diff options
Diffstat (limited to 'src/libnm-platform/nm-linux-platform.c')
| -rw-r--r-- | src/libnm-platform/nm-linux-platform.c | 588 |
1 files changed, 438 insertions, 150 deletions
diff --git a/src/libnm-platform/nm-linux-platform.c b/src/libnm-platform/nm-linux-platform.c index b798d12d..d4ab36f9 100644 --- a/src/libnm-platform/nm-linux-platform.c +++ b/src/libnm-platform/nm-linux-platform.c @@ -222,6 +222,16 @@ G_STATIC_ASSERT(RTA_MAX == (__RTA_MAX - 1)); /*****************************************************************************/ +#define IFLA_VTI_UNSPEC 0 +#define IFLA_VTI_LINK 1 +#define IFLA_VTI_IKEY 2 +#define IFLA_VTI_OKEY 3 +#define IFLA_VTI_LOCAL 4 +#define IFLA_VTI_REMOTE 5 +#define IFLA_VTI_FWMARK 6 + +/*****************************************************************************/ + #define WG_CMD_GET_DEVICE 0 #define WG_CMD_SET_DEVICE 1 @@ -580,25 +590,29 @@ NM_LINUX_PLATFORM_FROM_PRIVATE(NMLinuxPlatformPrivate *priv) #define _NMLOG2(level, ...) _LOG(level, _NMLOG2_DOMAIN, NULL, __VA_ARGS__) #define _NMLOG2_err(errsv, level, ...) _LOG_err(errsv, level, _NMLOG2_DOMAIN, NULL, __VA_ARGS__) -#define _LOG_print(__level, __domain, __errsv, self, ...) \ - G_STMT_START \ - { \ - char __prefix[32]; \ - const char *__p_prefix = _NMLOG_PREFIX_NAME; \ - NMPlatform *const __self = (self); \ - \ - if (__self && nm_platform_get_log_with_ptr(__self)) { \ - g_snprintf(__prefix, sizeof(__prefix), "%s[%p]", _NMLOG_PREFIX_NAME, __self); \ - __p_prefix = __prefix; \ - } \ - _nm_log(__level, \ - __domain, \ - __errsv, \ - NULL, \ - NULL, \ - "%s: " _NM_UTILS_MACRO_FIRST(__VA_ARGS__), \ - __p_prefix _NM_UTILS_MACRO_REST(__VA_ARGS__)); \ - } \ +#define _LOG_print(__level, __domain, __errsv, self, ...) \ + G_STMT_START \ + { \ + char __prefix[64]; \ + const char *__p_prefix = _NMLOG_PREFIX_NAME; \ + NMPlatform *const __self = (self); \ + \ + if (__self && nm_platform_get_log_with_ptr(__self)) { \ + g_snprintf(__prefix, \ + sizeof(__prefix), \ + "%s[" NM_HASH_OBFUSCATE_PTR_FMT "]", \ + _NMLOG_PREFIX_NAME, \ + NM_HASH_OBFUSCATE_PTR(__self)); \ + __p_prefix = __prefix; \ + } \ + _nm_log(__level, \ + __domain, \ + __errsv, \ + NULL, \ + NULL, \ + "%s: " _NM_UTILS_MACRO_FIRST(__VA_ARGS__), \ + __p_prefix _NM_UTILS_MACRO_REST(__VA_ARGS__)); \ + } \ G_STMT_END #define _LOG(level, domain, self, ...) \ @@ -811,6 +825,8 @@ static const LinkDesc link_descs[] = { [NM_LINK_TYPE_VETH] = {"veth", "veth", NULL}, [NM_LINK_TYPE_VLAN] = {"vlan", "vlan", "vlan"}, [NM_LINK_TYPE_VRF] = {"vrf", "vrf", "vrf"}, + [NM_LINK_TYPE_VTI] = {"vti", "vti", NULL}, + [NM_LINK_TYPE_VTI6] = {"vti6", "vti6", NULL}, [NM_LINK_TYPE_VXLAN] = {"vxlan", "vxlan", "vxlan"}, [NM_LINK_TYPE_WIREGUARD] = {"wireguard", "wireguard", "wireguard"}, @@ -853,6 +869,8 @@ _link_type_from_rtnl_type(const char *name) NM_LINK_TYPE_VETH, /* "veth" */ NM_LINK_TYPE_VLAN, /* "vlan" */ NM_LINK_TYPE_VRF, /* "vrf" */ + NM_LINK_TYPE_VTI, /* "vti" */ + NM_LINK_TYPE_VTI6, /* "vti6" */ NM_LINK_TYPE_VXLAN, /* "vxlan" */ NM_LINK_TYPE_WIMAX, /* "wimax" */ NM_LINK_TYPE_WIREGUARD, /* "wireguard" */ @@ -1158,7 +1176,7 @@ _linktype_read_devtype(int dirfd) end = strpbrk(cont, "\r\n"); if (end) *end++ = '\0'; - if (strncmp(cont, DEVTYPE_PREFIX, NM_STRLEN(DEVTYPE_PREFIX)) == 0) { + if (NM_STR_HAS_PREFIX(cont, DEVTYPE_PREFIX)) { cont += NM_STRLEN(DEVTYPE_PREFIX); memmove(contents, cont, strlen(cont) + 1); return g_steal_pointer(&contents); @@ -1289,14 +1307,36 @@ _linktype_get_type(NMPlatform *platform, } if (arptype == ARPHRD_ETHER) { - /* Misc non-upstream WWAN drivers. rmnet is Qualcomm's proprietary - * modem interface, ccmni is MediaTek's. FIXME: these drivers should - * really set devtype=WWAN. - */ - if (g_str_has_prefix(ifname, "rmnet") || g_str_has_prefix(ifname, "rev_rmnet") - || g_str_has_prefix(ifname, "ccmni")) + /* The USB gadget interfaces behave and look like ordinary ethernet devices + * aside from the DEVTYPE. */ + if (nm_streq0(devtype, "gadget")) + return NM_LINK_TYPE_ETHERNET; + + /* Distributed Switch Architecture