shaba/systemd-stable
1
1
Fork 0
mirror of https://github.com/systemd/systemd-stable.git synced 2025-01-03 01:17:45 +03:00
Code

Merge pull request #19594 from yuwata/network-nexthop-group

Browse Source 
network: nexthop: add Group= setting to configure multipath route with group nexthop
This commit is contained in:
Yu Watanabe 2021-05-22 16:01:13 +09:00 committed by GitHub
commit af92e46527
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
13 changed files with 552 additions and 182 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
man/systemd.network.xml
View File

@ -1386,6 +1386,19 @@ IPv6Token=prefixstable:2002:da8:1::</programlisting></para>
<literal>no</literal>.</para>
</listitem>
</varlistentry>
<varlistentry>
<term><varname>Group=</varname></term>
<listitem>
<para>Takes a whitespace separated list of nexthop IDs. Each ID must be in the range
1…4294967295. Optionally, each nexthop ID can take a weight after a colon
(<literal><replaceable>id</replaceable><optional>:<replaceable>weight</replaceable></optional></literal>).
The weight must be in the range 1…255. If the weight is not specified, then it is assumed
that the weight is 1. This setting cannot be specified with <varname>Gateway=</varname>,
<varname>Family=</varname>, <varname>Blackhole=</varname>. This setting can be specified
multiple times. If an empty string is assigned, then the all previous assignments are
cleared. Defaults to unset.</para>
</listitem>
</varlistentry>
</variablelist>
</refsect1>

2
src/libsystemd/sd-netlink/netlink-util.c
View File

@ -472,7 +472,7 @@ int rtattr_read_nexthop(const struct rtnexthop *rtnh, size_t size, int family, O
*m = (MultipathRoute) {
.ifindex = rtnh->rtnh_ifindex,
.weight = rtnh->rtnh_hops == 0 ? 0 : rtnh->rtnh_hops + 1,
.weight = rtnh->rtnh_hops,
};
if (rtnh->rtnh_len > sizeof(struct rtnexthop)) {

14
src/network/networkd-address.c
View File

@ -369,7 +369,6 @@ static int address_add_foreign(Link *link, const Address *in, Address **ret) {
}
static int address_add(Link *link, const Address *in, Address **ret) {
bool is_new = false;
Address *address;
int r;
@ -382,7 +381,6 @@ static int address_add(Link *link, const Address *in, Address **ret) {
r = address_add_internal(link, &link->addresses, in, &address);
if (r < 0)
return r;
is_new = true;
} else if (r == 0) {
/* Take over a foreign address */
r = set_ensure_put(&link->addresses, &address_hash_ops, address);
@ -398,7 +396,7 @@ static int address_add(Link *link, const Address *in, Address **ret) {
if (ret)
*ret = address;
return is_new;
return 0;
}
static int address_update(Address *address, const Address *src) {
@ -939,7 +937,7 @@ static int address_configure(
_cleanup_(sd_netlink_message_unrefp) sd_netlink_message *req = NULL;
Address *acquired_address, *a;
bool update;
int r, k;
int r;
assert(address);
assert(IN_SET(address->family, AF_INET, AF_INET6));
@ -1006,9 +1004,9 @@ static int address_configure(
if (r < 0)
return log_link_error_errno(link, r, "Could not append IFA_RT_PRIORITY attribute: %m");
k = address_add(link, address, &a);
if (k < 0)
return log_link_error_errno(link, k, "Could not add address: %m");
r = address_add(link, address, &a);
if (r < 0)
return log_link_error_errno(link, r, "Could not add address: %m");
r = address_set_masquerade(a, true);
if (r < 0)
@ -1031,7 +1029,7 @@ static int address_configure(
if (ret)
*ret = a;
return k;
return 0;
}
static int static_address_handler(sd_netlink *rtnl, sd_netlink_message *m, Link *link) {

4
src/network/networkd-neighbor.c
View File

@ -171,7 +171,6 @@ static int neighbor_add_internal(Link *link, Set **neighbors, const Neighbor *in
}
static int neighbor_add(Link *link, const Neighbor *in, Neighbor **ret) {
bool is_new = false;
Neighbor *neighbor;
int r;
@ -181,7 +180,6 @@ static int neighbor_add(Link *link, const Neighbor *in, Neighbor **ret) {
r = neighbor_add_internal(link, &link->neighbors, in, &neighbor);
if (r < 0)
return r;
is_new = true;
} else if (r == 0) {
/* Neighbor is foreign, claim it as recognized */
r = set_ensure_put(&link->neighbors, &neighbor_hash_ops, neighbor);
@ -197,7 +195,7 @@ static int neighbor_add(Link *link, const Neighbor *in, Neighbor **ret) {
if (ret)
*ret = neighbor;
return is_new;
return 0;
}
static int neighbor_add_foreign(Link *link, const Neighbor *in, Neighbor **ret) {

1
src/network/networkd-network-gperf.gperf
View File

@ -194,6 +194,7 @@ NextHop.Gateway, config_parse_nexthop_gateway,
NextHop.Family, config_parse_nexthop_family, 0, 0
NextHop.OnLink, config_parse_nexthop_onlink, 0, 0
NextHop.Blackhole, config_parse_nexthop_blackhole, 0, 0
NextHop.Group, config_parse_nexthop_group, 0, 0
DHCPv4.ClientIdentifier, config_parse_dhcp_client_identifier, 0, offsetof(Network, dhcp_client_identifier)
DHCPv4.UseDNS, config_parse_dhcp_use_dns, 0, 0
DHCPv4.RoutesToDNS, config_parse_bool, 0, offsetof(Network, dhcp_routes_to_dns)

