Add graph node to the IP lookup feature arc. See attached code, from a
known-to-work example. Program src routes into a non-default FIB.
HTH... Dave
From: [email protected] <[email protected]> On Behalf Of
[email protected]
Sent: Monday, August 27, 2018 1:28 AM
To: [email protected]
Subject: [vpp-dev] Source Based Routing #vpp
Hi VPP experts,
I would like to shed light on a particular scenario/usecase that I'm trying to
implement in VPP.
The scenario demands egress packet routing to be done based on the source IP of
the packet rather than the destination IP. I found a similar discussion in the
VPP mail
archive(https://www.mail-archive.com/[email protected]/msg06886.html), but
the solution discussed there would not be applicable for routing multiple
source IP packets having different routes.
I understand that we can configure multiple routing table in VPP with unique
routes in them, but would I be able to make the FIB lookup towards a particular
table based on source IP?
If I take an analogy from Linux kernel, does VPP support functionality similar
to 'ip rule' command, wherein we can specify a routing table to be used for a
particular source IP.
Looking forward to your advice and support in finding a solution to this.
Thanks & Regards,
Georgi
/* *INDENT-OFF* */
VNET_FEATURE_INIT (ip4_sdp_slookup, static) =
{
.arc_name = "ip4-unicast",
.node_name = "ip4-sdp-slookup",
.runs_before = VNET_FEATURES ("ip4-lookup"),
};
/* *INDENT-ON */
/* *INDENT-OFF* */
VNET_FEATURE_INIT (ip6_sdp_slookup, static) =
{
.arc_name = "ip6-unicast",
.node_name = "ip6-sdp-slookup",
.runs_before = VNET_FEATURES ("ip6-lookup"),
};
/* *INDENT-ON */
int
sdp_src_lookup_enable_disable (sdp_main_t * sm,
u32 sw_if_index,
u8 enable_ip4,
u8 enable_ip6, u32 ip4_vrf_id, u32 ip6_vrf_id)
{
vnet_sw_interface_t *sw;
int rv = 0;
u32 ip4_fib_index = 0, ip6_fib_index = 0;
/* Not a physical port? */
sw = vnet_get_sw_interface (sm->vnet_main, sw_if_index);
if (sw->type != VNET_SW_INTERFACE_TYPE_HARDWARE)
return VNET_API_ERROR_INVALID_SW_IF_INDEX;
if (enable_ip4)
{
ip4_fib_index = fib_table_find (FIB_PROTOCOL_IP4, ip4_vrf_id);
if (ip4_fib_index == ~0)
return VNET_API_ERROR_NO_SUCH_FIB;
vec_validate (sm->ip4_fib_index_by_sw_if_index, sw_if_index);
sm->ip4_fib_index_by_sw_if_index[sw_if_index] = ip4_fib_index;
}
vnet_feature_enable_disable ("ip4-unicast", "ip4-sdp-slookup",
sw_if_index, enable_ip4, 0, 0);
if (enable_ip6)
{
ip6_fib_index = fib_table_find (FIB_PROTOCOL_IP6, ip4_vrf_id);
if (ip6_fib_index == ~0)
return VNET_API_ERROR_NO_SUCH_INNER_FIB;
vec_validate (sm->ip6_fib_index_by_sw_if_index, sw_if_index);
sm->ip6_fib_index_by_sw_if_index[sw_if_index] = ip6_fib_index;
}
vnet_feature_enable_disable ("ip6-unicast", "ip6-sdp-slookup",
sw_if_index, enable_ip6, 0, 0);
return rv;
}
/*
* node.c - sdp plugin dst lookup, policy enforcement
*
* Copyright (c) <current-year> <your-organization>
* 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 <vppinfra/error.h>
#include <sdp/sdp.h>
#include <vnet/fib/ip4_fib.h>
#include <vnet/fib/ip6_fib.h>
typedef struct
{
u32 next_index;
u32 policy;
} sdp_slookup_trace_t;
/* packet trace format function */
#ifndef CLIB_MARCH_VARIANT
static u8 *
format_sdp_slookup_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 *);
sdp_slookup_trace_t *t = va_arg (*args, sdp_slookup_trace_t *);
s = format (s, "SDP src lookup: next index %d policy 0x%08x (%d)",
t->next_index, t->policy, t->policy);
return s;
}
#endif
vlib_node_registration_t sdp_slookup_node;
#define foreach_sdp_slookup_error \
_(SWAPPED, "Mac swap packets processed")
typedef enum
{
#define _(sym,str) SDP_SLOOKUP_ERROR_##sym,
foreach_sdp_slookup_error
#undef _
SDP_SLOOKUP_N_ERROR,
} sdp_slookup_error_t;
#ifndef CLIB_MARCH_VARIANT
static char *sdp_slookup_error_strings[] = {
#define _(sym,string) string,
foreach_sdp_slookup_error
#undef _
};
#endif
/**
* Add the drop arc for packets that don't pass muster.
