Bug 15613 - ipw2100 Fatal interrupt. Scheduling firmware restart.
Summary: ipw2100 Fatal interrupt. Scheduling firmware restart.
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: Drivers
Classification: Unclassified
Component: network-wireless (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: drivers_network-wireless@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-03-23 01:38 UTC by Charles Samuels
Modified: 2010-10-04 19:30 UTC (History)
2 users (show)

See Also:
Kernel Version: 2.6.32-3-686
Subsystem:
Regression: No
Bisected commit-id:


Attachments

Description Charles Samuels 2010-03-23 01:38:00 UTC
When using the ipw2100 network driver with WPA and transmitting a lot of data, I get the following errors on dmesg:

[  161.908391] ipw2100: exit - failed to send CARD_DISABLE command
[  161.926576] ipw2100: Fatal interrupt. Scheduling firmware restart.

Seems like the receive queue buffers run empty, causing the device to shutdown, and then get restarted by the driver. As a result, the driver continues to work, but gets pretty slow when it needs to restart the firmware.

My specific device is the "Network controller: Intel Corporation PRO/Wireless LAN 2100 3B Mini PCI Adapter (rev 04)", although it seems like it would affect all ipw2100 devices.

Here is what I believe to be the fix for the same problem that was made to the ipw2200 driver (unfortunately, the code for the two drivers is so different that the change is far more invasive than my user-space mind can handle): http://marc.info/?l=linux-wireless&m=120302948521611&w=2

You can see some associated bug reports here:
https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.15/+bug/24776
http://www.spinics.net/lists/linux-wireless/msg47020.html

This problem is also evident on debian unstable's shipping kernel (which I use).

Let me know if I can provide more information, or test patches.
Comment 1 Zhu Yi 2010-03-26 08:11:59 UTC
I doubt this is the root cause for your problem since the 2200 patch is fixing the rx queue while your problem is tx command sending failure. But it worth trying. Unfortunately, I don't have the bandwidth to port the patch at this time. Maybe you can ask for some help on the ML.

Regarding on the firmware error, we cannot find the correct 2100 firmware engineer any more. So it will be unlikely to have new 2100 firmwares released in the future. Sorry.
Comment 2 Charles Samuels 2010-04-07 22:56:40 UTC
Actually, the problem seems to occur when receiving lots of data, not necessarily transmitting. Logically, the firmware might just be running out of space and aborting for that reason.
Comment 3 Charles Samuels 2010-04-07 23:20:14 UTC
I also want to add that this problem affects all networks regardless of security settings, not just WPA, with or without a supplicant.
Comment 4 John W. Linville 2010-08-18 20:05:31 UTC
I think we need some better debugging info here.  Please be sure to build your kernel with CONFIG_IPW2100_DEBUG.  Then do this (as root):

echo 32 > /proc/net/ipw2100/debug_level

Finally, recreate the issue and attach full /var/log/messages after the failure.  Hopefully that points closer to the problem.
Comment 5 John W. Linville 2010-10-04 19:30:29 UTC
Closing due to lack of response -- please reopen if/when requested information becomes available...

Note You need to log in before you can comment on or make changes to this bug.