adapt SOF to use snd-hda-codec-hdmi
Subsystems (4)
Kai Vehmanen
maintainerv4 cover adapt SOF to use snd-hda-codec-hdmi
Hi all,
here's the 4th round for this series that adapts SOF to use snd-hda-codec-hdmi (patch_hdmi.c) codec driver instead of hdac_hdmi (soc/codecs/hdac_hdmi.c). The primary goal is to unify the HDMI codec implementation between DSP and non-DSP HDA configurations, offer same interface to user-space and reduce maintenance load for all.
v4 changes:
- Change order of patches to not break bisect (Pierre's feedback).
- Improve the explanation in commit message for mst_no_extra_pcms, patch 1. (Pierre's feedback).
- Fix errors in PCM constraints for HDMI (Pierre's feedback).
- Fix an issue on Ice Lake platforms (patch 3).
Feature and testing info:
- Tested on multiple Intel platforms supported by SOF.
- Tested with ALSA console tools as well as with Pulseaudio.
- requires Pulseaudio 12.x or newer, see
https://lists.freedesktop.org/archives/pulseaudio-discuss/2019-August/031358.html
- HDMI, DP, DP-MST with multi-monitor use-scenarios work ok.
- New feature for SOF: ELD /proc fs works just like in DSP-less mode.
- New feature for SOF: jack detection works out-of-the-box with Pulseaudio (no need for card specific UCM for HDMI)
Kai Vehmanen (9):
ALSA: hda/hdmi - implement mst_no_extra_pcms flag
ASoC: hdac_hda: add support for HDMI/DP as a HDA codec
ASoC: Intel: skl-hda-dsp-generic: use snd-hda-codec-hdmi
ASoC: Intel: skl-hda-dsp-generic: fix include guard name
ASoC: SOF: Intel: add support for snd-hda-codec-hdmi
ASoC: Intel: bxt-da7219-max98357a: common hdmi codec support
ASoC: Intel: glk_rt5682_max98357a: common hdmi codec support
ASoC: intel: sof_rt5682: common hdmi codec support
ASoC: Intel: bxt_rt298: common hdmi codec support
include/sound/hda_codec.h | 1 + include/sound/soc-acpi.h | 2 + sound/pci/hda/patch_hdmi.c | 19 +++- sound/soc/codecs/hdac_hda.c | 100 +++++++++++++++--- sound/soc/codecs/hdac_hda.h | 12 ++- sound/soc/intel/boards/bxt_da7219_max98357a.c | 11 ++ sound/soc/intel/boards/bxt_rt298.c | 11 ++ sound/soc/intel/boards/glk_rt5682_max98357a.c | 11 ++ sound/soc/intel/boards/hda_dsp_common.h | 93 ++++++++++++++++ sound/soc/intel/boards/skl_hda_dsp_common.c | 10 +- sound/soc/intel/boards/skl_hda_dsp_common.h | 27 ++++- sound/soc/intel/boards/skl_hda_dsp_generic.c | 1 + sound/soc/intel/boards/sof_rt5682.c | 11 ++ sound/soc/sof/intel/Kconfig | 10 ++ sound/soc/sof/intel/hda-codec.c | 19 +++- sound/soc/sof/intel/hda.c | 6 ++ sound/soc/sof/intel/hda.h | 6 +- 17 files changed, 323 insertions(+), 27 deletions(-) create mode 100644 sound/soc/intel/boards/hda_dsp_common.h
Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerPATCH v4 1/9 ALSA: hda/hdmi - implement mst_no_extra_pcms flag
To support the DP-MST multiple streams via single connector feature, the HDMI driver was extended with the concept of backup PCMs. See commit 9152085defb6 ("ALSA: hda - add DP MST audio support").
This implementation works fine with snd_hda_intel.c as PCM topology is fully managed within the single driver.
When the HDA codec driver is used from ASoC components, the concept of backup PCMs no longer fits. For ASoC topologies, the physical HDMI converters are presented as backend DAIs and these should match with hardware capabilities. The ASoC topology may define arbitrary PCMs (i.e. frontend DAIs) and have processing elements before eventual routing to the HDMI BE DAIs. With backup PCMs, the link between FE and BE DAIs would become dynamic and change when monitors are (un)plugged. This would lead to modifying the topology every time hotplug events happen, which is not currently possible in ASoC and there does not seem to be any obvious benefits from this design.
To overcome above problems and enable the HDMI driver to be used from ASoC, this patch adds a new mode (mst_no_extra_pcms flags) to patch_hdmi.c. In this mode, the codec driver does not assume the backup PCMs to be created.