* All other arcs are automatically added when the SDP DPO is stacked onto
* the various parent DPO types
*/
typedef enum
{
SDP_SLOOKUP_NEXT_DROP,
SDP_SLOOKUP_N_NEXT,
} sdp_slookup_next_t;
always_inline uword
ip46_sdp_slookup_inline (vlib_main_t * vm,
vlib_node_runtime_t * node, vlib_frame_t * frame,
int is_ip4, int is_traced)
{
u32 n_left_from, *from;
vlib_buffer_t *bufs[VLIB_FRAME_SIZE], **b;
u16 nexts[VLIB_FRAME_SIZE], *next;
sdp_main_t *sm = &sdp_main;
ip4_main_t *im4 = &ip4_main;
ip_lookup_main_t *lm4 = &im4->lookup_main;
u8 arc4 = lm4->ucast_feature_arc_index;
vnet_feature_config_main_t *fcm4;
ASSERT (is_ip4);
fcm4 = vnet_feature_get_config_main (arc4);
from = vlib_frame_vector_args (frame);
n_left_from = frame->n_vectors;
vlib_get_buffers (vm, from, bufs, n_left_from);
b = bufs;
next = nexts;
while (n_left_from >= 4)
{
sdp_dpo_t *sdp0, *sdp1, *sdp2, *sdp3;
ip4_header_t *ip0, *ip1, *ip2, *ip3;
u32 next0, next1, next2, next3;
const load_balance_t *lb0, *lb1, *lb2, *lb3;
ip4_fib_mtrie_t *mtrie0, *mtrie1, *mtrie2, *mtrie3;
ip4_fib_mtrie_leaf_t leaf0, leaf1, leaf2, leaf3;
ip4_address_t *src_addr0, *src_addr1, *src_addr2, *src_addr3;
u32 lbi0, lbi1, lbi2, lbi3;
u32 fib_index0, fib_index1, fib_index2, fib_index3;
flow_hash_config_t flow_hash_config0, flow_hash_config1,
flow_hash_config2, flow_hash_config3;
const dpo_id_t *dpo0, *dpo1, *dpo2, *dpo3;
u32 hash_c0, hash_c1, hash_c2, hash_c3;
if (PREDICT_TRUE (n_left_from >= 8))
{
vlib_prefetch_buffer_header (b[4], STORE);
vlib_prefetch_buffer_header (b[5], STORE);
vlib_prefetch_buffer_header (b[6], STORE);
vlib_prefetch_buffer_header (b[7], STORE);
CLIB_PREFETCH (b[4]->data, sizeof (ip0[0]), LOAD);
CLIB_PREFETCH (b[5]->data, sizeof (ip0[0]), LOAD);
CLIB_PREFETCH (b[6]->data, sizeof (ip0[0]), LOAD);
CLIB_PREFETCH (b[7]->data, sizeof (ip0[0]), LOAD);
}
/* Send pkt to next feature */
vnet_get_config_data (&fcm4->config_main,
&b[0]->current_config_index, &next0,
/* # bytes of config data */ 0);
vnet_get_config_data (&fcm4->config_main,
&b[1]->current_config_index, &next1,
/* # bytes of config data */ 0);
vnet_get_config_data (&fcm4->config_main,
&b[2]->current_config_index, &next2,
/* # bytes of config data */ 0);
vnet_get_config_data (&fcm4->config_main,
&b[3]->current_config_index, &next3,
/* # bytes of config data */ 0);
/* This sucks slightly */
next[0] = next0;
next[1] = next1;
next[2] = next2;
next[3] = next3;
