Transient fatal firmware errors in wpi

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Transient fatal firmware errors in wpi

Gregor Best
Hi there,

when using wpi, sometimes after waking up from APM suspend, the driver
fails with a fatal firmware error while associating to my access point.

Every time this has happened, I suspended the machine in one location
(such as university) and woke it up in another (e.g. home), so the AP
"environment" changed considerably, with the AP the NIC was associated
to disappearing completely.

dmesg shows the following:

    wpi0: fatal firmware error
    firmware error log (count=1):
      error type = "SYSASSERT" (0x00000005)
      error data      = 0x00000074
      branch link     = 0x000008B600000274
      interrupt link  = 0x0000032000003560
      time            = 230497304
    driver status:
      tx ring  0: qid=0  cur=2   queued=0  
      tx ring  1: qid=1  cur=0   queued=0  
      tx ring  2: qid=2  cur=0   queued=0  
      tx ring  3: qid=3  cur=0   queued=0  
      tx ring  4: qid=4  cur=21  queued=0  
      tx ring  5: qid=5  cur=0   queued=0  
      rx ring: cur=47
      802.11 state 4

After bringing the interface down via ifconfig wpi0 down, resetting
all relevant IEEE802.11x settings (such as network ID and WPA key) to
currently relevant information (such as the new nwid I'm trying to
connect to) and bringing the interface up again, it associates in no
time and works flawlessly (that is, until the next time I suspend and it
errors out).

While this doesn't exactly affect performance, it's an annoyance
nonetheless.

I attached a complete dmesg, collected over the time of about 3 days.
The fatal firmware error occurs multiple times during the dmesg (I
suspended/woke my machine up about two or three times a day).

--
    Gregor Best

wpi_dmesg (17K) Download Attachment
attachment1 (203 bytes) Download Attachment