--- include/sound/hda_codec.h | 1 + sound/pci/hda/patch_hdmi.c | 19 ++++++++++++++----- 2 files changed, 15 insertions(+), 5 deletions(-)
include/sound/hda_codec.h +1 -0
@@ -254,6 +254,7 @@ struct hda_codec { unsigned int force_pin_prefix:1; /* Add location prefix */ unsigned int link_down_at_suspend:1; /* link down at runtime suspend */ unsigned int relaxed_resume:1; /* don't resume forcibly for jack */ + unsigned int mst_no_extra_pcms:1; /* no backup PCMs for DP-MST */#ifdef CONFIG_PM unsigned long power_on_acct;
sound/pci/hda/patch_hdmi.c +14 -5
@@ -2072,15 +2072,24 @@ static bool is_hdmi_pcm_attached(struct hdac_device *hdac, int pcm_idx)static int generic_hdmi_build_pcms (struct hda_codec *codec) {struct hdmi_spec *spec = codec->spec; -int idx; +int idx, pcm_num;/* * for non-mst mode, pcm number is the same as before -* for DP MST mode, pcm number is (nid number + dev_num - 1) -* dev_num is the device entry number in a pin -* +* for DP MST mode without extra PCM, pcm number is same +* for DP MST mode with extra PCMs, pcm number is +* (nid number + dev_num - 1) +* dev_num is the device entry number in a pin */ -for (idx =0 ; idx < spec->num_nids + spec->dev_num -1 ; idx++) { + +if (codec->mst_no_extra_pcms) + pcm_num = spec->num_nids; +else + pcm_num = spec->num_nids + spec->dev_num -1 ; + +codec_dbg (codec,"hdmi: pcm_num set to %d\n" , pcm_num); + +for (idx =0 ; idx < pcm_num; idx++) {struct hda_pcm *info;struct hda_pcm_stream *pstr;
Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Re: [alsa-devel] [PATCH v4 1/9] ALSA: hda/hdmi - implement mst_no_extra_pcms flag
Quoted 19 lines · Kai Vehmanen
To support the DP-MST multiple streams via single connector feature, the HDMI driver was extended with the concept of backup PCMs. See commit
9152085defb6("ALSA: hda - add DP MST audio support").This implementation works fine with
snd_hda_intel.cas PCM topology is fully managed within the single driver.When the HDA codec driver is used from ASoC components, the concept of backup PCMs no longer fits. For ASoC topologies, the physical HDMI converters are presented as backend DAIs and these should match with hardware capabilities. The ASoC topology may define arbitrary PCMs (i.e. frontend DAIs) and have processing elements before eventual routing to the HDMI BE DAIs. With backup PCMs, the link between FE and BE DAIs would become dynamic and change when monitors are (un)plugged. This would lead to modifying the topology every time hotplug events happen, which is not currently possible in ASoC and there does not seem to be any obvious benefits from this design.
Quoted 6 lines
To overcome above problems and enable the HDMI driver to be used from ASoC, this patch adds a new mode (
mst_no_extra_pcmsflags) topatch_hdmi.c. In this mode, the codec driver does not assume the backup PCMs to be created.Signed-off-by Kai Vehmanen <[email protected]>
Looks good to me.
thanks,
Takashi
Quoted 59 lines
--- include/sound/hda_codec.h | 1 + sound/pci/hda/patch_hdmi.c | 19 ++++++++++++++----- 2 files changed, 15 insertions(+), 5 deletions(-)include/sound/hda_codec.h +1
-0@@ -254,6 +254,7 @@ struct hda_codec { unsigned int force_pin_prefix:1; /* Add location prefix */ unsigned int link_down_at_suspend:1; /* link down at runtime suspend */ unsigned int relaxed_resume:1; /* don't resume forcibly for jack */ + unsigned int mst_no_extra_pcms:1; /* no backup PCMs for DP-MST */#ifdef CONFIG_PM unsigned long power_on_acct;sound/pci/hda/patch_hdmi.c +14
-5@@ -2072,15 +2072,24 @@ static bool is_hdmi_pcm_attached(struct hdac_device *hdac, int pcm_idx)static int generic_hdmi_build_pcms (struct hda_codec *codec) {struct hdmi_spec *spec = codec->spec; -int idx; +int idx, pcm_num;/* * for non-mst mode, pcm number is the same as before -* for DP MST mode, pcm number is (nid number + dev_num - 1) -* dev_num is the device entry number in a pin -* +* for DP MST mode without extra PCM, pcm number is same +* for DP MST mode with extra PCMs, pcm number is +* (nid number + dev_num - 1) +* dev_num is the device entry number in a pin */ -for (idx =0 ; idx < spec->num_nids + spec->dev_num -1 ; idx++) { + +if (codec->mst_no_extra_pcms) + pcm_num = spec->num_nids; +else + pcm_num = spec->num_nids + spec->dev_num -1 ; + +codec_dbg (codec,"hdmi: pcm_num set to %d\n" , pcm_num); + +for (idx =0 ; idx < pcm_num; idx++) {struct hda_pcm *info;struct hda_pcm_stream *pstr;Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
_______________________________________________
Alsa-devel mailing list
[email protected]
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerPATCH v4 2/9 ASoC: hdac_hda: add support for HDMI/DP as a HDA codec
In reply to Kai VehmanenHandle all HDA codecs using same logic, including HDMI/DP.
