Re: Connman crashes when loosing network
Hi Marcel, Your patch works but that basically skips all multi line headers, so I think you have some coding to do;-) If you want to support multi line headers of course. Sure, the patch was just to verify the crash origin. There is indeed a bit of work to get it right now. Thanks for reporting this bug Tomasz ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
Re: Connman crashes when loosing network
Hi Marcel, connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] HTMLmeta http-equiv=Cache-Control content=no-cache connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] meta http-equiv=Pragma content=no-cache connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] meta http-equiv=Expires content=0 connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] meta http-equiv=Content-Type content= connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] text/html; charset=iso-8859-1 connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] script language=JavaScript type=text/javascript connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] function go_now(){ window.location.href = connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header]http://www.mywifiext.net/index.htm;} Program received signal SIGSEGV, Segmentation fault. 0x45a74f60 in g_str_hash () from /usr/lib/libglib-2.0.so.0 This is without the patch I sent I guess. g_str_hash and g_str_equal are not NULL-proof thus this crash here. Can you retry with the patch I sent? Thanks, Tomasz ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
RE: Connman crashes when loosing network
Hi Tomasz, This is without the patch indeed. Your patch works but that basically skips all multi line headers, so I think you have some coding to do ;-) If you want to support multi line headers of course. My connman is not crashing anymore. Thanks for the replies. Marcel -Original Message- From: connman [mailto:connman-boun...@connman.net] On Behalf Of Tomasz Bursztyka Sent: vrijdag 16 mei 2014 9:19 To: connman@connman.net Subject: Re: Connman crashes when loosing network Hi Marcel, connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] HTMLmeta http-equiv=Cache-Control content=no-cache connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] meta http-equiv=Pragma content=no-cache connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] meta http-equiv=Expires content=0 connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] meta http-equiv=Content-Type content= connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] text/html; charset=iso-8859-1 connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] script language=JavaScript type=text/javascript connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header] function go_now(){ window.location.href = connmand[30944]: WEB: gweb/gweb.c:received_data() web 0xd5ae0 [header]http://www.mywifiext.net/index.htm;} Program received signal SIGSEGV, Segmentation fault. 0x45a74f60 in g_str_hash () from /usr/lib/libglib-2.0.so.0 This is without the patch I sent I guess. g_str_hash and g_str_equal are not NULL-proof thus this crash here. Can you retry with the patch I sent? Thanks, Tomasz ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
RE: Connman crashes when loosing network
Hi, Does nobody know what is going when connman does an abort with signal 11? It is reproducible en a very simple setup. It is a connection via a WiFi extender from which the wireless connection is lost. I made a new build with HEAD e96a2cb241debd7a65971d30f28e903b9a1da49b but that does not change anything. Help is much appreciated. Cheers! May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {RX} 73 packets 10400 bytes May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {TX} 125 packets 19213 bytes May 12 07:40:20 CCGX user.info kernel: [ 204.764953] libphy: davinci_mdio-0:00 - Link is Down May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {update} flags 36867 UP May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {newlink} index 3 operstate 2 DOWN May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {del} address 192.168.0.101/24 label eth0 May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {del} route 192.168.0.0 gw 0.0.0.0 scope 253 LINK May 12 07:40:30 CCGX daemon.info connmand[1123]: eth0 {RX} 73 packets 10400 bytes May 12 07:40:30 CCGX user.info kernel: [ 214.767120] libphy: davinci_mdio-0:00 - Link is Up - 100/Full May 12 07:40:30 CCGX daemon.info connmand[1123]: eth0 {TX} 125 packets 19213 bytes May 12 07:40:30 CCGX daemon.info connmand[1123]: eth0 {update} flags 102467 UP,RUNNING,LOWER_UP May 12 07:40:30 CCGX daemon.info connmand[1123]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 May 12 07:40:30 CCGX daemon.info connmand[1123]: eth0 {newlink} index 3 operstate 