Bug#383169: Doesn't reassociate after resume

[Kel Modderman]

On Sunday 15 October 2006 19:57, Reinhard Tartler wrote:
> Kel Modderman <kelrin at tpg.com.au> writes:
> >> I think http://madwifi.org/ticket/671 may be related, but 'sudo iwconfig
> >> ath0 FAU-VPN' in another shell as suggested in that ticket  didn't help
> >> me :(
> >
> > And #201. Its a pretty old, known issue. I've made a few attempts at
> > looking at it with no result so far. Will make another effort soon, but I
> > cannot promise that anything can be fixed in a driver that is partially
> > opaque.
>
> since 0.9.2+r1710.20060914-1, this bug is fixed at least for me. In the
> referenced upstream bug the issue seems still not solved, so I leave it
> to the madwifi maintainers to close this.

Thanks for the feedback. IIRC, last time I tested madwifi/suspend, it is not 
working with my hardware. So it seems it has improved, but is not "fixed".

>
> >> My current workaround is to rmmod ath-hal and reload ath-pci after
> >> resuming, which is suboptimal and shouldn't be necessary. In fact, this
> >> seems to be a regression from the madwifi-old branch.
> >
> > Yep, it is a regression that is biting me too, although most good suspend
> > managers (ie powersave) hide it quite well by automating module removal
> > and reinsertion,
>
> They do this because of bugs in device drivers like this one. I don't
> think the ability to reload device drivers on suspend is a good excuse
> for not caring about these bugs, since ...
>
> > however this will require the networking system(s) to be restarted,
> > which is no good . . .
>
> is no good and causes other problems.

Full ACK.

Kel.