This is minor, and qute old issue, but botting in ubuntu kernel, 'fixes'
it so it must be a regression.
Sorry for not reporting this earlier.
Best regards,
Maxim Levitsky
On Mon, 2009-05-04 at 12:27 -0700, Cahill, Ben M wrote:
>
> >-----Original Message-----
> >From: Maxim Levitsky [mailto:[email protected]]
> >Sent: Monday, May 04, 2009 3:04 PM
> >To: Cahill, Ben M
> >Cc: iwlwifi maling list; linux-wireless
> >Subject: RE: [ipw3945-devel] IWL3945 always reports noise
> >levels of -127
> >
> >On Mon, 2009-05-04 at 11:53 -0700, Cahill, Ben M wrote:
> >> When one of our devices cannot figure out what the actual
> >noise level is, it reports -127.
> >>
> >> Newer devices are better at old devices at figuring noise
> >level. IIRC, 3945 cannot figure noise for CCK signals, and
> >may not be possible with OFDM as well.
> >
> >Why then it did work on older kernel ?
> >(and noise levels seemed to be correct)
>
> Hmmm, good point, and I don't know the answer! :-(
>
And I do know it now (this is what I suspected...)
This is firmware issue.
I forced the driver to use older firmware iwlwifi-3945-1.ucode,
and noise levels are back.
Regards,
Maxim Levitsky
On Tue, 2009-05-05 at 11:45 -0700, Cahill, Ben M wrote:
> Excellent information. Thank you.
>
> We just issued new 3945 uCode within the past week or two ... Is that also showing the problem (noise_diff = 0)?
>
Yes, I have just installed it to be sure.
Exactly same looking output.
Best regards,
Maxim Levitsky
Maxim,
On Tue, 2009-05-05 at 11:56 -0700, Maxim Levitsky wrote:
> On Tue, 2009-05-05 at 11:45 -0700, Cahill, Ben M wrote:
> > Excellent information. Thank you.
> >
> > We just issued new 3945 uCode within the past week or two ... Is that also showing the problem (noise_diff = 0)?
> >
> Yes, I have just installed it to be sure.
> Exactly same looking output.
>
This will take some coordination to resolve. Could you please open a new
bug at http://www.intellinuxwireless.org/bugzilla/ ?
Thank you
Reinette
On Mon, 2009-05-04 at 11:53 -0700, Cahill, Ben M wrote:
> When one of our devices cannot figure out what the actual noise level is, it reports -127.
>
> Newer devices are better at old devices at figuring noise level. IIRC, 3945 cannot figure noise for CCK signals, and may not be possible with OFDM as well.
Why then it did work on older kernel ?
(and noise levels seemed to be correct)
Regards,
Maxim Levitsky
On Fri, 2009-05-08 at 16:25 -0700, reinette chatre wrote:
> Maxim,
>
> On Tue, 2009-05-05 at 11:56 -0700, Maxim Levitsky wrote:
> > On Tue, 2009-05-05 at 11:45 -0700, Cahill, Ben M wrote:
> > > Excellent information. Thank you.
> > >
> > > We just issued new 3945 uCode within the past week or two ... Is that also showing the problem (noise_diff = 0)?
> > >
> > Yes, I have just installed it to be sure.
> > Exactly same looking output.
> >
>
> This will take some coordination to resolve. Could you please open a new
> bug at http://www.intellinuxwireless.org/bugzilla/ ?
>
> Thank you
>
> Reinette
>
Done.
http://www.intellinuxwireless.org/bugzilla/show_bug.cgi?id=1993
Best regards,
Maxim Levitsky
On Tue, 2009-05-05 at 17:50 +0300, Maxim Levitsky wrote:
> On Mon, 2009-05-04 at 12:27 -0700, Cahill, Ben M wrote:
> >
> > >-----Original Message-----
> > >From: Maxim Levitsky [mailto:[email protected]]
> > >Sent: Monday, May 04, 2009 3:04 PM
> > >To: Cahill, Ben M
> > >Cc: iwlwifi maling list; linux-wireless
> > >Subject: RE: [ipw3945-devel] IWL3945 always reports noise
> > >levels of -127
> > >
> > >On Mon, 2009-05-04 at 11:53 -0700, Cahill, Ben M wrote:
> > >> When one of our devices cannot figure out what the actual
> > >noise level is, it reports -127.
> > >>
> > >> Newer devices are better at old devices at figuring noise
> > >level. IIRC, 3945 cannot figure noise for CCK signals, and
> > >may not be possible with OFDM as well.
> > >
> > >Why then it did work on older kernel ?
> > >(and noise levels seemed to be correct)
> >
> > Hmmm, good point, and I don't know the answer! :-(
> >
>
> And I do know it now (this is what I suspected...)
> This is firmware issue.
>
> I forced the driver to use older firmware iwlwifi-3945-1.ucode,
> and noise levels are back.
Indeed old microcode reports values like that:
rssi:017 agc:000 sig_avg:5437 noise_diff:580
New microcode reports this instead:
rssi:018 agc:000 sig_avg:140 noise_diff:000
noise_diff is always zero now, and range for sig_avg changed.
I also noticed that with older firmware I couldn't set the tx speed
Thus what recently fixed it was a firmware update!
(But in both cases OFDM rated were used)
Thats all nothing to debug any more here.
Best regards,
Maxim Levitsky
Excellent information. Thank you.
We just issued new 3945 uCode within the past week or two ... Is that also showing the problem (noise_diff = 0)?
