netlink: Split nl_route() into separate operation functions

nl_route() can perform 3 quite different operations based on the 'op'
parameter.  Split this into separate functions for each one.  This requires
more lines of code, but makes the internal logic of each operation much
easier to follow.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
This commit is contained in:
David Gibson 2023-08-03 17:19:42 +10:00 committed by Stefano Brivio
parent eff3bcb245
commit 257a6b0b7e
4 changed files with 170 additions and 112 deletions

4
conf.c
View file

@ -649,7 +649,7 @@ static unsigned int conf_ip4(unsigned int ifi,
} }
if (IN4_IS_ADDR_UNSPECIFIED(&ip4->gw)) if (IN4_IS_ADDR_UNSPECIFIED(&ip4->gw))
nl_route(NL_GET, ifi, 0, AF_INET, &ip4->gw); nl_route_get_def(ifi, AF_INET, &ip4->gw);
if (IN4_IS_ADDR_UNSPECIFIED(&ip4->addr)) if (IN4_IS_ADDR_UNSPECIFIED(&ip4->addr))
nl_addr_get(ifi, AF_INET, &ip4->addr, &ip4->prefix_len, NULL); nl_addr_get(ifi, AF_INET, &ip4->addr, &ip4->prefix_len, NULL);
@ -700,7 +700,7 @@ static unsigned int conf_ip6(unsigned int ifi,
} }
if (IN6_IS_ADDR_UNSPECIFIED(&ip6->gw)) if (IN6_IS_ADDR_UNSPECIFIED(&ip6->gw))
nl_route(NL_GET, ifi, 0, AF_INET6, &ip6->gw); nl_route_get_def(ifi, AF_INET6, &ip6->gw);
nl_addr_get(ifi, AF_INET6, nl_addr_get(ifi, AF_INET6,
IN6_IS_ADDR_UNSPECIFIED(&ip6->addr) ? &ip6->addr : NULL, IN6_IS_ADDR_UNSPECIFIED(&ip6->addr) ? &ip6->addr : NULL,