switch chips */ + if (nm_streq0(devtype, "dsa")) + return NM_LINK_TYPE_ETHERNET; + } + + /* Misc non-upstream WWAN drivers. rmnet is Qualcomm's proprietary + * modem interface, ccmni is MediaTek's. FIXME: these drivers should + * really set devtype=WWAN. + * + * Value "530" is the out-of-tree version of ARPHRD_RAWIP before it's + * merged in Linux 4.14. For the mainline version, this has the value + * of "519". + * + * [1] https://github.com/LineageOS/android_kernel_google_msm-4.9/commit/54948008c293fdf48552a5c39c91c09c3eb76ed2 + */ + if (NM_IN_SET(arptype, + ARPHRD_ETHER, + 519 /* ARPHRD_RAWIP */, + 530 /* out-of-tree ARPHRD_RAWIP */)) { + if (NM_STR_HAS_PREFIX(ifname, "rmnet") || NM_STR_HAS_PREFIX(ifname, "rev_rmnet") + || NM_STR_HAS_PREFIX(ifname, "ccmni")) return NM_LINK_TYPE_WWAN_NET; + } + if (arptype == ARPHRD_ETHER) { /* Standard wired ethernet interfaces don't report an rtnl_link_type, so * only allow fallback to Ethernet if no type is given. This should * prevent future virtual network drivers from being treated as Ethernet @@ -1304,15 +1344,6 @@ _linktype_get_type(NMPlatform *platform, */ if (!kind && !devtype) return NM_LINK_TYPE_ETHERNET; - - /* The USB gadget interfaces behave and look like ordinary ethernet devices - * aside from the DEVTYPE. */ - if (nm_streq0(devtype, "gadget")) - return NM_LINK_TYPE_ETHERNET; - - /* Distributed Switch Architecture switch chips */ - if (nm_streq0(devtype, "dsa")) - return NM_LINK_TYPE_ETHERNET; } } @@ -1338,7 +1369,8 @@ typedef struct { }; } ParseNlmsgIter; -#define NLMSG_TAIL(nmsg) ((struct rtattr *) (((char *) (nmsg)) + NLMSG_ALIGN((nmsg)->nlmsg_len))) +#define NLMSG_TAIL(nmsg) \ + NM_CAST_ALIGN(struct rtattr, ((char *) (nmsg)) + NLMSG_ALIGN((nmsg)->nlmsg_len)) /* copied from iproute2's addattr_l(). */ static gboolean @@ -2244,6 +2276,11 @@ _parse_lnk_vlan(const char *kind, struct nlattr *info_data) obj = nmp_object_new(NMP_OBJECT_TYPE_LNK_VLAN, NULL); obj->lnk_vlan.id = nla_get_u16(tb[IFLA_VLAN_ID]); + if (tb[IFLA_VLAN_PROTOCOL]) + obj->lnk_vlan.protocol = ntohs(nla_get_u16(tb[IFLA_VLAN_PROTOCOL])); + else + obj->lnk_vlan.protocol = ETH_P_8021Q; + if (tb[IFLA_VLAN_FLAGS]) { struct ifla_vlan_flags flags; @@ -2389,6 +2426,76 @@ _parse_lnk_vxlan(const char *kind, struct nlattr *info_data) } static NMPObject * +_parse_lnk_vti(const char *kind, struct nlattr *info_data) +{ + static const struct nla_policy policy[] = { + [IFLA_VTI_LINK] = {.type = NLA_U32}, + [IFLA_VTI_LOCAL] = {.type = NLA_U32}, + [IFLA_VTI_REMOTE] = {.type = NLA_U32}, + [IFLA_VTI_IKEY] = {.type = NLA_U32}, + [IFLA_VTI_OKEY] = {.type = NLA_U32}, + [IFLA_VTI_FWMARK] = {.type = NLA_U32}, + }; + struct nlattr *tb[G_N_ELEMENTS(policy)]; + NMPObject *obj; + NMPlatformLnkVti *props; + + if (!info_data || !nm_streq0(kind, "vti")) + return NULL; + + if (nla_parse_nested_arr(tb, info_data, policy) < 0) + return NULL; + + obj = nmp_object_new(NMP_OBJECT_TYPE_LNK_VTI, NULL); + props = &obj->lnk_vti; + + props->parent_ifindex = tb[IFLA_VTI_LINK] ? nla_get_u32(tb[IFLA_VTI_LINK]) : 0; + props->local = tb[IFLA_VTI_LOCAL] ? nla_get_u32(tb[IFLA_VTI_LOCAL]) : 0; + props->remote = tb[IFLA_VTI_REMOTE] ? nla_get_u32(tb[IFLA_VTI_REMOTE]) : 0; + props->ikey = tb[IFLA_VTI_IKEY] ? ntohl(nla_get_u32(tb[IFLA_VTI_IKEY])) : 0; + props->okey = tb[IFLA_VTI_OKEY] ? ntohl(nla_get_u32(tb[IFLA_VTI_OKEY])) : 0; + props->fwmark = tb[IFLA_VTI_FWMARK] ? nla_get_u32(tb[IFLA_VTI_FWMARK]) : 0; + + return obj; +} + +static NMPObject * +_parse_lnk_vti6(const char *kind, struct nlattr *info_data) +{ + static const struct nla_policy policy[] = { + [IFLA_VTI_LINK] = {.type = NLA_U32}, + [IFLA_VTI_LOCAL] = {.minlen = sizeof(struct in6_addr)}, + [IFLA_VTI_REMOTE] = {.minlen = sizeof(struct in6_addr)}, + [IFLA_VTI_IKEY] = {.type = NLA_U32}, + [IFLA_VTI_OKEY] = {.type = NLA_U32}, + [IFLA_VTI_FWMARK] = {.type = NLA_U32}, + }; + struct nlattr *tb[G_N_ELEMENTS(policy)]; + NMPObject *obj; + NMPlatformLnkVti6 *props; + + if (!info_data || !nm_streq0(kind, "vti6")) + return NULL; + + if (nla_parse_nested_arr(tb, info_data, policy) < 0) + return NULL; + + obj = nmp_object_new(NMP_OBJECT_TYPE_LNK_VTI6, NULL); + props = &obj->lnk_vti6; + + props->parent_ifindex = tb[IFLA_VTI_LINK] ? nla_get_u32(tb[IFLA_VTI_LINK]) : 0; + if (tb[IFLA_VTI_LOCAL]) + props->local = *nla_data_as(struct in6_addr, tb[IFLA_VTI_LOCAL]); + if (tb[IFLA_VTI_REMOTE]) + props->remote = *nla_data_as(struct in6_addr, tb[IFLA_VTI_REMOTE]); + props->ikey = tb[IFLA_VTI_IKEY] ? ntohl(nla_get_u32(tb[IFLA_VTI_IKEY])) : 0; + props->okey = tb[IFLA_VTI_OKEY] ? ntohl(nla_get_u32(tb[IFLA_VTI_OKEY])) : 0; + props->fwmark = tb[IFLA_VTI_FWMARK] ? nla_get_u32(tb[IFLA_VTI_FWMARK]) : 0; + + return obj; +} + +static NMPObject * _parse_lnk_vrf(const char *kind, struct nlattr *info_data) { static const struct nla_policy policy[] = { @@ -2540,16 +2647,15 @@ _wireguard_update_from_peers_nla(CList *peers, GArray **p_allowed_ips, struct nl GArray *allowed_ips = *p_allowed_ips; nla_for_each_nested (attr, tb[WGPEER_A_ALLOWEDIPS], rem) { + NMPWireGuardAllowedIP *new; + if (!allowed_ips) { allowed_ips = g_array_new(FALSE, FALSE, sizeof(NMPWireGuardAllowedIP)); *p_allowed_ips = allowed_ips; - g_array_set_size(allowed_ips, 1); - } else - g_array_set_size(allowed_ips, allowed_ips->len + 1); + } - if (!_wireguard_update_from_allowed_ips_nla( - &g_array_index(allowed_ips, NMPWireGuardAllowedIP, allowed_ips->len - 1), - attr)) { + new = nm_g_array_append_new(allowed_ips, NMPWireGuardAllowedIP); + if (!_wireguard_update_from_allowed_ips_nla(new, attr)) { /* we ignore the error of parsing one allowed-ip. */ g_array_set_size(allowed_ips, allowed_ips->len - 1); continue; @@ -3379,6 +3485,12 @@ _new_from_nl_link(NMPlatform *platform, case NM_LINK_TYPE_VRF: lnk_data = _parse_lnk_vrf(nl_info_kind, nl_info_data); break; + case NM_LINK_TYPE_VTI: + lnk_data = _parse_lnk_vti(nl_info_kind, nl_info_data); + break; + case NM_LINK_TYPE_VTI6: + lnk_data = _parse_lnk_vti6(nl_info_kind, nl_info_data); + break; case NM_LINK_TYPE_VXLAN: lnk_data = _parse_lnk_vxlan(nl_info_kind, nl_info_data); break; @@ -3631,21 +3743,27 @@ _new_from_nl_route(const struct nlmsghdr *nlh, gboolean id_only, ParseNlmsgIter struct { gboolean found; gboolean has_more; + guint8 weight; int ifindex; NMIPAddr gateway; } nh = { .found = FALSE, .has_more = FALSE, }; - guint32 mss; - guint32 window = 0; - guint32 cwnd = 0; - guint32 initcwnd = 0; - guint32 initrwnd = 0; - guint32 mtu = 0; - guint32 rto_min = 0; - guint32 lock = 0; - gboolean quickack = FALSE; + guint v4_n_nexthops = 0; + NMPlatformIP4RtNextHop v4_nh_extra_nexthops_stack[10]; + gs_free NMPlatformIP4RtNextHop *v4_nh_extra_nexthops_heap = NULL; + NMPlatformIP4RtNextHop *v4_nh_extra_nexthops = v4_nh_extra_nexthops_stack; + guint v4_nh_extra_alloc = G_N_ELEMENTS(v4_nh_extra_nexthops_stack); + guint32 mss; + guint32 window = 0; + guint32 cwnd = 0; + guint32 initcwnd = 0; + guint32 initrwnd = 0; + guint32 mtu = 0; + guint32 rto_min = 0; + guint32 lock = 0; + gboolean quickack = FALSE; nm_assert((parse_nlmsg_iter->iter_more && parse_nlmsg_iter->ip6_route.next_multihop > 0) || (!parse_nlmsg_iter->iter_more && parse_nlmsg_iter->ip6_route.next_multihop == 0)); @@ -3669,15 +3787,6 @@ _new_from_nl_route(const struct nlmsghdr *nlh, gboolean id_only, ParseNlmsgIter else return NULL; - if (!NM_IN_SET(rtm->rtm_type, - RTN_UNICAST, - RTN_LOCAL, - RTN_BLACKHOLE, - RTN_UNREACHABLE, - RTN_PROHIBIT, - RTN_THROW)) - return NULL; - if (nlmsg_parse_arr(nlh, sizeof(struct rtmsg), tb, policy) < 0) return NULL; @@ -3703,9 +3812,57 @@ _new_from_nl_route(const struct nlmsghdr *nlh, gboolean id_only, ParseNlmsgIter idx = 0; while (TRUE) { - if (idx == multihop_idx) { + if (nh.found && IS_IPv4) { + NMPlatformIP4RtNextHop *new_nexthop; + + /* we parsed the first IPv4 nexthop in "nh", let's parse the following ones. + * + * At this point, v4_n_nexthops still counts how many hops we already added, + * now we are about to add the (v4_n_nexthops+1) hop. + * + * Note that the first hop (of then v4_n_nexthops) is