318
src/network/networkd-nexthop.c
View File

@ -2,6 +2,7 @@
* Copyright © 2019 VMware, Inc.
*/
#include <net/if.h>
#include <linux/nexthop.h>
#include "alloc-util.h"
@ -43,6 +44,8 @@ NextHop *nexthop_free(NextHop *nexthop) {
hashmap_remove(nexthop->manager->nexthops_by_id, UINT32_TO_PTR(nexthop->id));
}
hashmap_free_free(nexthop->group);
return mfree(nexthop);
}
@ -163,17 +166,41 @@ static bool nexthop_equal(const NextHop *a, const NextHop *b) {
return nexthop_compare_func(a, b) == 0;
}
static void nexthop_copy(NextHop *dest, const NextHop *src) {
assert(dest);
static int nexthop_dup(const NextHop *src, NextHop **ret) {
_cleanup_(nexthop_freep) NextHop *dest = NULL;
struct nexthop_grp *nhg;
int r;
assert(src);
assert(ret);
/* This only copies entries used in the above hash and compare functions. */
dest = newdup(NextHop, src, 1);
if (!dest)
return -ENOMEM;
dest->protocol = src->protocol;
dest->id = src->id;
dest->blackhole = src->blackhole;
dest->family = src->family;
dest->gw = src->gw;
/* unset all pointers */
dest->manager = NULL;
dest->link = NULL;
dest->network = NULL;
dest->section = NULL;
dest->group = NULL;
HASHMAP_FOREACH(nhg, src->group) {
_cleanup_free_ struct nexthop_grp *g = NULL;
g = newdup(struct nexthop_grp, nhg, 1);
if (!g)
return -ENOMEM;
r = hashmap_ensure_put(&dest->group, NULL, UINT32_TO_PTR(g->id), g);
if (r < 0)
return r;
if (r > 0)
TAKE_PTR(g);
}
*ret = TAKE_PTR(dest);
return 0;
}
int manager_get_nexthop_by_id(Manager *manager, uint32_t id, NextHop **ret) {
@ -224,12 +251,10 @@ static int nexthop_add_internal(Manager *manager, Link *link, Set **nexthops, co
assert(nexthops);
assert(in);
r = nexthop_new(&nexthop);
r = nexthop_dup(in, &nexthop);
if (r < 0)
return r;
nexthop_copy(nexthop, in);
r = set_ensure_put(nexthops, &nexthop_hash_ops, nexthop);
if (r < 0)
return r;
@ -251,35 +276,40 @@ static int nexthop_add_foreign(Manager *manager, Link *link, const NextHop *in,
return nexthop_add_internal(manager, link, link ? &link->nexthops_foreign : &manager->nexthops_foreign, in, ret);
}
static bool nexthop_has_link(const NextHop *nexthop) {
return !nexthop->blackhole && hashmap_isempty(nexthop->group);
}
static int nexthop_add(Link *link, const NextHop *in, NextHop **ret) {
bool is_new = false;
bool by_manager;
NextHop *nexthop;
int r;
assert(link);
assert(in);
if (in->blackhole)
by_manager = !nexthop_has_link(in);
if (by_manager)
r = nexthop_get(link->manager, NULL, in, &nexthop);
else
r = nexthop_get(NULL, link, in, &nexthop);
if (r == -ENOENT) {
/* NextHop does not exist, create a new one */
r = nexthop_add_internal(link->manager,
in->blackhole ? NULL : link,
in->blackhole ? &link->manager->nexthops : &link->nexthops,
by_manager ? NULL : link,
by_manager ? &link->manager->nexthops : &link->nexthops,
in, &nexthop);
if (r < 0)
return r;
is_new = true;
} else if (r == 0) {
/* Take over a foreign nexthop */
r = set_ensure_put(in->blackhole ? &link->manager->nexthops : &link->nexthops,
r = set_ensure_put(by_manager ? &link->manager->nexthops : &link->nexthops,
&nexthop_hash_ops, nexthop);
if (r < 0)
return r;
set_remove(in->blackhole ? link->manager->nexthops_foreign : link->nexthops_foreign, nexthop);
set_remove(by_manager ? link->manager->nexthops_foreign : link->nexthops_foreign, nexthop);
} else if (r == 1) {
/* NextHop exists, do nothing */
;
@ -288,7 +318,7 @@ static int nexthop_add(Link *link, const NextHop *in, NextHop **ret) {
if (ret)
*ret = nexthop;
return is_new;
return 0;
}
static int nexthop_update(Manager *manager, Link *link, NextHop *nexthop, const NextHop *in) {
@ -338,7 +368,8 @@ set_manager:
}
static void log_nexthop_debug(const NextHop *nexthop, uint32_t id, const char *str, const Link *link) {
_cleanup_free_ char *gw = NULL;
_cleanup_free_ char *gw = NULL, *new_id = NULL, *group = NULL;
struct nexthop_grp *nhg;
assert(nexthop);
assert(str);
@ -348,14 +379,16 @@ static void log_nexthop_debug(const NextHop *nexthop, uint32_t id, const char *s
if (!DEBUG_LOGGING)
return;
if (nexthop->id != id)
(void) asprintf(&new_id, "→%"PRIu32, id);
(void) in_addr_to_string(nexthop->family, &nexthop->gw, &gw);
if (nexthop->id == id)
log_link_debug(link, "%s nexthop: id: %"PRIu32", gw: %s, blackhole: %s",
str, nexthop->id, strna(gw), yes_no(nexthop->blackhole));
else
log_link_debug(link, "%s nexthop: id: %"PRIu32"→%"PRIu32", gw: %s, blackhole: %s",
str, nexthop->id, id, strna(gw), yes_no(nexthop->blackhole));
HASHMAP_FOREACH(nhg, nexthop->group)
(void) strextendf_with_separator(&group, ",", "%"PRIu32":%"PRIu32, nhg->id, nhg->weight+1);
log_link_debug(link, "%s nexthop: id: %"PRIu32"%s, gw: %s, blackhole: %s, group: %s",
str, nexthop->id, strempty(new_id), strna(gw), yes_no(nexthop->blackhole), strna(group));
}
static int link_nexthop_remove_handler(sd_netlink *rtnl, sd_netlink_message *m, Link *link) {