6 UP May 12 07:40:30 CCGX daemon.warn connmand[1123]: Skipping disconnect of carrier, network is connecting. May 12 07:40:34 CCGX daemon.info connmand[1123]: eth0 {add} address 192.168.0.128/24 label eth0 family 2 May 12 07:40:34 CCGX daemon.info connmand[1123]: eth0 {add} route 192.168.0.0 gw 0.0.0.0 scope 253 LINK May 12 07:40:34 CCGX daemon.info connmand[1123]: eth0 {add} route 192.168.0.103 gw 0.0.0.0 scope 253 LINK May 12 07:40:34 CCGX daemon.info connmand[1123]: eth0 {add} route 0.0.0.0 gw 192.168.0.103 scope 0 UNIVERSE May 12 07:40:34 CCGX daemon.info connmand[1123]: eth0 {add} route 192.168.0.103 gw 192.168.0.103 scope 0 UNIVERSE May 12 07:40:34 CCGX daemon.err connmand[1123]: Aborting (signal 11) [/usr/sbin/connmand] -Original Message- From: connman [mailto:connman-boun...@connman.net] On Behalf Of Marcel Mulder Sent: vrijdag 2 mei 2014 10:00 To: connman@connman.net Subject: RE: Connman crashes when loosing network Hi Patrik, Thanks for the reply! I made a build with dd96d41ea182f0d1e0b55686adf793fa72e0e4d3 and started it with 'CONNMAN_DHCP_DEBUG=1 connmand -n -d src/dhcp.c' Connman now stops with an abort signal 11 which I also have seen before. Depending on what I do and how things are going (timing I presume) I see signals 11 or seg faults. The log of this session is below starting from the beginning when connman starts. Regards, Marcel root@CCGX:~# CONNMAN_DHCP_DEBUG=1 connmand -n -d src/dhcp.c connmand[1602]: Connection Manager version 1.23 connmand[1602]: iptables support missing error 2 (No such file or directory) connmand[1602]: Failed to flush table 'mangle': Invalid argument connmand[1602]: No IPv6 support connmand[1602]: No IPv6 support connmand[1602]: iptables support missing error 2 (No such file or directory) connmand[1602]: Failed to install iptables rules: Invalid argument connmand[1602]: Checking loopback interface settings connmand[1602]: System hostname is CCGX connmand[1602]: System domainname is almere.victronenergy.local connmand[1602]: src/dhcp.c:__connman_dhcp_init() connmand[1602]: Failed to open RFKILL control device connmand[1602]: lo {newlink} index 1 operstate 0 UNKNOWN connmand[1602]: can0 {newlink} index 2 operstate 0 UNKNOWN connmand[1602]: eth0 {create} index 3 type 1 ETHER connmand[1602]: eth0 {RX} 312 packets 31884 bytes connmand[1602]: eth0 {TX} 133 packets 19049 bytes connmand[1602]: eth0 {update} flags 36866 DOWN connmand[1602]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 connmand[1602]: eth0 {newlink} index 3 operstate 2 DOWN connmand[1602]: Adding interface eth0 [ ethernet ] connmand[1602]: eth0 {RX} 312 packets 31884 bytes connmand[1602]: eth0 {TX} 133 packets 19049 bytes connmand[1602]: eth0 {update} flags 36867 UP connmand[1602]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 connmand[1602]: eth0 {newlink} index 3 operstate 2 DOWN connmand[1602]: eth0 {RX} 312 packets 31884 bytes connmand[1602]: eth0 {TX} 133 packets 19049 bytes connmand[1602]: eth0 {update} flags 102467 UP,RUNNING,LOWER_UP connmand[1602]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 connmand[1602]: eth0 {newlink} index 3 operstate 6 UP connmand[1602]: Skipping disconnect of carrier, network is connecting. connmand[1602]: src/dhcp.c:__connman_dhcp_start() connmand[1602]: src/dhcp.c:dhcp_request() dhcp
RE: Connman crashes when loosing network
Hi Il 15/mag/2014 08:15 Marcel Mulder mmul...@victronenergy.com ha scritto: Hi, Does nobody know what is going when connman does an abort with signal 11? It is reproducible en a very simple setup. It is a connection via a WiFi extender from which the wireless connection is lost. I made a new build with HEAD e96a2cb241debd7a65971d30f28e903b9a1da49b but that does not change anything. Help is much appreciated. You can try to connect gdb and give a backtrace . we have done some testing but we don't have this crashing. Michael Cheers! May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {RX} 73 packets 10400 bytes May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {TX} 125 packets 19213 bytes May 12 07:40:20 CCGX user.info kernel: [ 204.764953] libphy: davinci_mdio-0:00 - Link is Down May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {update} flags 36867 UP May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {newlink} index 3 operstate 2 DOWN May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {del} address 192.168.0.101/24 label eth0 May 12 07:40:20 CCGX daemon.info connmand[1123]: eth0 {del} route 192.168.0.0 gw 0.0.0.0 scope 253 LINK May 12 07:40:30 CCGX daemon.info connmand[1123]: eth0 {RX} 73 packets 10400 bytes May 12 07:40:30 CCGX user.info kernel: [ 214.767120] libphy: davinci_mdio-0:00 - Link is Up - 100/Full May 12 07:40:30 CCGX daemon.info connmand[1123]: eth0 {TX} 125 packets 19213 bytes May 12 07:40:30 CCGX daemon.info