ASoC: SOF: Intel: use NHLT information for ES83x6 devices #3338

plbossart · 2021-12-14T19:37:40Z

We have to figure out the presence of DMICs and which SSP is used without DMI quirks, otherwise we'll drown in support requests.

Completely untested since I don't have hardware...

@hli25 could you take this for a spin and report what you find on GLK/TGL/JSL devices? There should be dev_dbg messages that you can enable by copying this file
sof-dyndbg.conf.txt to /etc/modprobe.d/sof-dyndbg.conf
Enhancement tracked with #3337

Support requests with active users in
#3336
#3248

plbossart · 2021-12-14T20:58:17Z

@kv2019i @juimonen this NHLT information could be used for BT offload and figure out which SSP is used if we ever support this for non-Chrome devices designed for Windows.

yangxiaohua2009 · 2021-12-16T07:58:34Z

@plbossart Here is log glk_nhlt.log on glk.
Currently we define SSP number in the machine driver and use three tplg for SSP0,1,2.
It would be better for the Intel platform driver to read SSP number from NHLT and choose the corresponding tplg file.

plbossart · 2021-12-16T13:34:14Z

@plbossart Here is log glk_nhlt.log on glk. Currently we define SSP number in the machine driver and use three tplg for SSP0,1,2. It would be better for the Intel platform driver to read SSP number from NHLT and choose the corresponding tplg file.

Thanks @yangxiaohua2009 indeed that's the plan to autodetect the SSP, but in your example the mask is zero

[ 8.746047] sof-audio-pci-intel-apl 0000:00:0e.0: intel_nhlt_get_dmic_geo: dmic number 0 max_ch 0
[ 8.746054] sof-audio-pci-intel-apl 0000:00:0e.0: NHLT_LINK_SSP detected, ssp_mask 0x0

That makes no sense really, something is wrong here.
Can you please try to trace ssp_mask in intel_nhlt_ssp_endpoint_mask(), I wonder what happens here.

plbossart · 2021-12-16T13:40:58Z

@yangxiaohua2009 I screwed-up with the initial code, please try the updated branch and let me know what you detect. thanks!

yangxiaohua2009 · 2021-12-17T02:20:32Z

@plbossart glk_2.log Here's the new log.

plbossart · 2022-01-10T22:15:02Z

@yangxiaohua2009 @hli25 can you take a look? This should remove the need for the SSP quirk and DMIC quirk. For the GPIO I have no solution.

yangxiaohua2009 · 2022-01-11T06:37:06Z

@plbossart The kernel works well on glk glk_111.log. However on CML it doesn't select sof in intel-dsp-config.c cml_111.log and cannot recognize GPIO.
After force probing sof and GPIO, I see machine driver still using SSP2 instead of SSP0 on CML.

cml_111_2.log

plbossart · 2022-01-11T16:14:04Z

Thanks for testing @yangxiaohua2009

@plbossart The kernel works well on glk glk_111.log. However on CML it doesn't select sof in intel-dsp-config.c [cml_111.log]
(https://github.com/thesofproject/linux/files/7844183/cml_111.log) and cannot recognize GPIO.

Sorry, I don't get how the SOF driver is not selected. I see two quirks for CML PCI IDs 0x06c8 and 0x02c8. Can you provide the result of 'cat /sys/bus/pci/devices/0000:00:1f.3/device'

Or did you take the patches and applied them on a different tree that's missing the CML quirks?

After force probing sof and GPIO, I see machine driver still using SSP2 instead of SSP0 on CML.

cml_111_2.log

The logs show SSP0 is used and passed to the machine driver?

