wpa_supplicant fails to initiate WEP128 connection over ipw2200 driver/device

[Peter Breuer]

On Wed, 9 Jun 2021 08:17:12 +0100
Peter Breuer <Peter.T.Breuer at gmail.com> wrote:
> Recently I have not been able to connect over WEP (128) through my
> intel ipw2200 wifi device (venerable ibm tp x40 laptop, cerca
> 2006, 1.2GHz 686M processor, 1GB ram) via two or three new routers


To follow up on this, debugging on in the kernel module shows
disassociation from the device with "CMAS_INIT"??

 [ 6171.318959] ipw2200: ipw_best_network Network 'TP-Link_0B42 (c0:c9:e3:2f:0b:42)' is a viable match.
 [ 6171.318981] ipw2200: ipw_associate_network Asociation attempt: 'TP-Link_0B42', channel 1, 802.11g [12], long[:long], enc=on (open) key=1.
 [ 6171.319015] 00000000 54 50 2D 4C 69 6E 6B 5F  30 42 34 32               TP-Link_ 0B42    
 [ 6171.319909] ipw2200: ipw_associate_network associating: 'TP-Link_0B42' c0:c9:e3:2f:0b:42

Good. But 0.13s later CMAS_INIT is received by the driver, and ...

 [ 6171.448296] ipw2200: ipw_rx_notification disassociated: 'TP-Link_0B42' c0:c9:e3:2f:0b:42

Why would this happen?  Imbetween those two points there is

  [ 6467.685231] libipw: libipw_rx_mgt received BEACON (128)
  [ 6467.685237] libipw: libipw_rx_mgt Beacon

Then some probe responses showing other sources on the airwaves:

  [ 6467.685241] libipw: libipw_process_probe_response 'MiFibra-00CA' (e0:51:63:39:00:cc): 0000 0100-0001 0001
  [ 6467.685250] libipw: libipw_parse_info_param WLAN_EID_SSID: 'MiFibra-00CA' len=12.
  ...
  [ 6467.685311] libipw: libipw_process_probe_response Updating 'MiFibra-00CA' (e0:51:63:39:00:cc) via BEACON.

Then we're back in ipw2200:

  [ 6467.791322] ipw2200: ipw_rx_notification type = 11 (4 bytes)
  [ 6467.791346] ipw2200: ipw_rx_notification AUTH_SEQ_1
  [ 6467.792997] ipw2200: ipw_rx_notification type = 10 (4 bytes)

and boom!

  [ 6467.793009] ipw2200: ipw_rx_notification disassociated: 'TP-Link_0B42' c0:c9:e3:2f:0b:42

Why? And what to do? I'd be grateful for any pointers from this list.
*I* don't know what the finite state machine in the device is supposed
to be doing, so I can't possibly know why CMAS_INIT is issued or why
that is taken as a _disassociation_ by the driver!  It sounds to me like
it is intended to be the start of something, not the end.

Thanks in advance for any indication someone is alive out there ...

PTB