@ -443,13 +476,13 @@ static int nexthop_configure(
NextHop **ret) {
_cleanup_(sd_netlink_message_unrefp) sd_netlink_message *req = NULL;
int r, k;
int r;
assert(link);
assert(link->manager);
assert(link->manager->rtnl);
assert(link->ifindex > 0);
assert(IN_SET(nexthop->family, AF_INET, AF_INET6));
assert(IN_SET(nexthop->family, AF_UNSPEC, AF_INET, AF_INET6));
assert(callback);
log_nexthop_debug(nexthop, nexthop->id, "Configuring", link);
@ -466,7 +499,23 @@ static int nexthop_configure(
return log_link_error_errno(link, r, "Could not append NHA_ID attribute: %m");
}
if (nexthop->blackhole) {
if (!hashmap_isempty(nexthop->group)) {
_cleanup_free_ struct nexthop_grp *group = NULL;
struct nexthop_grp *p, *nhg;
group = new(struct nexthop_grp, hashmap_size(nexthop->group));
if (!group)
return log_oom();
p = group;
HASHMAP_FOREACH(nhg, nexthop->group)
*p++ = *nhg;
r = sd_netlink_message_append_data(req, NHA_GROUP, group, sizeof(struct nexthop_grp) * hashmap_size(nexthop->group));
if (r < 0)
return log_link_error_errno(link, r, "Could not append NHA_GROUP attribute: %m");
} else if (nexthop->blackhole) {
r = sd_netlink_message_append_flag(req, NHA_BLACKHOLE);
if (r < 0)
return log_link_error_errno(link, r, "Could not append NHA_BLACKHOLE attribute: %m");
@ -488,9 +537,9 @@ static int nexthop_configure(
}
}
k = nexthop_add(link, nexthop, ret);
if (k < 0)
return log_link_error_errno(link, k, "Could not add nexthop: %m");
r = nexthop_add(link, nexthop, ret);
if (r < 0)
return log_link_error_errno(link, r, "Could not add nexthop: %m");
r = netlink_call_async(link->manager->rtnl, NULL, req, callback,
link_netlink_destroy_callback, link);
@ -499,7 +548,7 @@ static int nexthop_configure(
link_ref(link);
return k;
return r;
}
static int static_nexthop_handler(sd_netlink *rtnl, sd_netlink_message *m, Link *link) {
@ -694,15 +743,26 @@ int link_drop_nexthops(Link *link) {
}
static bool nexthop_is_ready_to_configure(Link *link, const NextHop *nexthop) {
struct nexthop_grp *nhg;
assert(link);
assert(nexthop);
if (nexthop->blackhole) {
if (!link_is_ready_to_configure(link, false))
return false;
if (!nexthop_has_link(nexthop)) {
if (link->manager->nexthop_remove_messages > 0)
return false;
} else {
Link *l;
/* TODO: fdb nexthop does not require IFF_UP. The condition below needs to be updated
* when fdb nexthop support is added. See rtm_to_nh_config() in net/ipv4/nexthop.c of
* kernel. */
if (!FLAGS_SET(link->flags, IFF_UP))
return false;
HASHMAP_FOREACH(l, link->manager->links) {
if (l->address_remove_messages > 0)
return false;
@ -713,6 +773,11 @@ static bool nexthop_is_ready_to_configure(Link *link, const NextHop *nexthop) {
}
}
/* All group members must be configured first. */
HASHMAP_FOREACH(nhg, nexthop->group)
if (manager_get_nexthop_by_id(link->manager, nhg->id, NULL) < 0)
return false;
if (nexthop->id == 0) {
Request *req;
@ -741,9 +806,6 @@ int request_process_nexthop(Request *req) {
assert(req->nexthop);
assert(req->type == REQUEST_TYPE_NEXTHOP);
if (!link_is_ready_to_configure(req->link, false))
return 0;
if (!nexthop_is_ready_to_configure(req->link, req->nexthop))
return 0;
@ -765,7 +827,9 @@ int request_process_nexthop(Request *req) {
int manager_rtnl_process_nexthop(sd_netlink *rtnl, sd_netlink_message *message, Manager *m) {
_cleanup_(nexthop_freep) NextHop *tmp = NULL;
_cleanup_free_ void *raw_group = NULL;
NextHop *nexthop = NULL;
size_t raw_group_size;
uint32_t ifindex;
uint16_t type;
Link *link = NULL;
@ -818,7 +882,7 @@ int manager_rtnl_process_nexthop(sd_netlink *rtnl, sd_netlink_message *message,
if (r < 0) {
log_link_warning_errno(link, r, "rtnl: could not get nexthop family, ignoring: %m");
return 0;
} else if (!IN_SET(tmp->family, AF_INET, AF_INET6)) {
} else if (!IN_SET(tmp->family, AF_UNSPEC, AF_INET, AF_INET6)) {
log_link_debug(link, "rtnl: received nexthop message with invalid family %d, ignoring.", tmp->family);
return 0;
}
@ -829,11 +893,57 @@ int manager_rtnl_process_nexthop(sd_netlink *rtnl, sd_netlink_message *message,
return 0;
}
r = sd_netlink_message_read_data(message, NHA_GROUP, &raw_group_size, &raw_group);
if (r < 0 && r != -ENODATA) {
log_link_warning_errno(link, r, "rtnl: could not get NHA_GROUP attribute, ignoring: %m");
return 0;
} else if (r >= 0) {
struct nexthop_grp *group = raw_group;
size_t n_group;
if (raw_group_size == 0 || raw_group_size % sizeof(struct nexthop_grp) != 0) {
log_link_warning(link, "rtnl: received nexthop message with invalid nexthop group size, ignoring.");
return 0;
}
assert((uintptr_t) group % __alignof__(struct nexthop_grp) == 0);
n_group = raw_group_size / sizeof(struct nexthop_grp);
for (size_t i = 0; i < n_group; i++) {
_cleanup_free_ struct nexthop_grp *nhg = NULL;
if (group[i].id == 0) {
log_link_warning(link, "rtnl: received nexthop message with invalid ID in group, ignoring.");
