colin/NetworkManager
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

device: don't complain about repeated schedules of the same activation stage

Browse Source 
Can easily happend with a storm of DHCP responses or RAs before the idle
handler has a chance to run.

https://bugzilla.redhat.com/show_bug.cgi?id=1269520
This commit is contained in:
Lubomir Rintel
2015-10-07 15:19:21 +02:00
parent 76726b9922
commit 3abe1bb21a
1 changed files with 16 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
src/devices/nm-device.c
View File

@@ -2814,19 +2814,23 @@ activation_source_schedule (NMDevice *self, GSourceFunc func, int family)
act_source_func = &priv->act_source_func;
}
if (*act_source_id)
_LOGE (LOGD_DEVICE, "activation stage already scheduled");
/* Don't bother rescheduling the same function that's about to
* run anyway. Fixes issues with crappy wireless drivers sending
* streams of associate events before NM has had a chance to process
* the first one.
*/
if (!*act_source_id || (*act_source_func != func)) {
activation_source_clear (self, TRUE, family);
*act_source_id = g_idle_add (func, self);
*act_source_func = func;
if (*act_source_id) {
if (*act_source_func == func) {
/* Don't bother rescheduling the same function that's about to
* run anyway. Fixes issues with crappy wireless drivers sending
* streams of associate events before NM has had a chance to process
* the first one.
*/
_LOGD (LOGD_DEVICE, "activation stage already scheduled");
return;
} else {
_LOGW (LOGD_DEVICE, "a different activation stage already scheduled");
activation_source_clear (self, TRUE, family);
}
}
*act_source_id = g_idle_add (func, self);
*act_source_func = func;
}
static gboolean