On Fri, Jan 1, 2021 at 4:07 PM Takashi Iwai <tiwai@xxxxxxx> wrote:
>
> On Thu, 31 Dec 2020 19:06:43 +0100,
> Kai-Heng Feng wrote:
> >
> > On Thu, Dec 31, 2020 at 6:55 PM Takashi Iwai <tiwai@xxxxxxx> wrote:
> > >
> > > On Tue, 29 Dec 2020 14:38:15 +0100,
> > > Kai-Heng Feng wrote:
> > > >
> > > > System takes a very long time to suspend after commit 215a22ed31a1
> > > > ("ALSA: hda: Refactor codec PM to use direct-complete optimization"):
> > > > [ 90.065964] PM: suspend entry (s2idle)
> > > > [ 90.067337] Filesystems sync: 0.001 seconds
> > > > [ 90.185758] Freezing user space processes ... (elapsed 0.002 seconds) done.
> > > > [ 90.188713] OOM killer disabled.
> > > > [ 90.188714] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
> > > > [ 90.190024] printk: Suspending console(s) (use no_console_suspend to debug)
> > > > [ 90.904912] intel_pch_thermal 0000:00:12.0: CPU-PCH is cool [49C], continue to suspend
> > > > [ 321.262505] snd_hda_codec_realtek ehdaudio0D0: Unable to sync register 0x2b8000. -5
> > > > [ 328.426919] snd_hda_codec_realtek ehdaudio0D0: Unable to sync register 0x2b8000. -5
> > > > [ 329.490933] ACPI: EC: interrupt blocked
> > > >
> > > > That commit keeps codec suspended during the system suspend. However,
> > > > SOF driver's runtime resume, which is for system suspend, calls
> > > > hda_codec_jack_check() and schedules jackpoll_work. The jackpoll
> > > > work uses snd_hda_power_up_pm() which tries to resume the codec in
> > > > system suspend path, hence blocking the suspend process.
> > > >
> > > > So only check jack status if it's not in system PM process.
> > >
> > > After your previous patch set, the legacy HDA does queue the
> > > jackpoll_work only if jackpoll_interval is set. I suppose rather the
> > > same rule would be applied?
> >
> > It's queued in hda_codec_pm_complete(), which happens at the end of PM process.
> > This one is queued in the middle of PM suspend, so it's not the same here.
>
> But why do we need the jack status check explicitly there if
> hda_codec_pm_complete() already does it (via re-queuing the resume)?
Because hda_resume() is called for both system and runtime resume, but
hda_codec_pm_complete() only gets called for system resume. So we
still need to cover the runtime resume case.
However, we can use pm_request_resume() to wakeup the codec and do the
jack detection, instead of queueing jackpoll_work directly. I'll do
the change in v2.
Kai-Heng
>
> thanks,
>
> Takashi
>
> >
> > Kai-Heng
> >
> > >
> > >
> > > thanks,
> > >
> > > Takashi
> > >
> > > >
> > > > Fixes: 215a22ed31a1 ("ALSA: hda: Refactor codec PM to use direct-complete optimization")
> > > > Signed-off-by: Kai-Heng Feng <kai.heng.feng@xxxxxxxxxxxxx>
> > > > ---
> > > > sound/soc/sof/intel/hda-dsp.c | 3 ++-
> > > > 1 file changed, 2 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/sound/soc/sof/intel/hda-dsp.c b/sound/soc/sof/intel/hda-dsp.c
> > > > index 7d00107cf3b2..1c5e05b88a90 100644
> > > > --- a/sound/soc/sof/intel/hda-dsp.c
> > > > +++ b/sound/soc/sof/intel/hda-dsp.c
> > > > @@ -685,7 +685,8 @@ static int hda_resume(struct snd_sof_dev *sdev, bool runtime_resume)
> > > > /* check jack status */
> > > > if (runtime_resume) {
> > > > hda_codec_jack_wake_enable(sdev, false);
> > > > - hda_codec_jack_check(sdev);
> > > > + if (sdev->system_suspend_target == SOF_SUSPEND_NONE)
> > > > + hda_codec_jack_check(sdev);
> > > > }
> > > >
> > > > /* turn off the links that were off before suspend */
> > > > --
> > > > 2.29.2
> > > >
> >