ip0 = vlib_buffer_get_current (b[0]);
ip1 = vlib_buffer_get_current (b[1]);
ip2 = vlib_buffer_get_current (b[2]);
ip3 = vlib_buffer_get_current (b[3]);
/* Do a src address lookup */
src_addr0 = &ip0->src_address;
src_addr1 = &ip1->src_address;
src_addr2 = &ip2->src_address;
src_addr3 = &ip3->src_address;
/*
* Src lookup in the configured fib, almost certainly NOT
* the normal forwarding fib for the interface
*/
fib_index0 = sm->ip4_fib_index_by_sw_if_index
[vnet_buffer (b[0])->sw_if_index[VLIB_RX]];
fib_index1 = sm->ip4_fib_index_by_sw_if_index
[vnet_buffer (b[1])->sw_if_index[VLIB_RX]];
fib_index2 = sm->ip4_fib_index_by_sw_if_index
[vnet_buffer (b[2])->sw_if_index[VLIB_RX]];
fib_index3 = sm->ip4_fib_index_by_sw_if_index
[vnet_buffer (b[3])->sw_if_index[VLIB_RX]];
mtrie0 = &ip4_fib_get (fib_index0)->mtrie;
leaf0 = ip4_fib_mtrie_lookup_step_one (mtrie0, src_addr0);
leaf0 = ip4_fib_mtrie_lookup_step (mtrie0, leaf0, src_addr0, 2);
leaf0 = ip4_fib_mtrie_lookup_step (mtrie0, leaf0, src_addr0, 3);
lbi0 = ip4_fib_mtrie_leaf_get_adj_index (leaf0);
mtrie1 = &ip4_fib_get (fib_index1)->mtrie;
leaf1 = ip4_fib_mtrie_lookup_step_one (mtrie1, src_addr1);
leaf1 = ip4_fib_mtrie_lookup_step (mtrie1, leaf1, src_addr1, 2);
leaf1 = ip4_fib_mtrie_lookup_step (mtrie1, leaf1, src_addr1, 3);
lbi1 = ip4_fib_mtrie_leaf_get_adj_index (leaf1);
mtrie2 = &ip4_fib_get (fib_index2)->mtrie;
leaf2 = ip4_fib_mtrie_lookup_step_one (mtrie2, src_addr2);
leaf2 = ip4_fib_mtrie_lookup_step (mtrie2, leaf2, src_addr2, 2);
leaf2 = ip4_fib_mtrie_lookup_step (mtrie2, leaf2, src_addr2, 3);
lbi2 = ip4_fib_mtrie_leaf_get_adj_index (leaf2);
mtrie3 = &ip4_fib_get (fib_index3)->mtrie;
leaf3 = ip4_fib_mtrie_lookup_step_one (mtrie3, src_addr3);
leaf3 = ip4_fib_mtrie_lookup_step (mtrie3, leaf3, src_addr3, 2);
leaf3 = ip4_fib_mtrie_lookup_step (mtrie3, leaf3, src_addr3, 3);
lbi3 = ip4_fib_mtrie_leaf_get_adj_index (leaf3);
lb0 = load_balance_get (lbi0);
lb1 = load_balance_get (lbi1);
lb2 = load_balance_get (lbi2);
lb3 = load_balance_get (lbi3);
/* Use flow hash to compute multipath adjacency. */
hash_c0 = vnet_buffer (b[0])->ip.flow_hash = 0;
if (PREDICT_FALSE (lb0->lb_n_buckets > 1))
{
flow_hash_config0 = lb0->lb_hash_config;
hash_c0 = vnet_buffer (b[0])->ip.flow_hash =
ip4_compute_flow_hash (ip0, flow_hash_config0);
dpo0 =
load_balance_get_fwd_bucket (lb0,
(hash_c0 &
(lb0->lb_n_buckets_minus_1)));
}
else