Call to snd_hda_codec_build_controls() is delayed for HDMI/DP HDA devices. This is needed to discover the PCM device numbers as defined in topology.
--- sound/soc/codecs/hdac_hda.c | 100 +++++++++++++++++++++++++++++++----- sound/soc/codecs/hdac_hda.h | 12 ++++- 2 files changed, 99 insertions(+), 13 deletions(-)
sound/soc/codecs/hdac_hda.c +88 -12
@@ -16,11 +16,8 @@#include <sound/hdaudio_ext.h> #include <sound/hda_codec.h> #include <sound/hda_register.h> -#include "hdac_hda.h" -#define HDAC_ANALOG_DAI_ID 0 -#define HDAC_DIGITAL_DAI_ID 1 -#define HDAC_ALT_ANALOG_DAI_ID 2 +#include "hdac_hda.h" #define STUB_FORMATS (SNDRV_PCM_FMTBIT_S8 | \ SNDRV_PCM_FMTBIT_U8 | \
@@ -32,6 +29,11 @@ SNDRV_PCM_FMTBIT_U32_LE | \ SNDRV_PCM_FMTBIT_IEC958_SUBFRAME_LE) +#define STUB_HDMI_RATES (SNDRV_PCM_RATE_32000 | SNDRV_PCM_RATE_44100 |\ + SNDRV_PCM_RATE_48000 | SNDRV_PCM_RATE_88200 |\ + SNDRV_PCM_RATE_96000 | SNDRV_PCM_RATE_176400 |\ + SNDRV_PCM_RATE_192000) +static int hdac_hda_dai_open (struct snd_pcm_substream *substream,struct snd_soc_dai *dai); static void hdac_hda_dai_close(struct snd_pcm_substream *substream,
@@ -121,7 +123,46 @@ static struct snd_soc_dai_driver hdac_hda_dais[] = { .formats = STUB_FORMATS, .sig_bits = 24, }, -} +}, +{ + .id = HDAC_HDMI_0_DAI_ID, + .name = "intel-hdmi-hifi1", + .ops = &hdac_hda_dai_ops, + .playback = { + .stream_name = "hifi1", + .channels_min = 1, + .channels_max = 32, + .rates = STUB_HDMI_RATES, + .formats = STUB_FORMATS, + .sig_bits = 24, + }, +}, +{ + .id = HDAC_HDMI_1_DAI_ID, + .name = "intel-hdmi-hifi2", + .ops = &hdac_hda_dai_ops, + .playback = { + .stream_name = "hifi2", + .channels_min = 1, + .channels_max = 32, + .rates = STUB_HDMI_RATES, + .formats = STUB_FORMATS, + .sig_bits = 24, + }, +}, +{ + .id = HDAC_HDMI_2_DAI_ID, + .name = "intel-hdmi-hifi3", + .ops = &hdac_hda_dai_ops, + .playback = { + .stream_name = "hifi3", + .channels_min = 1, + .channels_max = 32, + .rates = STUB_HDMI_RATES, + .formats = STUB_FORMATS, + .sig_bits = 24, + }, +}, };
@@ -135,10 +176,11 @@ static int hdac_hda_dai_set_tdm_slot(struct snd_soc_dai *dai, hda_pvt =snd_soc_component_get_drvdata (component); pcm = &hda_pvt->pcm[dai->id]; +if (tx_mask) - pcm[dai->id].stream_tag[SNDRV_PCM_STREAM_PLAYBACK ] = tx_mask; + pcm->stream_tag[SNDRV_PCM_STREAM_PLAYBACK ] = tx_mask;else - pcm[dai->id].stream_tag[SNDRV_PCM_STREAM_CAPTURE ] = rx_mask; + pcm->stream_tag[SNDRV_PCM_STREAM_CAPTURE ] = rx_mask;return 0 ; }
@@ -278,6 +320,12 @@ static struct hda_pcm *snd_soc_find_pcm_from_dai(struct hdac_hda_priv *hda_pvt,struct hda_pcm *cpcm;const char *pcm_name; +/* +* map DAI ID to the closest matching PCM name, using the naming +* scheme used by hda-codec snd_hda_gen_build_pcms() and for +* HDMI in hda_codec patch_hdmi.c) +*/ +switch (dai->id) {case HDAC_ANALOG_DAI_ID : pcm_name ="Analog" ;
