable ARP-term for non-ARP and non-ICMP6 packet */
+      if (ethertype != ETHERNET_TYPE_ARP &&
+	  (ethertype != ETHERNET_TYPE_IP6 || protocol != IP_PROTOCOL_ICMP6))
+	feat_mask &= ~(L2INPUT_FEAT_ARP_TERM);
+      /*
+       * For packet from BVI - set SHG of ARP request or ICMPv6 neighbor
+       * solicitation packet from BVI to 0 so it can also flood to VXLAN
+       * tunnels or other ports with the same SHG as that of the BVI.
+       */
+      else if (PREDICT_FALSE (vnet_buffer (b0)->sw_if_index[VLIB_TX] ==
+			      L2INPUT_BVI))
+	{
+	  if (ethertype == ETHERNET_TYPE_ARP)
+	    {
+	      ethernet_arp_header_t *arp0 = (ethernet_arp_header_t *) l3h0;
+	      if (arp0->opcode ==
+		  clib_host_to_net_u16 (ETHERNET_ARP_OPCODE_request))
+		vnet_buffer (b0)->l2.shg = 0;
+	    }
+	  else			/* must be ICMPv6 */
+	    {
+	      ip6_header_t *iph0 = (ip6_header_t *) l3h0;
+	      icmp6_neighbor_solicitation_or_advertisement_header_t *ndh0;
+	      ndh0 = ip6_next_header (iph0);
+	      if (ndh0->icmp.type == ICMP6_neighbor_solicitation)
+		vnet_buffer (b0)->l2.shg = 0;
+	    }
+	}
+    }
+  else
+    {
+      /*
+       * For packet from BVI - set SHG of unicast packet from BVI to 0 so it
+       * is not dropped on output to VXLAN tunnels or other ports with the
+       * same SHG as that of the BVI.
+       */
+      if (PREDICT_FALSE (vnet_buffer (b0)->sw_if_index[VLIB_TX] ==
+			 L2INPUT_BVI))
+	vnet_buffer (b0)->l2.shg = 0;
+    }
+
+
+  if (l2_input_is_bridge (config))
+    {
+      /* Do bridge-domain processing */
+      /* save BD ID for next feature graph nodes */
+      vnet_buffer (b0)->l2.bd_index = config->bd_index;
+
+      /* Save bridge domain and interface seq_num */
+      vnet_buffer (b0)->l2.l2fib_sn = l2_fib_mk_seq_num
+	(config->bd_seq_num, config->seq_num);
+      vnet_buffer (b0)->l2.bd_age = config->bd_mac_age;
+
+      /*
+       * Process bridge domain feature enables.
+       * To perform learning/flooding/forwarding, the corresponding bit
+       * must be enabled in both the input interface config and in the
+       * bridge domain config. In the bd_bitmap, bits for features other
+       * than learning/flooding/forwarding should always be set.
+       */
+      feat_mask = feat_mask & config->bd_feature_bitmap;
+    }
+  else if (l2_input_is_xconnect (config))
+    {
+      /* Set the output interface */
+      vnet_buffer (b0)->sw_if_index[VLIB_TX] = config->output_sw_if_index;
+    }
+  else
+    feat_mask = L2INPUT_FEAT_DROP;
+
+  /* mask out features from bitmap using packet type and bd config */
+  u32 feature_bitmap = config->feature_bitmap & feat_mask;
+
+  /* save for next feature graph nodes */
+  vnet_buffer (b0)->l2.feature_bitmap = feature_bitmap;
+
+  /* Determine the next node */
+  *next0 = feat_bitmap_get_next_node_index (msm->feat_next_node_index,
+					    feature_bitmap);
+}
+
+static_always_inline uword
+l2input_node_inline (vlib_main_t * vm,
+		     vlib_node_runtime_t * node, vlib_frame_t * frame,
+		     int do_trace)
+{
+  u32 n_left, *from;
+  l2input_next_t next_index;
+  l2input_main_t *msm = &l2input_main;
+  vlib_buffer_t *bufs[VLIB_FRAME_SIZE], **b = bufs;
+  u16 nexts[VLIB_FRAME_SIZE], *next = nexts;
+
+  from = vlib_frame_vector_args (frame);
+  n_left = frame->n_vectors;	/* number of packets to process */
+  next_index = node->cached_next_index;
+
+  vlib_get_buffers (vm, from, bufs, n_left);
+
+  while (n_left > 0)
+    {
+      while (n_left >= 8)
+	{
+	  u32 sw_if_index0, sw_if_index1, sw_if_index2, sw_if_index3;
+
+	  /* Prefetch next iteration. */
+	  {
+	    /* Prefetch the buffer header and packet for the N+2 loop iteration */
+	    vlib_prefetch_buffer_header (b[4], LOAD);
+	    vlib_prefetch_buffer_header (b[5], LOAD);
+	    vlib_prefetch_buffer_header (b[6], LOAD);
+	    vlib_prefetch_buffer_header (b[7], LOAD);
+
+	    CLIB_PREFETCH (b[4]->data, CLIB_CACHE_LINE_BYTES, STORE);
+	    CLIB_PREFETCH (b[5]->data, CLIB_CACHE_LINE_BYTES, STORE);
+	    CLIB_PREFETCH (b[6]->data, CLIB_CACHE_LINE_BYTES, STORE);
+	    CLIB_PREFETCH (b[7]->data, CLIB_CACHE_LINE_BYTES, STORE);
+	  }
+