251
netlink.c
View file

@ -185,15 +185,71 @@ unsigned int nl_get_ext_if(sa_family_t af)
} }
/** /**
* nl_route() - Get/set/copy routes for given interface and address family * nl_route_get_def() - Get default route for given interface and address family
* @op: Requested operation * @ifi: Interface index
* @ifi: Interface index in outer network namespace
* @ifi_ns: Interface index in target namespace for NL_SET, NL_DUP
* @af: Address family * @af: Address family
* @gw: Default gateway to fill on NL_GET, to set on NL_SET * @gw: Default gateway to fill on NL_GET
*/ */
void nl_route(enum nl_op op, unsigned int ifi, unsigned int ifi_ns, void nl_route_get_def(unsigned int ifi, sa_family_t af, void *gw)
sa_family_t af, void *gw) {
struct req_t {
struct nlmsghdr nlh;
struct rtmsg rtm;
struct rtattr rta;
unsigned int ifi;
} req = {
.nlh.nlmsg_type = RTM_GETROUTE,
.nlh.nlmsg_len = sizeof(req),
.nlh.nlmsg_flags = NLM_F_REQUEST | NLM_F_DUMP,
.nlh.nlmsg_seq = nl_seq++,
.rtm.rtm_family = af,
.rtm.rtm_table = RT_TABLE_MAIN,
.rtm.rtm_scope = RT_SCOPE_UNIVERSE,
.rtm.rtm_type = RTN_UNICAST,
.rta.rta_type = RTA_OIF,
.rta.rta_len = RTA_LENGTH(sizeof(unsigned int)),
.ifi = ifi,
};
struct nlmsghdr *nh;
char buf[NLBUFSIZ];
ssize_t n;
if ((n = nl_req(0, buf, &req, req.nlh.nlmsg_len)) < 0)
return;
for (nh = (struct nlmsghdr *)buf;
NLMSG_OK(nh, n) && nh->nlmsg_type != NLMSG_DONE;
nh = NLMSG_NEXT(nh, n)) {
struct rtmsg *rtm = (struct rtmsg *)NLMSG_DATA(nh);
struct rtattr *rta;
size_t na;
if (nh->nlmsg_type != RTM_NEWROUTE)
continue;
if (rtm->rtm_dst_len)
continue;
for (rta = RTM_RTA(rtm), na = RTM_PAYLOAD(nh); RTA_OK(rta, na);
rta = RTA_NEXT(rta, na)) {
if (rta->rta_type != RTA_GATEWAY)
continue;
memcpy(gw, RTA_DATA(rta), RTA_PAYLOAD(rta));
return;
}
}
}
/**
* nl_route_set_def() - Set default route for given interface and address family
* @ifi: Interface index in target namespace
* @af: Address family
* @gw: Default gateway to set
*/
void nl_route_set_def(unsigned int ifi, sa_family_t af, void *gw)
{ {
struct req_t { struct req_t {
struct nlmsghdr nlh; struct nlmsghdr nlh;
@ -215,8 +271,69 @@ void nl_route(enum nl_op op, unsigned int ifi, unsigned int ifi_ns,
} r4; } r4;
} set; } set;
} req = { } req = {
.nlh.nlmsg_type = op == NL_SET ? RTM_NEWROUTE : RTM_GETROUTE, .nlh.nlmsg_type = RTM_NEWROUTE,
.nlh.nlmsg_flags = NLM_F_REQUEST, .nlh.nlmsg_flags = NLM_F_REQUEST | NLM_F_ACK |
NLM_F_CREATE | NLM_F_EXCL,
.nlh.nlmsg_seq = nl_seq++,
.rtm.rtm_family = af,
.rtm.rtm_table = RT_TABLE_MAIN,
.rtm.rtm_scope = RT_SCOPE_UNIVERSE,
.rtm.rtm_type = RTN_UNICAST,
.rtm.rtm_protocol = RTPROT_BOOT,
.rta.rta_type = RTA_OIF,
.rta.rta_len = RTA_LENGTH(sizeof(unsigned int)),
.ifi = ifi,
};
char buf[NLBUFSIZ];
if (af == AF_INET6) {
size_t rta_len = RTA_LENGTH(sizeof(req.set.r6.d));
req.nlh.nlmsg_len = offsetof(struct req_t, set.r6)
+ sizeof(req.set.r6);
req.set.r6.rta_dst.rta_type = RTA_DST;
req.set.r6.rta_dst.rta_len = rta_len;
memcpy(&req.set.r6.a, gw, sizeof(req.set.r6.a));
req.set.r6.rta_gw.rta_type = RTA_GATEWAY;
req.set.r6.rta_gw.rta_len = rta_len;
} else {
size_t rta_len = RTA_LENGTH(sizeof(req.set.r4.d));
req.nlh.nlmsg_len = offsetof(struct req_t, set.r4)
+ sizeof(req.set.r4);
req.set.r4.rta_dst.rta_type = RTA_DST;
req.set.r4.rta_dst.rta_len = rta_len;
req.set.r4.a = *(uint32_t *)gw;
req.set.r4.rta_gw.rta_type = RTA_GATEWAY;
req.set.r4.rta_gw.rta_len = rta_len;
}
nl_req(1, buf, &req, req.nlh.nlmsg_len);
}
/**
* nl_route_dup() - Copy routes for given interface and address family
* @ifi: Interface index in outer network namespace
* @ifi_ns: Interface index in target namespace for NL_SET, NL_DUP
* @af: Address family
*/
void nl_route_dup(unsigned int ifi, unsigned int ifi_ns, sa_family_t af)
{
struct req_t {
struct nlmsghdr nlh;
struct rtmsg rtm;
struct rtattr rta;
unsigned int ifi;
} req = {
.nlh.nlmsg_type = RTM_GETROUTE,
.nlh.nlmsg_len = sizeof(req),
.nlh.nlmsg_flags = NLM_F_REQUEST | NLM_F_DUMP,
.nlh.nlmsg_seq = nl_seq++, .nlh.nlmsg_seq = nl_seq++,
.rtm.rtm_family = af, .rtm.rtm_family = af,