@@ -288,13 +336,22 @@ static struct hda_pcm *snd_soc_find_pcm_from_dai(struct hdac_hda_priv *hda_pvt,case HDAC_ALT_ANALOG_DAI_ID : pcm_name ="Alt Analog" ;break ; +case HDAC_HDMI_0_DAI_ID : + pcm_name ="HDMI 0" ; +break ; +case HDAC_HDMI_1_DAI_ID : + pcm_name ="HDMI 1" ; +break ; +case HDAC_HDMI_2_DAI_ID : + pcm_name ="HDMI 2" ; +break ;default :dev_err (&hcodec->core.dev,"invalid dai id %d\n" , dai->id);return NULL ; }list_for_each_entry (cpcm, &hcodec->pcm_list_head, list) { -if (strpbrk (cpcm->name, pcm_name)) +if (strstr (cpcm->name, pcm_name))return cpcm; }
@@ -302,6 +359,18 @@ static struct hda_pcm *snd_soc_find_pcm_from_dai(struct hdac_hda_priv *hda_pvt,return NULL ; } +static bool is_hdmi_codec (struct hda_codec *hcodec) +{ +struct hda_pcm *cpcm; + +list_for_each_entry (cpcm, &hcodec->pcm_list_head, list) { +if (cpcm->pcm_type ==HDA_PCM_TYPE_HDMI ) +return true; + } + +return false; +} +static int hdac_hda_codec_probe (struct snd_soc_component *component) { struct hdac_hda_priv *hda_pvt =
@@ -366,16 +435,23 @@ static int hdac_hda_codec_probe(struct snd_soc_component *component)dev_dbg (&hdev->dev,"no patch file found\n" ); } +/* configure codec for 1:1 PCM:DAI mapping */ + hcodec->mst_no_extra_pcms =1 ; + ret =snd_hda_codec_parse_pcms (hcodec);if (ret <0 ) {dev_err (&hdev->dev,"unable to map pcms to dai %d\n" , ret); goto error; } - ret =snd_hda_codec_build_controls (hcodec); -if (ret <0 ) { -dev_err (&hdev->dev,"unable to create controls %d\n" , ret); - goto error; +/* HDMI controls need to be created in machine drivers */ +if (!is_hdmi_codec (hcodec)) { + ret =snd_hda_codec_build_controls (hcodec); +if (ret <0 ) { +dev_err (&hdev->dev,"unable to create controls %d\n" , + ret); + goto error; + } } hcodec->core.lazy_cache = true;
sound/soc/codecs/hdac_hda.h +11 -1
@@ -6,6 +6,16 @@ #ifndef __HDAC_HDA_H__ #define __HDAC_HDA_H__ +enum { + HDAC_ANALOG_DAI_ID = 0, + HDAC_DIGITAL_DAI_ID, + HDAC_ALT_ANALOG_DAI_ID, + HDAC_HDMI_0_DAI_ID, + HDAC_HDMI_1_DAI_ID, + HDAC_HDMI_2_DAI_ID, + HDAC_LAST_DAI_ID = HDAC_HDMI_2_DAI_ID, +}; + struct hdac_hda_pcm { int stream_tag[2]; unsigned int format_val[2];
@@ -13,7 +23,7 @@ struct hdac_hda_pcm {struct hdac_hda_priv {struct hda_codec codec; -struct hdac_hda_pcm pcm[2 ]; +struct hdac_hda_pcm pcm[HDAC_LAST_DAI_ID ]; };#define hdac_to_hda_priv (_hdac) \
Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerPATCH v4 3/9 ASoC: Intel: skl-hda-dsp-generic: use snd-hda-codec-hdmi
In reply to Kai VehmanenAdd support for using snd-hda-codec-hdmi driver for HDMI/DP instead of ASoC hdac-hdmi. This is aligned with how other HDA codecs are already handled.
When snd-hda-codec-hdmi is used, the PCM device numbers are parsed from card topology and passed to the codec driver. This needs to be done at runtime as topology changes may affect PCM device allocation.