connmand[1123]: eth0 {update} flags 102467 UP,RUNNING,LOWER_UP May 12 07:40:30 CCGX daemon.info connmand[1123]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 May 12 07:40:30 CCGX daemon.info connmand[1123]: eth0 {newlink} index 3 operstate 6 UP May 12 07:40:30 CCGX daemon.warn connmand[1123]: Skipping disconnect of carrier, network is connecting. May 12 07:40:34 CCGX daemon.info connmand[1123]: eth0 {add} address 192.168.0.128/24 label eth0 family 2 May 12 07:40:34 CCGX daemon.info connmand[1123]: eth0 {add} route 192.168.0.0 gw 0.0.0.0 scope 253 LINK May 12 07:40:34 CCGX daemon.info connmand[1123]: eth0 {add} route 192.168.0.103 gw 0.0.0.0 scope 253 LINK May 12 07:40:34 CCGX daemon.info connmand[1123]: eth0 {add} route 0.0.0.0 gw 192.168.0.103 scope 0 UNIVERSE May 12 07:40:34 CCGX daemon.info connmand[1123]: eth0 {add} route 192.168.0.103 gw 192.168.0.103 scope 0 UNIVERSE May 12 07:40:34 CCGX daemon.err connmand[1123]: Aborting (signal 11) [/usr/sbin/connmand] -Original Message- From: connman [mailto:connman-boun...@connman.net] On Behalf Of Marcel Mulder Sent: vrijdag 2 mei 2014 10:00 To: connman@connman.net Subject: RE: Connman crashes when loosing network Hi Patrik, Thanks for the reply! I made a build with dd96d41ea182f0d1e0b55686adf793fa72e0e4d3 and started it with 'CONNMAN_DHCP_DEBUG=1 connmand -n -d src/dhcp.c' Connman now stops with an abort signal 11 which I also have seen before. Depending on what I do and how things are going (timing I presume) I see signals 11 or seg faults. The log of this session is below starting from the beginning when connman starts. Regards, Marcel root@CCGX:~# CONNMAN_DHCP_DEBUG=1 connmand -n -d src/dhcp.c connmand[1602]: Connection Manager version 1.23 connmand[1602]: iptables support missing error 2 (No such file or directory) connmand[1602]: Failed to flush table 'mangle': Invalid argument connmand[1602]: No IPv6 support connmand[1602]: No IPv6 support connmand[1602]: iptables support missing error 2 (No such file or directory) connmand[1602]: Failed to install iptables rules: Invalid argument connmand[1602]: Checking loopback interface settings connmand[1602]: System hostname is CCGX connmand[1602]: System domainname is almere.victronenergy.local connmand[1602]: src/dhcp.c:__connman_dhcp_init() connmand[1602]: Failed to open RFKILL control device connmand[1602]: lo {newlink} index 1 operstate 0 UNKNOWN connmand[1602]: can0 {newlink} index 2 operstate 0 UNKNOWN connmand[1602]: eth0 {create} index 3 type 1 ETHER connmand[1602]: eth0 {RX} 312 packets 31884 bytes connmand[1602]: eth0 {TX} 133 packets 19049 bytes connmand[1602]: eth0 {update} flags 36866 DOWN connmand[1602]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 connmand[1602]: eth0 {newlink} index 3 operstate 2 DOWN connmand[1602]: Adding interface eth0 [ ethernet ] connmand[1602]: eth0 {RX} 312 packets 31884 bytes connmand[1602]: eth0 {TX} 133 packets 19049 bytes connmand[1602]: eth0 {update} flags 36867 UP connmand[1602]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 connmand[1602]: eth0 {newlink} index 3 operstate 2 DOWN connmand[1602]: eth0 {RX} 312 packets 31884 bytes connmand[1602]: eth0 {TX} 133 packets 19049 bytes connmand[1602]: eth0 {update} flags 102467 UP,RUNNING,LOWER_UP connmand[1602]: eth0 {newlink} index 3 address 50:56:63:1D:05
RE: Connman crashes when loosing network
Hi Michael, I did not do any changes. It is a git clone/pull and cross compiled for ARM. In gweb.c DEFAULT_BUFFER_SIZE is defined as: #define DEFAULT_BUFFER_SIZE 2048 Marcel -Original Message- From: connman [mailto:connman-boun...@connman.net] On Behalf Of Michael Trimarchi Sent: donderdag 15 mei 2014 10:25 To: connman@connman.net Subject: Re: Connman crashes when loosing network Hi On Thu, May 15, 2014 at 10:02 AM, Marcel Mulder Do you have a local change? DEFAULT_BUFFER_SIZE should be 2048 bytes and I think but I'm not sure that g_io_channel_read_chars () should be read up to this. Michael ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
RE: Connman crashes when loosing network
Hi, How can I disable the gweb stuff? I also disabled wispr but it is stil in the log! It is not a fix using connman without gweb but I will have a workaround for the moment. Marcel -Original Message- From: connman [mailto:connman-boun...@connman.net] On Behalf Of Marcel Mulder Sent: donderdag 15 mei 2014 10:38 To: connman@connman.net Subject: RE: Connman crashes when loosing network Hi Michael, I did not do any changes. It is a git clone/pull and cross compiled for ARM. In gweb.c DEFAULT_BUFFER_SIZE is defined as: #define DEFAULT_BUFFER_SIZE 2048 Marcel -Original Message- From: connman [mailto:connman-boun...@connman.net] On Behalf Of Michael Trimarchi Sent: donderdag 15 mei 2014 10:25 To: connman@connman.net Subject: Re: Connman crashes when loosing network Hi On Thu, May 15, 2014 at 10:02 AM, Marcel Mulder Do you have a local change? DEFAULT_BUFFER_SIZE should be 2048 bytes and I think but I'm not sure that g_io_channel_read_chars () should be read up to this. Michael ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
