usb network device stops working at random times

Catatonic Porpoise <[email protected]>
Newsgroups gmane.linux.drivers.at76c503a.user
Message-ID <[email protected]>
I'm having severe problems with the at76c503 driver - the network device
works, but every once in a while, at seemingly arbitrary times, it'll stop
working, and this is often accompanied by the keyboard locking up. Even
when the keyboard doesn't lock up, I have to reboot because ifconfig will
freeze (the entire system) anytime I run it after that.

The way I have it set up is, after each boot I connect the USB device, su
to root, and run the following commands:

iwconfig wlan0 mode managed essid any channel 6
iwconfig wlan0 ap 00:04:5a:db:a3:37
iwconfig wlan0 key [1] restricted 72A0DEB3ABAA6FA02D398D7279
iwconfig wlan0 key on

ifconfig wlan0 192.168.1.101

echo I used to do "ifconfig wlan0 promisc" here, but no longer; it doesn't
seem to make a difference though.

route add -v default gw 192.168.1.1

The reason I don't have the device in while booting is that, if I do that,
while it's in the process of "checking for new hardware" it prints a bunch
of garbage and something about the kernel having an unhandled exception.

I'm using Fedora Core 2 with kernel 2.6.5-1358, the driver is at76c503
v0.12beta11. The device is a Linksys wireless USB access point, the newer
version (2.6 I think). When the network goes down and keyboard locks,
/var/log/messages shows something resembling this:

May 31 20:30:10 localhost kernel: uhci_hcd 0000:00:10.1: host system
error, PCI problems?
May 31 20:30:10 localhost kernel: uhci_hcd 0000:00:10.1: host system
error, PCI problems?
May 31 20:30:10 localhost kernel: uhci_hcd 0000:00:10.1: host controller
halted, very bad!
May 31 20:30:10 localhost kernel: uhci_hcd 0000:00:10.1: host controller
halted, very bad!
May 31 20:30:20 localhost kernel:
/usr/src/redhat/BUILD/at76c503a/at76c503.c: wlan0: lost beacon bssid
00:04:5a:db:a3:37
May 31 20:30:20 localhost kernel:
/usr/src/redhat/BUILD/at76c503a/at76c503.c: wlan0: lost beacon bssid
00:04:5a:db:a3:37
May 31 20:30:21 localhost kernel: usb 3-1: control timeout on ep0out
May 31 20:30:21 localhost kernel: usb 3-1: control timeout on ep0out

When the keyboard doesn't lock, it doesn't show the "uhci_hcd" lines. It
seems however that on times the keyboard does lock it doesn't do so until
the time the "usb 3-1: control timeout" line shows up. When the keyboard
does not lock it shows lines like this, after the lost beacon and control
timeout:

Jun  1 00:32:36 localhost kernel:
/usr/src/redhat/BUILD/at76c503a/at76c503.c: wlan0: start_scan failed with
-110
Jun  1 00:32:36 localhost kernel:
/usr/src/redhat/BUILD/at76c503a/at76c503.c: wlan0: start_scan failed with
-110

A quick scan of /var/log/messages shows that the start_scan message never
seems to come up the times that the keyboard locks. Like I said, the
occurrences are mostly random. However, on a few occasions I've noticed
they happen at about the same time I do something requiring lots of CPU
for an extended time, such as starting Mozilla, or encoding an ogg file.
This has happened with both the keyboard-locking and non-keyboard-locking
varieties, and I should point out, the majority of the times this happens,
nothing out of the ordinary is going on at all (it's happened when I
wasn't even there at the computer).

Please let me know if there's more information I should provide.
Assistance would be greatly appreciated, since this is the only thing
preventing me from having a happy life at the moment. Thanks!

-graue
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.