[alsa-devel] [PATCH] ALSA: hda/hdmi - Don't report Jack event if no need to do that

[Hui Wang]

On 2019/4/30 下午3:35, Takashi Iwai wrote:
> On Tue, 30 Apr 2019 08:57:11 +0200,
> Hui Wang wrote:
>> On the machines with AMD GPU or Nvidia GPU, we often meet this issues:
>> after s3, there are 4 HDMI/DP audio devices in the gnome-sound-setting
>> even there is no any monitors plugged.
>>
>> When this problem happens, we check the /proc/asound/cardX/eld#N.M, we
>> will find the monitor_present=1, eld_valid=0.
>>
>> The root cause is somehow the pin_sense reports the monitor is present
>> and eld is valid when there is no monitor plugged.
>>
>> The current driver will read the eld data if the pin_sense reports the
>> eld is valid, because of no monitor is plugged, there is no valid eld
>> data, then the eld->valid is set to 0.
>>
>> If we don't let driver report Jack event when monitor_present=1 while
>> eld_valid=0, there will be no this issue.
>>
>> After this change, the driver only reports Jack event with one of the
>> below 2 conditons:
>>   eld->monitor_present=1 and eld->eld_valid=1 (a valid monitor detect)
>>   eld->monitor_present=0 (a monitor is unplugged)
>>
>> Signed-off-by: Hui Wang <hui.wang at canonical.com>
> Well, if the eld_valid=1 is mandatory, basically we can use it as the
> condition of jack=1, like the patch below.  The return value from
> hdmi_present_sense() indicates only whether we may sync jack state or
> not, and it's not about the jack state itself.
>
>
> thanks,
>
> Takashi
>
> ---
> --- a/sound/pci/hda/patch_hdmi.c
> +++ b/sound/pci/hda/patch_hdmi.c
> @@ -1625,7 +1625,7 @@ static void sync_eld_via_acomp(struct hda_codec *codec,
>   	if (jack == NULL)
>   		goto unlock;
>   	snd_jack_report(jack,
> -			eld->monitor_present ? SND_JACK_AVOUT : 0);
> +			(eld->monitor_present && eld->eld_valid) ? SND_JACK_AVOUT : 0);
>    unlock:
>   	mutex_unlock(&per_pin->lock);
>   }

All Intel machines which get eld via acomp don't have this issue, so no 
need to change this function.

For those machines (with nv gpu card or amd gpu card)  which need to 
call hdmi_present_sense_via_verbs(), they have this issue. if 
hdmi_present_sense_via_verbs() returns true, the hdmi_present_sense() 
returns true, then snd_hda_jack_report_sync() will be called. So if we 
want to fix this issue, we need to let hdmi_present_sense_via_verbs() 
not return true or we change the read_pin_sense() in the hda_jack.c



> _______________________________________________
> Alsa-devel mailing list
> Alsa-devel at alsa-project.org
> https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
>