RE: Connman crashes when loosing network
Hi, I narrowed the problem down to this. In function handle_multi_line (gweb.c:809) a call is made to g_hash_table_lookup but session-result.last_key is NULL. (gdb) print session-result.last_key $1 = (gchar *) 0x0 Is there anything I can do to help you guys further? Marcel -Original Message- From: connman [mailto:connman-boun...@connman.net] On Behalf Of Marcel Mulder Sent: donderdag 1 mei 2014 10:06 To: connman@connman.net Subject: Connman crashes when loosing network Hi All, I experience reproducible connman crashes with version v1.20, v1.21, v1.22, v1.23 and HEAD of git. Older version I did not try. The situation is as follow: I have a controller (AM3547 from TI, ARM Cortex-A8) attached to a Wifi extender by wire. The Wifi extender is connected to a Wifi access point which is connect to our network. If the connection are up and running everything works well. When I however turn off the Wifi access point connman crashes with a segmentation fault (see log below). I have seen more connman crashes when networks are coming and going but there were not so nicely reproducible as this one. For cross compiling I use Yocto-Poky (OpenEmbedded). If you need more information, e.g. Wireshark logs or other kind of information I am happy to provide this. Connman was started with: connmand -d -n --nobacktrace -r But also with dnsproxy connman crashes. When I restart connman in this situation it ends in a seg fault again. When the connection to the Wifi access point is restored connman starts without problems. Hope someone can help me because this is a serious problem for us. Best regards, Marcel 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv4() 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv4config() 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv6config() 05/01/14 09:56:37 AM connmand[26030]: src/service.c:append_nameservers() servers[0] 192.168.4.13 available 1 05/01/14 09:56:37 AM connmand[26030]: src/service.c:append_nameservers() servers[1] 192.168.1.26 available 1 05/01/14 09:56:37 AM connmand[26030]: src/storage.c:storage_load() Loading /var/lib/connman/settings 05/01/14 09:56:45 AM connmand[26030]: src/rtnl.c:rtnl_message() buf 0xbea8ab18 len 576 05/01/14 09:56:45 AM connmand[26030]: src/rtnl.c:rtnl_message() NEWLINK len 576 type 16 flags 0x seq 0 pid 0 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_newlink() index 3 05/01/14 09:56:45 AM connmand[26030]: eth0 {RX} 38268 packets 4322051 bytes 05/01/14 09:56:45 AM connmand[26030]: eth0 {TX} 31213 packets 3869059 bytes 05/01/14 09:56:45 AM connmand[26030]: src/service.c:stats_update() service 0xad800 05/01/14 09:56:45 AM connmand[26030]: eth0 {update} flags 36867 UP 05/01/14 09:56:45 AM connmand[26030]: src/service.c:service_lower_down() eth0 lower down 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_disable() ipconfig 0xac3e0 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_unref_debug() 0xac3e0 ref 1 by src/ipconfig.c:1694:__connman_ipconfig_disable() 05/01/14 09:56:45 AM connmand[26030]: src/service.c:stats_stop() service 0xad800 05/01/14 09:56:45 AM connmand[26030]: src/service.c:service_save() service 0xad800 new 0 05/01/14 09:56:45 AM connmand[26030]: src/storage.c:storage_load() Loading /var/lib/connman/ethernet_5056631d05ad_cable/settings 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_save() ipconfig 0xac3e0 identifier ethernet_5056631d05ad_cable method dhcp 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_save() ipconfig 0xa9fd8 identifier ethernet_5056631d05ad_cable method off 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_lower_down() ipconfig ipv4 (nil) ipv6 0xa9fd8 05/01/14 09:56:45 AM connmand[26030]: src/inet.c:connman_inet_clear_ipv6_address() index 3 address (null) prefix_len 0 05/01/14 09:56:45 AM connmand[26030]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 05/01/14 09:56:45 AM connmand[26030]: eth0 {newlink} index 3 operstate 2 DOWN 05/01/14 09:56:45 AM connmand[26030]: src/detect.c:detect_newlink() type 1 index 3 05/01/14 09:56:45 AM connmand[26030]: plugins/ethernet.c:ethernet_newlink() index 3 flags 36867 change 0 05/01/14 09:56:45 AM connmand[26030]: plugins/ethernet.c:ethernet_newlink() carrier off 05/01/14 09:56:45 AM connmand[26030]: src/device.c:connman_device_remove_network() device 0xacf88 network 0xaaf10 05/01/14 09:56:45 AM connmand[26030
Re: Connman crashes when loosing network