--- include/sound/soc-acpi.h | 2 + sound/soc/intel/boards/hda_dsp_common.h | 93 ++++++++++++++++++++ sound/soc/intel/boards/skl_hda_dsp_common.c | 10 ++- sound/soc/intel/boards/skl_hda_dsp_common.h | 23 +++++ sound/soc/intel/boards/skl_hda_dsp_generic.c | 1 + 5 files changed, 128 insertions(+), 1 deletion(-) create mode 100644 sound/soc/intel/boards/hda_dsp_common.h
include/sound/soc-acpi.h +2 -0
@@ -60,12 +60,14 @@ static inline struct snd_soc_acpi_mach *snd_soc_acpi_codec_list(void *arg)* @acpi_ipc_irq_index: used for BYT-CR detection * @platform: string used for HDaudio codec support * @codec_mask: used for HDAudio support +* @common_hdmi_codec_drv: use commom HDAudio HDMI codec driver */ struct snd_soc_acpi_mach_params {u32 acpi_ipc_irq_index;const char *platform;u32 codec_mask;u32 dmic_num; +bool common_hdmi_codec_drv; };/**
sound/soc/intel/boards/hda_dsp_common.h new file +93 -0
93 lines not on this page · raw message
sound/soc/intel/boards/skl_hda_dsp_common.c +9 -1
@@ -14,6 +14,9 @@#include "../../codecs/hdac_hdmi.h" #include "skl_hda_dsp_common.h" +#include <sound/hda_codec.h> +#include "../../codecs/hdac_hda.h" +#define NAME_SIZE 32int skl_hda_hdmi_add_pcm (struct snd_soc_card *card,int device)
@@ -136,14 +139,19 @@ int skl_hda_hdmi_jack_init(struct snd_soc_card *card)char jack_name[NAME_SIZE ];int err; +if (ctx->common_hdmi_codec_drv) +return skl_hda_hdmi_build_controls (card); +list_for_each_entry (pcm, &ctx->hdmi_pcm_list, head) { +if (!pcm) +continue ; + component = pcm->codec_dai->component;snprintf (jack_name,sizeof (jack_name),"HDMI/DP, pcm=%d Jack" , pcm->device); err =snd_soc_card_jack_new (card, jack_name,SND_JACK_AVOUT , &pcm->hdmi_jack,NULL ,0 ); -if (err)return err;
sound/soc/intel/boards/skl_hda_dsp_common.h +23 -0
@@ -14,6 +14,9 @@#include <linux/platform_device.h> #include <sound/core.h> #include <sound/jack.h> +#include <sound/hda_codec.h> +#include "../../codecs/hdac_hda.h" +#include "hda_dsp_common.h" #define HDA_DSP_MAX_BE_DAI_LINKS 7
@@ -29,10 +32,30 @@ struct skl_hda_private {int pcm_count;int dai_index;const char *platform_name; +bool common_hdmi_codec_drv; };extern struct snd_soc_dai_link skl_hda_be_dai_links[HDA_DSP_MAX_BE_DAI_LINKS ];int skl_hda_hdmi_jack_init (struct snd_soc_card *card);int skl_hda_hdmi_add_pcm (struct snd_soc_card *card,int device); +/* +* Search card topology and register HDMI PCM related controls +* to codec driver. +*/ +static inline int skl_hda_hdmi_build_controls (struct snd_soc_card *card) +{ +struct skl_hda_private *ctx =snd_soc_card_get_drvdata (card); +struct snd_soc_component *component; +struct skl_hda_hdmi_pcm *pcm; + + pcm = list_first_entry(&ctx->hdmi_pcm_list, struct skl_hda_hdmi_pcm, + head); + component = pcm->codec_dai->component; +if (!component) +return -EINVAL ; + +return hda_dsp_hdmi_build_controls (card, component); +} +#endif /* __SOUND_SOC_HDA_DSP_COMMON_H */
sound/soc/intel/boards/skl_hda_dsp_generic.c +1 -0
@@ -178,6 +178,7 @@ static int skl_hda_audio_probe(struct platform_device *pdev) ctx->pcm_count = hda_soc_card.num_links; ctx->dai_index =1 ;/* hdmi codec dai name starts from index 1 */ ctx->platform_name = mach->mach_params.platform; + ctx->common_hdmi_codec_drv = mach->mach_params.common_hdmi_codec_drv; hda_soc_card.dev = &pdev->dev;snd_soc_card_set_drvdata (&hda_soc_card, ctx);
Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerPATCH v4 4/9 ASoC: Intel: skl-hda-dsp-generic: fix include guard name
In reply to Kai VehmanenMatch the include guard define to actual filename. The source directory now has an actual hda_dsp_common.h header, so the old include guard may cause confusion.
--- sound/soc/intel/boards/skl_hda_dsp_common.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
sound/soc/intel/boards/skl_hda_dsp_common.h +2 -2
@@ -8,8 +8,8 @@* platforms with HDA Codecs. */ -#ifndef __SOUND_SOC_HDA_DSP_COMMON_H -#define __SOUND_SOC_HDA_DSP_COMMON_H +#ifndef __SKL_HDA_DSP_COMMON_H +#define __SKL_HDA_DSP_COMMON_H#include <linux/module.h> #include <linux/platform_device.h> #include <sound/core.h>
Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerPATCH v4 5/9 ASoC: SOF: Intel: add support for snd-hda-codec-hdmi
In reply to Kai VehmanenAdd support to implement HDMI/DP audio by using the common snd-hda-codec-hdmi driver.
