[v6,2/2] ALSA: hda: Allow HDA to be runtime suspended when dGPU is not bound to a driver

Nvidia proprietary driver doesn't support runtime power management, so
when a user only wants to use the integrated GPU, it's a common practice
to let dGPU not to bind any driver, and let its upstream port to be
runtime suspended. At the end of runtime suspension the port uses
platform power management to disable power through _OFF method of power
resource, which is listed by _PR3.

After commit b516ea586d71 ("PCI: Enable NVIDIA HDA controllers"), when
the dGPU comes with an HDA function, the HDA won't be suspended if the
dGPU is unbound, so the power resource can't be turned off by its
upstream port driver.

Commit 37a3a98ef601 ("ALSA: hda - Enable runtime PM only for
discrete GPU") only allows HDA to be runtime suspended once GPU is
bound, to keep APU's HDA working.

However, HDA on dGPU isn't that useful if dGPU is not bound to any
driver.  So let's relax the runtime suspend requirement for dGPU's HDA
function, to disable the power source to save lots of power.

BugLink: https://bugs.launchpad.net/bugs/1840835
Fixes: b516ea586d71 ("PCI: Enable NVIDIA HDA controllers")
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
---
v5, v6:
- No change.
v4:
- Find upstream port, it's callee's responsibility now.
v3:
- Make changelog more clear.
v2:
- Change wording.
- Rebase to Tiwai's branch.
 sound/pci/hda/hda_intel.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

Dne 18. 10. 19 v 9:38 Kai-Heng Feng napsal(a):
> Nvidia proprietary driver doesn't support runtime power management, so
> when a user only wants to use the integrated GPU, it's a common practice
> to let dGPU not to bind any driver, and let its upstream port to be
> runtime suspended. At the end of runtime suspension the port uses
> platform power management to disable power through _OFF method of power
> resource, which is listed by _PR3.
> 
> After commit b516ea586d71 ("PCI: Enable NVIDIA HDA controllers"), when
> the dGPU comes with an HDA function, the HDA won't be suspended if the
> dGPU is unbound, so the power resource can't be turned off by its
> upstream port driver.
> 
> Commit 37a3a98ef601 ("ALSA: hda - Enable runtime PM only for
> discrete GPU") only allows HDA to be runtime suspended once GPU is
> bound, to keep APU's HDA working.
> 
> However, HDA on dGPU isn't that useful if dGPU is not bound to any
> driver.  So let's relax the runtime suspend requirement for dGPU's HDA
> function, to disable the power source to save lots of power.

This patch breaks the HDMI audio detection at least on some platforms (Lenovo 
P50 for example) with nouveau and the proprietary nvidia driver. Those laptops 
have the external HDMI/DP ports connected to dGPU instead the iGPU. The ACPI 
PR3 is set.

The runtime PM off fixes this problem:

echo on > /sys/bus/pci/devices/0000\:01\:00.1/power/control

But I don't think that it's the best solution. My proposal is to create a pr3 
check blacklist to keep power for the HDMI audio for those machines. Also we 
may add a new module parameter for snd-hda-intel to control this. Other ideas?

					Jaroslav


> BugLink: https://bugs.launchpad.net/bugs/1840835
> Fixes: b516ea586d71 ("PCI: Enable NVIDIA HDA controllers")
> Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
> ---
> v5, v6:
> - No change.
> v4:
> - Find upstream port, it's callee's responsibility now.
> v3:
> - Make changelog more clear.
> v2:
> - Change wording.
> - Rebase to Tiwai's branch.
>   sound/pci/hda/hda_intel.c | 8 +++++++-
>   1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c
> index 240f4ca76391..e63b871343e5 100644
> --- a/sound/pci/hda/hda_intel.c
> +++ b/sound/pci/hda/hda_intel.c
> @@ -1280,11 +1280,17 @@ static void init_vga_switcheroo(struct azx *chip)
>   {
>   	struct hda_intel *hda = container_of(chip, struct hda_intel, chip);
>   	struct pci_dev *p = get_bound_vga(chip->pci);
> +	struct pci_dev *parent;
>   	if (p) {
>   		dev_info(chip->card->dev,
>   			 "Handle vga_switcheroo audio client\n");
>   		hda->use_vga_switcheroo = 1;
> -		chip->bus.keep_power = 1; /* cleared in either gpu_bound op or codec probe */
> +
> +		/* cleared in either gpu_bound op or codec probe, or when its
> +		 * upstream port has _PR3 (i.e. dGPU).
> +		 */
> +		parent = pci_upstream_bridge(p);
> +		chip->bus.keep_power = parent ? !pci_pr3_present(parent) : 1;
>   		chip->driver_caps |= AZX_DCAPS_PM_RUNTIME;
>   		pci_dev_put(p);
>   	}
>

On Fri, 10 Jan 2020 10:43:26 +0100,
Jaroslav Kysela wrote:
> 
> Dne 18. 10. 19 v 9:38 Kai-Heng Feng napsal(a):
> > Nvidia proprietary driver doesn't support runtime power management, so
> > when a user only wants to use the integrated GPU, it's a common practice
> > to let dGPU not to bind any driver, and let its upstream port to be
> > runtime suspended. At the end of runtime suspension the port uses
> > platform power management to disable power through _OFF method of power
> > resource, which is listed by _PR3.
> >
> > After commit b516ea586d71 ("PCI: Enable NVIDIA HDA controllers"), when
> > the dGPU comes with an HDA function, the HDA won't be suspended if the
> > dGPU is unbound, so the power resource can't be turned off by its
> > upstream port driver.
> >
> > Commit 37a3a98ef601 ("ALSA: hda - Enable runtime PM only for
> > discrete GPU") only allows HDA to be runtime suspended once GPU is
> > bound, to keep APU's HDA working.
> >
> > However, HDA on dGPU isn't that useful if dGPU is not bound to any
> > driver.  So let's relax the runtime suspend requirement for dGPU's HDA
> > function, to disable the power source to save lots of power.
> 
> This patch breaks the HDMI audio detection at least on some platforms
> (Lenovo P50 for example) with nouveau and the proprietary nvidia
> driver. Those laptops have the external HDMI/DP ports connected to
> dGPU instead the iGPU. The ACPI PR3 is set.
> 
> The runtime PM off fixes this problem:
> 
> echo on > /sys/bus/pci/devices/0000\:01\:00.1/power/control

But this will keep the power of the graphics chip on, and that's what
the patch was supposed to "fix".

> But I don't think that it's the best solution. My proposal is to
> create a pr3 check blacklist to keep power for the HDMI audio for
> those machines. Also we may add a new module parameter for
> snd-hda-intel to control this. Other ideas?

For nouveau, the best fix is to merge the audio component patch.
This will make things working without fiddling with the power
up/down.  The patch has been pending over months under review in DRM
side, unfortunately...  Please pinging them for driving ahead.

For Nvidia, though, it's no path a binary-only stuff can go with, due
to the GPL symbol of the component framework.  Those guys know of it
well, and they seem adding the temporary power up/down procedure by
poking the proc file from the user-space side at the HDMI connection.

About a module option: I don't think it's much better than the sysfs
toggle.  You can set up a simple udev rule if needed, too.


thanks,

Takashi

> 
> 					Jaroslav
> 
> 
> > BugLink: https://bugs.launchpad.net/bugs/1840835
> > Fixes: b516ea586d71 ("PCI: Enable NVIDIA HDA controllers")
> > Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
> > ---
> > v5, v6:
> > - No change.
> > v4:
> > - Find upstream port, it's callee's responsibility now.
> > v3:
> > - Make changelog more clear.
> > v2:
> > - Change wording.
> > - Rebase to Tiwai's branch.
> >   sound/pci/hda/hda_intel.c | 8 +++++++-
> >   1 file changed, 7 insertions(+), 1 deletion(-)
> >
> > diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c
> > index 240f4ca76391..e63b871343e5 100644
> > --- a/sound/pci/hda/hda_intel.c
> > +++ b/sound/pci/hda/hda_intel.c
> > @@ -1280,11 +1280,17 @@ static void init_vga_switcheroo(struct azx *chip)
> >   {
> >   	struct hda_intel *hda = container_of(chip, struct hda_intel, chip);
> >   	struct pci_dev *p = get_bound_vga(chip->pci);
> > +	struct pci_dev *parent;
> >   	if (p) {
> >   		dev_info(chip->card->dev,
> >   			 "Handle vga_switcheroo audio client\n");
> >   		hda->use_vga_switcheroo = 1;
> > -		chip->bus.keep_power = 1; /* cleared in either gpu_bound op or codec probe */
> > +
> > +		/* cleared in either gpu_bound op or codec probe, or when its
> > +		 * upstream port has _PR3 (i.e. dGPU).
> > +		 */
> > +		parent = pci_upstream_bridge(p);
> > +		chip->bus.keep_power = parent ? !pci_pr3_present(parent) : 1;
> >   		chip->driver_caps |= AZX_DCAPS_PM_RUNTIME;
> >   		pci_dev_put(p);
> >   	}
> >
> 
> 
> -- 
> Jaroslav Kysela <perex@perex.cz>
> Linux Sound Maintainer; ALSA Project; Red Hat, Inc.
>

Dne 10. 01. 20 v 10:56 Takashi Iwai napsal(a):
> On Fri, 10 Jan 2020 10:43:26 +0100,
> Jaroslav Kysela wrote:
>>
>> Dne 18. 10. 19 v 9:38 Kai-Heng Feng napsal(a):
>>> Nvidia proprietary driver doesn't support runtime power management, so
>>> when a user only wants to use the integrated GPU, it's a common practice
>>> to let dGPU not to bind any driver, and let its upstream port to be
>>> runtime suspended. At the end of runtime suspension the port uses
>>> platform power management to disable power through _OFF method of power
>>> resource, which is listed by _PR3.
>>>
>>> After commit b516ea586d71 ("PCI: Enable NVIDIA HDA controllers"), when
>>> the dGPU comes with an HDA function, the HDA won't be suspended if the
>>> dGPU is unbound, so the power resource can't be turned off by its
>>> upstream port driver.
>>>
>>> Commit 37a3a98ef601 ("ALSA: hda - Enable runtime PM only for
>>> discrete GPU") only allows HDA to be runtime suspended once GPU is
>>> bound, to keep APU's HDA working.
>>>
>>> However, HDA on dGPU isn't that useful if dGPU is not bound to any
>>> driver.  So let's relax the runtime suspend requirement for dGPU's HDA
>>> function, to disable the power source to save lots of power.
>>
>> This patch breaks the HDMI audio detection at least on some platforms
>> (Lenovo P50 for example) with nouveau and the proprietary nvidia
>> driver. Those laptops have the external HDMI/DP ports connected to
>> dGPU instead the iGPU. The ACPI PR3 is set.
>>
>> The runtime PM off fixes this problem:
>>
>> echo on > /sys/bus/pci/devices/0000\:01\:00.1/power/control
> 
> But this will keep the power of the graphics chip on, and that's what
> the patch was supposed to "fix".
> 
>> But I don't think that it's the best solution. My proposal is to
>> create a pr3 check blacklist to keep power for the HDMI audio for
>> those machines. Also we may add a new module parameter for
>> snd-hda-intel to control this. Other ideas?
> 
> For nouveau, the best fix is to merge the audio component patch.
> This will make things working without fiddling with the power
> up/down.  The patch has been pending over months under review in DRM
> side, unfortunately...  Please pinging them for driving ahead.

Adding Cc: to dri-devel. You probably mean this thread:

https://lists.freedesktop.org/archives/dri-devel/2019-July/thread.html#227423

> For Nvidia, though, it's no path a binary-only stuff can go with, due
> to the GPL symbol of the component framework.  Those guys know of it
> well, and they seem adding the temporary power up/down procedure by
> poking the proc file from the user-space side at the HDMI connection.

Wow.

> About a module option: I don't think it's much better than the sysfs
> toggle.  You can set up a simple udev rule if needed, too.

Ok, it's a bit nightmare to maintain those extra settings in the distribution.

					Jaroslav

> 
> 
> thanks,
> 
> Takashi
> 
>>
>> 					Jaroslav
>>
>>
>>> BugLink: https://bugs.launchpad.net/bugs/1840835
>>> Fixes: b516ea586d71 ("PCI: Enable NVIDIA HDA controllers")
>>> Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
>>> ---
>>> v5, v6:
>>> - No change.
>>> v4:
>>> - Find upstream port, it's callee's responsibility now.
>>> v3:
>>> - Make changelog more clear.
>>> v2:
>>> - Change wording.
>>> - Rebase to Tiwai's branch.
>>>    sound/pci/hda/hda_intel.c | 8 +++++++-
>>>    1 file changed, 7 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c
>>> index 240f4ca76391..e63b871343e5 100644
>>> --- a/sound/pci/hda/hda_intel.c
>>> +++ b/sound/pci/hda/hda_intel.c
>>> @@ -1280,11 +1280,17 @@ static void init_vga_switcheroo(struct azx *chip)
>>>    {
>>>    	struct hda_intel *hda = container_of(chip, struct hda_intel, chip);
>>>    	struct pci_dev *p = get_bound_vga(chip->pci);
>>> +	struct pci_dev *parent;
>>>    	if (p) {
>>>    		dev_info(chip->card->dev,
>>>    			 "Handle vga_switcheroo audio client\n");
>>>    		hda->use_vga_switcheroo = 1;
>>> -		chip->bus.keep_power = 1; /* cleared in either gpu_bound op or codec probe */
>>> +
>>> +		/* cleared in either gpu_bound op or codec probe, or when its
>>> +		 * upstream port has _PR3 (i.e. dGPU).
>>> +		 */
>>> +		parent = pci_upstream_bridge(p);
>>> +		chip->bus.keep_power = parent ? !pci_pr3_present(parent) : 1;
>>>    		chip->driver_caps |= AZX_DCAPS_PM_RUNTIME;
>>>    		pci_dev_put(p);
>>>    	}
>>>
>>
>>
>> -- 
>> Jaroslav Kysela <perex@perex.cz>
>> Linux Sound Maintainer; ALSA Project; Red Hat, Inc.
>>