both MTs have address setup on bridge interfaces. Everything works fine up to 4 or 5 mbits full duplex traffic but when exceed that ,cpu goes on 100% and RB532a restarts.
I don't think that using WDS is the problem causing reboots. I have a number of PtP wireless links in service using wds and bridging. Most are running RB532 or RB532a Routerboards and R52 cards or similar Atheros radios. My heaviest user pulls a lot of bandwidth with about 1000 computers on their network. They have been running on the same Routerboards and radios for more than 1 year. Their bw peaks are at 16Mbps down and 7Mbps up. They average about 7Mbps down and 2Mbps up. Their Routerboards have never rebooted except for RouterOS upgrades and once with a power outage. They are RB532s and powered POE at 48 volts using 3com 802.3af supplies. I use nstream on and turn off conntrack. There is nothing else special with the link.
Are you monitoring with dude? Could that be overloading it? I am using nagios and cacti.