return 0;
}
if (group[i].weight > 254) {
log_link_warning(link, "rtnl: received nexthop message with invalid weight in group, ignoring.");
return 0;
}
nhg = newdup(struct nexthop_grp, group + i, 1);
if (!nhg)
return log_oom();
r = hashmap_ensure_put(&tmp->group, NULL, UINT32_TO_PTR(nhg->id), nhg);
if (r == -ENOMEM)
return log_oom();
if (r < 0) {
log_link_warning_errno(link, r, "Failed to store nexthop group, ignoring: %m");
return 0;
}
if (r > 0)
TAKE_PTR(nhg);
}
}
if (tmp->family != AF_UNSPEC) {
r = netlink_message_read_in_addr_union(message, NHA_GATEWAY, tmp->family, &tmp->gw);
if (r < 0 && r != -ENODATA) {
log_link_warning_errno(link, r, "rtnl: could not get NHA_GATEWAY attribute, ignoring: %m");
return 0;
}
}
r = sd_netlink_message_has_flag(message, NHA_BLACKHOLE);
if (r < 0) {
@ -854,9 +964,9 @@ int manager_rtnl_process_nexthop(sd_netlink *rtnl, sd_netlink_message *message,
return 0;
}
/* All blackhole nexthops are managed by Manager. Note that the linux kernel does not set
* NHA_OID attribute when NHA_BLACKHOLE is set. Just for safety. */
if (tmp->blackhole)
/* All blackhole or group nexthops are managed by Manager. Note that the linux kernel does not
* set NHA_OID attribute when NHA_BLACKHOLE or NHA_GROUP is set. Just for safety. */
if (!nexthop_has_link(tmp))
link = NULL;
r = nexthop_get(m, link, tmp, &nexthop);
@ -908,8 +1018,26 @@ static int nexthop_section_verify(NextHop *nh) {
if (section_is_invalid(nh->section))
return -EINVAL;
if (nh->family == AF_UNSPEC)
/* When no Gateway= is specified, assume IPv4. */
if (!hashmap_isempty(nh->group)) {
if (in_addr_is_set(nh->family, &nh->gw))
return log_warning_errno(SYNTHETIC_ERRNO(EINVAL),
"%s: nexthop group cannot have gateway address. "
"Ignoring [NextHop] section from line %u.",
nh->section->filename, nh->section->line);
if (nh->family != AF_UNSPEC)
return log_warning_errno(SYNTHETIC_ERRNO(EINVAL),
"%s: nexthop group cannot have Family= setting. "
"Ignoring [NextHop] section from line %u.",
nh->section->filename, nh->section->line);
if (nh->blackhole && in_addr_is_set(nh->family, &nh->gw))
return log_warning_errno(SYNTHETIC_ERRNO(EINVAL),
"%s: nexthop group cannot be a blackhole. "
"Ignoring [NextHop] section from line %u.",
nh->section->filename, nh->section->line);
} else if (nh->family == AF_UNSPEC)
/* When neither Family=, Gateway=, nor Group= is specified, assume IPv4. */
nh->family = AF_INET;
if (nh->blackhole && in_addr_is_set(nh->family, &nh->gw))
@ -1185,3 +1313,107 @@ int config_parse_nexthop_blackhole(
TAKE_PTR(n);
return 0;
}
int config_parse_nexthop_group(
const char *unit,
const char *filename,
unsigned line,
const char *section,
unsigned section_line,
const char *lvalue,
int ltype,
const char *rvalue,
void *data,
void *userdata) {
_cleanup_(nexthop_free_or_set_invalidp) NextHop *n = NULL;
Network *network = userdata;
int r;
assert(filename);
assert(section);
assert(lvalue);
assert(rvalue);
assert(data);
r = nexthop_new_static(network, filename, section_line, &n);
if (r < 0)
return log_oom();
if (isempty(rvalue)) {
n->group = hashmap_free_free(n->group);
TAKE_PTR(n);
return 0;
}
for (const char *p = rvalue;;) {
_cleanup_free_ struct nexthop_grp *nhg = NULL;
_cleanup_free_ char *word = NULL;
uint32_t w;
char *sep;
r = extract_first_word(&p, &word, NULL, 0);
if (r == -ENOMEM)
return log_oom();
if (r < 0) {
log_syntax(unit, LOG_WARNING, filename, line, r,
"Invalid %s=, ignoring assignment: %s", lvalue, rvalue);
return 0;
}
if (r == 0)
break;
nhg = new0(struct nexthop_grp, 1);
if (!nhg)
return log_oom();
sep = strchr(word, ':');
if (sep) {
*sep++ = '\0';
r = safe_atou32(sep, &w);
if (r < 0) {
log_syntax(unit, LOG_WARNING, filename, line, r,
"Failed to parse weight for nexthop group, ignoring assignment: %s:%s",
word, sep);
continue;
}
if (w == 0 || w > 256) {
log_syntax(unit, LOG_WARNING, filename, line, 0,
"Invalid weight for nexthop group, ignoring assignment: %s:%s",
word, sep);
continue;
}
/* See comments in config_parse_multipath_route(). */
nhg->weight = w - 1;
}
r = safe_atou32(word, &nhg->id);
if (r < 0) {
log_syntax(unit, LOG_WARNING, filename, line, r,
"Failed to parse nexthop ID in %s=, ignoring assignment: %s%s%s",
lvalue, word, sep ? ":" : "", strempty(sep));
continue;
}
if (nhg->id == 0) {
log_syntax(unit, LOG_WARNING, filename, line, 0,
"Nexthop ID in %s= must be positive, ignoring assignment: %s%s%s",
lvalue, word, sep ? ":" : "", strempty(sep));
continue;
}
r = hashmap_ensure_put(&n->group, NULL, UINT32_TO_PTR(nhg->id), nhg);
if (r == -ENOMEM)
return log_oom();
if (r == -EEXIST) {
log_syntax(unit, LOG_WARNING, filename, line, r,
"Nexthop ID %"PRIu32" is specified multiple times in %s=, ignoring assignment: %s%s%s",
nhg->id, lvalue, word, sep ? ":" : "", strempty(sep));
continue;
}
assert(r > 0);
TAKE_PTR(nhg);
}
TAKE_PTR(n);
return 0;
}

