RE: 3.18-rc0: iwlegacy failed after a while

RE: 3.18-rc0: iwlegacy failed after a while

[Grumbach, Emmanuel]

[Changed subject]
This is really iwlegacy and not iwlwifi.


> 
> Hi!
> 
> Full dmesg is in the attachment, I guess the troubles started with:
> 
> iwl3945 0000:03:00.0: Queue 4 stuck for 2004 ms.
> iwl3945 0000:03:00.0: On demand firmware reload
> iwl3945 0000:03:00.0: Master Disable Timed Out, 100 usec
> ieee80211 phy0: Hardware restart was requested
> iwl3945 0000:03:00.0: BSM uCode verification failed at addr
> 0x00003800+0 (of 900), is 0xa5\
> a5a5a2, s/b 0xf802020
> iwl3945 0000:03:00.0: Unable to set up bootstrap uCode: -5
> 
> Best regards,
> 								Pavel
> 
> --

Huh - I don't remember any change there, but you should ask the right person - iwlegacy maintainer :)

Re: 3.18-rc0: iwlegacy failed after a while

[Stanislaw Gruszka]

On Thu, Oct 23, 2014 at 01:02:37PM +0000, Grumbach, Emmanuel wrote:
> [Changed subject]
> This is really iwlegacy and not iwlwifi.
> 
> 
> > 
> > Hi!
> > 
> > Full dmesg is in the attachment, I guess the troubles started with:
> > 
> > iwl3945 0000:03:00.0: Queue 4 stuck for 2004 ms.
> > iwl3945 0000:03:00.0: On demand firmware reload
> > iwl3945 0000:03:00.0: Master Disable Timed Out, 100 usec
> > ieee80211 phy0: Hardware restart was requested
> > iwl3945 0000:03:00.0: BSM uCode verification failed at addr
> > 0x00003800+0 (of 900), is 0xa5\
> > a5a5a2, s/b 0xf802020
> > iwl3945 0000:03:00.0: Unable to set up bootstrap uCode: -5
> > 
> > Best regards,
> > 								Pavel
> > 
> > --
> 
> Huh - I don't remember any change there, but you should ask the right person - iwlegacy maintainer :)

Since 3.17 we have only 2 minor iwlegacy changes on current linus tree.
If this is a regression it's probably caused by different subsystem
changes, most likely by PCI changes. If this is not regression, such
errors could be caused by enabling PS, but that has to be done
explicitly, Pavel did you enable power_save ? 

Stanislaw

Re: 3.18-rc0: iwlegacy failed after a while

[Pavel Machek]

On Fri 2014-10-24 14:12:40, Stanislaw Gruszka wrote:
> On Thu, Oct 23, 2014 at 01:02:37PM +0000, Grumbach, Emmanuel wrote:
> > [Changed subject]
> > This is really iwlegacy and not iwlwifi.
> > 
> > 
> > > 
> > > Hi!
> > > 
> > > Full dmesg is in the attachment, I guess the troubles started with:
> > > 
> > > iwl3945 0000:03:00.0: Queue 4 stuck for 2004 ms.
> > > iwl3945 0000:03:00.0: On demand firmware reload
> > > iwl3945 0000:03:00.0: Master Disable Timed Out, 100 usec
> > > ieee80211 phy0: Hardware restart was requested
> > > iwl3945 0000:03:00.0: BSM uCode verification failed at addr
> > > 0x00003800+0 (of 900), is 0xa5\
> > > a5a5a2, s/b 0xf802020
> > > iwl3945 0000:03:00.0: Unable to set up bootstrap uCode: -5
> > > 
> > 
> > Huh - I don't remember any change there, but you should ask the right person - iwlegacy maintainer :)
> 
> Since 3.17 we have only 2 minor iwlegacy changes on current linus tree.
> If this is a regression it's probably caused by different subsystem
> changes, most likely by PCI changes. If this is not regression, such
> errors could be caused by enabling PS, but that has to be done
> explicitly, Pavel did you enable power_save ? 

This is running pretty standart MATE desktop, and no, I don't think I
enabled that.
									Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: 3.18-rc0: iwlegacy failed after a while

[Stanislaw Gruszka]

On Sat, Oct 25, 2014 at 01:42:46PM +0200, Pavel Machek wrote:
> > Since 3.17 we have only 2 minor iwlegacy changes on current linus tree.
> > If this is a regression it's probably caused by different subsystem
> > changes, most likely by PCI changes. If this is not regression, such
> > errors could be caused by enabling PS, but that has to be done
> > explicitly, Pavel did you enable power_save ? 
> 
> This is running pretty standart MATE desktop, and no, I don't think I
> enabled that.

It can be checked by "iw dev wlan0 get power_save" .

More important quiestions: Is this regression, did you see this error
before? How reproducible is this problem?

There is this message in the dmesg:

iwl3945 0000:03:00.0: BSM uCode verification failed at addr 0x00003800+0 (of 900), is 0xa5a5a5a2, s/b 0xf802020

0xa5a5a5a2 value suggest that we disconnect from the PCI bus, hence
this could be also a H/W problem.

