19.01 Release Notes

Change-Id: I43fd3aac9039c6d551fac8607374a0ebfdc6f74a
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>

.gitreview

@@ -2,3 +2,4 @@
 host=gerrit.fd.io
 port=29418
 project=vpp
+defaultbranch=stable/1901

doxygen/test_framework_doc.md

@@ -4,6 +4,7 @@ Test Framework Documentation    {#test_framework_doc}
 PyDoc generated documentation for the "make test" framework is available for
 the following releases:
 
+- [Test framework documentation for VPP 19.01](https://docs.fd.io/vpp/19.01/vpp_make_test/html)
 - [Test framework documentation for VPP 18.10](https://docs.fd.io/vpp/18.10/vpp_make_test/html)
 - [Test framework documentation for VPP 18.07](https://docs.fd.io/vpp/18.07/vpp_make_test/html)
 - [Test framework documentation for VPP 18.04](https://docs.fd.io/vpp/18.04/vpp_make_test/html)

extras/scripts/list_api_changes.py

@@ -1,8 +1,8 @@
 #!/usr/bin/env python
 import os, fnmatch, subprocess
 
-starttag = 'v18.10-rc0'
-endtag = 'v18.10'
+starttag = 'v19.01-rc0'
+endtag = 'v19.01'
 emit_md = True
 apifiles = []
 

extras/vom/vom/route.cpp

@@ -632,7 +632,8 @@ ip_mroute::dump(std::ostream& os)
 ip_mroute::event_handler::event_handler()
 {
   OM::register_listener(this);
-  inspect::register_handler({ "ip-route" }, "ip route configurations", this);
+  inspect::register_handler({ "ip-mroute" },
+                            "ip multicast route configurations", this);
 }
 
 void

src/plugins/avf/device.c

@@ -653,14 +653,19 @@ clib_error_t *
 avf_op_enable_queues (vlib_main_t * vm, avf_device_t * ad, u32 rx, u32 tx)
 {
   virtchnl_queue_select_t qs = { 0 };
-  int i;
+  int i = 0;
   qs.vsi_id = ad->vsi_id;
   qs.rx_queues = rx;
   qs.tx_queues = tx;
-  for (i = 0; i < ad->n_rx_queues; i++)
+  while (rx)
     {
-      avf_rxq_t *rxq = vec_elt_at_index (ad->rxqs, i);
-      avf_reg_write (ad, AVF_QRX_TAIL (i), rxq->n_enqueued);
+      if (rx & (1 << i))
+	{
+	  avf_rxq_t *rxq = vec_elt_at_index (ad->rxqs, i);
+	  avf_reg_write (ad, AVF_QRX_TAIL (i), rxq->n_enqueued);
+	  rx &= ~(1 << i);
+	}
+      i++;
     }
   return avf_send_to_pf (vm, ad, VIRTCHNL_OP_ENABLE_QUEUES, &qs,
 			 sizeof (virtchnl_queue_select_t), 0, 0);
@@ -850,10 +855,8 @@ avf_device_init (vlib_main_t * vm, avf_main_t * am, avf_device_t * ad,
   if ((error = avf_op_add_eth_addr (vm, ad, 1, ad->hwaddr)))
     return error;
 
-  if ((error = avf_op_enable_queues (vm, ad, ad->n_rx_queues, 0)))
-    return error;
-
-  if ((error = avf_op_enable_queues (vm, ad, 0, ad->n_tx_queues)))
+  if ((error = avf_op_enable_queues (vm, ad, pow2_mask (ad->n_rx_queues),
+				     pow2_mask (ad->n_tx_queues))))
     return error;
 
   ad->flags |= AVF_DEVICE_F_INITIALIZED;

src/vcl/ldp.c

@@ -1700,6 +1700,11 @@ getsockopt (int fd, int level, int optname,
 	      else
 		rv = -EFAULT;
 	      break;
+	    case TCP_CONGESTION:
+	      strcpy (optval, "cubic");
+	      *optlen = strlen ("cubic");
+	      rv = 0;
+	      break;
 	    default:
 	      LDBG (0, "ERROR: fd %d: getsockopt SOL_TCP: sid %u, "
 		    "optname %d unsupported!", fd, vlsh, optname);
@@ -1808,6 +1813,10 @@ setsockopt (int fd, int level, int optname,
 	      rv = vls_attr (vlsh, VPPCOM_ATTR_SET_TCP_KEEPINTVL,
 			     (void *) optval, &optlen);
 	      break;
+	    case TCP_CONGESTION:
+	      /* Ignore */
+	      rv = 0;
+	      break;
 	    default:
 	      LDBG (0, "ERROR: fd %d: setsockopt() SOL_TCP: vlsh %u"
 		    "optname %d unsupported!", fd, vlsh, optname);

src/vnet/bonding/cli.c

@@ -531,11 +531,13 @@ bond_enslave (vlib_main_t * vm, bond_enslave_args_t * args)
       ethernet_set_rx_redirect (vnm, sif_hw, 1);
     }
 
