3.4.0-rc7 regression: bluetooth vs. s2ram

3.4.0-rc7 regression: bluetooth vs. s2ram

[Pavel Machek]

Hi!

In 3.4-rc7, if I suspend while bnep0 over btusb is connected, then
resume, system dies with a backtrace. (Unfortunately, it is not
exactly easy to capture).

If I suspend without bnep0 active, resume works. 

(This has worked before, but "before" may mean 3.2 in this case).

Can someone test/reproduce this? Ideas?

									Pavel

PS: Under normal suspend/resume, I see these warnings in the
log. Maybe relevant?

ehci_hcd 0000:00:1d.7: GetStatus port:7 status 003002 0  ACK POWER OWNER sig=se0 CSC
usb 5-2: finish reset-resume
uhci_hcd 0000:00:1d.3: port 1 portsc 0093,00
usb 5-1: finish reset-resume
usb 5-2: reset full-speed USB device number 3 using uhci_hcd
usb 5-2: ep0 maxpacket = 8
usb 5-1: reset full-speed USB device number 2 using uhci_hcd
btusb 5-1:1.0: no reset_resume for driver btusb?
btusb 5-1:1.1: no reset_resume for driver btusb?
btusb 5-1:1.0: forced unbind
usb usb2: suspend_rh (auto-stop)
usb usb3: suspend_rh (auto-stop)
usb usb4: suspend_rh (auto-stop)
ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 300)

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

Re: 3.4.0-rc7 regression: bluetooth vs. s2ram

[Oliver Neukum]

Am Dienstag, 15. Mai 2012, 15:14:57 schrieb Pavel Machek:
> If I suspend without bnep0 active, resume works. 
> 
> (This has worked before, but "before" may mean 3.2 in this case).
> 
> Can someone test/reproduce this? Ideas?
> 
>                                                                         Pavel
> 
> PS: Under normal suspend/resume, I see these warnings in the
> log. Maybe relevant?

Most likely. Something has introduced a bug that crashes if the controller
over which a bnep is run is getting away. Does it survive a surprise removal
of the BT device while bnep is up?

	Regards
		Oliver

Re: 3.4.0-rc7 regression: bluetooth vs. s2ram

[Pavel Machek]

On Tue 2012-05-15 15:29:01, Oliver Neukum wrote:
> Am Dienstag, 15. Mai 2012, 15:14:57 schrieb Pavel Machek:
> > If I suspend without bnep0 active, resume works. 
> > 
> > (This has worked before, but "before" may mean 3.2 in this case).
> > 
> > Can someone test/reproduce this? Ideas?
> > 
> > PS: Under normal suspend/resume, I see these warnings in the
> > log. Maybe relevant?
> 
> Most likely. Something has introduced a bug that crashes if the controller
> over which a bnep is run is getting away. Does it survive a surprise removal
> of the BT device while bnep is up?

No, it does not. If I turn off hardware radio kill switch, machine
dies in the same way. So not just sleep related...
									Pavel

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

Re: 3.4.0-rc7 regression: bluetooth vs. s2ram

[Oliver Neukum]

Am Dienstag, 15. Mai 2012, 20:01:06 schrieb Pavel Machek:
> On Tue 2012-05-15 15:29:01, Oliver Neukum wrote:
> > Am Dienstag, 15. Mai 2012, 15:14:57 schrieb Pavel Machek:
> > > If I suspend without bnep0 active, resume works. 
> > > 
> > > (This has worked before, but "before" may mean 3.2 in this case).
> > > 
> > > Can someone test/reproduce this? Ideas?
> > > 
> > > PS: Under normal suspend/resume, I see these warnings in the
> > > log. Maybe relevant?
> > 
> > Most likely. Something has introduced a bug that crashes if the controller
> > over which a bnep is run is getting away. Does it survive a surprise removal
> > of the BT device while bnep is up?
> 
> No, it does not. If I turn off hardware radio kill switch, machine
> dies in the same way. So not just sleep related...

You lose power, btusb cannot deal with that and you get an unplug/replug cycle.
Can you get an oops when you are hitting the kill switch?

	Regards
		Oliver