dpo0 = load_balance_get_bucket_i (lb0, 0);
hash_c1 = vnet_buffer (b[1])->ip.flow_hash = 0;
if (PREDICT_FALSE (lb1->lb_n_buckets > 1))
{
flow_hash_config1 = lb1->lb_hash_config;
hash_c1 = vnet_buffer (b[1])->ip.flow_hash =
ip4_compute_flow_hash (ip1, flow_hash_config1);
dpo1 =
load_balance_get_fwd_bucket (lb1,
(hash_c1 &
(lb1->lb_n_buckets_minus_1)));
}
else
dpo1 = load_balance_get_bucket_i (lb1, 0);
hash_c2 = vnet_buffer (b[2])->ip.flow_hash = 2;
if (PREDICT_FALSE (lb2->lb_n_buckets > 1))
{
flow_hash_config2 = lb2->lb_hash_config;
hash_c2 = vnet_buffer (b[2])->ip.flow_hash =
ip4_compute_flow_hash (ip2, flow_hash_config2);
dpo2 =
load_balance_get_fwd_bucket (lb2,
(hash_c2 &
(lb2->lb_n_buckets_minus_1)));
}
else
dpo2 = load_balance_get_bucket_i (lb2, 0);
hash_c3 = vnet_buffer (b[3])->ip.flow_hash = 0;
if (PREDICT_FALSE (lb3->lb_n_buckets > 1))
{
flow_hash_config3 = lb3->lb_hash_config;
hash_c3 = vnet_buffer (b[3])->ip.flow_hash =
ip4_compute_flow_hash (ip3, flow_hash_config3);
dpo3 =
load_balance_get_fwd_bucket (lb3,
(hash_c3 &
(lb3->lb_n_buckets_minus_1)));
}
else
dpo3 = load_balance_get_bucket_i (lb3, 0);
/*
* If the lookup returns something other than our
* interposed dpo, supply default src policy bits
* $$$ configurable default policy.
*/
if (PREDICT_FALSE (dpo0->dpoi_type != sdp_dpo_type))
vnet_buffer (b[0])->ip.sdp.src_policy = 0xFFFFFFFF;
else
{
/* Save the src policy in buffer metadata */
sdp0 = sdp_dpo_get (dpo0->dpoi_index);
vnet_buffer (b[0])->ip.sdp.src_policy = sdp0->sdp_policy;
}
if (PREDICT_FALSE (dpo1->dpoi_type != sdp_dpo_type))
vnet_buffer (b[1])->ip.sdp.src_policy = 0xFFFFFFFF;
else
{
/* Save the src policy in buffer metadata */
sdp1 = sdp_dpo_get (dpo1->dpoi_index);
vnet_buffer (b[1])->ip.sdp.src_policy = sdp1->sdp_policy;
}
if (PREDICT_FALSE (dpo2->dpoi_type != sdp_dpo_type))
vnet_buffer (b[2])->ip.sdp.src_policy = 0xFFFFFFFF;
else
{
/* Save the src policy in buffer metadata */
sdp2 = sdp_dpo_get (dpo2->dpoi_index);
vnet_buffer (b[2])->ip.sdp.src_policy = sdp2->sdp_policy;
}
if (PREDICT_FALSE (dpo3->dpoi_type != sdp_dpo_type))
vnet_buffer (b[3])->ip.sdp.src_policy = 0xFFFFFFFF;
else
{
/* Save the src policy in buffer metadata */
sdp3 = sdp_dpo_get (dpo3->dpoi_index);
vnet_buffer (b[3])->ip.sdp.src_policy = sdp3->sdp_policy;
}
if (is_traced)
{
if (b[0]->flags & VLIB_BUFFER_IS_TRACED)
{
sdp_slookup_trace_t *t =
