DAWN: wrong type of rrm array

Installing and Using OpenWrt Network and Wireless Configuration
1

Starting with June 4th master snapshot my syslog is flooded with wrong type of rrm array error messages. That is the first napshot to exhibit such a behaviour.
I am wondering what might be the reason. Any ideas @anom3 ?

2

@PolynomialDivision might be able to help. There have been several version updates to DAWN.

3

Whups. I accepted a PR that checks if a client supports rrm before sending 802.11k frames. I only tested if it compiles and DAWN still runs. I always run DAWN with a lot of debug output so that I did not see that. I will look how I can fix that. Workround is just to turn off stderr in the init file of dawn.

4

try the latest snapshot.

this bug is due to a typo and should be fixed days ago.

5

Thanks! :slight_smile: Was enough to update dawn alone. Now it is with some new EMs but definitelly less of them:

Jun 9 09:33:12 router dawn[7285]: not complete msg, len:4096, expected len:4196
Jun 9 09:33:12 router dawn[7285]: not complete msg, len:4096, expected len:4196
Jun 9 09:33:12 router dawn[7285]: not complete msg, len:4096, expected len:4196
Jun 9 09:33:12 router dawn[7285]: not complete msg, len:4096, expected len:4196
Jun 9 09:33:12 router dawn[7285]: not complete msg, len:100, expected len:250531845
Jun 9 09:33:12 router dawn[7285]: not complete msg, len:100, expected len:250531845
Jun 9 09:33:12 router dawn[7285]: not complete msg, len:100, expected len:250531845
Jun 9 09:33:12 router dawn[7285]: not complete msg, len:100, expected len:250531845

6

It seems that DAWN has some bug in send/recv msg between APs.

But I have no devices in testing it.

If DAWN could run properly, it would be better to just ignore such error msg.

7

Yeah. With the current amount of error messages that's not a big deal. And lately finally client tables on both my devices are properly populated.
Still might need to test handover. So far connection at the edge of the networks range seems to be very stable lately.

8

Should finally be fixed with https://github.com/berlin-open-wireless-lab/DAWN/commit/f7343a053c22f9af1d7a843e04b22bacd4a8ebf1 :slight_smile:

1 Like
9

There was a "huge" mistake in the 802.11v logic. The rrm string was not parsed correctly. This could cause wrong behavior when steering a client, and the nr script was wrong. Please update to the latest version which was released, today.