Change of codec driver affects user-space as the the two drivers expose different mixer controls. A new kernel module option "use_common_hdmi" is added to user-space to indicate which interface should be used. The default driver can be selected via a Kconfig option.
--- sound/soc/sof/intel/Kconfig | 10 ++++++++++ sound/soc/sof/intel/hda-codec.c | 19 +++++++++++++++---- sound/soc/sof/intel/hda.c | 6 ++++++ sound/soc/sof/intel/hda.h | 6 ++++-- 4 files changed, 35 insertions(+), 6 deletions(-)
sound/soc/sof/intel/Kconfig +10 -0
@@ -273,6 +273,16 @@ config SND_SOC_SOF_HDA_AUDIO_CODEC Say Y if you want to enable HDAudio codecs with SOF. If unsure select "N". +config SND_SOC_SOF_HDA_COMMON_HDMI_CODEC + bool "SOF common HDA HDMI codec driver" + depends on SND_SOC_SOF_HDA_LINK + depends on SND_HDA_CODEC_HDMI + help + This adds support for HDMI audio by using the common HDA + HDMI/DisplayPort codec driver. + Say Y if you want to use the common codec driver with SOF. + If unsure select "Y". + endif ## SND_SOC_SOF_HDA_COMMON config SND_SOC_SOF_HDA_LINK_BASELINE
sound/soc/sof/intel/hda-codec.c +15 -4
@@ -84,6 +84,8 @@ static int hda_codec_probe(struct snd_sof_dev *sdev, int address) {#if IS_ENABLED (CONFIG_SND_SOC_SOF_HDA_AUDIO_CODEC )struct hdac_hda_priv *hda_priv; +struct snd_soc_acpi_mach_params *mach_params =0 ; +struct snd_sof_pdata *pdata = sdev->pdata;#endif struct hda_bus *hbus =sof_to_hbus (sdev);struct hdac_device *hdev;
@@ -113,8 +115,16 @@ static int hda_codec_probe(struct snd_sof_dev *sdev, int address)if (ret <0 )return ret; -/* use legacy bus only for HDA codecs, idisp uses ext bus */ -if ((resp &0xFFFF0000 ) !=IDISP_VID_INTEL ) { +if (pdata->machine) + mach_params = (struct snd_soc_acpi_mach_params *) + &pdata->machine->mach_params; + +/* +* if common HDMI codec driver is not used, codec load +* is skipped here and hdac_hdmi is used instead +*/ +if ((mach_params && mach_params->common_hdmi_codec_drv) || + (resp &0xFFFF0000 ) !=IDISP_VID_INTEL ) { hdev->type =HDA_DEV_LEGACY ;hda_codec_load_module (&hda_priv->codec); }
@@ -155,7 +165,8 @@ int hda_codec_probe_bus(struct snd_sof_dev *sdev) }EXPORT_SYMBOL (hda_codec_probe_bus); -#if IS_ENABLED (CONFIG_SND_SOC_HDAC_HDMI ) +#if IS_ENABLED (CONFIG_SND_HDA_CODEC_HDMI ) || \ +IS_ENABLED (CONFIG_SND_SOC_HDAC_HDMI )void hda_codec_i915_get (struct snd_sof_dev *sdev) {
@@ -204,6 +215,6 @@ int hda_codec_i915_exit(struct snd_sof_dev *sdev) }EXPORT_SYMBOL (hda_codec_i915_exit); -#endif /* CONFIG_SND_SOC_HDAC_HDMI */ +#endif MODULE_LICENSE ("Dual BSD/GPL" );
sound/soc/sof/intel/hda.c +6 -0
@@ -54,6 +54,11 @@ MODULE_PARM_DESC(use_msi, "SOF HDA use PCI MSI mode");static int hda_dmic_num =-1 ;module_param_named (dmic_num, hda_dmic_num, int,0444 );MODULE_PARM_DESC (dmic_num,"SOF HDA DMIC number" ); + +static bool hda_codec_use_common_hdmi = +IS_ENABLED (CONFIG_SND_SOC_SOF_HDA_COMMON_HDMI_CODEC ); +module_param_named (use_common_hdmi, hda_codec_use_common_hdmi, bool,0444 ); +MODULE_PARM_DESC (use_common_hdmi,"SOF HDA use common HDMI codec driver" );#endif static const struct hda_dsp_msg_code hda_dsp_rom_msg[] = {
