December 2020 - connman

[PATCH] Fix hang up when connecting a secure network with an empty passphrase

by Emmanuel Vautrin

Hello everyone, Please find below the patch fixing this issue: diff --git a/src/service.c b/src/service.c index 8d236749..629db995 100644 --- a/src/service.c +++ b/src/service.c @@ -4021,6 +4021,7 @@ static void disconnect_on_last_session(enum connman_service_type type) static int active_sessions[MAX_CONNMAN_SERVICE_TYPES] = {}; static int always_connect[MAX_CONNMAN_SERVICE_TYPES] = {}; static int active_count = 0; +static bool passphrase_requested = false; void __connman_service_set_active_session(bool enable, GSList *list) { @@ -4187,6 +4188,9 @@ static bool auto_connect_service(GList *services, continue; } + if (passphrase_requested) + return true; + if (service->pending || is_connecting(service->state) || is_connected(service->state)) { @@ -5721,6 +5725,7 @@ static void request_input_cb(struct connman_service *service, err = __connman_service_set_passphrase(service, passphrase); done: + passphrase_requested = false; if (err >= 0) { /* We forget any previous error. */ set_error(service, CONNMAN_SERVICE_ERROR_UNKNOWN); @@ -6664,6 +6669,9 @@ int __connman_service_connect(struct connman_service *service, if (service->hidden && err != -EINPROGRESS) service->pending = pending; + if (err == -EINPROGRESS) + passphrase_requested = true; + return err; } } -- Best Regards, Emmanuel

1 year, 1 month

3
2
0 / 0

[PATCH] Fix time synchronization after updating timeservers

by Emmanuel Vautrin

Hello everyone, Please find below the patch fixing this issue: --- src/timeserver.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/timeserver.c b/src/timeserver.c index 417f20d2..c574b36a 100644 --- a/src/timeserver.c +++ b/src/timeserver.c @@ -354,7 +354,7 @@ int __connman_timeserver_sync(struct connman_service *default_service) if (!service) return -EINVAL; - if (service == ts_service) + if (default_service && service == ts_service) return -EALREADY; if (!resolv) -- Best Regards, Emmanuel

