2009-05-05 04:57:48

by Maxim Levitsky

[permalink] [raw]
Subject: [BUG] Connection to an access point is dropped shortly after a association

This is a recent regression.
For some access points, eg @ my home, everything is fine as before.
But here @ university (open access points) this happens.

Looking at output of 'iw dev wlan 0 station dump' reveals very large
'inactive time'.

Reconnecting restores the connection briefly for yet another 30 seconds
or so.

Best regards,
Maxim Levitsky

PS: this is iwlwifi.git commit:
5a94b6d38100b7056a5a347e5c51359d924d305d





2009-05-06 15:36:21

by Maxim Levitsky

[permalink] [raw]
Subject: Re: [BUG] Connection to an access point is dropped shortly after a association

On Tue, 2009-05-05 at 07:57 +0300, Maxim Levitsky wrote:
> This is a recent regression.
> For some access points, eg @ my home, everything is fine as before.
> But here @ university (open access points) this happens.
>
> Looking at output of 'iw dev wlan 0 station dump' reveals very large
> 'inactive time'.
>
> Reconnecting restores the connection briefly for yet another 30 seconds
> or so.
>
> Best regards,
> Maxim Levitsky
>
> PS: this is iwlwifi.git commit:
> 5a94b6d38100b7056a5a347e5c51359d924d305d
>
>
>
can't seem to reproduce it with wireless-testing of yestarday.
But I bet that this bug continues to exist.

Best regards,
Maxim Levitsky


2009-05-13 23:15:51

by Maxim Levitsky

[permalink] [raw]
Subject: Re: [BUG] Connection to an access point is dropped shortly after a association

On Wed, 2009-05-06 at 13:48 +0300, Maxim Levitsky wrote:
> On Tue, 2009-05-05 at 07:57 +0300, Maxim Levitsky wrote:
> > This is a recent regression.
> > For some access points, eg @ my home, everything is fine as before.
> > But here @ university (open access points) this happens.
> >
> > Looking at output of 'iw dev wlan 0 station dump' reveals very large
> > 'inactive time'.
> >
> > Reconnecting restores the connection briefly for yet another 30 seconds
> > or so.
> >
> > Best regards,
> > Maxim Levitsky
> >
> > PS: this is iwlwifi.git commit:
> > 5a94b6d38100b7056a5a347e5c51359d924d305d
> >
> >
> >
> can't seem to reproduce it with wireless-testing of yestarday.
> But I bet that this bug continues to exist.
Yep, this still happens.

I can easily reproduce that, but only at a specific spot, in my
University.

Happens with both old and new firmware.

I continue to investigate, maybe bisect.
Works fine with ubuntu kernel from 9.04.

It can be related to roaming.

Best regards,
Maxim Levitsky