[    2.847963] sof-audio-pci-intel-cnl 0000:00:1f.3: NHLT_DEVICE_I2S detected, ssp_mask 0x1

Are you sure you didn't modify the quirk module parameter?

yangxiaohua2009 · 2022-01-12T04:42:43Z

Since !dmi_check_system(sof_es8336_quirk_table) always return 1, machine driver still uses SSP2. After removing it and making several changes the machine driver works #24. For SSP2 the i2s_link_mask is 5, for SSP1 i2s_link_mask is 3 and for SSP0 i2s_link_mask is 1 apl_112.log.

yangxiaohua2009 · 2022-01-12T04:47:38Z

Also, since the dsp driver already know which SSP to use from NHLT, can it select the corresponding tplg file? For example, on GLK I see OEMs using SSP2, SSP1 and SSP0.
For the GPIO, can the machine driver neglect PA control instead of returing error so that at least the user can get mic, headphone and HDMI working. @plbossart

Add missing dmic_num mention and clarify that 'links' mean 'SoundWire links', not to be used for other links. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

The platform driver may have information on which I2S/TDM link(s) to enable in the machine driver. In the case of Intel devices, this may be extracted from NHLT tables in platform firmware. This link information is necessary to make sure machine driver and topology are aligned. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

We currently extract the DMIC number only for HDaudio or SoundWire platforms. For I2S/TDM platforms, this wasn't necessary until now, but with devices with ES8336 we need to find a solution to detect dmics more reliably than with a DMI quirk. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

The NHLT information can be used to figure out which SSPs are enabled in a platform. The 'SSP' link type is too broad for machine drivers, since it can cover the Bluetooth sideband and the analog audio codec connections, so this helper exposes a parameter to filter with the device type (DEVICE_I2S refers to analog audio codec in NHLT parlance). The helper returns a mask, since more than one SSP may be used for analog audio, e.g. the NHLT spec describes the use of SSP0 for amplifiers and SSP1 for headset codec. Note that if more than one bit is set, it's impossible to determine which SSP is connected to what external component. Additional platform-specific information based on e.g. DMI quirks would still be required in the machine driver to configure the relevant dailinks. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

For devices designed for Windows, the SSP information should be listed in the NHLT, and when present can be used to set quirks automatically in the machine driver. The NHLT information exposes BT and analog audio connections separately, for now we are only interested in the analog audio parts. The use of dev_info() for the SSP mask is intentional so that we can immediately flag devices with an ES8336 codec. Since NHLT is not used for recent Chromebooks these messages should be rare. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

Different topology filenames may be required depending on which SSP is used, and whether or not digital mics are present. This patch adds a tplg_quirk_mask and in the case of the SOF driver adds the relevant configurations. This is a short-term solution to the ES8336 support issues. In a long-term solution, we would need an interface where the machine driver or platform driver have the ability to alter the topology hard-coded low-level hardware support, e.g. by substituting an interface for another, or disabling an interface that is not supported on a given skew. BugLink: thesofproject#3248 Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

We only saw ESSX8336 so far, but now with reports of 'ESSX8326' we need to expand to a full list. Let's reuse the 'snd_soc_acpi_codecs' structure to store the information. Reported-by: anthony tonitch <d.tonitch@gmail.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

We only saw ESSX8336 so far, but now with reports of 'ESSX8326' we need to expand to a full list. Let's reuse the 'snd_soc_acpi_codecs' structure to store the information. Note that ES8326 will need a dedicated codec driver, but the plan is to use the same machine driver for all Everest Audio devices. Reported-by: anthony tonitch <d.tonitch@gmail.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

We're missing this check for the CNL PCI id Reported-by: Nikolai Kostrigin <nickel@altlinux.org> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

Add the support for all ES83x6 devices Signed-off-by: Nikolai Kostrigin <nickel@altlinux.org> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

Do not fail if the GPIO used for speakers is not present, at least the headphone, headset and internal mics should work. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

…arch We have an existing 'adev' handle from which we can find the codec device, no need for an I2C bus search. This change aligns this driver will all other I2S-based machine drivers. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

…5 2021" This reverts commit ce6a70b. The next patch will add run-time detection of the required SSP and this hard-coded quirk is not needed. Acked-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