1 year, 1 month

3
2
0 / 0

[PATCH 00/15] Support SplitRouting on vpnd, drop vpnd route management

by Jussi Laakkonen

SplitRouting is a VPN only setting that was managed by connmand and is also saved along the vpn_*/settings. This change exposes SplitRouting value also to vpnd, making it accessible over VPN D-Bus API as well. This allows to control the split routing value also on VPNs that are not connected yet. And being a VPN only value it seems natural to have it controllable over VPN D-Bus API. This does not affect the service moving, but extends it for internal use. In order to make this happen a lot of changes were required. First of all it is required to have synchronization of the value changes between connmand and vpnd. This is achieved by using service specific property changed signals sent always when the value changes, or is being set (e.g., loading the service), on both sides. This will result in vpnd informing about the change twice when the SplitRouting value is changed over VPN D-Bus API. This happens because VPN changes to connmand are propagated using property changed signals and after connmand has processed the SplitRouting signal connmand will notify the value again with a property changed signal to keep vpnd in sync in case of an error, for instance. Since service moving is utilized here for making the SplitRouting value changes on a connected VPN the route checks in plugins/vpn.c is enhanced to handle the routes better. If a VPN that is being set as a split routed does not have any other than the default route a network route will be attempted to be added. If that fails an error is printed and the resulting property changed signal will contain the current, not affected value of SplitRouting But if a VPN has a default route configured it will be prevented for SplitRouted VPNs. Also to make sure that a non-SplitRouted VPN will have a default route set it will be added if missing. And to make this happen inet.c was amended with new functionality to get route addresses. This resulted in some other changes to clean up the code and fix the issue of getifaddrs() not being able to interpret the destination address if the P-t-P connection has broadcast address also set. Combined many different getifaddrs() uses behind one function to have it a multi-purpose function for different uses. The added, inet.c internal get_interface_addresses() uses struct interface_address that allows AF specific addresses to be retrieved to their respective by copying the struct ifaddrs content that matches the given criteria. Also some minor change to service.c was required in order to propagate the SplitRouting value at all times. When loading a service it is imperative to apply the service settings after registering the D-Bus service because otherwise any signals cannot be sent about the loaded settings, if needed. This boils down to the fact that the internal gdbus implementation does not allow signals to be sent if the path given is not registered. Route management is also dropped from vpnd as a separate commit so it may be reverted easily if need arises. This was deemed to be unnecessary and confusing in discussions with Daniel Wagner on the ConnMan mailing list. Jussi Laakkonen (15): connection: Add getter for the phy index of a VPN transport service inet: Add function for detecting a default route inet: Refactor with getifaddrs() and add network route getter function inet: Do not add broadcast address for P2P/VPNs dbus: Report back the return value of g_dbus_send_message() service: Split service move functionality for internal use service: Expose set_split_routing() for internal use service: Add property changed signal for SplitRouting value service: Load and apply service settings after D-Bus registration provider: Add support for managing SplitRouting vpn: Support SplitRouting in D-Bus variables, improve route code vpn-config: Implement function to get boolean from keyfile vpn-provider: Drop route management from vpnd vpn-provider: Support SplitRouting option from connmand doc: Document VPN connection SplitRouting boolean doc/connman-vpn.8.in | 4 - doc/vpn-connection-api.txt | 9 +- include/inet.h | 10 +- include/ipaddress.h | 4 + include/provider.h | 4 + plugins/dundee.c | 2 +- plugins/ofono.c | 4 +- plugins/vpn.c | 228 +++++++++++-- src/bridge.c | 3 +- src/connection.c | 23 ++ src/connman.h | 10 +- src/dbus.c | 24 +- src/inet.c | 642 ++++++++++++++++++++++--------------- src/ipaddress.c | 10 + src/ipconfig.c | 42 ++- src/peer.c | 2 +- src/provider.c | 101 ++++++ src/service.c | 113 +++++-- src/tethering.c | 3 +- vpn/main.c | 5 +- vpn/plugins/l2tp.c | 2 + vpn/plugins/openconnect.c | 2 + vpn/plugins/openvpn.c | 1 + vpn/plugins/pptp.c | 1 + vpn/plugins/vpnc.c | 1 + vpn/plugins/wireguard.c | 2 + vpn/vpn-config.c | 15 + vpn/vpn-ipconfig.c | 8 +- vpn/vpn-provider.c | 276 +++++++++------- vpn/vpn-provider.h | 3 + vpn/vpn.h | 12 +- 31 files changed, 1096 insertions(+), 470 deletions(-) -- 2.20.1

1 year, 1 month

2
21
0 / 0

strange connection problem

by Thomas Green

I'm having an issue that I can't currently figure out. In my application I have two wireless access points that I've connected to previously. I've looked in their settings file, and they have the proper passwords in them. I'm currently attached to AP 1. I disconnect it, then attempt to connect to AP 2. I receive back to my registered agent a connection error, and the error that the service reports is Input/Output error. This is typically what I see when I enter a bad password when I first connect to the AP. So, I remove the service and try to connect to the original AP and I get the same Input/Output error on the AP that I was previously connected to. I've been trying to figure this out for days, but have not yet followed why this is happening. I've turned on debugging, and attached that log. If you could kindly help me to understand what's going on, I would appreciate it. TIA Tom