tracked in "nh". + * v4_nh_extra_nexthops tracks the additional hops. + * + * v4_nh_extra_alloc is how many space is allocated for + * v4_nh_extra_nexthops (note that in the end we will only add (v4_n_nexthops-1) + * hops in this list). */ + nm_assert(v4_n_nexthops > 0u); + if (v4_n_nexthops - 1u >= v4_nh_extra_alloc) { + v4_nh_extra_alloc = NM_MAX(4, v4_nh_extra_alloc * 2u); + if (!v4_nh_extra_nexthops_heap) { + v4_nh_extra_nexthops_heap = + g_new(NMPlatformIP4RtNextHop, v4_nh_extra_alloc); + memcpy(v4_nh_extra_nexthops_heap, + v4_nh_extra_nexthops_stack, + G_N_ELEMENTS(v4_nh_extra_nexthops_stack)); + } else { + v4_nh_extra_nexthops_heap = g_renew(NMPlatformIP4RtNextHop, + v4_nh_extra_nexthops_heap, + v4_nh_extra_alloc); + } + v4_nh_extra_nexthops = v4_nh_extra_nexthops_heap; + } + nm_assert(v4_n_nexthops - 1u < v4_nh_extra_alloc); + new_nexthop = &v4_nh_extra_nexthops[v4_n_nexthops - 1u]; + new_nexthop->ifindex = rtnh->rtnh_ifindex; + new_nexthop->weight = NM_MAX(((guint) rtnh->rtnh_hops) + 1u, 1u); + if (rtnh->rtnh_len > sizeof(*rtnh)) { + struct nlattr *ntb[RTA_MAX + 1]; + + if (nla_parse_arr(ntb, + (struct nlattr *) RTNH_DATA(rtnh), + rtnh->rtnh_len - sizeof(*rtnh), + NULL) + < 0) + return NULL; + + if (_check_addr_or_return_null(ntb, RTA_GATEWAY, addr_len)) + memcpy(&new_nexthop->gateway, nla_data(ntb[RTA_GATEWAY]), addr_len); + } + } else if (IS_IPv4 || idx == multihop_idx) { nh.found = TRUE; nh.ifindex = rtnh->rtnh_ifindex; + nh.weight = NM_MAX(((guint) rtnh->rtnh_hops) + 1u, 1u); if (rtnh->rtnh_len > sizeof(*rtnh)) { struct nlattr *ntb[RTA_MAX + 1]; @@ -3722,15 +3879,6 @@ _new_from_nl_route(const struct nlmsghdr *nlh, gboolean id_only, ParseNlmsgIter } else if (nh.found) { /* we just parsed a nexthop, but there is yet another hop afterwards. */ nm_assert(idx == multihop_idx + 1); - if (IS_IPv4) { - /* for IPv4, multihop routes are currently not supported. - * - * If we ever support them, then the next-hop list is part of the NMPlatformIPRoute, - * that is, for IPv4 we truly have multihop routes. Unlike for IPv6. - * - * For now, just error out. */ - return NULL; - } /* For IPv6 multihop routes, we need to remember to iterate again. * For each next-hop, we will create a distinct single-hop NMPlatformIP6Route. */ @@ -3738,6 +3886,9 @@ _new_from_nl_route(const struct nlmsghdr *nlh, gboolean id_only, ParseNlmsgIter break; } + if (IS_IPv4) + v4_n_nexthops++; + if (tlen < RTNH_ALIGN(rtnh->rtnh_len) + sizeof(*rtnh)) break; @@ -3771,6 +3922,9 @@ rta_multipath_done: nh.ifindex = ifindex; nh.gateway = gateway; nh.found = TRUE; + nm_assert(v4_n_nexthops == 0); + if (IS_IPv4) + v4_n_nexthops = 1; } else { /* Kernel supports new style nexthop configuration, * verify that it is a duplicate and ignore old-style nexthop. */ @@ -3861,6 +4015,23 @@ rta_multipath_done: obj->ip_route.ifindex = nh.ifindex; + if (IS_IPv4) { + nm_assert((!!nh.found) == (v4_n_nexthops > 0u)); + obj->ip4_route.n_nexthops = v4_n_nexthops; + if (v4_n_nexthops > 1) { + /* We only set the weight for multihop routes. I think that corresponds to what kernel + * does. The weight is mostly undefined for single-hop. */ + obj->ip4_route.weight = NM_MAX(nh.weight, 1u); + + obj->_ip4_route.extra_nexthops = + (v4_nh_extra_alloc == v4_n_nexthops - 1u + && v4_nh_extra_nexthops == v4_nh_extra_nexthops_heap) + ? g_steal_pointer(&v4_nh_extra_nexthops_heap) + : nm_memdup(v4_nh_extra_nexthops, + sizeof(v4_nh_extra_nexthops[0]) * (v4_n_nexthops - 1u)); + } + } + if (_check_addr_or_return_null(tb, RTA_DST, addr_len)) memcpy(obj->ip_route.network_ptr, nla_data(tb[RTA_DST]), addr_len); @@ -4580,11 +4751,13 @@ _nl_msg_new_link_set_linkinfo(struct nl_msg *msg, NMLinkType link_type, gconstpo const NMPlatformLnkVlan *props = extra_data; nm_assert(extra_data); + nm_assert(props->protocol != 0); if (!