[Prism54-users] SMC2835 stopped working, how to troubleshoot?

[Dirk]

Hi all. Out of the blue my wlan stopped working. I don't have a clue
what the cause can be. Until now it worked without a problem. (almost)

Saturday morning it worked. Monday it didn't. I tried to find the
problem monday night but coudln't find anything. Suddenly it started
working, I tried an update (debian, apt-get update) but before the
update finished the connection was gone. At that moment my usual routine
was to eject the card, unload the drivers, insert the card and do an
"ifup eth1". Normally this works. I tried the other pcmcia slot, but
that didn't. Then, suddenly it did work again, it worked for maybe 15 or
30 minutes, I left the machine alone, came back, connection is down
again. 

What I did mostly was ejecting and reinserting the card (un)loading the
drivers, both prism54 and firmware_class. I checked the AP of course,
dis/en abled WEP, SSID broadcast, MAC filtering, it didn't matter at
all. 

System is debian testing with 2.6.1 kernel, prism54 driver is fresh
from monday. The problem is the same for 2.6.0, that did'nt work either. 

How do you troubleshoot such a problem? I only have 1 AP and 1 wlan
card. I only use linux on the laptop, so there is no way I can test it
with another system or OS. 

But...I read my logs just to make sure there is nothing wrong and I see
some things that I didn't saw yesterday. 
This is a part of the kern.log:
Feb  3 19:01:13 CB1 kernel: Assuming someone else called the IRQ
Feb  3 19:35:45 CB1 kernel: eth1: timeout waiting for mgmt response 1000, trigging device
Feb  3 19:35:46 CB1 kernel: eth1: timeout waiting for mgmt response
Feb  3 19:39:55 CB1 kernel: eth1: resetting device...
Feb  3 19:39:59 CB1 kernel: eth1: mgmt response not collected
Feb  3 19:48:14 CB1 kernel: eth1: timeout waiting for mgmt response 1000, trigging device

debug:
Feb  3 18:55:37 CB1 kernel: eth1: prism54 driver detected card model: SMC2835W
Feb  3 18:56:06 CB1 kernel: eth1: islpci_open()
Feb  3 18:56:06 CB1 kernel: eth1: resetting device...
Feb  3 18:56:06 CB1 kernel: eth1: uploading firmware...
Feb  3 18:56:06 CB1 kernel: eth1: firmware uploaded done, now triggering reset...
Feb  3 18:56:53 CB1 kernel: Assuming someone else called the IRQ
Feb  3 18:56:55 CB1 last message repeated 20 times
Feb  3 19:01:13 CB1 kernel: Assuming someone else called the IRQ
Feb  3 19:01:13 CB1 kernel: Assuming someone else called the IRQ
Feb  3 19:35:45 CB1 kernel: eth1: timeout waiting for mgmt response 1000, trigging device
Feb  3 19:39:55 CB1 kernel: eth1: resetting device...
Feb  3 19:48:14 CB1 kernel: eth1: timeout waiting for mgmt response 1000, trigging device

Especially the assuming line is new to me. 

Somewhere in messages this line:

Feb  3 17:33:21 CB1 kernel: PCI: IRQ 0 for device 0000:00:0b.0 doesn't match PIRQ mask - try pci=usepirqmask

But this is also present in an older message file. But could it be
related or just a hint for a solution?

Dirk