Re: at76c503a network stalls after 30-60MB of traffic on MA101B usb wlan

Nick Steeves <[email protected]> Mon, 08 Nov 2004 18:22:24 -0700
Newsgroups gmane.linux.drivers.at76c503a.user
Message-ID <[email protected]>
I have a similar situation here.  It seems to be related to the USB chipset.  
I have had success with Intel and ALi chipsets.  I have had this problem with 
VIA, and AMD(the original 756 to nForce) chips.  I think what would happen, 
is that the chipset would hiccup, the USB NIC would "loose beacon", and then 
the network would stall.  Taking down the offending interface, unloading the 
modules, then re-loading, then re-iwconfig'ing and re-ifconfig wlan0 would 
fix it.  Rebooting would *not* fix it.  Neither would a full power down, or 
re-plug.  I have been plagued by this problem for almost a year and a half 
now.  It doesn't affect the mouse or printer.  Unplugging the mouse or 
printer don't help.  No amount of kernel contortions helped either.  The 
advice I received was "turn on debugging, and email us the relevant output".

On a more positive note, I only have to take down the interface & bring it 
back up, etc.  Only once every 700mb or so of (any kind of) traffic.  About 
once a week, I need to power down, because the USB chipset gets dirty or 
something, and then the NIC doesn't work...  Maybe someday it'll be ok, for 
my old Slot-A.

On my Intel box, this is the (bad, error, etc) output I get from the latest 
CVS:

Nov  8 18:19:14 DigitalMercury kernel: /root/at76c503a/at76c503.c: registered 
wlan0
Nov  8 18:19:14 DigitalMercury kernel: usbcore: registered new driver 
at76c503-rfmd
Nov  8 18:19:20 DigitalMercury 
kernel: /root/at76c503a/at76c503.c: /root/at76c503a/at76c503.c:2924 assertion 
dev->istate == INIT failed
Nov  8 18:19:21 DigitalMercury 
kernel: /root/at76c503a/at76c503.c: /root/at76c503a/at76c503.c:2895 assertion 
dev->istate == SCANNING failed
Nov  8 18:19:21 DigitalMercury kernel: /root/at76c503a/at76c503.c: using BSSID 
02:00:e0:e5:5b:81
Nov  8 18:19:21 DigitalMercury 
kernel: /root/at76c503a/at76c503.c: /root/at76c503a/at76c503.c:2203 assertion 
0 failed

I'm not sure if the two bugs are related.  I am running an unpatched 2.6.9 
kernel.

--
Nick


On November 8, 2004 1:22, Alexander Steppke wrote:
> Hi,
>
> the Netgear MA101 Adapter is working fine here, I just transfered
> several hundred MBs with WEP enabled. My computer has a similiar
> configuration but I am using a newer kernel. Maybe you can try changing
> the way linux talks to your usb controller. I am using "uhci_hcd" and I
> recognized you are using "uhci-alt". But that is just a guess.
> Are there any error messages in your syslog after the connection dies?
>
> Alexander Steppke
>
> BC Flyer wrote:
> > Hi,
> > I have a Netgear MA101 Rev B wireless usb transceiver plugged into a
> > Toshiba
> > laptop running Debian 3r2 (woody). The berlios at76c503 drivers
> > installed without
> > a hitch and seemed to be working fine. However, i've noticed that after
> > somewhere
> > between 30 and 60 MB of traffic has passed, the interface stops working.
> > I can
> > consistently reproduce this by transferring a large file to or from the
> > system, and
> > in every case, the file transfer stalls somewhere between 30 and 60 MB,
> > and no more network traffic can get through except for an occasional ping
> > packet.
>
> _______________________________________________
> at76c503a-user mailing list
> [email protected]
> http://lists.berlios.de/mailman/listinfo/at76c503a-user