From: Daniel Wagner <[email protected]>
There are cases where technology->pending_reply is set but the
corresponding technology->pending_timeout not. The technology enabled
was called from the Manager and the device was enabled right away.
connmand[27545]: src/manager.c:enable_technology() conn 0x6a0770
connmand[27545]: src/notifier.c:__connman_notifier_is_registered() type 6
connmand[27545]: src/notifier.c:__connman_notifier_is_enabled() type 6
connmand[27545]: src/technology.c:__connman_technology_enable() type 6 enable
connmand[27545]: src/technology.c:technology_find() type 6
connmand[27545]: src/technology.c:save_state() technology 0x6af570
connmand[27545]: src/storage.c:storage_load() Loading /var/lib/connman/settings
connmand[27545]: src/rfkill.c:__connman_rfkill_block() type 6 block 0
connmand[27545]: src/rfkill.c:rfkill_process()
connmand[27545]: src/rfkill.c:rfkill_process() idx 5 type 5 op 2 soft 0 hard 0
connmand[27545]: src/technology.c:__connman_technology_update_rfkill() index 5
soft 0 hard 0
connmand[27545]: src/technology.c:technology_find() type 6
connmand[27545]: plugins/ofono.c:cm_changed() path /hso0
connmand[27545]: plugins/ofono.c:update_cm_attached() path /hso0 attached 1
connmand[27545]: plugins/ofono.c:create_device() path /hso0
connmand[27545]: src/device.c:connman_device_create() node 262021007716562 type
5
connmand[27545]: src/device.c:connman_device_create() device 0x6aaec0
connmand[27545]: src/device.c:connman_device_set_string() device 0x6aaec0 key
Path value /hso0
connmand[27545]: src/device.c:connman_device_register() device 0x6aaec0 name
Cellular
connmand[27545]: src/device.c:connman_device_register() driver 0x695d20 name
modem
connmand[27545]: plugins/ofono.c:modem_probe() device 0x6aaec0
connmand[27545]: src/technology.c:__connman_technology_add_device() device
0x6aaec0
connmand[27545]: src/notifier.c:__connman_notifier_register() type 6
connmand[27545]: src/technology.c:technology_get() type 6
connmand[27545]: src/technology.c:technology_find() type 6
connmand[27545]: src/technology.c:technology_get() technology 0x6af570
connmand[27545]: src/device.c:__connman_device_enable() device 0x6aaec0
connmand[27545]: plugins/ofono.c:modem_enable() device 0x6aaec0
connmand[27545]: src/device.c:connman_device_set_powered() driver 0x6aaec0
powered 1
connmand[27545]: src/technology.c:technology_find() type 6
connmand[27545]: src/notifier.c:__connman_notifier_enable() type 6
connmand[27545]: src/notifier.c:technology_enabled() type 6 enabled 1
(connmand:27545): GLib-CRITICAL **: g_source_remove: assertion `tag > 0' failed
---
Hi,
I'm fixing the symptons with this patch. But I am not sure if the
whole handling of the pending msg is correct.
cheers,
daniel
src/technology.c | 16 ++++++++++++----
1 files changed, 12 insertions(+), 4 deletions(-)
diff --git a/src/technology.c b/src/technology.c
index c6fd826..c581791 100644
--- a/src/technology.c
+++ b/src/technology.c
@@ -817,10 +817,14 @@ int __connman_technology_enabled(enum
connman_service_type type)
if (technology->pending_reply != NULL) {
g_dbus_send_reply(connection, technology->pending_reply,
DBUS_TYPE_INVALID);
+
dbus_message_unref(technology->pending_reply);
- g_source_remove(technology->pending_timeout);
technology->pending_reply = NULL;
- technology->pending_timeout = 0;
+
+ if (technology->pending_timeout > 0) {
+ g_source_remove(technology->pending_timeout);
+ technology->pending_timeout = 0;
+ }
}
return 0;
@@ -908,10 +912,14 @@ int __connman_technology_disabled(enum
connman_service_type type)
if (technology->pending_reply != NULL) {
g_dbus_send_reply(connection, technology->pending_reply,
DBUS_TYPE_INVALID);
+
dbus_message_unref(technology->pending_reply);
- g_source_remove(technology->pending_timeout);
technology->pending_reply = NULL;
- technology->pending_timeout = 0;
+
+ if (technology->pending_timeout > 0) {
+ g_source_remove(technology->pending_timeout);
+ technology->pending_timeout = 0;
+ }
}
if (g_atomic_int_dec_and_test(&technology->enabled) == TRUE) {
--
1.7.8.rc0.35.gee6df.dirty
_______________________________________________
connman mailing list
[email protected]
http://lists.connman.net/listinfo/connman