Hi Marcel, I narrowed the problem down to this. In function handle_multi_line (gweb.c:809) a call is made to g_hash_table_lookup but session-result.last_key is NULL. (gdb) print session-result.last_key $1 = (gchar *) 0x0 Is there anything I can do to help you guys further? Interesting. It would be nice to see the actual content it's parsing (which you can get in requesting the same url as wispr with wget) Might be a corner case we never stumble into. Anyway, I'll check the code Thanks, Tomasz ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
RE: Connman crashes when loosing network
[1602]: src/dhcp.c:__connman_dhcp_start() connmand[1602]: src/dhcp.c:dhcp_request() dhcp 0xab9d0 connmand[1602]: DHCP index 3: switch listening mode (0 == 1) connmand[1602]: DHCP index 3: DHCP client start with state init_reboot connmand[1602]: DHCP index 3: sending DHCP request connmand[1602]: DHCP index 3: received DHCP packet xid 0x (current state 1) connmand[1602]: DHCP index 3: processed DHCP packet (new state 1) connmand[1602]: DHCP index 3: restart DHCP timeout connmand[1602]: DHCP index 3: restart DHCP (retries 0) connmand[1602]: DHCP index 3: sending DHCP discover request connmand[1602]: DHCP index 3: received DHCP packet xid 0x (current state 0) connmand[1602]: DHCP index 3: start request (retries 0) connmand[1602]: DHCP index 3: sending DHCP request connmand[1602]: DHCP index 3: received DHCP packet xid 0x (current state 2) connmand[1602]: DHCP index 3: switch listening mode (1 == 0) connmand[1602]: src/dhcp.c:lease_available_cb() Lease available connmand[1602]: src/dhcp.c:lease_available_cb() last address 192.168.0.128 connmand[1602]: src/dhcp.c:lease_available_cb() c_address (null) connmand[1602]: DHCP index 3: start bound connmand[1602]: DHCP index 3: processed DHCP packet (new state 3) connmand[1602]: eth0 {add} address 192.168.0.128/24 label eth0 family 2 connmand[1602]: eth0 {add} route 192.168.0.0 gw 0.0.0.0 scope 253 LINK connmand[1602]: eth0 {add} route 192.168.0.100 gw 0.0.0.0 scope 253 LINK connmand[1602]: eth0 {add} route 0.0.0.0 gw 192.168.0.100 scope 0 UNIVERSE connmand[1602]: eth0 {add} route 192.168.0.100 gw 192.168.0.100 scope 0 UNIVERSE connmand[1602]: Aborting (signal 11) [connmand] -Original Message- From: connman [mailto:connman-boun...@connman.net] On Behalf Of Patrik Flykt Sent: vrijdag 2 mei 2014 7:55 To: connman@connman.net Subject: Re: Connman crashes when loosing network On Thu, 2014-05-01 at 08:05 +, Marcel Mulder wrote: I experience reproducible connman crashes with version v1.20, v1.21, v1.22, v1.23 and HEAD of git. Older version I did not try. See if dd96d41ea182f0d1e0b55686adf793fa72e0e4d3 helps. At least it was an attempt to solve a very similar sounding problem. When debugging this, run 'CONNMAN_DHCP_DEBUG=1 connmand -n -d src/dhcp.c' so that all DHCP logs are available. Cheers, Patrik ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
RE: Connman crashes when loosing network
Hi Eduardo, The log was made with connman HEAD e22ad5f161a487afc41d12ab630a536cbfa3e2a7 Regards, Marcel -Original Message- From: connman [mailto:connman-boun...@connman.net] On Behalf Of Eduardo Abinader Sent: donderdag 1 mei 2014 17:37 To: connman@connman.net Subject: Re: Connman crashes when loosing network Could you please try with HEAD version and pastebin the logs? Saw a refcount issue in: 05/01/14 09:56:55 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_unref_debug() 0xa9fd8 ref -1 by src/ipconfig.c:1707:__connman_ipconfig_disable() But I don't know which version your are actually running. On Thu, May 1, 2014 at 4:05 AM, Marcel Mulder mmul...@victronenergy.comwrote: Hi All, I experience reproducible connman crashes with version v1.20, v1.21, v1.22, v1.23 and HEAD of git. Older version I did not try. The situation is as follow: I have a controller (AM3547 from TI, ARM Cortex-A8) attached to a Wifi extender by wire. The Wifi extender is connected to a Wifi access point which is connect to our network. If the connection are up and running everything works well. When I however turn off the Wifi access point connman crashes with a segmentation fault (see log below). I have seen more connman crashes when networks are coming and going but there were not so nicely reproducible as this one. For cross compiling I use Yocto-Poky (OpenEmbedded). If you need more information, e.g. Wireshark logs or other kind of information I am happy to provide this. Connman was started with: connmand -d -n --nobacktrace -r But also with dnsproxy connman crashes. When I restart connman in this situation it ends in a seg fault again. When the connection to the Wifi access point is restored connman starts without problems. Hope someone can help me because this is a serious problem for us. Best regards, Marcel 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv4() 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv4config() 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv6config() 