(data = nla_nest_start(msg, IFLA_INFO_DATA))) goto nla_put_failure; NLA_PUT_U16(msg, IFLA_VLAN_ID, props->id); + NLA_PUT_U16(msg, IFLA_VLAN_PROTOCOL, htons(props->protocol)); { struct ifla_vlan_flags flags = { @@ -4835,6 +5008,44 @@ _nl_msg_new_link_set_linkinfo(struct nl_msg *msg, NMLinkType link_type, gconstpo NLA_PUT_U16(msg, IFLA_MACVLAN_FLAGS, props->no_promisc ? MACVLAN_FLAG_NOPROMISC : 0); break; } + case NM_LINK_TYPE_VTI: + { + const NMPlatformLnkVti *props = extra_data; + + nm_assert(props); + + if (!(data = nla_nest_start(msg, IFLA_INFO_DATA))) + goto nla_put_failure; + + if (props->parent_ifindex > 0) + NLA_PUT_U32(msg, IFLA_VTI_LINK, props->parent_ifindex); + NLA_PUT_U32(msg, IFLA_VTI_LOCAL, props->local); + NLA_PUT_U32(msg, IFLA_VTI_REMOTE, props->remote); + NLA_PUT_U32(msg, IFLA_VTI_IKEY, htonl(props->ikey)); + NLA_PUT_U32(msg, IFLA_VTI_OKEY, htonl(props->okey)); + NLA_PUT_U32(msg, IFLA_VTI_FWMARK, props->fwmark); + break; + } + case NM_LINK_TYPE_VTI6: + { + const NMPlatformLnkVti6 *props = extra_data; + + nm_assert(props); + + if (!(data = nla_nest_start(msg, IFLA_INFO_DATA))) + goto nla_put_failure; + + if (props->parent_ifindex > 0) + NLA_PUT_U32(msg, IFLA_VTI_LINK, props->parent_ifindex); + if (!IN6_IS_ADDR_UNSPECIFIED(&props->local)) + NLA_PUT(msg, IFLA_VTI_LOCAL, sizeof(props->local), &props->local); + if (!IN6_IS_ADDR_UNSPECIFIED(&props->remote)) + NLA_PUT(msg, IFLA_VTI_REMOTE, sizeof(props->remote), &props->remote); + NLA_PUT_U32(msg, IFLA_VTI_IKEY, htonl(props->ikey)); + NLA_PUT_U32(msg, IFLA_VTI_OKEY, htonl(props->okey)); + NLA_PUT_U32(msg, IFLA_VTI_FWMARK, props->fwmark); + break; + } default: nm_assert(!extra_data); break; @@ -5088,7 +5299,7 @@ ip_route_get_lock_flag(const NMPlatformIPRoute *route) } static gboolean -ip_route_ignored_protocol(const NMPlatformIPRoute *route) +ip_route_is_alive(const NMPlatformIPRoute *route) { guint8 prot; @@ -5100,12 +5311,29 @@ ip_route_ignored_protocol(const NMPlatformIPRoute *route) nm_assert(nmp_utils_ip_config_source_from_rtprot(prot) == route->rt_source); - /* We ignore all routes outside a certain subest of rtm_protocol. NetworkManager - * itself wouldn't configure those, so they are always configured by somebody - * external. We thus ignore them to avoid the overhead that processing them brings. - * For example, the BGP daemon "bird" might configure a huge number of RTPROT_BIRD routes. */ + if (prot > RTPROT_STATIC && !NM_IN_SET(prot, RTPROT_DHCP, RTPROT_RA)) { + /* We ignore certain rtm_protocol, because NetworkManager would only ever + * configure certain protocols. Other routes are not configured by NetworkManager + * and we don't track them in the platform cache. + * + * This is to help with the performance overhead of a huge number of + * routes, for example with the bird BGP software, that adds routes + * with RTPROT_BIRD protocol. */ + return FALSE; + } - return prot > RTPROT_STATIC && !NM_IN_SET(prot, RTPROT_DHCP, RTPROT_RA); + if (!NM_IN_SET(nm_platform_route_type_uncoerce(route->type_coerced), + RTN_UNICAST, + RTN_LOCAL, + RTN_BLACKHOLE, + RTN_UNREACHABLE, + RTN_PROHIBIT, + RTN_THROW)) { + /* Certain route types are ignored and not placed into the cache. */ + return FALSE; + } + + return TRUE; } /* Copied and modified from libnl3's build_route_msg() and rtnl_route_build_msg(). */ @@ -5118,7 +5346,7 @@ _nl_msg_new_route(uint16_t nlmsg_type, uint16_t nlmsg_flags, const NMPObject *ob const guint32 lock = ip_route_get_lock_flag(NMP_OBJECT_CAST_IP_ROUTE(obj)); const guint32 table = nm_platform_route_table_uncoerce(NMP_OBJECT_CAST_IP_ROUTE(obj)->table_coerced, TRUE); - const struct rtmsg rtmsg = { + struct rtmsg rtmsg = { .rtm_family = klass->addr_family, .rtm_tos = IS_IPv4 ? obj->ip4_route.tos : 0, .rtm_table = table <= 0xFF ? table : RT_TABLE_UNSPEC, @@ -5126,7 +5354,7 @@ _nl_msg_new_route(uint16_t nlmsg_type, uint16_t nlmsg_flags, const NMPObject *ob .rtm_scope = IS_IPv4 ? nm_platform_route_scope_inv(obj->ip4_route.scope_inv) : RT_SCOPE_NOWHERE, .rtm_type = nm_platform_route_type_uncoerce(NMP_OBJECT_CAST_IP_ROUTE(obj)->type_coerced), - .rtm_flags = obj->ip_route.r_rtm_flags & ((unsigned) (RTNH_F_ONLINK)), + .rtm_flags = 0, .rtm_dst_len = obj->ip_route.plen, .rtm_src_len = IS_IPv4 ? 