1 year, 1 month

1
0
0 / 0

[PATCH] vpn: Export vpn_ipconfig_foreach as linker symbol

by Daniel Wagner

Commit 95b25140bec7 ("vpn: Add WireGuard support") introduced a plugin dependency to __vpn_ipconfig_foreach. Because the linker does not export the prefixed functions, we need to rename it to vpn_ipconfig_foreach in order to export the linker symbol. --- vpn/plugins/wireguard.c | 2 +- vpn/vpn-ipconfig.c | 2 +- vpn/vpn-rtnl.c | 2 +- vpn/vpn.h | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) diff --git a/vpn/plugins/wireguard.c b/vpn/plugins/wireguard.c index a0449b36aa43..0a86acd04a0e 100644 --- a/vpn/plugins/wireguard.c +++ b/vpn/plugins/wireguard.c @@ -245,7 +245,7 @@ static char *get_ifname(void) for (i = 0; i < 256; i++) { data.ifname = g_strdup_printf("wg%d", i); data.found = false; - __vpn_ipconfig_foreach(ifname_check_cb, &data); + vpn_ipconfig_foreach(ifname_check_cb, &data); if (!data.found) return data.ifname; diff --git a/vpn/vpn-ipconfig.c b/vpn/vpn-ipconfig.c index 8b792c1b9929..825b43c40ff5 100644 --- a/vpn/vpn-ipconfig.c +++ b/vpn/vpn-ipconfig.c @@ -108,7 +108,7 @@ unsigned int __vpn_ipconfig_get_flags_from_index(int index) return ipdevice->flags; } -void __vpn_ipconfig_foreach(void (*function) (int index, +void vpn_ipconfig_foreach(void (*function) (int index, void *user_data), void *user_data) { GList *list, *keys; diff --git a/vpn/vpn-rtnl.c b/vpn/vpn-rtnl.c index 6ddfd8325252..295c05ce5363 100644 --- a/vpn/vpn-rtnl.c +++ b/vpn/vpn-rtnl.c @@ -184,7 +184,7 @@ int vpn_rtnl_register(struct vpn_rtnl *rtnl) rtnl_list = g_slist_insert_sorted(rtnl_list, rtnl, compare_priority); - __vpn_ipconfig_foreach(trigger_rtnl, rtnl); + vpn_ipconfig_foreach(trigger_rtnl, rtnl); return 0; } diff --git a/vpn/vpn.h b/vpn/vpn.h index fa968f602989..477cb2272c11 100644 --- a/vpn/vpn.h +++ b/vpn/vpn.h @@ -35,7 +35,7 @@ struct vpn_ipconfig; struct connman_ipaddress *__vpn_ipconfig_get_address(struct vpn_ipconfig *ipconfig); unsigned short __vpn_ipconfig_get_type_from_index(int index); unsigned int __vpn_ipconfig_get_flags_from_index(int index); -void __vpn_ipconfig_foreach(void (*function) (int index, +void vpn_ipconfig_foreach(void (*function) (int index, void *user_data), void *user_data); void __vpn_ipconfig_set_local(struct vpn_ipconfig *ipconfig, const char *address); -- 2.29.2

