Several months ago my Linksys BEFW11S4 wireless router crapped out. It had a history of locking up after several days of heavy use (read about other people’s experiences with this device at SeattleWireless) and this was coincident with a couple of denial of service attacks being announced for it so I figured someone hosed it. Except it would not reset. Even upgrading to a then recent firmware and doing a hard-reset to clear out any remaining settings failed to revive it. Oddly, that did cause it to emit UPnP and dhcp requests on the LAN and WAN ports but the internal dhcp server and the embedded http server for the administration pages remained inaccessible. Maria and I both ran out of patience after a few days so I put my Lucent RG-1000 in its place and tossed the dead blue box into a storage tub beside my desk.

Today, for no particular reason, I tried again. I snagged a recent firmware image and a hacked tftp client that supports the password extension that Linksys uses. After fiddling with a static IP, I remembered that these units have a touchy IP stack so after editing /etc/sysctl.conf net/ipv4/tcp_ecn=0 and doing a sysctl -p to turn off ECN, I could ping it. I updated the firmware and did the reset-power on-hold 30 seconds-power off routine and the damn thing works.

A few minutes later, I had it reconfigured and put it back in place outside the firewall. I’ll see how long it runs before something happens. It’s a shame it isn’t as reliable as the RG-1000 as it has much better range and signal.