Re: iwn(4) in -HEAD supporting Centrino Wireless-N 135

I still don't have any ideas here. I do however want to try hacking
the driver to transmit EAPOL frames at the management rate, and then
ensure the management rate is non-MCS.


-a


On 28 February 2014 15:14, Adrian Chadd <adrian_at_freebsd.org> wrote:
> Hi,
>
> the interesting bits:
>
>
> Feb 28 22:55:22 kernel: iwn_tx_data: qid 3 idx 3 len 129 nsegs 2 rate
> 0002 plcp 0x0000420a
> Feb 28 22:55:22 kernel: iwn_tx_data: qid 3 idx 4 len 129 nsegs 2 rate
> 0002 plcp 0x0000420a
> Feb 28 22:55:22 kernel: iwn5000_tx_done: qid 3 idx 3 retries 16 nkill
> 0 rate 80006902 duration 2815 status 83
> Feb 28 22:55:22 kernel: iwn5000_tx_done: qid 3 idx 4 retries 16 nkill
> 0 rate 80006902 duration 2815 status 83
> Feb 28 22:55:23 kernel: iwn_tx_data: qid 3 idx 5 len 129 nsegs 2 rate
> 0002 plcp 0x0000420a
> Feb 28 22:55:23 kernel: iwn_tx_data: qid 3 idx 6 len 129 nsegs 2 rate
> 0002 plcp 0x0000420a
> Feb 28 22:55:23 kernel: iwn5000_tx_done: qid 3 idx 5 retries 16 nkill
> 0 rate 80006902 duration 2815 status 83
> Feb 28 22:55:23 kernel: iwn5000_tx_done: qid 3 idx 6 retries 16 nkill
> 0 rate 80006902 duration 2815 status 83
> Feb 28 22:55:24 kernel: iwn_tx_data: qid 3 idx 7 len 129 nsegs 2 rate
> 0002 plcp 0x0000420a
> Feb 28 22:55:24 kernel: iwn_tx_data: qid 3 idx 8 len 129 nsegs 2 rate
> 0002 plcp 0x0000420a
> Feb 28 22:55:24 kernel: iwn5000_tx_done: qid 3 idx 7 retries 16 nkill
> 0 rate 80006902 duration 2815 status 83
> Feb 28 22:55:24 kernel: iwn5000_tx_done: qid 3 idx 8 retries 16 nkill
> 0 rate 80006902 duration 2815 status 83
>
> .. so it's failing to transmit the management frames after association
> - they're being transmitted at MCS0 and the AP is just plain not
> ACKing them.
>
> Now, I don't know why this is. It's trying to transmit the initial
> frame at non-MCS rates, but I have a feeling the multi-rate retry
> table thing is confusing it and it's trying to send it as MCS. So
> maybe the AP doesn't like management frames at MCS rates.
>
> I'll have to think about this a little.
>
> -a
>
>
> On 28 February 2014 15:07, Tom Murphy <freebsd_at_pertho.net> wrote:
>> I've attached my iwn debug messages to this email starting
>> with the point I tried to associate to the Wifi.
>>
>> Thanks again for looking at this!
>>
>> Kind regards,
>> Tom
>>
>> On Thu, Feb 27, 2014 at 12:13:51PM -0800, Adrian Chadd wrote:
>>> On 26 February 2014 23:52, Alexandr <shuriku_at_shurik.kiev.ua> wrote:
>>> > Tom, could you:
>>> >
>>> > 1. compile kernel WITH_IWNDEBUG
>>> > 2. sysctl dev.iwn.0.debug=0x1
>>> > 3. wlandebug -i wlan0 auth+assoc
>>> > 4. Associate with AP in 11n mode
>>> > 5. Send us appropriate /var/log/messages
>>> >
>>> > Then I try to compare it with my log.
>>>
>>> Please do. I've been trying to track down the source of this "ht just
>>> doesn't work!" but it works fine with all of the Intel NICs I have
>>> here.
>>>
>>> Can someone see if they can find a mtaching NIC online (amazon,ebay?)
>>> Owning one that I can whack in a laptop is likely going ot help things
>>> a lot.
>>>
>>> Thanks,
>>>
>>>
>>> -a