SIOCSIWAUTH not supported?

Live forum: http://rt2x00.serialmonkey.com/viewtopic.php?t=4612

Fjodor

10-02-2008 19:38:49

Pulled the git rt2x00 git tree today, but am surprised to find that wpa_supplicant just spits out

ioctl[SIOCSIWAUTH] Operation not supported
WEXT auth param 4 value0x0 -

Any ideas as to why this is?

Best regards,

F

IvD

10-02-2008 21:30:45

It's normal and can be ignored.
It's trying to set some vague encryption parameter which isn't implemented in mac80211 (or any non-mac80211 drivers). I think it will disappear in a future version of wpa_supplicant.

Fjodor

11-02-2008 11:04:51

Thanks for the answer, but the problem is that this is all that it does, and hence I get no connection...

Any ideas?

F

Fjodor

11-02-2008 12:16:15

Ah, and enabling debugging reveals lots of the following

phy0 -> rt2500pci_bbp_read Error - BBPCSR register busy. Read failed
phy0 -> rt2500pci_init_bbp Notice - Waiting for BBP register.

followed by

phy0 -> rt2500pci_enable_radio Error - Register initialization failed.

Hope this gives something to work with ;-)

Best regards,

F

Fjodor

11-02-2008 12:30:10

No, wait, it seems this was caused by not unloading all relevant modules...

Last 2 messages now are

phy2 -> rt2x00_set_chip Info - Chipset detected - rt 0201, rf 0003, rev 00000004.
phy2 Selected rate control algorithm 'pid'

Shouldn't there be more after this?

Best regards,

F

Fjodor

11-02-2008 12:57:00

A little more info wpa_supplicant -dd seems to suggest that it scans for my AP in vain, and indeed I get no scan results whatsoever with iwlist...

Best regards,

F

IvD

11-02-2008 15:13:39

Current rt2x00.git master branch is known to be broken. Together with Adam I am looking for a solution.

Fjodor

11-02-2008 15:16:21

Great that you are on it. Please let me know if I can help in any way -)

Best regards,

F

IvD

15-02-2008 14:11:08

Please retest with latest rt2x00.git. The issue should be fixed.

Fjodor

15-02-2008 15:06:32

Have initiated compilation. Will install and test when I get home -)

Thanks!

F

Fjodor

15-02-2008 20:12:52

I'm sorry to report that that didn't do it -(

What info do you need?

Best regards,

F

AdamBaker

16-02-2008 15:27:36

First info is exactly what version you tested with, as the tree was only intermittently usable yesterday.

Fjodor

16-02-2008 15:50:55

Well, I'll cg-update -f right now then. The answer would then be "from today around tea-time", I suppose -)

Best regards,

F

Fjodor

16-02-2008 16:21:58

But if rt2500pci doesn't work when I boot up the new kernel, which test do you want me to run, and which dumps should I make?

Best regards,

F

IvD

16-02-2008 16:40:13

At the moment I am lacking ideas about this issue, so I really am hoping it is working. If it is not, I don't have any backup plans ready. (

Fjodor

16-02-2008 19:29:37

Sorry, no go -(

Not knowing what to do, I issued 'dmesg |grep rt2 > dmesg.txt' and wpa_supplicant with some options > wpa_supplicant.txt (both commands are at the top of the respective files)...

Please let me know what else I might contribute with...

Best regards, and sorry about the bad news,

F

IvD

16-02-2008 21:34:38

Hmm, small idea
Could you try enabling debugfs and check what the file contents is of <debugfsroot>/<mac80211 root>/rt2500pci/queue/queue

check the contents after the TX/RX are failing.

AdamBaker

17-02-2008 00:00:37

It would also be useful to know if scan works either before or after it has failed and if you can switch to monitor mode and then if receive works, e.g. in wireshark.

My current belief is that Tx works for a while on rt61 and then fails whereas on rt73 Tx never works (checked by monitoring with another adapter whilst doing a scan). After Tx has failed on rt61 receive is still OK.

Fjodor

17-02-2008 11:28:26

Hmm, small idea
Could you try enabling debugfs and check what the file contents is of <debugfsroot>/<mac80211 root>/rt2500pci/queue/queue

check the contents after the TX/RX are failing.[/quote26zaxzxd]

Hmmm, the mac80211 dir exists as /mnt/debug/ieee80211/phy0, however there is no rt2500pci dir, and no dir called queue whatsoever...

/F

IvD

17-02-2008 12:30:43

Did you enable debugfs support for rt2x00?

Fjodor

17-02-2008 12:47:27

I was 99% sure that i did, but actually, it wasn't set. Do I feel silly -$

Rebuilding checking...

/F

Fjodor

17-02-2008 16:26:53

Here we go -)

/F

IvD

19-02-2008 11:46:33

Please test attached patch to see if that helps.

IvD

19-02-2008 18:11:12

Instead of the patch, please update to latest rt2x00.git version. That one contains above patch + 2nd fix.

Fjodor

22-02-2008 00:45:40

Sorry for the late reply, but I had a dying harddisk to attend to -(

And sorry to report that it didn't work. Note that no packages whatsoever seem to have been transferred on wlan0...

What else might we try?

Best regards,

F

IvD

25-02-2008 14:59:42

Please retest with latest rt2x00.git, I think latest patch might have solved this issue.

Fjodor

25-02-2008 16:25:46

Sounds great!

Compiling as we speak -)

Best regards,

F

Fjodor

25-02-2008 20:05:26

Sorry, but no cigar -(

Btw., try to compare theese two files...

IvD

25-02-2008 22:09:08

Hmm, strange. Current success rate of the current version is now 31 (3 success, 1 failure).

Those logs aren't a real good indication what is going on, as long as debugfs reports all queues (6 for rt61pci) then all queues are present.

AdamBaker

26-02-2008 00:07:44

Fjodor said earlier that he had an rt2500pci and I don't think I've seen any other success reports for that.

I'm still seeing rt61 fail under moderate load with a git that includes up to rt2x00 Fix rt2x00lib_reset_link_tuner() - I think the Tx queue processing fails when it dies and this is a card that suffered badly from the non free entry in non full queue problem. Strangely when it dies the queue debugfs indicates that receive is still operating but after an ifdown then ifup it doesn't receive beacon's during a scan. It will only recover with an ifdown followed by unplugging and plugging back in, if you don't ifdown unplugging isn't enough.

I'm also still not seeing any transmit activity on rt73 even though the queue done counter increases and the activity LED flashes.