vlib_add_trace (vm, node, b[0], sizeof (*t));
t->next_index = next0;
t->policy = vnet_buffer (b[0])->ip.sdp.src_policy;
}
if (b[1]->flags & VLIB_BUFFER_IS_TRACED)
{
sdp_slookup_trace_t *t =
vlib_add_trace (vm, node, b[1], sizeof (*t));
t->next_index = next1;
t->policy = vnet_buffer (b[1])->ip.sdp.src_policy;
}
if (b[2]->flags & VLIB_BUFFER_IS_TRACED)
{
sdp_slookup_trace_t *t =
vlib_add_trace (vm, node, b[2], sizeof (*t));
t->next_index = next2;
t->policy = vnet_buffer (b[2])->ip.sdp.src_policy;
}
if (b[3]->flags & VLIB_BUFFER_IS_TRACED)
{
sdp_slookup_trace_t *t =
vlib_add_trace (vm, node, b[3], sizeof (*t));
t->next_index = next3;
t->policy = vnet_buffer (b[3])->ip.sdp.src_policy;
}
}
b += 4;
next += 4;
n_left_from -= 4;
}
while (n_left_from > 0)
{
sdp_dpo_t *sdp0;
ip4_header_t *ip0;
u32 next0;
const load_balance_t *lb0;
ip4_fib_mtrie_t *mtrie0;
ip4_fib_mtrie_leaf_t leaf0;
ip4_address_t *src_addr0;
u32 lbi0;
u32 fib_index0;
flow_hash_config_t flow_hash_config0;
const dpo_id_t *dpo0;
u32 hash_c0;
/* Send pkt to next feature */
vnet_get_config_data (&fcm4->config_main,
&b[0]->current_config_index, &next0,
/* # bytes of config data */ 0);
next[0] = next0;
ip0 = vlib_buffer_get_current (b[0]);
/* Do a src address lookup */
src_addr0 = &ip0->src_address;
/*
* Src lookup in the configured fib, almost certainly NOT
* the normal forwarding fib for the interface
*/
fib_index0 = sm->ip4_fib_index_by_sw_if_index
[vnet_buffer (b[0])->sw_if_index[VLIB_RX]];
mtrie0 = &ip4_fib_get (fib_index0)->mtrie;
leaf0 = ip4_fib_mtrie_lookup_step_one (mtrie0, src_addr0);
leaf0 = ip4_fib_mtrie_lookup_step (mtrie0, leaf0, src_addr0, 2);
leaf0 = ip4_fib_mtrie_lookup_step (mtrie0, leaf0, src_addr0, 3);
lbi0 = ip4_fib_mtrie_leaf_get_adj_index (leaf0);
ASSERT (lbi0);
lb0 = load_balance_get (lbi0);
ASSERT (lb0->lb_n_buckets > 0);
ASSERT (is_pow2 (lb0->lb_n_buckets));
/* Use flow hash to compute multipath adjacency. */
hash_c0 = vnet_buffer (b[0])->ip.flow_hash = 0;
if (PREDICT_FALSE (lb0->lb_n_buckets > 1))
{
flow_hash_config0 = lb0->lb_hash_config;
hash_c0 = vnet_buffer (b[0])->ip.flow_hash =
ip4_compute_flow_hash (ip0, flow_hash_config0);
dpo0 =
load_balance_get_fwd_bucket (lb0,
(hash_c0 &
(lb0->lb_n_buckets_minus_1)));
}
else
{
dpo0 = load_balance_get_bucket_i (lb0, 0);
}
/*
* If the lookup returns something other than our
* interposed dpo, supply default src policy bits
* $$$ configurable default policy.
