You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[CML][PM] White noise when being runtime suspended after playback #603
Summary:
when playback using alc5682 codec on cometlake, there is a about 2s white noise when entering runtime suspend.
Test steps:
Play some music, eg. aplay -vv -Dhw:0,0 -c2 -d2 48California_Dreaming.wav
when it ends, check the runtime status cat /sys/bus/pci/devices/0000:00:1f.3/power/runtime_status
The noise occurs when entering runtime suspended.
However, this noise only happens in suspending and if we trigger new playback, there will be no noise during playback.
Test env:
sof cml-001-drop-stable: 1f5a965
kernel topic/cros-sof-v4.19: f69d618
tplg: sof-cnl-rt5682.tplg
changed the title [-][CML][PM] white noise when playback stops before runtime suspend[/-][+][CML][PM] White noise when being runtime suspended after playback[/+]on Jan 29, 2019
MCLK is turned off in runtime suspended. However, sof enter runtime suspend before DAPM off. In other word, MCLK is turned off before codec is powered off. And noise will occur in that period.
MCLK is turned off in runtime suspended. However, sof enter runtime
suspend before DAPM off. In other word, MCLK is turned off before
codec is powered off. And noise will occur in that period.
I believe we will have to make the MCLK/BCLK controllable from the
machine/codec driver. This is a feature that was implemented in the cAVS
driver and that we know we need to avoid issues due to clock
transitions, but haven't done any work on. There are some fields in the
IPC that could be used to leave the MCLK on, but they are not set. see
SOF issue #629thesofproject/sof#629
We'd also need the pm_runtime to suspend the DSP when there is no need
for any clocks.
In other words it's not going to be fixed overnight.
Checked also on WHL RVP and there is also no reproduction with "while noises":
kernel: cros-sof-v4.19 / 8492020d5967 + PR#651 applied
FW: cml-002-drop-stable / dcd825b
tplg for hda: sof-hda-generic.tplg
No, I tested it by hdmi.
On headphone I haven't any sound although there is no errors in dmesg log and aplay show playback progress without any issues. I haven't went deeper for debugging it yet - I will try it again with deeper analysis and I will give a feedback.
@plbossart We still have this issue. Bard says he's trying to set a ignore_pmdown_time flag, which can shut down rt5682 immediately after playback. I'll look into it and help to check with Bard.
Activity
[-][CML][PM] white noise when playback stops before runtime suspend[/-][+][CML][PM] White noise when being runtime suspended after playback[/+]bardliao commentedon Jan 30, 2019
MCLK is turned off in runtime suspended. However, sof enter runtime suspend before DAPM off. In other word, MCLK is turned off before codec is powered off. And noise will occur in that period.
plbossart commentedon Jan 30, 2019
emilchudzik commentedon Feb 25, 2019
@wwittbrx bringup Hatch. thank you Wojtek.
We just checked flow from topic and we haven't repro. There is no "white noises" after end of the playback on each time.
Configuration:
Platform: Hatch CML
kernel: cros-sof-v4.19 / 8492020d5967 + PR#651 applied
FW: cml-002-drop-stable / dcd825b
tplg: sof-cnl-rt5682.tplg
Checked also on WHL RVP and there is also no reproduction with "while noises":
kernel: cros-sof-v4.19 / 8492020d5967 + PR#651 applied
FW: cml-002-drop-stable / dcd825b
tplg for hda: sof-hda-generic.tplg
bardliao commentedon Feb 26, 2019
@emilchudzik Did you test it by headphone?
emilchudzik commentedon Feb 26, 2019
No, I tested it by hdmi.
On headphone I haven't any sound although there is no errors in dmesg log and aplay show playback progress without any issues. I haven't went deeper for debugging it yet - I will try it again with deeper analysis and I will give a feedback.
emilchudzik commentedon Feb 28, 2019
I just tested flow from topic on headphone (thanks to @bardliao to help configure alsamixer).
On SSP0 (connected headphones) I have this observation:
aplay -vv -Dhw:0,0 sine.wav -d5Here is a dmesg from start aplay to going suspended
dmesg.txt
plbossart commentedon Mar 24, 2019
@xun2z is this issue still relevant? It's been 24 days without update.
xun2z commentedon Mar 25, 2019
@plbossart We still have this issue. Bard says he's trying to set a
ignore_pmdown_timeflag, which can shut down rt5682 immediately after playback. I'll look into it and help to check with Bard.xun2z commentedon Apr 2, 2019
close since #741 is the fix
coresight: Fix DEBUG_LOCKS_WARN_ON for uninitialized attribute
reset: whitespace fixes
net: bridge: keep ports without IFF_UNICAST_FLT in BR_PROMISC mode
soc: imx8m: Probe the SoC driver as platform driver