What steps will reproduce the problem? 1. Cross-compile madwimax for OpenWRT 2. Run /usr/sbin/madwimax -vv 3. Observe:

Bus 001 Device 002: ID xxxx:xxxx Device found Claimed interface Continuous async read start... Chip info: cmc730_v2.1 Firmware info: u200-rev1-2.3.0-BL22 MAC: xx:xx:xx:xx:xx:xx Allocated tap interfaces: wimax0 Starting if-create script... Network found. RSSI: -93 CINR: -8.000000 TX Power: 57344 Frequency: 2505000 BSID: 00:00:15:03:0b:14 Network found. short_write (0) RSSI: -93 CINR: -5.250000 TX Power: 57344 Frequency: 2505000 BSID: 00:00:15:03:0b:14 State: NEGO Number: 2 Response: 1 short write (0) Segmentation fault

What is the expected output? What do you see instead? Something like State: NORMAL ...

What version of the product are you using? On what operating system? madwimax-0.1.0 on OpenWRT (r15477).

Using the source from svn (r160) fixed the problem. The now correctly associates itslef of BS

Can you please find out which revision fixes subj :) You can do this using binary search: r130 is bad, r160 is good, so test r145 ((130 + 160) / 2) and then test r152 if it is bad or r137 otherwise, and so on... I currently suspect r153.

gordick, it's r154.

Well r153 still segfaults, but r154 works.

Interesting... Ok, I'm closing the bug.