1 year, 1 month

2
2
0 / 0

[PATCH] src: Test return value of inet_pton consistently

by Daniel Wagner

inet_pton return 1 on success and 0 or -1 on failure depending on the error it wants to return. Use either '== 1' or '!= 1' consistently to test if the call was succesfull. ConnMan doesn't destinguish between 0 and -1. --- src/dhcpv6.c | 2 +- src/dns-systemd-resolved.c | 4 ++-- src/inet.c | 22 +++++++++++----------- src/ipaddress.c | 2 +- src/iptables.c | 4 ++-- 5 files changed, 17 insertions(+), 17 deletions(-) diff --git a/src/dhcpv6.c b/src/dhcpv6.c index 2d5f8f6a601e..8b683599753f 100644 --- a/src/dhcpv6.c +++ b/src/dhcpv6.c @@ -945,7 +945,7 @@ static void do_dad(GDHCPClient *dhcp_client, struct connman_dhcpv6 *dhcp) ref_own_address(user_data); - if (inet_pton(AF_INET6, address, &addr) < 0) { + if (inet_pton(AF_INET6, address, &addr) != 1) { DBG("Invalid IPv6 address %s %d/%s", address, -errno, strerror(errno)); goto fail; diff --git a/src/dns-systemd-resolved.c b/src/dns-systemd-resolved.c index 5fe306c3445d..912ab3fe8a42 100644 --- a/src/dns-systemd-resolved.c +++ b/src/dns-systemd-resolved.c @@ -106,7 +106,7 @@ static void setlinkdns_append(DBusMessageIter *iter, void *user_data) if (type == AF_INET) { result = inet_pton(type, server, ipv4_bytes); - if (!result) { + if (result != 1) { DBG("Failed to parse IPv4 address: %s", server); return; @@ -128,7 +128,7 @@ static void setlinkdns_append(DBusMessageIter *iter, void *user_data) &byte_array); } else if (type == AF_INET6) { result = inet_pton(type, server, ipv6_bytes); - if (!result) { + if (result != 1) { DBG("Failed to parse IPv6 address: %s", server); return; } diff --git a/src/inet.c b/src/inet.c index 2d30c3b6f6dc..df94d1eebaa7 100644 --- a/src/inet.c +++ b/src/inet.c @@ -121,11 +121,11 @@ int __connman_inet_modify_address(int cmd, int flags, ifaddrmsg->ifa_index = index; if (family == AF_INET) { - if (inet_pton(AF_INET, address, &ipv4_addr) < 1) + if (inet_pton(AF_INET, address, &ipv4_addr) != 1) return -1; if (peer) { - if (inet_pton(AF_INET, peer, &ipv4_dest) < 1) + if (inet_pton(AF_INET, peer, &ipv4_dest) != 1) return -1; err = __connman_inet_rtnl_addattr_l(header, @@ -165,7 +165,7 @@ int __connman_inet_modify_address(int cmd, int flags, return err; } } else if (family == AF_INET6) { - if (inet_pton(AF_INET6, address, &ipv6_addr) < 1) + if (inet_pton(AF_INET6, address, &ipv6_addr) != 1) return -1; err = __connman_inet_rtnl_addattr_l(header, @@ -707,7 +707,7 @@ int connman_inet_del_ipv6_network_route(int index, const char *host, rt.rtmsg_dst_len = prefix_len; - if (inet_pton(AF_INET6, host, &rt.rtmsg_dst) < 0) { + if (inet_pton(AF_INET6, host, &rt.rtmsg_dst) != 1) { err = -errno; goto out; } @@ -757,7 +757,7 @@ int connman_inet_add_ipv6_network_route(int index, const char *host, rt.rtmsg_dst_len = prefix_len; - if (inet_pton(AF_INET6, host, &rt.rtmsg_dst) < 0) { + if (inet_pton(AF_INET6, host, &rt.rtmsg_dst) != 1) { err = -errno; goto out; } @@ -773,7 +773,7 @@ int connman_inet_add_ipv6_network_route(int index, const char *host, */ if (gateway && !__connman_inet_is_any_addr(gateway, AF_INET6) && - inet_pton(AF_INET6, gateway, &rt.rtmsg_gateway) > 0) + inet_pton(AF_INET6, gateway, &rt.rtmsg_gateway) == 1) rt.rtmsg_flags |= RTF_GATEWAY; rt.rtmsg_metric = 1; @@ -816,7 +816,7 @@ int connman_inet_clear_ipv6_gateway_address(int index, const char *gateway) memset(&rt, 0, sizeof(rt)); - if (inet_pton(AF_INET6, gateway, &rt.rtmsg_gateway) < 0) { + if (inet_pton(AF_INET6, gateway, &rt.rtmsg_gateway) != 1) { err = -errno; goto out; } @@ -1255,7 +1255,7 @@ bool connman_inet_compare_subnet(int index, const char *host) if (!host) return false; - if (inet_pton(AF_INET, host, &haddr) <= 0) + if (inet_pton(AF_INET, host, &haddr) != 1) return false; if_addr.index = index; @@ -1292,7 +1292,7 @@ bool connman_inet_compare_ipv6_subnet(int index, const char *host) struct in6_addr imask = { 0 }; struct in6_addr haddr = { 0 }; - if (inet_pton(AF_INET6, host, &haddr) <= 0) + if (inet_pton(AF_INET6, host, &haddr) != 1) return false; addr.index = index; @@ -3179,7 +3179,7 @@ static int iproute_default_modify(int cmd, uint32_t table_id, int ifindex, ret = inet_pton(family, dst ? dst : gateway, buf); g_free(dst); - if (ret <= 0) + if (ret != 1) return -EINVAL; memset(&rth, 0, sizeof(rth)); @@ -3415,7 +3415,7 @@ static int get_nfs_server_ip(const char *cmdline_file, const char *pnp_file, addrstr[len] = '\0'; err = inet_pton(AF_INET, addrstr, addr); - if (err <= 0) { + if (err != 1) { connman_error("%s: Cannot convert to numeric addr \"%s\"\n", __func__, addrstr); err = -1; diff --git a/src/ipaddress.c b/src/ipaddress.c index f728eab319da..201d83454136 100644 --- a/src/ipaddress.c +++ b/src/ipaddress.c @@ -104,7 +104,7 @@ static bool check_ipv6_address(const char *address) return false; err = inet_pton(AF_INET6, address, buf); - if (err > 0) + if (err == 1) return true; return false; diff --git a/src/iptables.c b/src/iptables.c index 6feb00747324..664b27f13d86 100644 --- a/src/iptables.c +++ b/src/iptables.c @@ -2931,7 +2931,7 @@ static int parse_ip_and_mask(const char *str, struct in_addr *ip, if (!tokens) return -1; - if (!inet_pton(AF_INET, tokens[0], ip)) { + if (inet_pton(AF_INET, tokens[0], ip) != 1) { err = -1; goto out; } @@ -2972,7 +2972,7 @@ static int parse_ipv6_and_mask(const char *str, struct in6_addr *ip, if (!tokens) return -1; - if (!inet_pton(AF_INET6, tokens[0], ip)) { + if (inet_pton(AF_INET6, tokens[0], ip) != 1) { err = -1; goto out; } -- 2.29.2

