summaryrefslogtreecommitdiffstats
path: root/src/vnet/ip/ip4_options.c
blob: 9b01151a1f0150fb5ef154065182a52b8c716e92 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
/*
 * Copyright (c) 2018 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.
 */

/**
 * @brief Handle IPv4 header options in the data-path
 */

#include <vnet/ip/ip.h>

typedef enum ip4_options_next_t_
{
  IP4_OPTIONS_NEXT_PUNT,
  IP4_OPTIONS_NEXT_LOCAL,
  IP4_OPTIONS_N_NEXT,
} ip4_options_next_t;

typedef struct ip4_options_trace_t_
{
  u8 option[4];
} ip4_options_trace_t;

VLIB_NODE_FN (ip4_options_node) (vlib_main_t * vm,
				 vlib_node_runtime_t * node,
				 vlib_frame_t * frame)
{
  uword n_left_from, n_left_to_next, next_index;
  u32 *from, *to_next;

  from = vlib_frame_vector_args (frame);
  n_left_from = frame->n_vectors;
  next_index = 0;

  while (n_left_from > 0)
    {
      vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);

      /*
       * IP options packets, when properly used, are very low rate,
       * so this code is not dual-looped for extra performance.
       */
      while (n_left_from > 0 && n_left_to_next > 0)
	{
	  ip4_options_next_t next;
	  ip4_header_t *ip4;
	  vlib_buffer_t *b;
	  u8 *options;
	  u32 bi;

	  bi = from[0];
	  from += 1;
	  n_left_from -= 1;
	  to_next[0] = bi;
	  to_next += 1;
	  n_left_to_next -= 1;

	  b = vlib_get_buffer (vm, bi);
	  ip4 = vlib_buffer_get_current (b);
	  next = IP4_OPTIONS_NEXT_PUNT;

	  options = (u8 *) (ip4 + 1);

	  /*
	   * mask out the copy flag to leave the option type
	   */
	  switch (options[0] & 0x7f)
	    {
	    case IP4_ROUTER_ALERT_OPTION:
	      /*
	       * if it's an IGMP packet, pass up the local stack
	       */
	      if (IP_PROTOCOL_IGMP == ip4->protocol)
		{
		  ip_lookup_set_buffer_fib_index (
		    ip4_main.fib_index_by_sw_if_index, b);
		  next = IP4_OPTIONS_NEXT_LOCAL;
		}
	      break;
	    default:
	      break;
	    }

	  if (b->flags & VLIB_BUFFER_IS_TRACED)
	    {
	      ip4_options_trace_t *t =
		vlib_add_trace (vm, node, b, sizeof (*t));

	      clib_memcpy_fast (t->option, options, 4);
	    }
	  vlib_validate_buffer_enqueue_x1 (vm, node, next_index, to_next,
					   n_left_to_next, bi, next);

	}

      vlib_put_next_frame (vm, node, next_index, n_left_to_next);
    }
  return frame->n_vectors;
}

u8 *
format_ip4_options_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 *);
  ip4_options_trace_t *t = va_arg (*args, ip4_options_trace_t *);
  u32 indent = format_get_indent (s);

  s = format (s, "%Uoption:[0x%x,0x%x,0x%x,0x%x]",
	      format_white_space, indent,
	      t->option[0], t->option[1], t->option[2], t->option[3]);
  return s;
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (ip4_options_node) = {
  .name = "ip4-options",
  .vector_size = sizeof (u32),

  .n_next_nodes = IP4_OPTIONS_N_NEXT,
  .next_nodes = {
    [IP4_OPTIONS_NEXT_PUNT] = "ip4-punt",
    [IP4_OPTIONS_NEXT_LOCAL] = "ip4-local",
  },
  .format_buffer = format_ip4_header,
  .format_trace = format_ip4_options_trace,
};
/* *INDENT-ON* */

