rt2500pci (git HEAD) fails to associate

Live forum: http://rt2x00.serialmonkey.com/viewtopic.php?t=4809

Hazzl

24-05-2008 09:03:09

Hi Ivo,

latest git HEAD boots correctly, however, I'm unable to associate to my WEP encrypted AP (

Bye
Hazzl

Hazzl

24-05-2008 09:04:11

Couldn't attach this to last post

Hazzl

24-05-2008 09:56:43

I found out that in the bad case, more information is dumped in dmseg than in the kernel log

Fjodor

25-05-2008 07:28:46

+1 on this. It's still better than the BUG-thingy, though ;-)

/F

IvD

25-05-2008 15:54:53

Is this with rt2x00.git which includes the patch "rt2x00 Clear ENTRY_OWNER_DEVICE_DATA on write_tx_data failure"?

Hazzl

25-05-2008 16:11:36

The original report isn't. But I've retested since and commit 735c4078530 doesn't fix the issue.

IvD

25-05-2008 16:44:11

Ok, could you try doing
git pull git//git.kernel.org/pub/scm/linux/kern ... rt2x00.git wirelesswireless
git checkout wireless

and compile that kernel? That corresponds to the latest wireless-testing git kernel,
and if that works that would seriously limit the number of patches that might have broken it.

Note that it is advised to do
git checkout master

after running the test to make sure you are back to latest rt2x00 version again.

Hazzl

25-05-2008 17:29:16

The wireless branch still has the bug that schedules while atomic. When I apply the fix for that bug to the wireless branch, the resulting kernel boots but can't associate cry Sorry

Hazzl

31-05-2008 05:14:18

I tried to capture the conversation between my laptop and the ap using wireshark. It seems, the ap never sends any replies. I'm attaching the critical section of the tcpdump, does that provide any clues as to what is going on?

Fjodor

02-06-2008 15:42:08

Just wanted to send dmesg output and dumps from latest git, in case someone wondered how it went (no association)...

/F