1 year, 1 month

1
1
0 / 0

[PATCH] vpn: Do not do mixed declerations and code

by Daniel Wagner

gcc complains 'struct vpn_route' is declared within the code section. By declaring it at the begin of the function we can also avoid the additional helper variables. --- plugins/vpn.c | 32 ++++++++++++++------------------ 1 file changed, 14 insertions(+), 18 deletions(-) diff --git a/plugins/vpn.c b/plugins/vpn.c index 94d6e8591d88..3ed9c32268b4 100644 --- a/plugins/vpn.c +++ b/plugins/vpn.c @@ -1523,32 +1523,29 @@ static int save_route(GHashTable *routes, int family, const char *network, static int add_network_route(struct connection_data *data) { - char *network = NULL; - char *gateway = NULL; - char *netmask = NULL; - int family; + struct vpn_route rt = { 0, }; int err; if (!data) return -EINVAL; - family = connman_provider_get_family(data->provider); - switch (family) { + rt.family = connman_provider_get_family(data->provider); + switch (rt.family) { case PF_INET: - err = connman_inet_get_route_addresses(data->index, &network, - &netmask, &gateway); + err = connman_inet_get_route_addresses(data->index, + &rt.network, &rt.netmask, &rt.gateway); break; case PF_INET6: err = connman_inet_ipv6_get_route_addresses(data->index, - &network, &netmask, &gateway); + &rt.network, &rt.netmask, &rt.gateway); break; default: - connman_error("invalid protocol family %d", family); + connman_error("invalid protocol family %d", rt.family); return -EINVAL; } DBG("network %s gateway %s netmask %s for provider %p", - network, gateway, netmask, + rt.network, rt.gateway, rt.netmask, data->provider); if (err) { @@ -1557,21 +1554,20 @@ static int add_network_route(struct connection_data *data) goto out; } - err = save_route(data->server_routes, family, network, netmask, - gateway); + err = save_route(data->server_routes, rt.family, rt.network, rt.netmask, + rt.gateway); if (err) { connman_warn("failed to add network route for provider" "%p", data->provider); goto out; } - struct vpn_route route = { family, network, netmask, gateway }; - set_route(data, &route); + set_route(data, &rt); out: - g_free(network); - g_free(gateway); - g_free(netmask); + g_free(rt.network); + g_free(rt.netmask); + g_free(rt.gateway); return 0; } -- 2.29.2

1 year, 1 month

2
3
0 / 0

Re: gdhcp: Make DHCP client timeouts suspend aware

by Holesch, Simon (GED-SDD1)