0 : NMP_OBJECT_CAST_IP6_ROUTE(obj)->src_plen, }; @@ -5137,6 +5365,17 @@ _nl_msg_new_route(uint16_t nlmsg_type, uint16_t nlmsg_flags, const NMPObject *ob NM_IN_SET(NMP_OBJECT_GET_TYPE(obj), NMP_OBJECT_TYPE_IP4_ROUTE, NMP_OBJECT_TYPE_IP6_ROUTE)); nm_assert(NM_IN_SET(nlmsg_type, RTM_NEWROUTE, RTM_DELROUTE)); + if (NM_FLAGS_HAS(obj->ip_route.r_rtm_flags, ((unsigned) (RTNH_F_ONLINK)))) { + if (IS_IPv4 && obj->ip4_route.gateway == 0) { + /* Kernel does not allow setting the onlink flag, if there is no gateway. + * We silently don't configure the flag. + * + * For multi-hop routes, we will set the flag for each next-hop (which + * has a gateway set). */ + } else + rtmsg.rtm_flags |= ((unsigned) RTNH_F_ONLINK); + } + msg = nlmsg_alloc_new(0, nlmsg_type, nlmsg_flags); if (nlmsg_append_struct(msg, &rtmsg) < 0) @@ -5171,6 +5410,54 @@ _nl_msg_new_route(uint16_t nlmsg_type, uint16_t nlmsg_flags, const NMPObject *ob NLA_PUT(msg, RTA_PREFSRC, addr_len, &obj->ip6_route.pref_src); } + if (IS_IPv4 && obj->ip4_route.n_nexthops > 1u) { + struct nlattr *multipath; + guint i; + + if (!(multipath = nla_nest_start(msg, RTA_MULTIPATH))) + goto nla_put_failure; + + for (i = 0u; i < obj->ip4_route.n_nexthops; i++) { + struct rtnexthop *rtnh; + in_addr_t gw; + + rtnh = nlmsg_reserve(msg, sizeof(*rtnh), NLMSG_ALIGNTO); + if (!rtnh) + goto nla_put_failure; + + /* For multihop routes, a valid weight must be in range 1-256 (on netlink's + * "rtnh_hops" this is 0-255). We allow that the caller leaves the value unset + * at zero. */ + if (i == 0u) { + rtnh->rtnh_hops = NM_MAX(obj->ip4_route.weight, 1u) - 1u; + rtnh->rtnh_ifindex = obj->ip4_route.ifindex; + gw = obj->ip4_route.gateway; + } else { + const NMPlatformIP4RtNextHop *n = &obj->_ip4_route.extra_nexthops[i - 1u]; + + rtnh->rtnh_hops = NM_MAX(n->weight, 1u) - 1u; + rtnh->rtnh_ifindex = n->ifindex; + gw = n->gateway; + } + NLA_PUT_U32(msg, RTA_GATEWAY, gw); + + rtnh->rtnh_flags = 0; + + if (obj->ip4_route.n_nexthops > 1 + && NM_FLAGS_HAS(obj->ip_route.r_rtm_flags, (unsigned) (RTNH_F_ONLINK)) && gw != 0) { + /* Unlike kernel, we only track the onlink flag per NMPlatformIP4Address, and + * not per nexthop. That is fine for NetworkManager configuring addresses. + * It is not fine for tracking addresses from kernel in platform cache, + * because the rtnh_flags of the nexthops need to be part of nmp_object_id_cmp(). */ + rtnh->rtnh_flags |= RTNH_F_ONLINK; + } + + rtnh->rtnh_len = (char *) nlmsg_tail(nlmsg_hdr(msg)) - (char *) rtnh; + } + + nla_nest_end(msg, multipath); + } + if (obj->ip_route.mss || obj->ip_route.window || obj->ip_route.cwnd || obj->ip_route.initcwnd || obj->ip_route.initrwnd || obj->ip_route.mtu || obj->ip_route.quickack || obj->ip_route.rto_min || lock) { @@ -5462,7 +5749,7 @@ _nl_msg_new_tfilter(uint16_t nlmsg_type, uint16_t nlmsg_flags, const NMPlatformT if (nm_streq(action->kind, NM_PLATFORM_ACTION_KIND_SIMPLE)) { const NMPlatformActionSimple *simple = &action->simple; struct tc_defact sel = { - 0, + 0, }; if (!(act_options = nla_nest_start(msg, TCA_ACT_OPTIONS))) @@ -5476,7 +5763,7 @@ _nl_msg_new_tfilter(uint16_t nlmsg_type, uint16_t nlmsg_flags, const NMPlatformT } else if (nm_streq(action->kind, NM_PLATFORM_ACTION_KIND_MIRRED)) { const NMPlatformActionMirred *mirred = &action->mirred; struct tc_mirred sel = { - 0, + 0, }; if (!(act_options = nla_nest_start(msg, TCA_ACT_OPTIONS))) @@ -5512,25 +5799,25 @@ nla_put_failure: /*****************************************************************************/ -#define ASSERT_SYSCTL_ARGS(pathid, dirfd, path) \ - G_STMT_START \ - { \ - const char *const _pathid = (pathid); \ - const int _dirfd = (dirfd); \ - const char *const _path = (path); \ - \ - nm_assert(_path &&_path[0]); \ - g_assert(!strstr(_path, "/../")); \ - if (_dirfd < 0) { \ - nm_assert(!