[alsa-devel] Fwd: Re: no reset_resume for driver snd-usb-audio for logitech headset H600

[Takashi Iwai]

At Tue, 25 Mar 2014 20:21:34 +0100,
<baumber at hotmail.com> wrote:
> 
> On 2014-01-22 18:16 CET (UTC+1), Takashi Iwai wrote:
> > At Wed, 22 Jan 2014 18:12:45 +0100,
> > <baumber at hotmail.com> wrote:
> >>
> >> On 2014-01-20 17:07, Takashi Iwai wrote:
> >>> At Mon, 20 Jan 2014 09:45:58 +0100,
> >>> Takashi Iwai wrote:
> >>>>
> >>>> At Sun, 19 Jan 2014 21:30:58 +0100,
> >>>> <baumber at hotmail.com> wrote:
> >>>>>
> >>>>> Hello,
> >>>>>
> >>>>> there must be a misunderstanding;
> >>>>>
> >>>>> There is a functional change, because with the line/patch ".reset_resume = usb_audio_resume," the mixer of the usb headset was not removed during suspend/resume and so the master channel in KDE Mixer was not changed (Headset), which is the correct behaviour.
> >>>>>
> >>>>> Without the line ".reset_resume = usb_audio_resume," in /sound/usb/card.c, the mixer was removed during suspend/resume and re-added, and so the master channel changed in KDE Mixer to the default onboard audio.
> >>>>>
> >>>>> I have tested kernel 3.13rc8, and as you mentioned, the "reset_resume"-messages are gone.
> >>>>> But the behaviour of removing the mixer of the usb_audio_headset occurs, which is, in my opinion, not the correct behaviour, because the chosen master channel/mixer should stay after suspend/resume.
> >>>>>
> >>>>> Could you please take a look at the issue again.
> >>
> >>>> Simply adding usb_audio-resume to reset_resume ops doesn't work
> >>>> properly as Clemens already suggested.  It misses the recovery of the
> >>>> current mixer values.  This might work in some cases like S3 (where
> >>>> the device is kept more or less powered on), but it'll be definitely
> >>>> broken in some cases like S4.
> >>
> >> For my USB Logitech Headset (ID 046d:0a29 Logitech, Inc. H600 [Wireless Headset]) S3 and S4 are working (mixer resume), when using the "reset_resume" line/patch.
> >
> > You didn't change the mixer value (e.g. mute), right?  Otherwise
> > you'll have to notice the difference after S4.
> 
> I always let the last used mixer value stay (no mute), but I turn off the headset (power switch) before hibernate.
> 
> 
> >>> ... and below is a quick fix with mixer resume code.
> >>> It still doesn't handle some quirks, so it might break some devices.
> >>
> >> Thank you for your patch/quick fix, I'll try it.
> >>
> >> When can I expect a more complete solution for the mixer resume code, in the next kernel releases 3.14 or 3.15?
> >
> > Only after I get positive test reports.  If I get them earlier, it can
> > be in 3.14-rc1, as it's no too intrusive change.  If not, it'll be in
> > 3.15.
> 
> Thank you for your patch, it is working fine.
> 
> It works with hibernate, the mixer value stays and the master channel "headset" (kmix) is selected after hibernate as it should be.
> 
> When will the patch be in mainline => 3.14? and backported to LTS-kernels?

It'll be in 3.15 kernel.  No stable backport is planned, though.

> 
> Do you know, what is the meaning of the following message in /var/log/Xorg.0.log ?
> 
> (EE) [dix] Logitech Logitech Wireless Headset: unable to find touch point 0

This implies that your device provides a HID but X driver couldn't
handle it.


Takashi