3
src/network/networkd-nexthop.h
View File

@ -9,6 +9,7 @@
#include "sd-netlink.h"
#include "conf-parser.h"
#include "hashmap.h"
#include "in-addr-util.h"
#include "networkd-util.h"
@ -31,6 +32,7 @@ typedef struct NextHop {
int family;
union in_addr_union gw;
int onlink;
Hashmap *group;
} NextHop;
NextHop *nexthop_free(NextHop *nexthop);
@ -51,3 +53,4 @@ CONFIG_PARSER_PROTOTYPE(config_parse_nexthop_gateway);
CONFIG_PARSER_PROTOTYPE(config_parse_nexthop_family);
CONFIG_PARSER_PROTOTYPE(config_parse_nexthop_onlink);
CONFIG_PARSER_PROTOTYPE(config_parse_nexthop_blackhole);
CONFIG_PARSER_PROTOTYPE(config_parse_nexthop_group);

236
src/network/networkd-route.c
View File

@ -2,6 +2,7 @@
#include <linux/icmpv6.h>
#include <linux/ipv6_route.h>
#include <linux/nexthop.h>
#include "alloc-util.h"
#include "netlink-util.h"
@ -468,7 +469,7 @@ static int route_get(const Manager *manager, const Link *link, const Route *in,
return -ENOENT;
}
static void route_copy(Route *dest, const Route *src, const MultipathRoute *m, const NextHop *nh) {
static void route_copy(Route *dest, const Route *src, const MultipathRoute *m, const NextHop *nh, uint8_t nh_weight) {
assert(dest);
assert(src);
@ -496,9 +497,11 @@ static void route_copy(Route *dest, const Route *src, const MultipathRoute *m, c
dest->nexthop_id = src->nexthop_id;
if (nh) {
assert(hashmap_isempty(nh->group));
dest->gw_family = nh->family;
dest->gw = nh->gw;
dest->gw_weight = src->gw_weight;
dest->gw_weight = nh_weight != UINT8_MAX ? nh_weight : src->gw_weight;
} else if (m) {
dest->gw_family = m->gateway.family;
dest->gw = m->gateway.address;
@ -560,7 +563,7 @@ static int route_add_internal(Manager *manager, Link *link, Set **routes, const
if (r < 0)
return r;
route_copy(route, in, NULL, NULL);
route_copy(route, in, NULL, NULL, UINT8_MAX);
r = set_ensure_put(routes, &route_hash_ops, route);
if (r < 0)
@ -584,9 +587,8 @@ static int route_add_foreign(Manager *manager, Link *link, const Route *in, Rout
return route_add_internal(manager, link, link ? &link->routes_foreign : &manager->routes_foreign, in, ret);
}
static int route_add(Manager *manager, Link *link, const Route *in, const MultipathRoute *m, const NextHop *nh, Route **ret) {
static int route_add(Manager *manager, Link *link, const Route *in, const MultipathRoute *m, const NextHop *nh, uint8_t nh_weight, Route **ret) {
_cleanup_(route_freep) Route *tmp = NULL;
bool is_new = false;
Route *route;
int r;
@ -594,11 +596,13 @@ static int route_add(Manager *manager, Link *link, const Route *in, const Multip
assert(in);
if (nh) {
assert(hashmap_isempty(nh->group));
r = route_new(&tmp);
if (r < 0)
return r;
route_copy(tmp, in, NULL, nh);
route_copy(tmp, in, NULL, nh, nh_weight);
in = tmp;
} else if (m) {
assert(link && (m->ifindex == 0 || m->ifindex == link->ifindex));
@ -607,7 +611,7 @@ static int route_add(Manager *manager, Link *link, const Route *in, const Multip
if (r < 0)
return r;
route_copy(tmp, in, m, NULL);
route_copy(tmp, in, m, NULL, UINT8_MAX);
in = tmp;
}
@ -617,7 +621,6 @@ static int route_add(Manager *manager, Link *link, const Route *in, const Multip
r = route_add_internal(manager, link, link ? &link->routes : &manager->routes, in, &route);
if (r < 0)
return r;
is_new = true;
} else if (r == 0) {
/* Take over a foreign route */
r = set_ensure_put(link ? &link->routes : &manager->routes, &route_hash_ops, route);
@ -633,7 +636,7 @@ static int route_add(Manager *manager, Link *link, const Route *in, const Multip
if (ret)
*ret = route;
return is_new;
return 0;
}
static bool route_type_is_reject(const Route *route) {
@ -642,6 +645,26 @@ static bool route_type_is_reject(const Route *route) {
return IN_SET(route->type, RTN_UNREACHABLE, RTN_PROHIBIT, RTN_BLACKHOLE, RTN_THROW);
}
static int link_has_route_one(Link *link, const Route *route, const NextHop *nh, uint8_t nh_weight) {
_cleanup_(route_freep) Route *tmp = NULL;
int r;
assert(link);
assert(route);
assert(nh);
r = route_new(&tmp);
if (r < 0)
return r;
route_copy(tmp, route, NULL, nh, nh_weight);
if (route_type_is_reject(route) || (nh && nh->blackhole))
return route_get(link->manager, NULL, tmp, NULL) >= 0;
else
return route_get(NULL, link, tmp, NULL) >= 0;
}
int link_has_route(Link *link, const Route *route) {
MultipathRoute *m;
int r;
@ -650,22 +673,27 @@ int link_has_route(Link *link, const Route *route) {
assert(route);
if (route->nexthop_id > 0) {
_cleanup_(route_freep) Route *tmp = NULL;
struct nexthop_grp *nhg;
NextHop *nh;
if (manager_get_nexthop_by_id(link->manager, route->nexthop_id, &nh) < 0)
return false;