-- Ben --
>-----Original Message-----
>From: Maxim Levitsky [mailto:[email protected]]
>Sent: Tuesday, May 05, 2009 12:17 PM
>To: Cahill, Ben M
>Cc: iwlwifi maling list; linux-wireless
>Subject: RE: [ipw3945-devel] IWL3945 always reports noise
>levels of -127
>
>On Tue, 2009-05-05 at 17:50 +0300, Maxim Levitsky wrote:
>> On Mon, 2009-05-04 at 12:27 -0700, Cahill, Ben M wrote:
>> >
>> > >-----Original Message-----
>> > >From: Maxim Levitsky [mailto:[email protected]]
>> > >Sent: Monday, May 04, 2009 3:04 PM
>> > >To: Cahill, Ben M
>> > >Cc: iwlwifi maling list; linux-wireless
>> > >Subject: RE: [ipw3945-devel] IWL3945 always reports noise
>> > >levels of -127
>> > >
>> > >On Mon, 2009-05-04 at 11:53 -0700, Cahill, Ben M wrote:
>> > >> When one of our devices cannot figure out what the actual
>> > >noise level is, it reports -127.
>> > >>
>> > >> Newer devices are better at old devices at figuring noise
>> > >level. IIRC, 3945 cannot figure noise for CCK signals, and
>> > >may not be possible with OFDM as well.
>> > >
>> > >Why then it did work on older kernel ?
>> > >(and noise levels seemed to be correct)
>> >
>> > Hmmm, good point, and I don't know the answer! :-(
>> >
>>
>> And I do know it now (this is what I suspected...)
>> This is firmware issue.
>>
>> I forced the driver to use older firmware iwlwifi-3945-1.ucode,
>> and noise levels are back.
>
>
>Indeed old microcode reports values like that:
>rssi:017 agc:000 sig_avg:5437 noise_diff:580
>
>
>New microcode reports this instead:
>rssi:018 agc:000 sig_avg:140 noise_diff:000
>
>
>noise_diff is always zero now, and range for sig_avg changed.
>
>I also noticed that with older firmware I couldn't set the tx speed
>Thus what recently fixed it was a firmware update!
>(But in both cases OFDM rated were used)
>
>Thats all nothing to debug any more here.
>
>
>Best regards,
> Maxim Levitsky
>
>
When one of our devices cannot figure out what the actual noise level is, it reports -127.
Newer devices are better at old devices at figuring noise level. IIRC, 3945 cannot figure noise for CCK signals, and may not be possible with OFDM as well.
-- Ben --
>-----Original Message-----
>From: Maxim Levitsky [mailto:[email protected]]
>Sent: Friday, May 01, 2009 5:31 PM
>To: iwlwifi maling list
>Cc: linux-wireless
>Subject: [ipw3945-devel] IWL3945 always reports noise levels of -127
>
>This is minor, and qute old issue, but botting in ubuntu
>kernel, 'fixes'
>it so it must be a regression.
>
>Sorry for not reporting this earlier.
>
>Best regards,
> Maxim Levitsky
>
>
>
>---------------------------------------------------------------
>---------------
>Register Now & Save for Velocity, the Web Performance & Operations
>Conference from O'Reilly Media. Velocity features a full day of
>expert-led, hands-on workshops and two days of sessions from industry
>leaders in dedicated Performance & Operations tracks. Use code
>vel09scf
>and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>_______________________________________________
>Ipw3945-devel mailing list
>[email protected]
>https://lists.sourceforge.net/lists/listinfo/ipw3945-devel
>
On Mon, 2009-05-04 at 12:27 -0700, Cahill, Ben M wrote:
>
> >-----Original Message-----
> >From: Maxim Levitsky [mailto:[email protected]]
> >Sent: Monday, May 04, 2009 3:04 PM
> >To: Cahill, Ben M
> >Cc: iwlwifi maling list; linux-wireless
> >Subject: RE: [ipw3945-devel] IWL3945 always reports noise
> >levels of -127
> >
> >On Mon, 2009-05-04 at 11:53 -0700, Cahill, Ben M wrote:
> >> When one of our devices cannot figure out what the actual
> >noise level is, it reports -127.
> >>
> >> Newer devices are better at old devices at figuring noise
> >level. IIRC, 3945 cannot figure noise for CCK signals, and
> >may not be possible with OFDM as well.
> >
> >Why then it did work on older kernel ?
> >(and noise levels seemed to be correct)
>
> Hmmm, good point, and I don't know the answer! :-(
>
> Guess it deserves a bugzilla.
>
> Would you be willing to look in the source to see if you can figure out why? Looks like a search for "noise" in iwl-3945.c will lead in the right direction.
>
> -- Ben --
Indeed I'll do.
One thing I need to check is whether this is related to firmware, as
there was a switch to newer one.
Anyway I'll see what I can do there.
I already did look at source, and I didn't yet find anything suspicios,
I guess few printk won't hurt there.
Thanks,
Maxim Levitsky
>-----Original Message-----
>From: Maxim Levitsky [mailto:[email protected]]
>Sent: Monday, May 04, 2009 3:04 PM
>To: Cahill, Ben M
>Cc: iwlwifi maling list; linux-wireless
>Subject: RE: [ipw3945-devel] IWL3945 always reports noise
>levels of -127
>
>On Mon, 2009-05-04 at 11:53 -0700, Cahill, Ben M wrote:
>> When one of our devices cannot figure out what the actual
>noise level is, it reports -127.
>>
>> Newer devices are better at old devices at figuring noise
>level. IIRC, 3945 cannot figure noise for CCK signals, and
>may not be possible with OFDM as well.
>
>Why then it did work on older kernel ?
>(and noise levels seemed to be correct)
Hmmm, good point, and I don't know the answer! :-(
Guess it deserves a bugzilla.
Would you be willing to look in the source to see if you can figure out why? Looks like a search for "noise" in iwl-3945.c will lead in the right direction.
-- Ben --
>
>Regards,
> Maxim Levitsky
>
>