/*
 * fd.io coding-style-patch-verification: ON
 *
 * Local Variables:
 * eval: (c-set-style "gnu")
 * End:
 */
n> int (*app_rx_evt) (transport_connection_t *tconn); /* * Connection retrieval */ transport_connection_t *(*get_connection) (u32 conn_idx, u32 thread_idx); transport_connection_t *(*get_listener) (u32 conn_index); transport_connection_t *(*get_half_open) (u32 conn_index); /* * Format */ u8 *(*format_connection) (u8 * s, va_list * args); u8 *(*format_listener) (u8 * s, va_list * args); u8 *(*format_half_open) (u8 * s, va_list * args); /* * Properties retrieval */ void (*get_transport_endpoint) (u32 conn_index, u32 thread_index, transport_endpoint_t *tep, u8 is_lcl); void (*get_transport_listener_endpoint) (u32 conn_index, transport_endpoint_t *tep, u8 is_lcl); /* * Properties */ transport_options_t transport_options; } transport_proto_vft_t; /* *INDENT-ON* */ extern transport_proto_vft_t *tp_vfts; #define transport_proto_foreach(VAR, BODY) \ do { \ for (VAR = 0; VAR < vec_len (tp_vfts); VAR++) \ if (tp_vfts[VAR].push_header != 0) \ do { BODY; } while (0); \ } while (0) int transport_connect (transport_proto_t tp, transport_endpoint_cfg_t * tep); void transport_close (transport_proto_t tp, u32 conn_index, u8 thread_index); void transport_reset (transport_proto_t tp, u32 conn_index, u8 thread_index); u32 transport_start_listen (transport_proto_t tp, u32 session_index, transport_endpoint_t * tep); u32 transport_stop_listen (transport_proto_t tp, u32 conn_index); void transport_cleanup (transport_proto_t tp, u32 conn_index, u8 thread_index); void transport_cleanup_half_open (transport_proto_t tp, u32 conn_index); void transport_get_endpoint (transport_proto_t tp, u32 conn_index, u32 thread_index, transport_endpoint_t * tep, u8 is_lcl); void transport_get_listener_endpoint (transport_proto_t tp, u32 conn_index, transport_endpoint_t * tep, u8 is_lcl); static inline transport_connection_t * transport_get_connection (transport_proto_t tp, u32 conn_index, u8 thread_index) { return tp_vfts[tp].get_connection (conn_index, thread_index); } static inline transport_connection_t * transport_get_listener (transport_proto_t tp, u32 conn_index) { return tp_vfts[tp].get_listener (conn_index); } static inline transport_connection_t * transport_get_half_open (transport_proto_t tp, u32 conn_index) { return tp_vfts[tp].get_half_open (conn_index); } static inline int transport_custom_tx (transport_proto_t tp, void *s, transport_send_params_t * sp) { return tp_vfts[tp].custom_tx (s, sp); } static inline int transport_app_rx_evt (transport_proto_t tp, u32 conn_index, u32 thread_index) { transport_connection_t *tc; if (!tp_vfts[tp].app_rx_evt) return 0; tc = transport_get_connection (tp, conn_index, thread_index); return tp_vfts[tp].app_rx_evt (tc); } /** * Get send parameters for transport connection * * These include maximum tx burst, mss, tx offset and other flags * transport might want to provide to sessin layer * * @param tc transport connection * @param sp send paramaters * */ static inline u32 transport_connection_snd_params (transport_connection_t * tc, transport_send_params_t * sp) { return tp_vfts[tc->proto].send_params (tc, sp); } static inline u8 transport_connection_is_descheduled (transport_connection_t * tc) { return ((tc->flags & TRANSPORT_CONNECTION_F_DESCHED) ? 1 : 0); } static inline void transport_connection_deschedule (transport_connection_t * tc) { tc->flags |= TRANSPORT_CONNECTION_F_DESCHED; } static inline u8 transport_connection_is_cless (transport_connection_t * tc) { return ((tc->flags & TRANSPORT_CONNECTION_F_CLESS) ? 1 : 0); } void transport_connection_reschedule (transport_connection_t * tc); /** * Register transport virtual function table. * * @param transport_proto - transport protocol type (i.e., TCP, UDP ..) * @param vft - virtual function table for transport proto * @param fib_proto - network layer protocol * @param output_node - output node index that session layer will hand off * buffers to, for requested fib proto */ void transport_register_protocol (transport_proto_t transport_proto, const transport_proto_vft_t * vft, fib_protocol_t fib_proto, u32 output_node); transport_proto_t transport_register_new_protocol (const transport_proto_vft_t * vft, fib_protocol_t fib_proto, u32 output_node); transport_proto_vft_t *transport_protocol_get_vft (transport_proto_t tp); void transport_update_time (clib_time_type_t time_now, u8 thread_index); int transport_alloc_local_port (u8 proto, ip46_address_t * ip); int transport_alloc_local_endpoint (u8 proto, transport_endpoint_cfg_t * rmt, ip46_address_t * lcl_addr, u16 * lcl_port); void transport_share_local_endpoint (u8 proto, ip46_address_t * lcl_ip, u16 port); void transport_endpoint_cleanup (u8 proto, ip46_address_t * lcl_ip, u16 port); void transport_enable_disable (vlib_main_t * vm, u8 is_en); void transport_init (void); always_inline u32 transport_elog_track_index (transport_connection_t * tc) { #if TRANSPORT_DEBUG return tc->elog_track.track_index_plus_one - 1; #else return ~0; #endif } void transport_connection_tx_pacer_reset (transport_connection_t * tc, u64 rate_bytes_per_sec, u32 initial_bucket, clib_us_time_t rtt); /** * Initialize tx pacer for connection * * @param tc transport connection * @param rate_bytes_per_second initial byte rate * @param burst_bytes initial burst size in bytes */ void transport_connection_tx_pacer_init (transport_connection_t * tc, u64 rate_bytes_per_sec, u32 initial_bucket); /** * Update tx pacer pacing rate * * @param tc transport connection * @param bytes_per_sec new pacing rate * @param rtt connection rtt that is used to compute * inactivity time after which pacer bucket is * reset to 1 mtu */ void transport_connection_tx_pacer_update (transport_connection_t * tc, u64 bytes_per_sec, clib_us_time_t rtt); /** * Get tx pacer max burst * * @param tc transport connection * @param time_now current cpu time * @return max burst for connection */ u32 transport_connection_tx_pacer_burst (transport_connection_t * tc); /** * Get tx pacer current rate * * @param tc transport connection * @return rate for connection in bytes/s */ u64 transport_connection_tx_pacer_rate (transport_connection_t * tc); /** * Reset tx pacer bucket * * @param tc transport connection * @param bucket value the bucket will be reset to */ void transport_connection_tx_pacer_reset_bucket (transport_connection_t * tc, u32 bucket); /** * Check if transport connection is paced */ always_inline u8 transport_connection_is_tx_paced (transport_connection_t * tc) { return (tc->flags & TRANSPORT_CONNECTION_F_IS_TX_PACED); } u8 *format_transport_pacer (u8 * s, va_list * args); /** * Update tx bytes for paced transport connection * * If tx pacing is enabled, this update pacer bucket to account for the * amount of bytes that have been sent. * * @param tc transport connection * @param bytes bytes recently sent */ void transport_connection_update_tx_bytes (transport_connection_t * tc, u32 bytes); void transport_connection_tx_pacer_update_bytes (transport_connection_t * tc, u32 bytes); #endif /* SRC_VNET_SESSION_TRANSPORT_H_ */ /* * fd.io coding-style-patch-verification: ON * * Local Variables: * eval: (c-set-style "gnu") * End: */