On Tue, Aug 18, 2009 at 01:10:49PM +0100, Matthew Garrett wrote:
> On Mon, Aug 17, 2009 at 10:32:27PM -0700, Dmitry Torokhov wrote:
>
> > dell-laptop module does not work on my Latitude d630 unless driver for
> > my wireless card (which happens to be iwl3945) is already loaded.
> > Otherwise dcsbas_smi_request() returns all zeroes which causes
> > dell-laptop to not create any rfkill devices nor install backlight
> > device:
>
> That's quite bizarre. I'd had one report of the backlight setup failing,
> but hadn't connected it with wireless driver loading. Long shot, but I
> don't suppose you can figure out which bit of the wifi config is needed
> to get dell-laptop working by just bailing at various bits of the probe
> and seeing if you get sensible answers?
>
Admittedly I have not spend much time on it but it looks like my box
_really_ wants wifi interface to be fully operational before
dcsbas_smi_request() starts returning proper data. I wonder if iwl3945's
rfkill support interferes somehow here...
Maybe Mario (being a Dell guy) has an idea why this might be happening?
--
Dmitry
Hi Dmitry:
Dmitry Torokhov wrote:
> On Tue, Aug 18, 2009 at 01:10:49PM +0100, Matthew Garrett wrote:
>
>
> Admittedly I have not spend much time on it but it looks like my box
> _really_ wants wifi interface to be fully operational before
> dcsbas_smi_request() starts returning proper data. I wonder if iwl3945's
> rfkill support interferes somehow here...
>
> Maybe Mario (being a Dell guy) has an idea why this might be happening?
>
>
I'm on 2.6.31-rc8 or so. I can't actually reproduce this on a d630. I
get all of the rfkill devices populated.
$ rfkill list
0: dell-wifi: Wireless LAN
Soft blocked: yes
Hard blocked: yes
1: dell-bluetooth: Bluetooth
Soft blocked: yes
Hard blocked: yes
2: phy0: Wireless LAN
Soft blocked: no
Hard blocked: yes
$ sudo rmmod iwl3945 iwlcore dell-laptop
$ rfkill list
$ sudo modprobe dell-laptop
$ rfkill list
3: dell-wifi: Wireless LAN
Soft blocked: yes
Hard blocked: yes
4: dell-bluetooth: Bluetooth
Soft blocked: yes
Hard blocked: yes
<hit the hardware killswitch>
$ rfkill list
3: dell-wifi: Wireless LAN
Soft blocked: yes
Hard blocked: yes
4: dell-bluetooth: Bluetooth
Soft blocked: yes
Hard blocked: yes
5: hci0: Bluetooth
Soft blocked: no
Hard blocked: no
$ sudo modprobe iwl3945
$ rfkill list
3: dell-wifi: Wireless LAN
Soft blocked: yes
Hard blocked: yes
4: dell-bluetooth: Bluetooth
Soft blocked: yes
Hard blocked: yes
5: hci0: Bluetooth
Soft blocked: no
Hard blocked: no
6: phy1: Wireless LAN
Soft blocked: no
Hard blocked: no
There *is* a problem here that the hardware killswitch status doesn't
get updated, but Matthew is aware of that. Make sure you're not getting
false data because of it.
--
Mario Limonciello
*Dell | Linux Engineering*
[email protected]
On Thursday 03 September 2009 09:42:35 am Mario Limonciello wrote:
> Hi Dmitry:
>
> Dmitry Torokhov wrote:
> > On Tue, Aug 18, 2009 at 01:10:49PM +0100, Matthew Garrett wrote:
> >
> >
> > Admittedly I have not spend much time on it but it looks like my box
> > _really_ wants wifi interface to be fully operational before
> > dcsbas_smi_request() starts returning proper data. I wonder if iwl3945's
> > rfkill support interferes somehow here...
> >
> > Maybe Mario (being a Dell guy) has an idea why this might be happening?
> >
> >
> I'm on 2.6.31-rc8 or so. I can't actually reproduce this on a d630. I
> get all of the rfkill devices populated.
>
Hmm, what is the BIOS version in your box? Also, mine used to have
broadcom-based wifi and I replaced it with iwl3945, would that make
any difference?
Thanks!
--
Dmitry
Hi Dmitry:
Dmitry Torokhov wrote:
> On Thursday 03 September 2009 09:42:35 am Mario Limonciello wrote:
>
>
> Hmm, what is the BIOS version in your box? Also, mine used to have
> broadcom-based wifi and I replaced it with iwl3945, would that make
> any difference?
I'm on BIOS A06, but I wouldn't /expect/ any of the stable (A**) BIOS
releases to change any behaviors like this.
What card the machine shipped with won't have a bearing on the behaviors
with the current card. There aren't any factory processes that change
the BIOS based on the card, just processes that would add newer firmware
directly to cards that support it.
--
Mario Limonciello
*Dell | Linux Engineering*
[email protected]