-  if ((bif->mode == BOND_MODE_LACP) && bm->lacp_enable_disable)
+  if (bif->mode == BOND_MODE_LACP)
     {
-      (*bm->lacp_enable_disable) (vm, bif, sif, 1);
+      if (bm->lacp_enable_disable)
+	(*bm->lacp_enable_disable) (vm, bif, sif, 1);
     }
-  else
+  else if (sif->port_enabled &&
+	   (sif_hw->flags & VNET_HW_INTERFACE_FLAG_LINK_UP))
     {
       bond_enable_collecting_distributing (vm, sif);
     }

src/vnet/bonding/node.c

@@ -396,19 +396,21 @@ bond_sw_interface_up_down (vnet_main_t * vnm, u32 sw_if_index, u32 flags)
   if (sif)
     {
       sif->port_enabled = flags & VNET_SW_INTERFACE_FLAG_ADMIN_UP;
+      if (sif->lacp_enabled)
+	return 0;
+
       if (sif->port_enabled == 0)
 	{
-	  if (sif->lacp_enabled == 0)
-	    {
-	      bond_disable_collecting_distributing (vm, sif);
-	    }
+	  bond_disable_collecting_distributing (vm, sif);
 	}
       else
 	{
-	  if (sif->lacp_enabled == 0)
-	    {
-	      bond_enable_collecting_distributing (vm, sif);
-	    }
+	  vnet_main_t *vnm = vnet_get_main ();
+	  vnet_hw_interface_t *hw =
+	    vnet_get_sup_hw_interface (vnm, sw_if_index);
+
+	  if (hw->flags & VNET_HW_INTERFACE_FLAG_LINK_UP)
+	    bond_enable_collecting_distributing (vm, sif);
 	}
     }
 
@@ -429,19 +431,16 @@ bond_hw_interface_up_down (vnet_main_t * vnm, u32 hw_if_index, u32 flags)
   sif = bond_get_slave_by_sw_if_index (sw->sw_if_index);
   if (sif)
     {
+      if (sif->lacp_enabled)
+	return 0;
+
       if (!(flags & VNET_HW_INTERFACE_FLAG_LINK_UP))
 	{
-	  if (sif->lacp_enabled == 0)
-	    {
-	      bond_disable_collecting_distributing (vm, sif);
-	    }
+	  bond_disable_collecting_distributing (vm, sif);
 	}
-      else
+      else if (sif->port_enabled)
 	{
-	  if (sif->lacp_enabled == 0)
-	    {
-	      bond_enable_collecting_distributing (vm, sif);
-	    }
+	  bond_enable_collecting_distributing (vm, sif);
 	}
     }
 

src/vnet/ip/ip_api.c

@@ -331,19 +331,17 @@ send_ip6_fib_details (vpe_api_main_t * am,
 
 typedef struct apt_ip6_fib_show_ctx_t_
 {
-  u32 fib_index;
   fib_node_index_t *entries;
 } api_ip6_fib_show_ctx_t;
 
-static void
-api_ip6_fib_table_put_entries (clib_bihash_kv_24_8_t * kvp, void *arg)
+static fib_table_walk_rc_t
+api_ip6_fib_table_put_entries (fib_node_index_t fei, void *arg)
 {
   api_ip6_fib_show_ctx_t *ctx = arg;
 
-  if ((kvp->key[2] >> 32) == ctx->fib_index)
-    {
-      vec_add1 (ctx->entries, kvp->value);
-    }
+  vec_add1 (ctx->entries, fei);
+
+  return (FIB_TABLE_WALK_CONTINUE);
 }
 
 static void
@@ -352,18 +350,15 @@ api_ip6_fib_table_get_all (vl_api_registration_t * reg,
 			   fib_table_t * fib_table)
 {
   vpe_api_main_t *am = &vpe_api_main;
-  ip6_main_t *im6 = &ip6_main;
   fib_node_index_t *fib_entry_index;
   api_ip6_fib_show_ctx_t ctx = {
-    .fib_index = fib_table->ft_index,
     .entries = NULL,
   };
   fib_route_path_encode_t *api_rpaths;
   const fib_prefix_t *pfx;
 