05/01/14 09:56:37 AM connmand[26030]: src/service.c:append_nameservers() servers[0] 192.168.4.13 available 1 05/01/14 09:56:37 AM connmand[26030]: src/service.c:append_nameservers() servers[1] 192.168.1.26 available 1 05/01/14 09:56:37 AM connmand[26030]: src/storage.c:storage_load() Loading /var/lib/connman/settings 05/01/14 09:56:45 AM connmand[26030]: src/rtnl.c:rtnl_message() buf 0xbea8ab18 len 576 05/01/14 09:56:45 AM connmand[26030]: src/rtnl.c:rtnl_message() NEWLINK len 576 type 16 flags 0x seq 0 pid 0 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_newlink() index 3 05/01/14 09:56:45 AM connmand[26030]: eth0 {RX} 38268 packets 4322051 bytes 05/01/14 09:56:45 AM connmand[26030]: eth0 {TX} 31213 packets 3869059 bytes 05/01/14 09:56:45 AM connmand[26030]: src/service.c:stats_update() service 0xad800 05/01/14 09:56:45 AM connmand[26030]: eth0 {update} flags 36867 UP 05/01/14 09:56:45 AM connmand[26030]: src/service.c:service_lower_down() eth0 lower down 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_disable() ipconfig 0xac3e0 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_unref_debug() 0xac3e0 ref 1 by src/ipconfig.c:1694:__connman_ipconfig_disable() 05/01/14 09:56:45 AM connmand[26030]: src/service.c:stats_stop() service 0xad800 05/01/14 09:56:45 AM connmand[26030]: src/service.c:service_save() service 0xad800 new 0 05/01/14 09:56:45 AM connmand[26030]: src/storage.c:storage_load() Loading /var/lib/connman/ethernet_5056631d05ad_cable/settings 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_save() ipconfig 0xac3e0 identifier ethernet_5056631d05ad_cable method dhcp 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_save() ipconfig 0xa9fd8 identifier ethernet_5056631d05ad_cable method off 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_lower_down() ipconfig ipv4 (nil) ipv6 0xa9fd8 05/01/14 09:56:45 AM connmand[26030]: src/inet.c:connman_inet_clear_ipv6_address() index 3 address (null) prefix_len 0 05/01/14 09:56:45 AM connmand[26030]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 05/01/14 09:56:45 AM connmand[26030]: eth0 {newlink} index 3 operstate 2 DOWN 05/01/14 09:56:45 AM connmand[26030]: src/detect.c:detect_newlink() type 1 index 3 05/01/14 09:56:45 AM connmand[26030]: plugins/ethernet.c:ethernet_newlink() index 3 flags 36867 change 0
Re: Connman crashes when loosing network
Hi Marcel, On pe, 2014-05-02 at 08:00 +, Marcel Mulder wrote: Hi Patrik, Thanks for the reply! I made a build with dd96d41ea182f0d1e0b55686adf793fa72e0e4d3 and started it with 'CONNMAN_DHCP_DEBUG=1 connmand -n -d src/dhcp.c' Connman now stops with an abort signal 11 which I also have seen before. Depending on what I do and how things are going (timing I presume) I see signals 11 or seg faults. The log of this session is below starting from the beginning when connman starts. Regards, Marcel root@CCGX:~# CONNMAN_DHCP_DEBUG=1 connmand -n -d src/dhcp.c Do you have any possibility to use valgrind in your target device? I have successfully used it to find out similar issues in the past. Cheers, Jukka ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
RE: Connman crashes when loosing network
Hi Jukka, I did not know about valgrind. It looks like a very useful tool and I have to possibility to put it on the target. So I will :-) Many Thanks for your suggestion and I will let you know if it was of any help. Regrads, Marcel -Original Message- From: Jukka Rissanen [mailto:jukka.rissa...@linux.intel.com] Sent: vrijdag 2 mei 2014 10:11 To: connman@connman.net Cc: Marcel Mulder Subject: Re: Connman crashes when loosing network Hi Marcel, On pe, 2014-05-02 at 08:00 +, Marcel Mulder wrote: Hi Patrik, Thanks for the reply! I made a build with dd96d41ea182f0d1e0b55686adf793fa72e0e4d3 and started it with 'CONNMAN_DHCP_DEBUG=1 connmand -n -d src/dhcp.c' Connman now stops with an abort signal 11 which I also have seen before. Depending on what I do and how things are going (timing I presume) I see signals 11 or seg faults. The log of this session is below starting from the beginning when connman starts. Regards, Marcel root@CCGX:~# CONNMAN_DHCP_DEBUG=1 connmand -n -d src/dhcp.c Do you have any possibility to use valgrind in your target device? I have successfully used it to find out similar issues in the past. Cheers, Jukka ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
Re: Connman crashes when loosing network
On pe, 2014-05-02 at 08:30 +, Marcel Mulder wrote: Hi Jukka, I did not know about valgrind. It looks like a very useful tool and I have to possibility to put it on the target. So I will :-) Many Thanks for your suggestion and I will let you know if it was of any help. By default valgrind reports also errors from various system libraries but you can suppress those using a suppression example file found in doc/valgrind.suppressions So one way to use valgrind is like this: $ cd connman source directory $ cp doc/valgrind.suppressions .valgrind.suppressions $ sudo sh -c export CONNMAN_DHCP_DEBUG=1; valgrind --leak-check=full ./src/connmand -n -d 21 | tee connman.log Cheers, Jukka ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