@ -226,111 +343,53 @@ void nl_route(enum nl_op op, unsigned int ifi, unsigned int ifi_ns,
.rta.rta_type = RTA_OIF, .rta.rta_type = RTA_OIF,
.rta.rta_len = RTA_LENGTH(sizeof(unsigned int)), .rta.rta_len = RTA_LENGTH(sizeof(unsigned int)),
.ifi = op == NL_SET ? ifi_ns : ifi, .ifi = ifi,
}; };
char buf[NLBUFSIZ], resp[NLBUFSIZ];
unsigned dup_routes = 0; unsigned dup_routes = 0;
ssize_t n, nlmsgs_size; ssize_t n, nlmsgs_size;
struct nlmsghdr *nh; struct nlmsghdr *nh;
struct rtattr *rta; unsigned i;
char buf[NLBUFSIZ];
struct rtmsg *rtm;
size_t na;
if (op == NL_SET) { if ((n = nl_req(0, buf, &req, req.nlh.nlmsg_len)) < 0)
if (af == AF_INET6) {
size_t rta_len = RTA_LENGTH(sizeof(req.set.r6.d));
req.nlh.nlmsg_len = offsetof(struct req_t, set.r6)
+ sizeof(req.set.r6);
req.set.r6.rta_dst.rta_type = RTA_DST;
req.set.r6.rta_dst.rta_len = rta_len;
memcpy(&req.set.r6.a, gw, sizeof(req.set.r6.a));
req.set.r6.rta_gw.rta_type = RTA_GATEWAY;
req.set.r6.rta_gw.rta_len = rta_len;
} else {
size_t rta_len = RTA_LENGTH(sizeof(req.set.r4.d));
req.nlh.nlmsg_len = offsetof(struct req_t, set.r4)
+ sizeof(req.set.r4);
req.set.r4.rta_dst.rta_type = RTA_DST;
req.set.r4.rta_dst.rta_len = rta_len;
req.set.r4.a = *(uint32_t *)gw;
req.set.r4.rta_gw.rta_type = RTA_GATEWAY;
req.set.r4.rta_gw.rta_len = rta_len;
}
req.rtm.rtm_protocol = RTPROT_BOOT;
req.nlh.nlmsg_flags |= NLM_F_ACK | NLM_F_EXCL | NLM_F_CREATE;
} else {
req.nlh.nlmsg_len = offsetof(struct req_t, set.r6);
req.nlh.nlmsg_flags |= NLM_F_DUMP;
}
if ((n = nl_req(op == NL_SET, buf, &req, req.nlh.nlmsg_len)) < 0)
return; return;
if (op == NL_SET)
return;
nh = (struct nlmsghdr *)buf;
nlmsgs_size = n; nlmsgs_size = n;
for ( ; NLMSG_OK(nh, n); nh = NLMSG_NEXT(nh, n)) { for (nh = (struct nlmsghdr *)buf;
NLMSG_OK(nh, n) && nh->nlmsg_type != NLMSG_DONE;
nh = NLMSG_NEXT(nh, n)) {
struct rtmsg *rtm = (struct rtmsg *)NLMSG_DATA(nh);
struct rtattr *rta;
size_t na;
if (nh->nlmsg_type != RTM_NEWROUTE) if (nh->nlmsg_type != RTM_NEWROUTE)
goto next;
if (op == NL_DUP) {
nh->nlmsg_seq = nl_seq++;
nh->nlmsg_pid = 0;
nh->nlmsg_flags &= ~NLM_F_DUMP_FILTERED;
nh->nlmsg_flags |= NLM_F_REQUEST | NLM_F_ACK |
NLM_F_CREATE;
dup_routes++;
}
rtm = (struct rtmsg *)NLMSG_DATA(nh);
if (op == NL_GET && rtm->rtm_dst_len)
continue; continue;
nh->nlmsg_seq = nl_seq++;
nh->nlmsg_pid = 0;
nh->nlmsg_flags &= ~NLM_F_DUMP_FILTERED;
nh->nlmsg_flags |= NLM_F_REQUEST | NLM_F_ACK |
NLM_F_CREATE;
dup_routes++;
for (rta = RTM_RTA(rtm), na = RTM_PAYLOAD(nh); RTA_OK(rta, na); for (rta = RTM_RTA(rtm), na = RTM_PAYLOAD(nh); RTA_OK(rta, na);
rta = RTA_NEXT(rta, na)) { rta = RTA_NEXT(rta, na)) {
if (op == NL_GET) { if (rta->rta_type == RTA_OIF)
if (rta->rta_type != RTA_GATEWAY)
continue;
memcpy(gw, RTA_DATA(rta), RTA_PAYLOAD(rta));
return;
}
if (op == NL_DUP && rta->rta_type == RTA_OIF)
*(unsigned int *)RTA_DATA(rta) = ifi_ns; *(unsigned int *)RTA_DATA(rta) = ifi_ns;
} }
next:
if (nh->nlmsg_type == NLMSG_DONE)
break;
} }
if (op == NL_DUP) { nh = (struct nlmsghdr *)buf;
char resp[NLBUFSIZ]; /* Routes might have dependencies between each other, and the kernel
unsigned i; * processes RTM_NEWROUTE messages sequentially. For n valid routes, we
* might need to send up to n requests to get all of them inserted.
nh = (struct nlmsghdr *)buf; * Routes that have been already inserted won't cause the whole request
/* Routes might have dependencies between each other, and the * to fail, so we can simply repeat the whole request. This approach
* kernel processes RTM_NEWROUTE messages sequentially. For n * avoids the need to calculate dependencies: let the kernel do that.
* valid routes, we might need to send up to n requests to get */
* all of them inserted. Routes that have been already inserted for (i = 0; i < dup_routes; i++)
* won't cause the whole request to fail, so we can simply nl_req(1, resp, nh, nlmsgs_size);
* repeat the whole request. This approach avoids the need to
* calculate dependencies: let the kernel do that.
*/
for (i = 0; i < dup_routes; i++)
nl_req(1, resp, nh, nlmsgs_size);
}
} }
/** /**

