Description of problem:
black screen of death from tcpdump (or wireshark)
Version-Release number of selected component (if applicable):
tcpdump -i wlan0 -w ipp.tr -s 1500
Steps to Reproduce:
normal packet trace
wlan0 is run by rt73usb, which works fine when not attempting to trace.
The oops is in some usb_* calls.
I realize a kdump would be needed here, but system-config-kdump doesn't work on Fedora 16, and I haven't got a manual configuration to work yet (not even with sysrq-trigger, the system crashes, but no crash kernel) see bug#769763
Can you provide a stack trace as recorded from a serial console?
This is the only computer I have with a serial port. But I can take a picture of the stack trace on the console and attach it. Will do that in a little while.
Created attachment 549230 [details]
photo of kernel oops screen
I have managed to collect a kdump vmcore. What should I extract from it to attach? I am a newbie to crash.
It seems that crash is broken on f16 also: bug#742260
Any other way to extract useful info from vmcore?
I have rt73usb, I'll see if I can reproduce that locally ...
Retested with kernel-3.1.6-1.fc16.i686
Still crashes, got new vmcore. The "crash" utility still crashes.
Created attachment 550844 [details]
I was able to reproduce, and this patch prevent the crash. But I have to investigate issue a bit more to find out why key is NULL and to see if perhaps some better fix for the problem exist.
I installed an Atheros Communications Inc. AR2413 802.11bg NIC, and tcpdump works with that. I think this means that the problem is with rt73usb, not wireless in general, despite the patch in comment#8 being for wpa.c.
That's way I'm still looking at that problem, I'm not sure if rt73usb driver do something wrong or mac80211 should be more resistant for rt73usb data.
I think this should be fixed in mac80211.
I posted fix here:
(In reply to comment #11)
> I think this should be fixed in mac80211.
> I posted fix here:
I added this to all 3 branches. Should be in the next respective builds of each. Thanks!
kernel-3.1.9-1.fc16 has been submitted as an update for Fedora 16.
kernel-3.1.9-1.fc16 has been pushed to the Fedora 16 stable repository. If problems still persist, please make note of it in this bug report.
We experiment with compat-wireless package. That mean there are two drivers installed, one in /lib/modules/`uname -r`/drivers/net/wireless/ second one in
/lib/modules/`uname -r`/backports/drivers/net/wireless/ . By default compat-wireless (backport) modules are used, which does not include that fix AFAIK.
To use in kernel modules, should be enough to remove "backport" entry from /etc/depmod.d/dist.conf and reboot the system.
(In reply to comment #15)
> We experiment with compat-wireless package. That mean there are two drivers
> installed, one in /lib/modules/`uname -r`/drivers/net/wireless/ second one in
> /lib/modules/`uname -r`/backports/drivers/net/wireless/ . By default
> compat-wireless (backport) modules are used, which does not include that fix
> To use in kernel modules, should be enough to remove "backport" entry from
> /etc/depmod.d/dist.conf and reboot the system.
Sigh. You are right. Knowing if and when we have to patch the backported drivers is a bit non-obvious. I'll reopen this for now and get it fixed, especially since we use those by default.
OK. Should be fixed now. THIS TIME WITH FEELING!
kernel-3.2.1-3.fc16 has been submitted as an update for Fedora 16.
* should fix your issue,
* was pushed to the Fedora 16 testing repository,
* should be available at your local mirror within two days.
Update it with:
# su -c 'yum update --enablerepo=updates-testing kernel-3.2.1-3.fc16'
as soon as you are able to, then reboot.
Please go to the following url:
then log in and leave karma (feedback).
kernel-3.2.1-3.fc16 has been pushed to the Fedora 16 stable repository. If problems still persist, please make note of it in this bug report.