Connman crashes when loosing network
Hi All, I experience reproducible connman crashes with version v1.20, v1.21, v1.22, v1.23 and HEAD of git. Older version I did not try. The situation is as follow: I have a controller (AM3547 from TI, ARM Cortex-A8) attached to a Wifi extender by wire. The Wifi extender is connected to a Wifi access point which is connect to our network. If the connection are up and running everything works well. When I however turn off the Wifi access point connman crashes with a segmentation fault (see log below). I have seen more connman crashes when networks are coming and going but there were not so nicely reproducible as this one. For cross compiling I use Yocto-Poky (OpenEmbedded). If you need more information, e.g. Wireshark logs or other kind of information I am happy to provide this. Connman was started with: connmand -d -n --nobacktrace -r But also with dnsproxy connman crashes. When I restart connman in this situation it ends in a seg fault again. When the connection to the Wifi access point is restored connman starts without problems. Hope someone can help me because this is a serious problem for us. Best regards, Marcel 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv4() 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv4config() 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv6config() 05/01/14 09:56:37 AM connmand[26030]: src/service.c:append_nameservers() servers[0] 192.168.4.13 available 1 05/01/14 09:56:37 AM connmand[26030]: src/service.c:append_nameservers() servers[1] 192.168.1.26 available 1 05/01/14 09:56:37 AM connmand[26030]: src/storage.c:storage_load() Loading /var/lib/connman/settings 05/01/14 09:56:45 AM connmand[26030]: src/rtnl.c:rtnl_message() buf 0xbea8ab18 len 576 05/01/14 09:56:45 AM connmand[26030]: src/rtnl.c:rtnl_message() NEWLINK len 576 type 16 flags 0x seq 0 pid 0 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_newlink() index 3 05/01/14 09:56:45 AM connmand[26030]: eth0 {RX} 38268 packets 4322051 bytes 05/01/14 09:56:45 AM connmand[26030]: eth0 {TX} 31213 packets 3869059 bytes 05/01/14 09:56:45 AM connmand[26030]: src/service.c:stats_update() service 0xad800 05/01/14 09:56:45 AM connmand[26030]: eth0 {update} flags 36867 UP 05/01/14 09:56:45 AM connmand[26030]: src/service.c:service_lower_down() eth0 lower down 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_disable() ipconfig 0xac3e0 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_unref_debug() 0xac3e0 ref 1 by src/ipconfig.c:1694:__connman_ipconfig_disable() 05/01/14 09:56:45 AM connmand[26030]: src/service.c:stats_stop() service 0xad800 05/01/14 09:56:45 AM connmand[26030]: src/service.c:service_save() service 0xad800 new 0 05/01/14 09:56:45 AM connmand[26030]: src/storage.c:storage_load() Loading /var/lib/connman/ethernet_5056631d05ad_cable/settings 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_save() ipconfig 0xac3e0 identifier ethernet_5056631d05ad_cable method dhcp 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_save() ipconfig 0xa9fd8 identifier ethernet_5056631d05ad_cable method off 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_lower_down() ipconfig ipv4 (nil) ipv6 0xa9fd8 05/01/14 09:56:45 AM connmand[26030]: src/inet.c:connman_inet_clear_ipv6_address() index 3 address (null) prefix_len 0 05/01/14 09:56:45 AM connmand[26030]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 05/01/14 09:56:45 AM connmand[26030]: eth0 {newlink} index 3 operstate 2 DOWN 05/01/14 09:56:45 AM connmand[26030]: src/detect.c:detect_newlink() type 1 index 3 05/01/14 09:56:45 AM connmand[26030]: plugins/ethernet.c:ethernet_newlink() index 3 flags 36867 change 0 05/01/14 09:56:45 AM connmand[26030]: plugins/ethernet.c:ethernet_newlink() carrier off 05/01/14 09:56:45 AM connmand[26030]: src/device.c:connman_device_remove_network() device 0xacf88 network 0xaaf10 05/01/14 09:56:45 AM connmand[26030]: src/device.c:free_network() network 0xaaf10 05/01/14 09:56:45 AM connmand[26030]: src/network.c:network_remove() network 0xaaf10 name Wired 05/01/14 09:56:45 AM connmand[26030]: src/network.c:set_disconnected() service 0xad800 ipv4 0xac3e0 ipv6 0xa9fd8 05/01/14 09:56:45 AM connmand[26030]: src/network.c:set_disconnected() method ipv4 4 ipv6 1 05/01/14 09:56:45 AM connmand[26030]: src/dhcp.c:__connman_dhcp_stop() network_table 0xa5528 network
Re: Connman crashes when loosing network