-  BV (clib_bihash_foreach_key_value_pair)
-    ((BVT (clib_bihash) *) & im6->ip6_table[IP6_FIB_TABLE_NON_FWDING].
-     ip6_hash, api_ip6_fib_table_put_entries, &ctx);
+  ip6_fib_table_walk (fib_table->ft_index,
+		      api_ip6_fib_table_put_entries, &ctx);
 
   vec_sort_with_function (ctx.entries, fib_entry_cmp_for_sort);
 

src/vnet/ipsec/ipsec.c

@@ -533,12 +533,22 @@ ipsec_rand_seed (void)
 }
 
 static clib_error_t *
-ipsec_check_support (ipsec_sa_t * sa)
+ipsec_check_ah_support (ipsec_sa_t * sa)
+{
+  if (sa->integ_alg == IPSEC_INTEG_ALG_NONE)
+    return clib_error_return (0, "unsupported none integ-alg");
+  return 0;
+}
+
+static clib_error_t *
+ipsec_check_esp_support (ipsec_sa_t * sa)
 {
   if (sa->crypto_alg == IPSEC_CRYPTO_ALG_AES_GCM_128)
     return clib_error_return (0, "unsupported aes-gcm-128 crypto-alg");
-  if (sa->integ_alg == IPSEC_INTEG_ALG_NONE)
-    return clib_error_return (0, "unsupported none integ-alg");
+  if (sa->crypto_alg == IPSEC_CRYPTO_ALG_AES_GCM_192)
+    return clib_error_return (0, "unsupported aes-gcm-192 crypto-alg");
+  if (sa->crypto_alg == IPSEC_CRYPTO_ALG_AES_GCM_256)
+    return clib_error_return (0, "unsupported aes-gcm-256 crypto-alg");
 
   return 0;
 }
@@ -730,7 +740,7 @@ ipsec_init (vlib_main_t * vm)
 				       "ah4-decrypt",
 				       "ah6-encrypt",
 				       "ah6-decrypt",
-				       ipsec_check_support,
+				       ipsec_check_ah_support,
 				       NULL);
 
   im->ah_default_backend = idx;
@@ -743,7 +753,7 @@ ipsec_init (vlib_main_t * vm)
 				    "esp4-decrypt",
 				    "esp6-encrypt",
 				    "esp6-decrypt",
-				    ipsec_check_support, NULL);
+				    ipsec_check_esp_support, NULL);
   im->esp_default_backend = idx;
 
   rv = ipsec_select_esp_backend (im, idx);

src/vnet/ipsec/ipsec_api.c

@@ -717,9 +717,10 @@ static void
 #if WITH_LIBSSL > 0
   vlib_main_t *vm = vlib_get_main ();
   clib_error_t *error;
+  int data_len = ntohl (mp->data_len);
   u8 *tmp = format (0, "%s", mp->name);
-  u8 *data = vec_new (u8, mp->data_len);
-  clib_memcpy (data, mp->data, mp->data_len);
+  u8 *data = vec_new (u8, data_len);
+  clib_memcpy (data, mp->data, data_len);
   error = ikev2_set_profile_auth (vm, tmp, mp->auth_method, data, mp->is_hex);
   vec_free (tmp);
   vec_free (data);
@@ -742,8 +743,9 @@ vl_api_ikev2_profile_set_id_t_handler (vl_api_ikev2_profile_set_id_t * mp)
   vlib_main_t *vm = vlib_get_main ();
   clib_error_t *error;
   u8 *tmp = format (0, "%s", mp->name);
-  u8 *data = vec_new (u8, mp->data_len);
-  clib_memcpy (data, mp->data, mp->data_len);
+  int data_len = ntohl (mp->data_len);
+  u8 *data = vec_new (u8, data_len);
+  clib_memcpy (data, mp->data, data_len);
   error = ikev2_set_profile_id (vm, tmp, mp->id_type, data, mp->is_local);
   vec_free (tmp);
   vec_free (data);