Stanislaw

Re: 3.18-rc0: iwlegacy failed after a while

[Pavel Machek]

On Sun 2014-10-26 11:33:15, Stanislaw Gruszka wrote:
> On Sat, Oct 25, 2014 at 01:42:46PM +0200, Pavel Machek wrote:
> > > Since 3.17 we have only 2 minor iwlegacy changes on current linus tree.
> > > If this is a regression it's probably caused by different subsystem
> > > changes, most likely by PCI changes. If this is not regression, such
> > > errors could be caused by enabling PS, but that has to be done
> > > explicitly, Pavel did you enable power_save ? 
> > 
> > This is running pretty standart MATE desktop, and no, I don't think I
> > enabled that.
> 
> It can be checked by "iw dev wlan0 get power_save" .
> 
> More important quiestions: Is this regression, did you see this error
> before? How reproducible is this problem?

I've never seen it before, and it only happened once after update to
3.18-rc.

> There is this message in the dmesg:
> 
> iwl3945 0000:03:00.0: BSM uCode verification failed at addr 0x00003800+0 (of 900), is 0xa5a5a5a2, s/b 0xf802020
> 
> 0xa5a5a5a2 value suggest that we disconnect from the PCI bus, hence
> this could be also a H/W problem.

That machine is 10 years old and will work for another 10. It does not
have h/w problems ;-).

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: 3.18-rc0: iwlegacy failed after a while

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 1560 bytes --]

On Sun 2014-10-26 12:05:54, Pavel Machek wrote:
> On Sun 2014-10-26 11:33:15, Stanislaw Gruszka wrote:
> > On Sat, Oct 25, 2014 at 01:42:46PM +0200, Pavel Machek wrote:
> > > > Since 3.17 we have only 2 minor iwlegacy changes on current linus tree.
> > > > If this is a regression it's probably caused by different subsystem
> > > > changes, most likely by PCI changes. If this is not regression, such
> > > > errors could be caused by enabling PS, but that has to be done
> > > > explicitly, Pavel did you enable power_save ? 
> > > 
> > > This is running pretty standart MATE desktop, and no, I don't think I
> > > enabled that.
> > 
> > It can be checked by "iw dev wlan0 get power_save" .

So it seems I have power save on. I guess someone enabled it for me :-).

root@duo:~# iwconfig wlan0
wlan0     IEEE 802.11abg  ESSID:off/any  
          Mode:Managed  Access Point: Not-Associated   Tx-Power=15 dBm   
          Retry short limit:7   RTS thr:off   Fragment thr:off
          Encryption key:off
          Power Management:on
          
root@duo:~# iw dev wlan0 get power_save
Power save: on
root@duo:~# 

> > More important quiestions: Is this regression, did you see this error
> > before? How reproducible is this problem?
> 
> I've never seen it before, and it only happened once after update to
> 3.18-rc.

Today wifi failed, again. Will reboot. dmesg is in the attachment,
looks like same fail.

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

[-- Attachment #2: delme.gz --]
[-- Type: application/gzip, Size: 20350 bytes --]

Re: 3.18-rc0: iwlegacy failed after a while

[Stanislaw Gruszka]

On Tue, Oct 28, 2014 at 11:20:15PM +0100, Pavel Machek wrote:
> > > It can be checked by "iw dev wlan0 get power_save" .
> 
> So it seems I have power save on. I guess someone enabled it for me :-).
> 
> root@duo:~# iwconfig wlan0
> wlan0     IEEE 802.11abg  ESSID:off/any  
>           Mode:Managed  Access Point: Not-Associated   Tx-Power=15 dBm   
>           Retry short limit:7   RTS thr:off   Fragment thr:off
>           Encryption key:off
>           Power Management:on
>           
> root@duo:~# iw dev wlan0 get power_save
> Power save: on
> root@duo:~# 

Could you confirm that problem happen with power save on and do not
happen without it ?

Stanislaw

Re: 3.18-rc0: iwlegacy failed after a while

[Pavel Machek]

On Sun 2014-11-02 13:25:45, Stanislaw Gruszka wrote:
> On Tue, Oct 28, 2014 at 11:20:15PM +0100, Pavel Machek wrote:
> > > > It can be checked by "iw dev wlan0 get power_save" .
> > 
> > So it seems I have power save on. I guess someone enabled it for me :-).
> > 
> > root@duo:~# iwconfig wlan0
> > wlan0     IEEE 802.11abg  ESSID:off/any  
> >           Mode:Managed  Access Point: Not-Associated   Tx-Power=15 dBm   
> >           Retry short limit:7   RTS thr:off   Fragment thr:off
> >           Encryption key:off
> >           Power Management:on
> >           
> > root@duo:~# iw dev wlan0 get power_save
> > Power save: on
> > root@duo:~# 
> 
> Could you confirm that problem happen with power save on and do not
> happen without it ?

It has not happened for a while, but I was not loading wifi as
heavily.

But even powersave=on case needs to be fixed, right?
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html