@@ -458,6 +463,7 @@ static int hda_init_caps(struct snd_sof_dev *sdev) &pdata->machine->mach_params; mach_params->codec_mask = bus->codec_mask; mach_params->platform =dev_name (sdev->dev); + mach_params->common_hdmi_codec_drv = hda_codec_use_common_hdmi; }/* create codec instances */
sound/soc/sof/intel/hda.h +4 -2
@@ -562,7 +562,9 @@ void hda_codec_jack_check(struct snd_sof_dev *sdev);#endif /* CONFIG_SND_SOC_SOF_HDA */ -#if IS_ENABLED (CONFIG_SND_SOC_SOF_HDA ) &&IS_ENABLED (CONFIG_SND_SOC_HDAC_HDMI ) +#if IS_ENABLED (CONFIG_SND_SOC_SOF_HDA ) && \ + (IS_ENABLED (CONFIG_SND_HDA_CODEC_HDMI ) || \ +IS_ENABLED (CONFIG_SND_SOC_HDAC_HDMI ))void hda_codec_i915_get (struct snd_sof_dev *sdev);void hda_codec_i915_put (struct snd_sof_dev *sdev);
@@ -576,7 +578,7 @@ static inline void hda_codec_i915_put(struct snd_sof_dev *sdev) { }static inline int hda_codec_i915_init (struct snd_sof_dev *sdev) {return 0 ; }static inline int hda_codec_i915_exit (struct snd_sof_dev *sdev) {return 0 ; } -#endif /* CONFIG_SND_SOC_SOF_HDA && CONFIG_SND_SOC_HDAC_HDMI */ +#endif /* * Trace Control.
Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerPATCH v4 6/9 ASoC: Intel: bxt-da7219-max98357a: common hdmi codec support
In reply to Kai VehmanenAdd support for using snd-hda-codec-hdmi driver for HDMI/DP instead of ASoC hdac-hdmi. This is aligned with how other HDA codecs are already handled.
--- sound/soc/intel/boards/bxt_da7219_max98357a.c | 11 +++++++++++ 1 file changed, 11 insertions(+)
sound/soc/intel/boards/bxt_da7219_max98357a.c +11 -0
@@ -21,6 +21,7 @@#include "../../codecs/da7219.h" #include "../../codecs/da7219-aad.h" #include "../common/soc-intel-quirks.h" +#include "hda_dsp_common.h" #define BXT_DIALOG_CODEC_DAI "da7219-hifi"#define BXT_MAXIM_CODEC_DAI "HiFi"
@@ -38,6 +39,7 @@ struct bxt_hdmi_pcm {struct bxt_card_private {struct list_head hdmi_pcm_list; +bool common_hdmi_codec_drv; };enum {
@@ -615,6 +617,13 @@ static int bxt_card_late_probe(struct snd_soc_card *card)snd_soc_dapm_add_routes (&card->dapm, broxton_map,ARRAY_SIZE (broxton_map)); + pcm = list_first_entry(&ctx->hdmi_pcm_list, struct bxt_hdmi_pcm, + head); + component = pcm->codec_dai->component; + +if (ctx->common_hdmi_codec_drv) +return hda_dsp_hdmi_build_controls (card, component); +list_for_each_entry (pcm, &ctx->hdmi_pcm_list, head) { component = pcm->codec_dai->component; snprintf(jack_name, sizeof(jack_name),
@@ -720,6 +729,8 @@ static int broxton_audio_probe(struct platform_device *pdev)if (ret)return ret; + ctx->common_hdmi_codec_drv = mach->mach_params.common_hdmi_codec_drv; +return devm_snd_soc_register_card (&pdev->dev, &broxton_audio_card); }
Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerPATCH v4 7/9 ASoC: Intel: glk_rt5682_max98357a: common hdmi codec support
In reply to Kai VehmanenAdd support for using snd-hda-codec-hdmi driver for HDMI/DP instead of ASoC hdac-hdmi. This is aligned with how other HDA codecs are already handled.