View file

@ -6,16 +6,11 @@
#ifndef NETLINK_H #ifndef NETLINK_H
#define NETLINK_H #define NETLINK_H
enum nl_op {
NL_GET,
NL_SET,
NL_DUP,
};
void nl_sock_init(const struct ctx *c, bool ns); void nl_sock_init(const struct ctx *c, bool ns);
unsigned int nl_get_ext_if(sa_family_t af); unsigned int nl_get_ext_if(sa_family_t af);
void nl_route(enum nl_op op, unsigned int ifi, unsigned int ifi_ns, void nl_route_get_def(unsigned int ifi, sa_family_t af, void *gw);
sa_family_t af, void *gw); void nl_route_set_def(unsigned int ifi, sa_family_t af, void *gw);
void nl_route_dup(unsigned int ifi, unsigned int ifi_ns, sa_family_t af);
void nl_addr_get(unsigned int ifi, sa_family_t af, void *addr, void nl_addr_get(unsigned int ifi, sa_family_t af, void *addr,
int *prefix_len, void *addr_l); int *prefix_len, void *addr_l);
void nl_addr_set(unsigned int ifi, sa_family_t af, void *addr, int prefix_len); void nl_addr_set(unsigned int ifi, sa_family_t af, void *addr, int prefix_len);

16
pasta.c
View file

@ -281,8 +281,6 @@ void pasta_ns_conf(struct ctx *c)
nl_link_set_mac(1, c->pasta_ifi, c->mac_guest); nl_link_set_mac(1, c->pasta_ifi, c->mac_guest);
if (c->pasta_conf_ns) { if (c->pasta_conf_ns) {
enum nl_op op_routes = c->no_copy_routes ? NL_SET : NL_DUP;
nl_link_up(1, c->pasta_ifi, c->mtu); nl_link_up(1, c->pasta_ifi, c->mtu);
if (c->ifi4) { if (c->ifi4) {
@ -293,8 +291,11 @@ void pasta_ns_conf(struct ctx *c)
nl_addr_dup(c->ifi4, c->pasta_ifi, AF_INET); nl_addr_dup(c->ifi4, c->pasta_ifi, AF_INET);
} }
nl_route(op_routes, c->ifi4, c->pasta_ifi, AF_INET, if (c->no_copy_routes)
&c->ip4.gw); nl_route_set_def(c->pasta_ifi, AF_INET,
&c->ip4.gw);
else
nl_route_dup(c->ifi4, c->pasta_ifi, AF_INET);
} }
if (c->ifi6) { if (c->ifi6) {
@ -305,8 +306,11 @@ void pasta_ns_conf(struct ctx *c)
nl_addr_dup(c->ifi6, c->pasta_ifi, AF_INET6); nl_addr_dup(c->ifi6, c->pasta_ifi, AF_INET6);
} }
nl_route(op_routes, c->ifi6, c->pasta_ifi, AF_INET6, if (c->no_copy_routes)
&c->ip6.gw); nl_route_set_def(c->pasta_ifi, AF_INET6,
&c->ip6.gw);
else
nl_route_dup(c->ifi6, c->pasta_ifi, AF_INET6);
} }
} }