Re: [Ilw] Re: 3.4-rc2, ilwagn still most of the time completely unusable

[Johannes Berg <johannes@sipsolutions.net>]

On Thu, 2012-07-12 at 17:14 +0900, Norbert Preining wrote:

> > I can actually work now wirelessly, and in case of hangs an rfkill block
> > rfkill unblock practically always, and mostly is not needed.
> 
> Ok, that was said *UNTIL* I reenabled 11n. Up to now I was loading the
> iwlwifi module with
> 	11n_disable=1
> 
> Now, latest git from today, I though, let us try to enable 11n again,
> and there we go, boot from cold machine and after short time everything
> is dead without reaction. Dmesg gives many many funny messages:
> 
> [   30.047943] iwlwifi 0000:06:00.0: L1 Enabled; Disabling L0S
> [   30.051933] iwlwifi 0000:06:00.0: Radio type=0x1-0x2-0x0
> [   30.164836] iwlwifi 0000:06:00.0: L1 Enabled; Disabling L0S
> [   30.167876] iwlwifi 0000:06:00.0: Radio type=0x1-0x2-0x0
> [   37.533619] wlan0: authenticate with 00:0a:79:eb:56:10
> [   37.536433] wlan0: send auth to 00:0a:79:eb:56:10 (try 1/3)
> [   37.538241] wlan0: authenticated
> [   37.538391] wlan0: associating with AP with corrupt beacon

That might be an indication of things already going wrong with your
AP ... :-(

[snip log]

Mostly from the log it seems that there's a lot of spurious aggregation
setup/teardown, some of it seems to originate from the AP.

> At that point I removed the iwlwifi module, and reloaded it with 
> 11n_disable=1 again, otherwise nothing would work at all...

With the latest, I've extended 11n_disable to have more bits.

11n_disable=1 will disable HT completely
11n_disable=2 will disable TX aggregation
11n_disable=4 will disable RX aggregation

It'd be interesting to see what happens for you with 11n_disable=2/4/6.

johannes