On Thu, Dec 10, 2020 at 09:23, Hugo Portisch wrote: > Hi, > I found no other way to create a issue like on GitHub. > It's about this commit:connman/connman.git - Connection Manager Hi, I guess this is the best way to reach out. > Why DHCP timeout alarm timer should run in suspend/power off status?? The issue this commit was trying to solve is, that the client kept the IP address even after the lease expired. The timer that was previously used, stops completely during suspend and resumes after wakeup. This behavior could cause the DHCP server to assign the same IP address again and you'd have two clients with the same address. Here are the relevant parts of RFC 2131: > 4.4.5 Reacquisition and expiration > If no DHCPACK arrives before time T2, the client moves to REBINDING > state and sends (via broadcast) a DHCPREQUEST message to extend its > lease. > [...] > If the lease expires before the client receives a DHCPACK, the client > moves to INIT state, MUST immediately stop any other network > processing and requests network initialization parameters as if the > client were uninitialized. One option would be to use CLOCK_BOOTTIME (instead of CLOCK_BOOTTIME_ALARM, see timer_create(2)), which continues to run during suspend, but doesn't wake up the system. But I'm not aware of a way to prevent the client to use the IP address right after wakeup. That's why I used the CLOCK_BOOTTIME_ALARM clock. > When a device is sleeping and or powered off it should remain sleeping and not wake-up by a DHCP timeout.Therefor is a WOL package to wake-up the system over network. There is no reason to request DHCP updates when sleeping! > We have now the issue if our users (+20k) do have a device with a RTC assembled he can not power off or send the device to suspend anymore.These alarm timers for DHCP wake up the device within 1-5 minutes. > This behavior is not acceptable! There is and was never a reason why keeping the timeout timer running when the device is in a low power mode. > The option to remove CAP_WAKE_ALARM in connman.service does not solve the issue. It even make it worse as there are no DHCP updates request send at all. Yes, without the capability, the client isn't allowed to create the CLOCK_BOOTTIME_ALARM timer. > So only way is to revert this commit. Another idea, not sure if acceptable: Can the client release the IP address before going to low power mode? Alternatively we could use CLOCK_BOOTTIME to not wake up, but this isn't completely conformant with the RFC. > Please feel free to add a issue on your tracker https://01.org/jira/browse/CM as I am not able to register and login! > > Thank you very much,Team CoreELECPortisch >

1 year, 1 month

3
5
0 / 0

[PATCH] wispr: check service before stopping portal detection

by Sergey Matyukevich

Do not stop wispr unconditionally. Make sure that wispr matches service before stopping it. Otherwise removal of unrelevant service may cancel ongoing online check. For instance, this is the case for WiFi networks, since their services share the same index equal to network device index. So when unused WiFi services are removed, online check for active WiFi service is canceled before completion and not rescheduled. --- src/wispr.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/src/wispr.c b/src/wispr.c index 41157580..138d9014 100644 --- a/src/wispr.c +++ b/src/wispr.c @@ -991,6 +991,7 @@ int __connman_wispr_start(struct connman_service *service, void __connman_wispr_stop(struct connman_service *service) { + struct connman_wispr_portal *wispr_portal; int index; DBG("service %p", service); @@ -1002,7 +1003,22 @@ void __connman_wispr_stop(struct connman_service *service) if (index < 0) return; - g_hash_table_remove(wispr_portal_list, GINT_TO_POINTER(index)); + wispr_portal = g_hash_table_lookup(wispr_portal_list, + GINT_TO_POINTER(index)); + if (!wispr_portal) + return; + + if (wispr_portal->ipv4_context) { + if (service == wispr_portal->ipv4_context->service) { + g_hash_table_remove(wispr_portal_list, GINT_TO_POINTER(index)); + } + } + + if (wispr_portal->ipv6_context) { + if (service == wispr_portal->ipv6_context->service) { + g_hash_table_remove(wispr_portal_list, GINT_TO_POINTER(index)); + } + } } int __connman_wispr_init(void) -- 2.29.2

1 year, 1 month

2
1
0 / 0

Re: [PATCH] Bugfixes

by Daniel Wagner

Hi Boleslaw, On Mon, Dec 07, 2020 at 09:06:26AM +0000, Boleslaw Tokarski wrote: > Hi, > > Bugfix patches along with their desciprions attached, Good job fixing those bugs. I've removed the JB reference in the commit message, as they don't have any meaning upstream. Also please consider to send the patches with 'git send-email' :) Anyway, all patches applied. Thanks, Daniel

1 year, 1 month

2
1
0 / 0

2022

2021

2020

2019

2018

2017

2016

2015

connman December 2020