src/vnet/ipsec/ipsec_if.c

@@ -299,6 +299,7 @@ ipsec_add_del_tunnel_if_internal (vnet_main_t * vnm,
       pool_get (im->sad, sa);
       clib_memset (sa, 0, sizeof (*sa));
       t->input_sa_index = sa - im->sad;
+      sa->protocol = IPSEC_PROTOCOL_ESP;
       sa->spi = args->remote_spi;
       sa->tunnel_src_addr.ip4.as_u32 = args->remote_ip.as_u32;
       sa->tunnel_dst_addr.ip4.as_u32 = args->local_ip.as_u32;
@@ -325,6 +326,7 @@ ipsec_add_del_tunnel_if_internal (vnet_main_t * vnm,
       pool_get (im->sad, sa);
       clib_memset (sa, 0, sizeof (*sa));
       t->output_sa_index = sa - im->sad;
+      sa->protocol = IPSEC_PROTOCOL_ESP;
       sa->spi = args->local_spi;
       sa->tunnel_src_addr.ip4.as_u32 = args->local_ip.as_u32;
       sa->tunnel_dst_addr.ip4.as_u32 = args->remote_ip.as_u32;

src/vnet/session/session_cli.c

@@ -100,12 +100,14 @@ uword
 unformat_stream_session_id (unformat_input_t * input, va_list * args)
 {
   u8 *proto = va_arg (*args, u8 *);
+  u32 *fib_index = va_arg (*args, u32 *);
   ip46_address_t *lcl = va_arg (*args, ip46_address_t *);
   ip46_address_t *rmt = va_arg (*args, ip46_address_t *);
   u16 *lcl_port = va_arg (*args, u16 *);
   u16 *rmt_port = va_arg (*args, u16 *);
   u8 *is_ip4 = va_arg (*args, u8 *);
   u8 tuple_is_set = 0;
+  u32 vrf = ~0;
 
   clib_memset (lcl, 0, sizeof (*lcl));
   clib_memset (rmt, 0, sizeof (*rmt));
@@ -114,10 +116,16 @@ unformat_stream_session_id (unformat_input_t * input, va_list * args)
     {
       *proto = TRANSPORT_PROTO_TCP;
     }
-  if (unformat (input, "udp"))
+  else if (unformat (input, "udp"))
     {
       *proto = TRANSPORT_PROTO_UDP;
     }
+  else
+    return 0;
+
+  if (unformat (input, "vrf %u", &vrf))
+    ;
+
   if (unformat (input, "%U:%d->%U:%d", unformat_ip4_address, &lcl->ip4,
 		lcl_port, unformat_ip4_address, &rmt->ip4, rmt_port))
     {
@@ -131,6 +139,13 @@ unformat_stream_session_id (unformat_input_t * input, va_list * args)
       tuple_is_set = 1;
     }
 
+  if (vrf != ~0)
+    {
+      fib_protocol_t fib_proto;
+      fib_proto = *is_ip4 ? FIB_PROTOCOL_IP4 : FIB_PROTOCOL_IP6;
+      *fib_index = fib_table_find (fib_proto, vrf);
+    }
+
   return tuple_is_set;
 }
 
@@ -138,14 +153,14 @@ uword
 unformat_stream_session (unformat_input_t * input, va_list * args)
 {
   stream_session_t **result = va_arg (*args, stream_session_t **);
+  u32 lcl_port = 0, rmt_port = 0, fib_index = 0;
+  ip46_address_t lcl, rmt;
   stream_session_t *s;
   u8 proto = ~0;
-  ip46_address_t lcl, rmt;
-  u32 lcl_port = 0, rmt_port = 0, fib_index = 0;
   u8 is_ip4 = 0;
 
-  if (!unformat (input, "%U", unformat_stream_session_id, &proto, &lcl, &rmt,
-		 &lcl_port, &rmt_port, &is_ip4))
+  if (!unformat (input, "%U", unformat_stream_session_id, &proto, &fib_index,
+		 &lcl, &rmt, &lcl_port, &rmt_port, &is_ip4))
     return 0;
 
   if (is_ip4)
@@ -176,8 +191,8 @@ unformat_transport_connection (unformat_input_t * input, va_list * args)
   u32 lcl_port = 0, rmt_port = 0, fib_index = 0;
   u8 is_ip4 = 0;
 