Since we see a proliferation of devices with various configurations, we want to automatically set the DMIC and SSP information. This patch relies on the information extracted from NHLT and partially reverts existing DMI quirks added by commit a164137 ("ASoC: Intel: add machine driver for SOF+ES8336") Note that NHLT can report multiple SSPs, choosing from the ssp_link_mask in an MSB-first manner was found experimentally to work fine. The only thing that cannot be detected is the GPIO type, and users may want to use the quirk override parameter if the 'wrong' solution is provided. Tested-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

We only logged the SSP quirk, make sure the GPIO and DMIC quirks are exposed. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

Additional code was added and the comment on the speaker GPIO needs to be moved before we actually try to get the GPIO. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

The codec driver exposes a set of properties that can be set from the machine driver - as done in bytcht_es8316.c Start by adding the JD_INVERTED quirk. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

The ES8326 requires a different codec driver than ES8316/8336, fixup the codec name and dai name depending on the ACPI _HID exposed in the DSDT. Also add the select in Kconfig Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

The presence of DMICs needs to be signaled to UCM, follow the HDaudio example and use the 'cfg-dmics' component string to report the number of dmics present on the platform. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

Additional code was added and the comment on the speaker GPIO needs to be moved before we actually try to get the GPIO. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

fredoh9 · 2022-03-03T19:54:54Z

SOFCI TEST

plbossart · 2022-03-07T15:49:16Z

@ujfalusi can I please ask you to take a look, you had initial objections and I don't want to merge without your feedback. thanks!

plbossart · 2022-03-08T15:04:46Z

https://sof-ci.01.org/linuxpr/PR3338/build7261/devicetest/ had CI issues, let's redo the tests

plbossart · 2022-03-08T15:04:57Z

SOFCI TEST

ujfalusi

Looks fine to my eyes.

ujfalusi · 2022-03-08T15:20:22Z

sound/hda/intel-dsp-config.c

 	u8 acpi_hid[ACPI_ID_LEN];
 	const struct dmi_system_id *dmi_table;
-	u8 codec_hid[ACPI_ID_LEN];
+	const struct snd_soc_acpi_codecs *codec_hid;


@plbossart, out of curiosity, what 'hid' stands for? Can this be simply named as 'codecs'?

It's an ACPI acronym for 'Hardware ID'. There's also 'Compatible ID'.
this is represented in DSDT, e.g.

Device (ESSX) { Name (_ADR, Zero) // _ADR: Address Name (_HID, "ESSX8336") // _HID: Hardware ID Name (_CID, "ESSX8336") // _CID: Compatible ID Name (_DDN, "ES8336") // _DDN: DOS Device Name

plbossart requested review from ajitkupandey, bardliao, dbaluta, keyonjie, kuanhsuncheng, kv2019i, lgirdwood, libinyang, ranj063, ujfalusi and yaochunhung as code owners December 14, 2021 19:37

plbossart marked this pull request as draft December 14, 2021 19:37

plbossart force-pushed the fix/sof-es8336-quirk branch from 5d2d3f9 to 5ecffb3 Compare December 14, 2021 20:28

This was referenced Jan 10, 2022

[FEATURE] Support for ES8336 on Intel Comet Lake #3248

Closed

No card sound deteted TECLAST F15 Intel Celeron/Pentium Silver Processor High Definition Audio #3357

Closed

plbossart force-pushed the fix/sof-es8336-quirk branch from 5e149dc to eef6c90 Compare January 10, 2022 22:12

plbossart changed the title ~~[TEST] [DO NOT MERGE] use NHLT information for ES8336 devices~~ ASoC: SOF: Intel: use NHLT information for ES8336 devices Jan 10, 2022

plbossart marked this pull request as ready for review January 10, 2022 22:13

This was referenced Jan 11, 2022

No Internal Speaker Sound on GML Kuu Yobook M with ES8336 #3365

Closed

No sound on Multimedia audio controller: Intel Corporation Celeron/Pentium Silver Processor High Definition Audio (rev 06) #3360