_pathid); \ - nm_assert(_path[0] == '/'); \ - nm_assert(g_str_has_prefix(_path, "/proc/sys/") || g_str_has_prefix(_path, "/sys/") \ - || g_str_has_prefix(_path, "/proc/net")); \ - } else { \ - nm_assert(_pathid &&_pathid[0] && _pathid[0] != '/'); \ - nm_assert(_path[0] != '/'); \ - } \ - } \ +#define ASSERT_SYSCTL_ARGS(pathid, dirfd, path) \ + G_STMT_START \ + { \ + const char *const _pathid = (pathid); \ + const int _dirfd = (dirfd); \ + const char *const _path = (path); \ + \ + nm_assert(_path &&_path[0]); \ + g_assert(!strstr(_path, "/../")); \ + if (_dirfd < 0) { \ + nm_assert(!_pathid); \ + nm_assert(_path[0] == '/'); \ + nm_assert(NM_STR_HAS_PREFIX(_path, "/proc/sys/") || NM_STR_HAS_PREFIX(_path, "/sys/") \ + || NM_STR_HAS_PREFIX(_path, "/proc/net")); \ + } else { \ + nm_assert(_pathid &&_pathid[0] && _pathid[0] != '/'); \ + nm_assert(_path[0] != '/'); \ + } \ + } \ G_STMT_END /*****************************************************************************/ @@ -5599,12 +5886,14 @@ sysctl_set_internal(NMPlatform *platform, const char *path, const char *value) { - int fd, tries; + int fd; + int tries; gssize nwrote; gssize len; char *actual; gs_free char *actual_free = NULL; int errsv; + int r; if (dirfd < 0) { pathid = path; @@ -5698,18 +5987,13 @@ sysctl_set_internal(NMPlatform *platform, _LOGE("sysctl: failed to set '%s' to '%s' after three attempts", path, value); } - if (nwrote < len - 1) { - if (nm_close(fd) != 0) { - if (errsv != 0) - errno = errsv; - } else if (errsv != 0) - errno = errsv; - else - errno = EIO; - return FALSE; - } - if (nm_close(fd) != 0) { - /* errno is already properly set. */ + r = nm_close_with_error(fd); + if (r < 0 || nwrote < len - 1) { + if (errsv == 0) { + /* propagate the error from nm_close_with_error(). */ + errsv = (r < 0) ? -r : EIO; + } + errno = errsv; return FALSE; } @@ -6260,11 +6544,11 @@ static NM_UTILS_LOOKUP_STR_DEFINE( NM_UTILS_LOOKUP_ITEM_IGNORE(DELAYED_ACTION_TYPE_REFRESH_ALL_RTNL_ROUTING_RULES_ALL), NM_UTILS_LOOKUP_ITEM_IGNORE(__DELAYED_ACTION_TYPE_MAX), ); -#define delayed_action_get_list_wait_for_resonse(priv, netlink_protocol, idx) \ - (&g_array_index((priv)->delayed_action.list_wait_for_response_x[nmp_netlink_protocol_check( \ - (netlink_protocol))], \ - DelayedActionWaitForNlResponseData, \ - (idx))) +#define delayed_action_get_list_wait_for_resonse(priv, netlink_protocol, idx) \ + (&nm_g_array_index((priv)->delayed_action.list_wait_for_response_x[nmp_netlink_protocol_check( \ + (netlink_protocol))], \ + DelayedActionWaitForNlResponseData, \ + (idx))) static const char * delayed_action_to_string_full(DelayedActionType action_type, @@ -6753,6 +7037,14 @@ cache_prune_one_type(NMPlatform *platform, const NMPLookup *lookup) obj = main_entry->obj; + if (NMP_OBJECT_GET_TYPE(obj) == NMP_OBJECT_TYPE_IP6_ADDRESS) { + const NMPlatformIP6Address *pladdr = NMP_OBJECT_CAST_IP6_ADDRESS(obj); + + if (pladdr->n_ifa_flags & IFA_F_TENTATIVE) { + nm_platform_ip6_dadfailed_set(platform, pladdr->ifindex, &pladdr->address, TRUE); + } + } + _LOGt("cache-prune: prune %s", nmp_object_to_string(obj, NMP_OBJECT_TO_STRING_ALL, sbuf, sizeof(sbuf))); @@ -7568,6 +7860,7 @@ _rtnl_handle_msg(NMPlatform *platform, const struct nl_msg_lite *msg) gboolean resync_required = FALSE; gboolean only_dirty = FALSE; gboolean is_ipv6; + gboolean route_is_alive; /* IPv4 routes that are a response to RTM_GETROUTE must have * the cloned flag while IPv6 routes don't have to. */ @@ -7597,24 +7890,13 @@ _rtnl_handle_msg(NMPlatform *platform, const struct nl_msg_lite *msg) } } - if (ip_route_ignored_protocol(NMP_OBJECT_CAST_IP_ROUTE(obj))) { - /* We ignore certain rtm_protocol, because NetworkManager would only ever - * configure certain protocols. Other routes were not added by NetworkManager - * and we don't need to track them in the platform cache. - * - * This is to help with the performance overhead of a huge number of - * routes, for example with the bird BGP software, that adds routes - * with RTPROT_BIRD protocol. - * - * Even if this is a IPv6 multipath route, we abort (parse_nlmsg_iter). There - * is nothing for us to do. */ - return; - } + route_is_alive = ip_route_is_alive(NMP_OBJECT_CAST_IP_ROUTE(obj)); cache_op = nmp_cache_update_netlink_route(cache, obj, is_dump, msghdr->nlmsg_flags, + route_is_alive, &obj_old, &obj_new, &obj_replace, @@ -7661,11 +7943,21 @@ _rtnl_handle_msg(NMPlatform *platform, const struct nl_msg_lite *msg) delayed_action_schedule(platform, delayed_action_refresh_from_needle_object(obj), NULL); + /* We are done here. */ + return; } break; } case RTM_DELADDR: + if (NMP_OBJECT_GET_TYPE(obj) == NMP_OBJECT_TYPE_IP6_ADDRESS) { + const NMPlatformIP6Address *ip6 = NMP_OBJECT_CAST_IP6_ADDRESS(obj); + + if (ip6->n_ifa_flags & IFA_F_DADFAILED) { + nm_platform_ip6_dadfailed_set(platform, ip6->ifindex, &ip6->address, TRUE); + } + } + /* fall-through */ case RTM_DELLINK: case RTM_DELQDISC: case RTM_DELROUTE: @@ -7948,7 +8240,8 @@ retry: } else if (NM_IN_SET(seq_result, -ESRCH, -ENOENT)) { log_detail = ", firmware not found"; result = -NME_PL_NO_FIRMWARE; - } else if (NM_IN_SET(seq_result, -ERANGE) && change_link_type == CHANGE_LINK_TYPE_SET_MTU) { + } else if (NM_IN_SET(seq_result, -ERANGE, -EINVAL) + && change_link_type == CHANGE_LINK_TYPE_SET_MTU) { log_detail = ", setting MTU to requested size is not possible"; result = -NME_PL_CANT_SET_MTU; } else if (NM_IN_SET(seq_result, -ENFILE) && change_link_type == CHANGE_LINK_TYPE_SET_ADDRESS @@ -8126,7 +8419,7 @@ static gboolean link_set_token(NMPlatform *platform, int ifindex, const NMUtilsIPv6IfaceId *iid) { nm_auto_nlmsg struct nl_msg *nlmsg = NULL; - char sbuf[NM_UTILS_INET_ADDRSTRLEN]; + char sbuf[NM_INET_ADDRSTRLEN]; _LOGD("link: change %d: token: set IPv6 address generation token to %s", ifindex, @@ -8202,7 +8495,7 @@ link_set_address(NMPlatform *platform, int ifindex, gconstpointer address, size_ nm_auto_nlmsg struct nl_msg *nlmsg = NULL; const ChangeLinkData d = { .set_address = - { + { .address = address, .length = length, }, @@ -9310,25 +9603,15 @@ ip6_address_delete(NMPlatform *platform, int ifindex, struct in6_addr addr, guin /*****************************************************************************/ static int -ip_route_add(NMPlatform *platform, - NMPNlmFlags flags, - int addr_family, - const NMPlatformIPRoute *route) +ip_route_add(NMPlatform *platform, NMPNlmFlags flags, NMPObject *obj_stack) { nm_auto_nlmsg struct nl_msg *nlmsg = NULL; - NMPObject obj; - - nmp_object_stackinit(&obj, - NMP_OBJECT_TYPE_IP_ROUTE(NM_IS_IPv4(addr_family)), - (const NMPlatformObject *) route); - - nm_platform_ip_route_normalize(addr_family, NMP_OBJECT_CAST_IP_ROUTE(&obj)); - nlmsg = _nl_msg_new_route(RTM_NEWROUTE, flags & NMP_NLM_FLAG_FMASK, &obj); + nlmsg = _nl_msg_new_route(RTM_NEWROUTE, flags & NMP_NLM_FLAG_FMASK, obj_stack); if (!nlmsg) g_return_val_if_reached(-NME_BUG); return do_add_addrroute(platform, - &obj, + obj_stack, nlmsg, NM_FLAGS_HAS(flags, NMP_NLM_FLAG_SUPPRESS_NETLINK_FAILURE)); } @@ -9820,7 +10103,7 @@ continue_reading: goto stop; } - hdr = (struct nlmsghdr *) priv->netlink_recv_buf.buf; + hdr = NM_CAST_ALIGN(struct nlmsghdr, priv->netlink_recv_buf.buf); while (nlmsg_ok(hdr, n)) { WaitForNlResponseResult seq_result; gboolean process_valid_msg = FALSE; @@ -10588,8 +10871,8 @@ constructed(GObject *_object) : (!nmp_netns_get_current() ? "no netns support" : nm_sprintf_bufa(100, - "in netns[%p]%s", - nmp_netns_get_current(), + "in netns[" NM_HASH_OBFUSCATE_PTR_FMT "]%s", + NM_HASH_OBFUSCATE_PTR(nmp_netns_get_current()), nmp_netns_get_current() == nmp_netns_get_initial() ? "/main" : "")), nm_platform_get_use_udev(platform) ? "use" : "no", @@ -10732,7 +11015,10 @@ path_is_read_only_fs(const char *path) } NMPlatform * -nm_linux_platform_new(gboolean log_with_ptr, gboolean netns_support, gboolean cache_tc) +nm_linux_platform_new(NMDedupMultiIndex *multi_idx, + gboolean log_with_ptr, + gboolean netns_support, + gboolean cache_tc) { gboolean use_udev = FALSE; @@ -10740,6 +11026,8 @@ nm_linux_platform_new(gboolean log_with_ptr, gboolean netns_support, gboolean ca use_udev = TRUE; return g_object_new(NM_TYPE_LINUX_PLATFORM, + NM_PLATFORM_MULTI_IDX, + multi_idx, NM_PLATFORM_LOG_WITH_PTR, log_with_ptr, NM_PLATFORM_USE_UDEV, |