r = route_new(&tmp);
if (r < 0)
if (hashmap_isempty(nh->group))
return link_has_route_one(link, route, nh, UINT8_MAX);
HASHMAP_FOREACH(nhg, nh->group) {
NextHop *h;
if (manager_get_nexthop_by_id(link->manager, nhg->id, &h) < 0)
return false;
r = link_has_route_one(link, route, h, nhg->weight);
if (r <= 0)
return r;
}
route_copy(tmp, route, NULL, nh);
if (route_type_is_reject(route) || (nh && nh->blackhole))
return route_get(link->manager, NULL, tmp, NULL) >= 0;
else
return route_get(NULL, link, tmp, NULL) >= 0;
return true;
}
if (ordered_set_isempty(route->multipath_routes)) {
@ -694,7 +722,7 @@ int link_has_route(Link *link, const Route *route) {
if (r < 0)
return r;
route_copy(tmp, route, m, NULL);
route_copy(tmp, route, m, NULL, UINT8_MAX);
if (route_get(NULL, l, tmp, NULL) < 0)
return false;
@ -1104,7 +1132,7 @@ int link_drop_foreign_routes(Link *link) {
continue;
if (link_has_static_route(link, route))
k = route_add(NULL, link, route, NULL, NULL, NULL);
k = route_add(NULL, link, route, NULL, NULL, UINT8_MAX, NULL);
else
k = route_remove(route, NULL, link);
if (k < 0 && r >= 0)
@ -1156,33 +1184,36 @@ static int route_expire_handler(sd_event_source *s, uint64_t usec, void *userdat
return 1;
}
static int route_add_and_setup_timer(Link *link, const Route *route, const MultipathRoute *m, Route **ret) {
static int route_add_and_setup_timer_one(Link *link, const Route *route, const MultipathRoute *m, const NextHop *nh, uint8_t nh_weight, Route **ret) {
_cleanup_(sd_event_source_unrefp) sd_event_source *expire = NULL;
NextHop *nh = NULL;
Route *nr;
int r, k;
int r;
assert(link);
assert(link->manager);
assert(route);
(void) manager_get_nexthop_by_id(link->manager, route->nexthop_id, &nh);
assert(!(m && nh));
assert(ret);
if (route_type_is_reject(route) || (nh && nh->blackhole))
k = route_add(link->manager, NULL, route, NULL, nh, &nr);
else if (!m || m->ifindex == 0 || m->ifindex == link->ifindex)
k = route_add(NULL, link, route, m, nh, &nr);
else {
r = route_add(link->manager, NULL, route, NULL, nh, nh_weight, &nr);
else if (nh) {
assert(nh->link);
assert(hashmap_isempty(nh->group));
r = route_add(NULL, nh->link, route, NULL, nh, nh_weight, &nr);
} else if (m && m->ifindex != 0 && m->ifindex != link->ifindex) {
Link *link_gw;
r = link_get(link->manager, m->ifindex, &link_gw);
if (r < 0)
return log_link_error_errno(link, r, "Failed to get link with ifindex %d: %m", m->ifindex);
k = route_add(NULL, link_gw, route, m, NULL, &nr);
}
if (k < 0)
return log_link_error_errno(link, k, "Could not add route: %m");
r = route_add(NULL, link_gw, route, m, NULL, UINT8_MAX, &nr);
} else
r = route_add(NULL, link, route, m, NULL, UINT8_MAX, &nr);
if (r < 0)
return log_link_error_errno(link, r, "Could not add route: %m");
/* TODO: drop expiration handling once it can be pushed into the kernel */
if (nr->lifetime != USEC_INFINITY && !kernel_route_expiration_supported()) {
@ -1195,10 +1226,79 @@ static int route_add_and_setup_timer(Link *link, const Route *route, const Multi
sd_event_source_unref(nr->expire);
nr->expire = TAKE_PTR(expire);
if (ret)
*ret = nr;
return 0;
}
return k;
static int route_add_and_setup_timer(Link *link, const Route *route, unsigned *ret_n_routes, Route ***ret_routes) {
_cleanup_free_ Route **routes = NULL;
unsigned n_routes;
NextHop *nh;
Route **p;
int r;
assert(link);
assert(route);
assert(ret_n_routes);
assert(ret_routes);
if (route->nexthop_id > 0) {
r = manager_get_nexthop_by_id(link->manager, route->nexthop_id, &nh);
if (r < 0)
return log_link_error_errno(link, r, "Could not get nexthop by ID %"PRIu32": %m", route->nexthop_id);
} else
nh = NULL;
if (nh && !hashmap_isempty(nh->group)) {
struct nexthop_grp *nhg;
n_routes = hashmap_size(nh->group);
p = routes = new(Route*, n_routes);
if (!routes)
return log_oom();
HASHMAP_FOREACH(nhg, nh->group) {
NextHop *h;
r = manager_get_nexthop_by_id(link->manager, nhg->id, &h);
if (r < 0)
return log_link_error_errno(link, r, "Could not get nexthop group member by ID %"PRIu32": %m", nhg->id);
/* The nexthop h may be a blackhole nexthop. In that case, h->link is NULL. */
r = route_add_and_setup_timer_one(h->link ?: link, route, NULL, h, nhg->weight, p++);
if (r < 0)
return r;
}
} else if (!ordered_set_isempty(route->multipath_routes)) {
MultipathRoute *m;
assert(!nh);
assert(!in_addr_is_set(route->gw_family, &route->gw));
n_routes = ordered_set_size(route->multipath_routes);
p = routes = new(Route*, n_routes);
if (!routes)
return log_oom();
ORDERED_SET_FOREACH(m, route->multipath_routes) {