+	  classify_and_dispatch (msm, b[0], &next[0]);
+	  classify_and_dispatch (msm, b[1], &next[1]);
+	  classify_and_dispatch (msm, b[2], &next[2]);
+	  classify_and_dispatch (msm, b[3], &next[3]);
+
+	  if (do_trace)
+	    {
+	      /* RX interface handles */
+	      sw_if_index0 = vnet_buffer (b[0])->sw_if_index[VLIB_RX];
+	      sw_if_index1 = vnet_buffer (b[1])->sw_if_index[VLIB_RX];
+	      sw_if_index2 = vnet_buffer (b[2])->sw_if_index[VLIB_RX];
+	      sw_if_index3 = vnet_buffer (b[3])->sw_if_index[VLIB_RX];
+
+	      if (b[0]->flags & VLIB_BUFFER_IS_TRACED)
+		{
+		  ethernet_header_t *h0 = vlib_buffer_get_current (b[0]);
+		  l2input_trace_t *t =
+		    vlib_add_trace (vm, node, b[0], sizeof (*t));
+		  t->sw_if_index = sw_if_index0;
+		  t->feat_mask = vnet_buffer (b[0])->l2.feature_bitmap;
+		  clib_memcpy_fast (t->dst_and_src, h0->dst_address,
+				    sizeof (h0->dst_address) +
+				    sizeof (h0->src_address));
+		}
+	      if (b[1]->flags & VLIB_BUFFER_IS_TRACED)
+		{
+		  ethernet_header_t *h1 = vlib_buffer_get_current (b[1]);
+		  l2input_trace_t *t =
+		    vlib_add_trace (vm, node, b[1], sizeof (*t));
+		  t->sw_if_index = sw_if_index1;
+		  t->feat_mask = vnet_buffer (b[1])->l2.feature_bitmap;
+		  clib_memcpy_fast (t->dst_and_src, h1->dst_address,
+				    sizeof (h1->dst_address) +
+				    sizeof (h1->src_address));
+		}
+	      if (b[2]->flags & VLIB_BUFFER_IS_TRACED)
+		{
+		  ethernet_header_t *h2 = vlib_buffer_get_current (b[2]);
+		  l2input_trace_t *t =
+		    vlib_add_trace (vm, node, b[2], sizeof (*t));
+		  t->sw_if_index = sw_if_index2;
+		  t->feat_mask = vnet_buffer (b[2])->l2.feature_bitmap;
+		  clib_memcpy_fast (t->dst_and_src, h2->dst_address,
+				    sizeof (h2->dst_address) +
+				    sizeof (h2->src_address));
+		}
+	      if (b[3]->flags & VLIB_BUFFER_IS_TRACED)
+		{
+		  ethernet_header_t *h3 = vlib_buffer_get_current (b[3]);
+		  l2input_trace_t *t =
+		    vlib_add_trace (vm, node, b[3], sizeof (*t));
+		  t->sw_if_index = sw_if_index3;
+		  t->feat_mask = vnet_buffer (b[3])->l2.feature_bitmap;
+		  clib_memcpy_fast (t->dst_and_src, h3->dst_address,
+				    sizeof (h3->dst_address) +
+				    sizeof (h3->src_address));
+		}
+	    }
+
+	  b += 4;
+	  n_left -= 4;
+	  next += 4;
+	}
+
+      while (n_left > 0)
+	{
+	  classify_and_dispatch (msm, b[0], &next[0]);
+
+	  if (do_trace && PREDICT_FALSE (b[0]->flags & VLIB_BUFFER_IS_TRACED))
+	    {
+	      ethernet_header_t *h0 = vlib_buffer_get_current (b[0]);
+	      l2input_trace_t *t =
+		vlib_add_trace (vm, node, b[0], sizeof (*t));
+	      t->sw_if_index = vnet_buffer (b[0])->sw_if_index[VLIB_RX];
+	      t->feat_mask = vnet_buffer (b[0])->l2.feature_bitmap;
+	      clib_memcpy_fast (t->dst_and_src, h0->dst_address,
+				sizeof (h0->dst_address) +
+				sizeof (h0->src_address));
+	    }
+
+	  b += 1;
+	  next += 1;
+	  n_left -= 1;
+	}
+    }
+
+  vlib_node_increment_counter (vm, l2input_node.index,
+			       L2INPUT_ERROR_L2INPUT, frame->n_vectors);
+
+  vlib_buffer_enqueue_to_next (vm, node, from, nexts, frame->n_vectors);
+
+  return frame->n_vectors;
+}
+
+VLIB_NODE_FN (l2input_node) (vlib_main_t * vm,
+			     vlib_node_runtime_t * node, vlib_frame_t * frame)
+{
+  if (PREDICT_FALSE ((node->flags & VLIB_NODE_FLAG_TRACE)))
+    return l2input_node_inline (vm, node, frame, 1 /* do_trace */ );
+  return l2input_node_inline (vm, node, frame, 0 /* do_trace */ );
+}
+
+/* *INDENT-OFF* */
+VLIB_REGISTER_NODE (l2input_node) = {
+  .name = "l2-input",
+  .vector_size = sizeof (u32),
+  .format_trace = format_l2input_trace,
+  .format_buffer = format_ethernet_header_with_length,
+  .type = VLIB_NODE_TYPE_INTERNAL,
+
+  .n_errors = ARRAY_LEN(l2input_error_strings),
+  .error_strings = l2input_error_strings,
+
+  .n_next_nodes = L2INPUT_N_NEXT,
+
+  /* edit / add dispositions here */
+  .next_nodes = {
+       [L2INPUT_NEXT_LEARN] = "l2-learn",
+       [L2INPUT_NEXT_FWD]   = "l2-fwd",
+       [L2INPUT_NEXT_DROP]  = "error-drop",
+  },
+};
+/* *INDENT-ON* */
+
+/*
+ * fd.io coding-style-patch-verification: ON
+ *
+ * Local Variables:
+ * eval: (c-set-style "gnu")
+ * End:
+ */