- Added milliseconds timestamp precision
- Data type changed to uint64 to hold millisecond timestamp
Signed-off-by: Yashvardhan <yashvardhan@netexperience.com>
- Removed unwanted nested client event structures.
Doing this not only reduced complexity in the code but also helped get
rid of multple memory leaks and crashes in SM and datapipeline.
- Simplified ubus invoke calls. Nested ubus invokes were resulting in some
- data loss in callbacks
- Fixed incorrect conversion of session_id data type. This was blocking
clearing of sessions.
- Got rid of unnecessary strdups/mallocs. Potential memory leaks.
- Fixed missing conversion of frequency to band
- Removed unwanted logging which was filling the syslog buffer.
- Resolved a race occuring on the shared list between ubus_collector and SM.
- Relevant changes in datapipeline/protobuf to support the modified interface.
Signed-off-by: Yashvardhan <yashvardhan@netexperience.com>
* Previous patch caused CI/CD to fail; this fix
fixes the issue.
Signed-off-by: Mario Senecic <mario.senecic@sartura.hr>
Signed-off-by: Damir Franusic <damir.franusic@sartura.hr>
* Bugfix in hostapd: fixed improper increment operator usage in os_realloc
function that caused a crash on multiple SSIDs
* Bugfix in ubus_collector: fixed the improper usage of blobmsg_for_each_attr
macro and the misplaced evsched_task_reschedule calls
* Reverted the previous client sessions hostapd patch
Signed-off-by: Mario Senecic <mario.senecic@sartura.hr>
* implementation of the following ubus methods:
1. "get_sessions"
2. "clear_sessions"
3. "clear_session"
4. "get_bss_list"
* tracking of STA sessions and events (AUTH, ASSOC, etc.)
* caching of events is handled by hostapd and resides in
AVL until it is processed by ubus_collector
* added hard-coded caching limit to prevent memory leaks;
this should be made configurable at some point in the
future
Signed-off-by: Damir Franusic <damir.franusic@netexperience.com>
We can now put "option rssi_ignore_probe_request -40" into uci. Probe requests
with a rssi worse then -40 will not be replied to.
Signed-off-by: John Crispin <john@phrozen.org>