r = route_add_and_setup_timer_one(link, route, m, NULL, UINT8_MAX, p++);
if (r < 0)
return r;
}
} else {
n_routes = 1;
routes = new(Route*, n_routes);
if (!routes)
return log_oom();
r = route_add_and_setup_timer_one(link, route, NULL, nh, UINT8_MAX, routes);
if (r < 0)
return r;
}
*ret_n_routes = n_routes;
*ret_routes = TAKE_PTR(routes);
return 0;
}
static int append_nexthop_one(const Route *route, const MultipathRoute *m, struct rtattr **rta, size_t offset) {
@ -1220,7 +1320,7 @@ static int append_nexthop_one(const Route *route, const MultipathRoute *m, struc
*rtnh = (struct rtnexthop) {
.rtnh_len = sizeof(*rtnh),
.rtnh_ifindex = m->ifindex,
.rtnh_hops = m->weight > 0 ? m->weight - 1 : 0,
.rtnh_hops = m->weight,
};
(*rta)->rta_len += sizeof(struct rtnexthop);
@ -1312,7 +1412,7 @@ static int route_configure(
_cleanup_(sd_netlink_message_unrefp) sd_netlink_message *req = NULL;
_cleanup_free_ Route **routes = NULL;
unsigned n_routes = 0; /* avoid false maybe-uninitialized warning */
int r, k;
int r;
assert(link);
assert(link->manager);
@ -1400,45 +1500,18 @@ static int route_configure(
if (r < 0)
return log_link_error_errno(link, r, "Could not append RTA_METRICS attribute: %m");
if (route->nexthop_id != 0 ||
in_addr_is_set(route->gw_family, &route->gw) ||
ordered_set_isempty(route->multipath_routes)) {
if (ret_routes) {
n_routes = 1;
routes = new(Route*, n_routes);
if (!routes)
return log_oom();
}
k = route_add_and_setup_timer(link, route, NULL, routes);
if (k < 0)
return k;
} else {
MultipathRoute *m;
Route **p;
if (!ordered_set_isempty(route->multipath_routes)) {
assert(route->nexthop_id == 0);
assert(!in_addr_is_set(route->gw_family, &route->gw));
r = append_nexthops(route, req);
if (r < 0)
return log_link_error_errno(link, r, "Could not append RTA_MULTIPATH attribute: %m");
if (ret_routes) {
n_routes = ordered_set_size(route->multipath_routes);
routes = new(Route*, n_routes);
if (!routes)
return log_oom();
}
k = 0;
p = routes;
ORDERED_SET_FOREACH(m, route->multipath_routes) {
r = route_add_and_setup_timer(link, route, m, ret_routes ? p++ : NULL);
r = route_add_and_setup_timer(link, route, &n_routes, &routes);
if (r < 0)
return r;
if (r > 0)
k = 1;
}
}
r = netlink_call_async(link->manager->rtnl, NULL, req, callback,
link_netlink_destroy_callback, link);
@ -1452,7 +1525,7 @@ static int route_configure(
*ret_routes = TAKE_PTR(routes);
}
return k;
return r;
}
static int static_route_handler(sd_netlink *rtnl, sd_netlink_message *m, Link *link) {
@ -1534,10 +1607,17 @@ static int route_is_ready_to_configure(const Route *route, Link *link) {
assert(route);
assert(link);
if (route->nexthop_id > 0 &&
manager_get_nexthop_by_id(link->manager, route->nexthop_id, &nh) < 0)
if (route->nexthop_id > 0) {
struct nexthop_grp *nhg;
if (manager_get_nexthop_by_id(link->manager, route->nexthop_id, &nh) < 0)
return false;
HASHMAP_FOREACH(nhg, nh->group)
if (manager_get_nexthop_by_id(link->manager, nhg->id, NULL) < 0)
return false;
}
if (route_type_is_reject(route) || (nh && nh->blackhole)) {
if (nh && link->manager->nexthop_remove_messages > 0)
return false;
@ -1642,18 +1722,19 @@ static int process_route_one(Manager *manager, Link *link, uint16_t type, const
(void) manager_get_nexthop_by_id(manager, tmp->nexthop_id, &nh);
if (nh) {
if (link && link != nh->link)
if (nh && hashmap_isempty(nh->group)) {
if (link && nh->link && link != nh->link)
return log_link_warning_errno(link, SYNTHETIC_ERRNO(EINVAL),
"rtnl: received RTA_OIF and ifindex of nexthop corresponding to RTA_NH_ID do not match, ignoring.");
if (nh->link)
link = nh->link;
r = route_new(&nr);
if (r < 0)
return log_oom();
route_copy(nr, tmp, NULL, nh);
route_copy(nr, tmp, NULL, nh, UINT8_MAX);
tmp = nr;
} else if (m) {
@ -1675,7 +1756,7 @@ static int process_route_one(Manager *manager, Link *link, uint16_t type, const
if (r < 0)
return log_oom();
route_copy(nr, tmp, m, NULL);
route_copy(nr, tmp, m, NULL, UINT8_MAX);
tmp = nr;
}
@ -2779,11 +2860,16 @@ int config_parse_multipath_route(
"Invalid multipath route weight, ignoring assignment: %s", p);
return 0;
}
/* ip command takes weight in the range 1…255, while kernel takes the value in the
* range 0…254. MultiPathRoute= setting also takes weight in the same range which ip
* command uses, then networkd decreases by one and stores it to match the range which
* kernel uses. */
if (m->weight == 0 || m->weight > 256) {
log_syntax(unit, LOG_WARNING, filename, line, 0,
"Invalid multipath route weight, ignoring assignment: %s", p);
return 0;
}
m->weight--;
}
r = ordered_set_ensure_put(&n->multipath_routes, NULL, m);