*/
if (PREDICT_FALSE (dpo0->dpoi_type != sdp_dpo_type))
vnet_buffer (b[0])->ip.sdp.src_policy = 0xFFFFFFFF;
else
{
/* Save the src policy in buffer metadata */
sdp0 = sdp_dpo_get (dpo0->dpoi_index);
vnet_buffer (b[0])->ip.sdp.src_policy = sdp0->sdp_policy;
}
if (is_traced)
{
if (b[0]->flags & VLIB_BUFFER_IS_TRACED)
{
sdp_slookup_trace_t *t =
vlib_add_trace (vm, node, b[0], sizeof (*t));
t->next_index = next0;
t->policy = vnet_buffer (b[0])->ip.sdp.src_policy;
}
}
b += 1;
next += 1;
n_left_from -= 1;
}
vlib_buffer_enqueue_to_next (vm, node, from, nexts, frame->n_vectors);
return frame->n_vectors;
}
VLIB_NODE_FN (ip4_sdp_slookup_node) (vlib_main_t * vm, vlib_node_runtime_t *
node,
vlib_frame_t * frame)
{
if (PREDICT_FALSE (node->flags & VLIB_NODE_FLAG_TRACE))
return ip46_sdp_slookup_inline (vm, node, frame, 1 /* is_ip4 */ ,
1 /* is_traced */ );
else
return ip46_sdp_slookup_inline (vm, node, frame, 1 /* is_ip4 */ ,
0 /* is_traced */ );
}
/* *INDENT-OFF* */
#ifndef CLIB_MARCH_VARIANT
VLIB_REGISTER_NODE (ip4_sdp_slookup_node) =
{
.name = "ip4-sdp-slookup",
.vector_size = sizeof (u32),
.format_trace = format_sdp_slookup_trace,
.type = VLIB_NODE_TYPE_INTERNAL,
.n_errors = ARRAY_LEN(sdp_slookup_error_strings),
.error_strings = sdp_slookup_error_strings,
.n_next_nodes = SDP_SLOOKUP_N_NEXT,
/* edit / add dispositions here */
.next_nodes =
{
[SDP_SLOOKUP_NEXT_DROP] = "ip4-drop",
},
};
#endif
/* *INDENT-ON* */
VLIB_NODE_FN(ip6_sdp_slookup_node) (vlib_main_t * vm, vlib_node_runtime_t *
node,
vlib_frame_t * frame)
{
if (PREDICT_FALSE (node->flags & VLIB_NODE_FLAG_TRACE))
return ip46_sdp_slookup_inline (vm, node, frame, 0 /* is_ip4 */ ,
1 /* is_traced */ );
else
return ip46_sdp_slookup_inline (vm, node, frame, 0 /* is_ip4 */ ,
0 /* is_traced */ );
}
/* *INDENT-OFF* */
#ifndef CLIB_MARCH_VARIANT
VLIB_REGISTER_NODE (ip6_sdp_slookup_node) =
{
.name = "ip6-sdp-slookup",
.vector_size = sizeof (u32),
.format_trace = format_sdp_slookup_trace,
.type = VLIB_NODE_TYPE_INTERNAL,
.n_errors = ARRAY_LEN(sdp_slookup_error_strings),
.error_strings = sdp_slookup_error_strings,
.n_next_nodes = SDP_SLOOKUP_N_NEXT,
/* edit / add dispositions here */
.next_nodes =
{
[SDP_SLOOKUP_NEXT_DROP] = "ip6-drop",
},
};
#endif
/* *INDENT-ON* */
/*
* fd.io coding-style-patch-verification: ON
*
* Local Variables:
* eval: (c-set-style "gnu")
* End:
*/
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#10296): https://lists.fd.io/g/vpp-dev/message/10296
Mute This Topic: https://lists.fd.io/mt/24970841/21656
Mute #vpp: https://lists.fd.io/mk?hashtag=vpp&subid=1480452
Group Owner: [email protected]
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-