-  if (!unformat (input, "%U", unformat_stream_session_id, &proto, &lcl, &rmt,
-		 &lcl_port, &rmt_port, &is_ip4))
+  if (!unformat (input, "%U", unformat_stream_session_id, &fib_index, &proto,
+		 &lcl, &rmt, &lcl_port, &rmt_port, &is_ip4))
     return 0;
 
   proto = (proto == (u8) ~ 0) ? suggested_proto : proto;

src/vnet/tcp/tcp_debug.h

@@ -879,6 +879,7 @@ if (tcp_cc_time_to_print_stats (_tc))					\
 }
 #else
 #define TCP_EVT_CC_STAT_HANDLER(_tc, ...)
+#define TCP_EVT_CC_STAT_PRINT(_tc)
 #endif
 
 /*

src/vnet/tcp/tcp_input.c

@@ -359,6 +359,7 @@ tcp_segment_validate (tcp_worker_ctx_t * wrk, tcp_connection_t * tc0,
   /* 4th: check the SYN bit */
   if (PREDICT_FALSE (tcp_syn (th0)))
     {
+      *error0 = tcp_ack (th0) ? TCP_ERROR_SYN_ACKS_RCVD : TCP_ERROR_SYNS_RCVD;
       /* TODO implement RFC 5961 */
       if (tc0->state == TCP_STATE_SYN_RCVD)
 	{
@@ -2712,8 +2713,7 @@ tcp46_rcv_process_inline (vlib_main_t * vm, vlib_node_runtime_t * node,
 	    }
 
 	  /* Make sure the ack is exactly right */
-	  if (tc0->rcv_nxt != vnet_buffer (b0)->tcp.seq_number || is_fin0
-	      || vnet_buffer (b0)->tcp.data_len)
+	  if (tc0->rcv_nxt != vnet_buffer (b0)->tcp.seq_number || is_fin0)
 	    {
 	      tcp_connection_reset (tc0);
 	      error0 = TCP_ERROR_SEGMENT_INVALID;
@@ -3144,6 +3144,7 @@ tcp46_listen_inline (vlib_main_t * vm, vlib_node_runtime_t * node,
       child0->snd_wl2 = vnet_buffer (b0)->tcp.ack_number;
 
       tcp_connection_init_vars (child0);
+      child0->rto = TCP_RTO_MIN;
       TCP_EVT_DBG (TCP_EVT_SYN_RCVD, child0, 1);
 
       if (stream_session_accept (&child0->connection, lc0->c_s_index,

src/vppinfra/string.c

@@ -385,7 +385,7 @@ strncat_s (char *__restrict__ dest, rsize_t dmax,
  *         s1max and ptr are modified to contain the state
  *         null      runtime constraint error or token is not found
  *
- * @example
+ * Example:
  *   char *str2 = " ";
  *   char str1[100];
  *   uword len;
@@ -457,7 +457,7 @@ strnlen_s (const char *s, size_t maxsize)
  *         EINVAL runtime constraint error
  *         ESRCH  no match
  *
- * @example
+ * Example:
  *   char *sub = 0;
  *   char *s1 = "success is not final, failure is not fatal.";
  *

test/requirements.txt

@@ -7,9 +7,7 @@ parameterized>=0.6.1                    # BSD
 pexpect                                 # ISC
 psutil                                  # BSD
 pycodestyle                             # MIT (Expat license)       https://pypi.org/project/pycodestyle/
-# used by syslog_rfc5424_parser
-pyparsing<2.3.1                         # MIT
 scapy==2.4.0; python_version >= '2.7' or python_version >= '3.4'    # GPL2  https://github.com/secdev/scapy/blob/master/LICENSE
 six                                     # MIT
 subprocess32                            # PSF
-syslog_rfc5424_parser                   # ISC
+syslog_rfc5424_parser>=0.2.0            # ISC

test/test_syslog.py

@@ -127,8 +127,8 @@ class TestSyslog(VppTestCase):
             self.assertEqual(capture[0][UDP].dport, 514)
             self.assert_packet_checksums_valid(capture[0], False)
         except:
-                self.logger.error(ppp("invalid packet:", capture[0]))
-                raise
+            self.logger.error(ppp("invalid packet:", capture[0]))
+            raise
         self.syslog_verify(capture[0][Raw].load,
                            SyslogFacility.local7,
                            SyslogSeverity.info,
@@ -182,8 +182,8 @@ class TestSyslog(VppTestCase):
         try:
             self.assertEqual(capture[0][UDP].dport, 12345)
         except:
-                self.logger.error(ppp("invalid packet:", capture[0]))
-                raise
+            self.logger.error(ppp("invalid packet:", capture[0]))
+            raise
         self.syslog_verify(capture[0][Raw].load,
                            SyslogFacility.local5,
                            SyslogSeverity.err,