4
src/network/networkd-routing-policy-rule.c
View File

@ -318,7 +318,6 @@ static int routing_policy_rule_get(Manager *m, const RoutingPolicyRule *rule, Ro
static int routing_policy_rule_add(Manager *m, const RoutingPolicyRule *in, RoutingPolicyRule **ret) {
_cleanup_(routing_policy_rule_freep) RoutingPolicyRule *rule = NULL;
RoutingPolicyRule *existing;
bool is_new = false;
int r;
assert(m);
@ -339,7 +338,6 @@ static int routing_policy_rule_add(Manager *m, const RoutingPolicyRule *in, Rout
rule->manager = m;
existing = TAKE_PTR(rule);
is_new = true;
} else if (r == 0) {
/* Take over a foreign rule. */
r = set_ensure_put(&m->rules, &routing_policy_rule_hash_ops, existing);
@ -356,7 +354,7 @@ static int routing_policy_rule_add(Manager *m, const RoutingPolicyRule *in, Rout
if (ret)
*ret = existing;
return is_new;
return 0;
}
static int routing_policy_rule_consume_foreign(Manager *m, RoutingPolicyRule *rule) {

1
test/fuzz/fuzz-network-parser/directives.network
View File

@ -372,6 +372,7 @@ Gateway=
Family=
OnLink=
Blackhole=
Group=
[QDisc]
Parent=
Handle=

13
test/test-network/conf/25-nexthop-dummy.network Normal file
View File

@ -0,0 +1,13 @@
[Match]
Name=dummy98
[Network]
Address=192.168.20.20/24
[NextHop]
Id=20
Gateway=192.168.20.1
[NextHop]
Id=21
Group=1:3 20:1

4
test/test-network/conf/25-nexthop.network
View File

@ -72,3 +72,7 @@ Destination=10.10.10.13
[Route]
NextHop=7
Destination=2001:1234:5:8f62::2
[Route]
NextHop=21
Destination=10.10.10.14

31
test/test-network/systemd-networkd-tests.py
View File

@ -1819,6 +1819,7 @@ class NetworkdNetworkTests(unittest.TestCase, Utilities):
'25-neighbor-ipv6.network',
'25-neighbor-ip-dummy.network',
'25-neighbor-ip.network',
'25-nexthop-dummy.network',
'25-nexthop-nothing.network',
'25-nexthop.network',
'25-qdisc-cake.network',
@ -2844,9 +2845,8 @@ class NetworkdNetworkTests(unittest.TestCase, Utilities):
@expectedFailureIfNexthopIsNotAvailable()
def test_nexthop(self):
copy_unit_to_networkd_unit_path('25-nexthop.network', '25-veth.netdev', '25-veth-peer.network')
start_networkd()
self.wait_online(['veth99:routable', 'veth-peer:routable'])
def check_nexthop(self):
self.wait_online(['veth99:routable', 'veth-peer:routable', 'dummy98:routable'])
output = check_output('ip nexthop list dev veth99')
print(output)
@ -2857,12 +2857,21 @@ class NetworkdNetworkTests(unittest.TestCase, Utilities):
self.assertRegex(output, 'id 5 via 192.168.10.1 dev veth99 .*onlink')
self.assertRegex(output, r'id [0-9]* via 192.168.5.2 dev veth99')
output = check_output('ip nexthop list dev dummy98')
print(output)
self.assertIn('id 20 via 192.168.20.1 dev dummy98', output)
# kernel manages blackhole nexthops on lo
output = check_output('ip nexthop list dev lo')
print(output)
self.assertIn('id 6 blackhole', output)
self.assertIn('id 7 blackhole', output)
# group nexthops are shown with -0 option
output = check_output('ip -0 nexthop list id 21')
print(output)
self.assertRegex(output, r'id 21 group (1,3/20|20/1,3)')
output = check_output('ip route show dev veth99 10.10.10.10')
print(output)
self.assertEqual('10.10.10.10 nhid 1 via 192.168.5.1 proto static', output)
@ -2887,6 +2896,18 @@ class NetworkdNetworkTests(unittest.TestCase, Utilities):
print(output)
self.assertEqual('blackhole 2001:1234:5:8f62::2 nhid 7 dev lo proto static metric 1024 pref medium', output)
output = check_output('ip route show 10.10.10.14')
print(output)
self.assertIn('10.10.10.14 nhid 21 proto static', output)
self.assertIn('nexthop via 192.168.20.1 dev dummy98 weight 1', output)
self.assertIn('nexthop via 192.168.5.1 dev veth99 weight 3', output)
copy_unit_to_networkd_unit_path('25-nexthop.network', '25-veth.netdev', '25-veth-peer.network',
'12-dummy.netdev', '25-nexthop-dummy.network')
start_networkd()
check_nexthop(self)
remove_unit_from_networkd_path(['25-nexthop.network'])
copy_unit_to_networkd_unit_path('25-nexthop-nothing.network')
rc = call(*networkctl_cmd, 'reload', env=env)
@ -2904,11 +2925,13 @@ class NetworkdNetworkTests(unittest.TestCase, Utilities):
remove_unit_from_networkd_path(['25-nexthop-nothing.network'])
copy_unit_to_networkd_unit_path('25-nexthop.network')
rc = call(*networkctl_cmd, 'reconfigure', 'dummy98', env=env)
self.assertEqual(rc, 0)
rc = call(*networkctl_cmd, 'reload', env=env)
self.assertEqual(rc, 0)
time.sleep(1)
self.wait_online(['veth99:routable', 'veth-peer:routable'])
check_nexthop(self)
rc = call('ip link del veth99')
self.assertEqual(rc, 0)