Could you please try with HEAD version and pastebin the logs? Saw a refcount issue in: 05/01/14 09:56:55 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_unref_debug() 0xa9fd8 ref -1 by src/ipconfig.c:1707:__connman_ipconfig_disable() But I don't know which version your are actually running. On Thu, May 1, 2014 at 4:05 AM, Marcel Mulder mmul...@victronenergy.comwrote: Hi All, I experience reproducible connman crashes with version v1.20, v1.21, v1.22, v1.23 and HEAD of git. Older version I did not try. The situation is as follow: I have a controller (AM3547 from TI, ARM Cortex-A8) attached to a Wifi extender by wire. The Wifi extender is connected to a Wifi access point which is connect to our network. If the connection are up and running everything works well. When I however turn off the Wifi access point connman crashes with a segmentation fault (see log below). I have seen more connman crashes when networks are coming and going but there were not so nicely reproducible as this one. For cross compiling I use Yocto-Poky (OpenEmbedded). If you need more information, e.g. Wireshark logs or other kind of information I am happy to provide this. Connman was started with: connmand -d -n --nobacktrace -r But also with dnsproxy connman crashes. When I restart connman in this situation it ends in a seg fault again. When the connection to the Wifi access point is restored connman starts without problems. Hope someone can help me because this is a serious problem for us. Best regards, Marcel 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv4() 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv4config() 05/01/14 09:56:37 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_append_ipv6config() 05/01/14 09:56:37 AM connmand[26030]: src/service.c:append_nameservers() servers[0] 192.168.4.13 available 1 05/01/14 09:56:37 AM connmand[26030]: src/service.c:append_nameservers() servers[1] 192.168.1.26 available 1 05/01/14 09:56:37 AM connmand[26030]: src/storage.c:storage_load() Loading /var/lib/connman/settings 05/01/14 09:56:45 AM connmand[26030]: src/rtnl.c:rtnl_message() buf 0xbea8ab18 len 576 05/01/14 09:56:45 AM connmand[26030]: src/rtnl.c:rtnl_message() NEWLINK len 576 type 16 flags 0x seq 0 pid 0 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_newlink() index 3 05/01/14 09:56:45 AM connmand[26030]: eth0 {RX} 38268 packets 4322051 bytes 05/01/14 09:56:45 AM connmand[26030]: eth0 {TX} 31213 packets 3869059 bytes 05/01/14 09:56:45 AM connmand[26030]: src/service.c:stats_update() service 0xad800 05/01/14 09:56:45 AM connmand[26030]: eth0 {update} flags 36867 UP 05/01/14 09:56:45 AM connmand[26030]: src/service.c:service_lower_down() eth0 lower down 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_disable() ipconfig 0xac3e0 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_unref_debug() 0xac3e0 ref 1 by src/ipconfig.c:1694:__connman_ipconfig_disable() 05/01/14 09:56:45 AM connmand[26030]: src/service.c:stats_stop() service 0xad800 05/01/14 09:56:45 AM connmand[26030]: src/service.c:service_save() service 0xad800 new 0 05/01/14 09:56:45 AM connmand[26030]: src/storage.c:storage_load() Loading /var/lib/connman/ethernet_5056631d05ad_cable/settings 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_save() ipconfig 0xac3e0 identifier ethernet_5056631d05ad_cable method dhcp 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_save() ipconfig 0xa9fd8 identifier ethernet_5056631d05ad_cable method off 05/01/14 09:56:45 AM connmand[26030]: src/ipconfig.c:__connman_ipconfig_lower_down() ipconfig ipv4 (nil) ipv6 0xa9fd8 05/01/14 09:56:45 AM connmand[26030]: src/inet.c:connman_inet_clear_ipv6_address() index 3 address (null) prefix_len 0 05/01/14 09:56:45 AM connmand[26030]: eth0 {newlink} index 3 address 50:56:63:1D:05:AD mtu 1500 05/01/14 09:56:45 AM connmand[26030]: eth0 {newlink} index 3 operstate 2 DOWN 05/01/14 09:56:45 AM connmand[26030]: src/detect.c:detect_newlink() type 1 index 3 05/01/14 09:56:45 AM connmand[26030]: plugins/ethernet.c:ethernet_newlink() index 3 flags 36867 change 0 05/01/14 09:56:45 AM connmand[26030]: plugins/ethernet.c:ethernet_newlink() carrier off 05/01/14 09:56:45 AM connmand[26030]: src/device.c:connman_device_remove_network() device 0xacf88 network 0xaaf10 05/01/14 09:56:45 AM connmand[26030]: src/device.c:free_network() network 0xaaf10
Re: Connman crashes when loosing network
On Thu, 2014-05-01 at 08:05 +, Marcel Mulder wrote: I experience reproducible connman crashes with version v1.20, v1.21, v1.22, v1.23 and HEAD of git. Older version I did not try. See if dd96d41ea182f0d1e0b55686adf793fa72e0e4d3 helps. At least it was an attempt to solve a very similar sounding problem. When debugging this, run 'CONNMAN_DHCP_DEBUG=1 connmand -n -d src/dhcp.c' so that all DHCP logs are available. Cheers, Patrik ___ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman