Skip to contentSkip to conversation

ASoC: codec: hdac_hdmi add device_link to card device

PATCH In review [email protected] 3 participants 6 messages Apr 13, 2019 – Apr 19, 2019

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 ASoC: codec: hdac_hdmi add device_link to card device

From: Libin Yang <[email protected]>

In resume from S3, HDAC HDMI codec driver dapm event callback may be operated before HDMI codec driver turns on the display audio power domain because of the contest between display driver and hdmi codec driver.

This patch adds the device_link between soc card device (consumer) and hdmi codec device (supplier) to make sure the sequence is always correct.

---
 sound/soc/codecs/hdac_hdmi.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
sound/soc/codecs/hdac_hdmi.c +11 -0
index 5eeb0fe..4de1fbf 100644
@@ -1855,6 +1855,17 @@ static int hdmi_codec_probe(struct snd_soc_component *component)
 	hdmi->card = dapm->card->snd_card;
 
 	/*
+	 * Setup a device_link between card device and HDMI codec device.
+	 * The card device is the consumer and the HDMI codec device is
+	 * the supplier. With this setting, we can make sure that the audio
+	 * domain in display power will be always turned on before operating
+	 * on the HDMI audio codec registers.
+	 * Let's use the flag DL_FLAG_AUTOREMOVE_CONSUMER. This can make
+	 * sure the device link is freed when the machine driver is removed.
+	 */
+	device_link_add(component->card->dev, &hdev->dev, DL_FLAG_RPM_ACTIVE |
+			DL_FLAG_AUTOREMOVE_CONSUMER);
+	/*
 	 * hdac_device core already sets the state to active and calls
 	 * get_noresume. So enable runtime and set the device to suspend.
 	 */
Signature · 1 line
2.7.4
Reviewed-by on v2 1/1
#1 lore raw
Quoted 11 lines · On Sat, 13 Apr 2019 15:18:12 +0200, [email protected] wrote:

From: Libin Yang <[email protected]>

In resume from S3, HDAC HDMI codec driver dapm event callback may be operated before HDMI codec driver turns on the display audio power domain because of the contest between display driver and hdmi codec driver.

This patch adds the device_link between soc card device (consumer) and hdmi codec device (supplier) to make sure the sequence is always correct.

Is it an issue that has been present for older released kernels? If so, it deserves for Cc-to-stable.

thanks,

Takashi

Quoted 29 lines
---
 sound/soc/codecs/hdac_hdmi.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
sound/soc/codecs/hdac_hdmi.c +11 -0
index 5eeb0fe..4de1fbf 100644
@@ -1855,6 +1855,17 @@ static int hdmi_codec_probe(struct snd_soc_component *component)
 	hdmi->card = dapm->card->snd_card;
 
 	/*
+	 * Setup a device_link between card device and HDMI codec device.
+	 * The card device is the consumer and the HDMI codec device is
+	 * the supplier. With this setting, we can make sure that the audio
+	 * domain in display power will be always turned on before operating
+	 * on the HDMI audio codec registers.
+	 * Let's use the flag DL_FLAG_AUTOREMOVE_CONSUMER. This can make
+	 * sure the device link is freed when the machine driver is removed.
+	 */
+	device_link_add(component->card->dev, &hdev->dev, DL_FLAG_RPM_ACTIVE |
+			DL_FLAG_AUTOREMOVE_CONSUMER);
+	/*
 	 * hdac_device core already sets the state to active and calls
 	 * get_noresume. So enable runtime and set the device to suspend.
 	 */
Signature · 1 line
2.7.4
#2 lore raw
Quoted 12 lines · Takashi Iwai
Quoted 10 lines · On Sat, 13 Apr 2019 15:18:12 +0200, [email protected] wrote:

From: Libin Yang <[email protected]>

In resume from S3, HDAC HDMI codec driver dapm event callback may be operated before HDMI codec driver turns on the display audio power domain because of the contest between display driver and hdmi codec driver.

This patch adds the device_link between soc card device (consumer) and hdmi codec device (supplier) to make sure the sequence is always correct.

Quoted 6 lines
Quoted 1 line

Is it an issue that has been present for older released kernels? If so, it deserves for Cc-to-stable.

Yes and no. In theory the same problem impacts the Skylake driver. However its support for HDMI has been flaky at best, see e.g. the probe/timing issues on Linus' laptop and a variety of devices in December, and the only real distribution using it - ChromeOS - does not resume from S3. Not to mention that products typically use 3.18 to 4.9.

#3 lore raw
Quoted 14 lines
Quoted 4 lines
Quoted 4 lines

In resume from S3, HDAC HDMI codec driver dapm event callback may be operated before HDMI codec driver turns on the display audio power domain because of the contest between display driver and hdmi codec

driver.

Quoted 3 lines
Quoted 3 lines

This patch adds the device_link between soc card device (consumer) and hdmi codec device (supplier) to make sure the sequence is always

correct.

Quoted 5 lines
Quoted 2 lines
Quoted 8 lines
Quoted 2 lines

Is it an issue that has been present for older released kernels? If so, it deserves for Cc-to-stable.

Yes and no. In theory the same problem impacts the Skylake driver. However its support for HDMI has been flaky at best, see e.g. the probe/timing issues on Linus' laptop and a variety of devices in December, and the only real distribution using it - ChromeOS - does not resume from S3. Not to mention that products typically use 3.18 to 4.9.

Agree with Pierre. I have tested on 4.1x before (I didn't remember the exact version number), it works well. On some later stable versions, for example 5.0.0, we can easily reproduce this issue.

And for other earlier version, I didn't have ever test on it.

Regards,
Libin

#4 lore raw
In reply to Takashi Iwai

Hi Takashi,

Quoted 17 lines

-----Original Message-----
From: Takashi Iwai [mailto:[email protected]]
Sent: Tuesday, April 16, 2019 6:24 PM
To: Yang, Libin <[email protected]>
Cc: [email protected]; [email protected]; pierre- [email protected]
Subject: Re: [alsa-devel] [PATCH v2] ASoC: codec: hdac_hdmi add device_link to card device

Quoted 6 lines · On Sat, 13 Apr 2019 15:18:12 +0200, [email protected] wrote:

From: Libin Yang <[email protected]>

In resume from S3, HDAC HDMI codec driver dapm event callback may be operated before HDMI codec driver turns on the display audio power domain because of the contest between display driver and hdmi codec

Quoted 8 lines

driver.

Quoted 5 lines

This patch adds the device_link between soc card device (consumer) and hdmi codec device (supplier) to make sure the sequence is always correct.

Thanks for the review.

Quoted 3 lines

Is it an issue that has been present for older released kernels? If so, it deserves for Cc-to-stable.

It is for all released kernels, and the rates of reproduce are different. It's graphics driver related.

Regards,
Libin

Quoted 38 lines

thanks,

Takashi

sound/soc/codecs/hdac_hdmi.c static int hdmi_codec_probe(struct

Quoted 9 lines
---
 sound/soc/codecs/hdac_hdmi.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
(quoted diff) +0 -0

b/sound/soc/codecs/hdac_hdmi.c index 5eeb0fe..4de1fbf 100644

sound/soc/codecs/hdac_hdmi.c +0 -0
@@ -1855,6 +1855,17 @@ static int hdmi_codec_probe(struct

snd_soc_component *component)

sound/soc/codecs/hdac_hdmi.c static int hdmi_codec_probe(struct snd_soc_component *component)

Quoted 9 lines
sound/soc/codecs/hdac_hdmi.c +6 -0
 	hdmi->card = dapm->card->snd_card;
 
 	/*
+	 * Setup a device_link between card device and HDMI codec device.
+	 * The card device is the consumer and the HDMI codec device is
+	 * the supplier. With this setting, we can make sure that the audio
+	 * domain in display power will be always turned on before operating
+	 * on the HDMI audio codec registers.
+	 * Let's use the flag DL_FLAG_AUTOREMOVE_CONSUMER. This can

make

sound/soc/codecs/hdac_hdmi.c static int hdmi_codec_probe(struct snd_soc_component *component)

Quoted 3 lines
sound/soc/codecs/hdac_hdmi.c +3 -0
+	 * sure the device link is freed when the machine driver is removed.
+	 */
+	device_link_add(component->card->dev, &hdev->dev,

DL_FLAG_RPM_ACTIVE |

Quoted 8 lines
sound/soc/codecs/hdac_hdmi.c +2 -1
+			DL_FLAG_AUTOREMOVE_CONSUMER);
+	/*
 	 * hdac_device core already sets the state to active and calls
 	 * get_noresume. So enable runtime and set the device to suspend.
 	 */
--

2.7.4

Acked-by on v2 1/1
#5 lore raw
In reply to [email protected]
Quoted 10 lines · On 4/13/19 8:18 AM, [email protected] wrote:

From: Libin Yang <[email protected]>

In resume from S3, HDAC HDMI codec driver dapm event callback may be operated before HDMI codec driver turns on the display audio power domain because of the contest between display driver and hdmi codec driver.

This patch adds the device_link between soc card device (consumer) and hdmi codec device (supplier) to make sure the sequence is always correct.

Libin, please follow-up on the question from Takashi on CC:ing to stable. I would guess this probably would be needed for 4.14+, older stuff is unlikely to work anyways due to other issues.

Quoted 27 lines
---
  sound/soc/codecs/hdac_hdmi.c | 11 +++++++++++
  1 file changed, 11 insertions(+)
sound/soc/codecs/hdac_hdmi.c +11 -0
index 5eeb0fe..4de1fbf 100644
@@ -1855,6 +1855,17 @@ static int hdmi_codec_probe(struct snd_soc_component *component)
  	hdmi->card = dapm->card->snd_card;
 
  	/*
+	 * Setup a device_link between card device and HDMI codec device.
+	 * The card device is the consumer and the HDMI codec device is
+	 * the supplier. With this setting, we can make sure that the audio
+	 * domain in display power will be always turned on before operating
+	 * on the HDMI audio codec registers.
+	 * Let's use the flag DL_FLAG_AUTOREMOVE_CONSUMER. This can make
+	 * sure the device link is freed when the machine driver is removed.
+	 */
+	device_link_add(component->card->dev, &hdev->dev, DL_FLAG_RPM_ACTIVE |
+			DL_FLAG_AUTOREMOVE_CONSUMER);
+	/*
  	 * hdac_device core already sets the state to active and calls
  	 * get_noresume. So enable runtime and set the device to suspend.
  	 */