We are also on 8.5.135. I want to say we hit this bug in December of last year with a single 3802. Slowly, more and more APs have been affected. Our network consists mainly of 3702 and 3802 APs. As of last week, we have had around 9 APs with the issue.
We have been able to resolve the issue, to a certain degree, by either performing a factory reset through the WLC (Clear All Config option in the AP's 'General' tab), or by flashing the AP and reloading the IOS manually. Both methods call for a reconfig, adding it to the proper AP groups, high availability, and all that, etc... It has worked on 7 of the 9 affected, and has shown stability for about a week. Most port bounces or manual reboots of the AP may fix the issue for a few hours to a max of 2 days before they lose their ACL.
Something I noticed, is that the affected APs will reboot on their own prior to coming back up with no ACL. Logs do not indicate any issue prior to the reboot, however. I was able to catch this happening on 3 of the affected APs, but cannot say for certain that all of them exhibit this behavior prior to ACL loss.