Closed

plbossart and others added 20 commits March 3, 2022 13:12

ASoC: soc-acpi: fix kernel-doc descriptor

33261ab

Add missing dmic_num mention and clarify that 'links' mean 'SoundWire links', not to be used for other links. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

ALSA: intel-dspconfig: add ES8336 support for CNL

f2cc1ea

We're missing this check for the CNL PCI id Reported-by: Nikolai Kostrigin <nickel@altlinux.org> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

ASoC: Intel: soc-acpi: add ESSX8336 support on Cannon Lake machines

6e883a6

Add the support for all ES83x6 devices Signed-off-by: Nikolai Kostrigin <nickel@altlinux.org> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

ASoC: Intel: sof_es8336: make gpio optional

c5efb96

Do not fail if the GPIO used for speakers is not present, at least the headphone, headset and internal mics should work. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

ASoC: Intel: sof_es8336: log all quirks

107c934

We only logged the SSP quirk, make sure the GPIO and DMIC quirks are exposed. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

ASoC: Intel: sof_es8336: move comment to the right place

c808077

Additional code was added and the comment on the speaker GPIO needs to be moved before we actually try to get the GPIO. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

ASoC: Intel: sof_es8336: add support for JD inverted quirk

14d1c3b

The codec driver exposes a set of properties that can be set from the machine driver - as done in bytcht_es8316.c Start by adding the JD_INVERTED quirk. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

ASoC: Intel: bytcht_es8316: move comment to the right place

92ba636

Additional code was added and the comment on the speaker GPIO needs to be moved before we actually try to get the GPIO. Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

plbossart force-pushed the fix/sof-es8336-quirk branch from 0434208 to 92ba636 Compare March 3, 2022 19:37

plbossart requested a review from bardliao March 3, 2022 20:13

bardliao approved these changes Mar 4, 2022

View reviewed changes

ujfalusi approved these changes Mar 8, 2022

View reviewed changes

plbossart mentioned this pull request Mar 8, 2022

No detect audio jack Intel Corporation Celeron/Pentium Silver Processor High Definition Audio (rev 06) #3460

Closed

plbossart merged commit ea9b114 into thesofproject:topic/sof-dev Mar 8, 2022

ASoC: SOF: Intel: use NHLT information for ES83x6 devices #3338

ASoC: SOF: Intel: use NHLT information for ES83x6 devices #3338

Uh oh!

Conversation

plbossart commented Dec 14, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

plbossart commented Dec 14, 2021

Uh oh!

yangxiaohua2009 commented Dec 16, 2021

Uh oh!

plbossart commented Dec 16, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

plbossart commented Dec 16, 2021

Uh oh!

yangxiaohua2009 commented Dec 17, 2021

Uh oh!

plbossart commented Jan 10, 2022

Uh oh!

yangxiaohua2009 commented Jan 11, 2022

Uh oh!

plbossart commented Jan 11, 2022

Uh oh!

yangxiaohua2009 commented Jan 12, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

yangxiaohua2009 commented Jan 12, 2022

Uh oh!

fredoh9 commented Mar 3, 2022

Uh oh!

plbossart commented Mar 7, 2022

Uh oh!

plbossart commented Mar 8, 2022

Uh oh!

plbossart commented Mar 8, 2022

Uh oh!

ujfalusi left a comment

Choose a reason for hiding this comment

Uh oh!

ujfalusi Mar 8, 2022

Choose a reason for hiding this comment

Uh oh!

plbossart Mar 8, 2022

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

13 participants

plbossart commented Dec 14, 2021 •

edited

Loading

plbossart commented Dec 16, 2021 •

edited

Loading

yangxiaohua2009 commented Jan 12, 2022 •

edited

Loading