--- sound/soc/intel/boards/glk_rt5682_max98357a.c | 11 +++++++++++ 1 file changed, 11 insertions(+)
sound/soc/intel/boards/glk_rt5682_max98357a.c +11 -0
@@ -19,6 +19,7 @@#include <sound/soc-acpi.h> #include "../../codecs/rt5682.h" #include "../../codecs/hdac_hdmi.h" +#include "hda_dsp_common.h" /* The platform clock outputs 19.2Mhz clock to codec as I2S MCLK */ #define GLK_PLAT_CLK_FREQ 19200000
@@ -41,6 +42,7 @@ struct glk_hdmi_pcm {struct glk_card_private {struct snd_soc_jack geminilake_headset;struct list_head hdmi_pcm_list; +bool common_hdmi_codec_drv; };enum {
@@ -545,6 +547,13 @@ static int glk_card_late_probe(struct snd_soc_card *card)int err =0 ;int i =0 ; + pcm = list_first_entry(&ctx->hdmi_pcm_list, struct glk_hdmi_pcm, + head); + component = pcm->codec_dai->component; + +if (ctx->common_hdmi_codec_drv) +return hda_dsp_hdmi_build_controls (card, component); +list_for_each_entry (pcm, &ctx->hdmi_pcm_list, head) { component = pcm->codec_dai->component; snprintf(jack_name, sizeof(jack_name),
@@ -612,6 +621,8 @@ static int geminilake_audio_probe(struct platform_device *pdev)if (ret)return ret; + ctx->common_hdmi_codec_drv = mach->mach_params.common_hdmi_codec_drv; +return devm_snd_soc_register_card (&pdev->dev, card); }
Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerPATCH v4 8/9 ASoC: intel: sof_rt5682: common hdmi codec support
In reply to Kai VehmanenAdd support for using snd-hda-codec-hdmi driver for HDMI/DP instead of ASoC hdac-hdmi. This is aligned with how other HDA codecs are already handled.
--- sound/soc/intel/boards/sof_rt5682.c | 11 +++++++++++ 1 file changed, 11 insertions(+)
sound/soc/intel/boards/sof_rt5682.c +11 -0
@@ -21,6 +21,7 @@#include "../../codecs/rt5682.h" #include "../../codecs/hdac_hdmi.h" #include "../common/soc-intel-quirks.h" +#include "hda_dsp_common.h" #define NAME_SIZE 32
@@ -53,6 +54,7 @@ struct sof_card_private {struct clk *mclk;struct snd_soc_jack sof_headset;struct list_head hdmi_pcm_list; +bool common_hdmi_codec_drv; };static int sof_rt5682_quirk_cb (const struct dmi_system_id *id)
@@ -274,6 +276,13 @@ static int sof_card_late_probe(struct snd_soc_card *card)if (is_legacy_cpu)return 0 ; + pcm = list_first_entry(&ctx->hdmi_pcm_list, struct sof_hdmi_pcm, + head); + component = pcm->codec_dai->component; + +if (ctx->common_hdmi_codec_drv) +return hda_dsp_hdmi_build_controls (card, component); +list_for_each_entry (pcm, &ctx->hdmi_pcm_list, head) { component = pcm->codec_dai->component; snprintf(jack_name, sizeof(jack_name),
@@ -615,6 +624,8 @@ static int sof_audio_probe(struct platform_device *pdev)if (ret)return ret; + ctx->common_hdmi_codec_drv = mach->mach_params.common_hdmi_codec_drv; +snd_soc_card_set_drvdata (&sof_audio_card_rt5682, ctx); return devm_snd_soc_register_card(&pdev->dev,
Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerPATCH v4 9/9 ASoC: Intel: bxt_rt298: common hdmi codec support
In reply to Kai VehmanenAdd support for using snd-hda-codec-hdmi driver for HDMI/DP instead of ASoC hdac-hdmi. This is aligned with how other HDA codecs are already handled.
--- sound/soc/intel/boards/bxt_rt298.c | 11 +++++++++++ 1 file changed, 11 insertions(+)
sound/soc/intel/boards/bxt_rt298.c +11 -0
@@ -18,6 +18,7 @@#include <sound/pcm_params.h> #include "../../codecs/hdac_hdmi.h" #include "../../codecs/rt298.h" +#include "hda_dsp_common.h" /* Headset jack detection DAPM pins */ static struct snd_soc_jack broxton_headset;
@@ -31,6 +32,7 @@ struct bxt_hdmi_pcm {struct bxt_rt286_private {struct list_head hdmi_pcm_list; +bool common_hdmi_codec_drv; };enum {
@@ -527,6 +529,13 @@ static int bxt_card_late_probe(struct snd_soc_card *card)int err, i =0 ;char jack_name[NAME_SIZE ]; + pcm = list_first_entry(&ctx->hdmi_pcm_list, struct bxt_hdmi_pcm, + head); + component = pcm->codec_dai->component; + +if (ctx->common_hdmi_codec_drv) +return hda_dsp_hdmi_build_controls (card, component); +list_for_each_entry (pcm, &ctx->hdmi_pcm_list, head) { component = pcm->codec_dai->component; snprintf(jack_name, sizeof(jack_name),
@@ -626,6 +635,8 @@ static int broxton_audio_probe(struct platform_device *pdev)if (ret)return ret; + ctx->common_hdmi_codec_drv = mach->mach_params.common_hdmi_codec_drv; +return devm_snd_soc_register_card (&pdev->dev, card); }
Signature · 6 lines
2.17.1 _______________________________________________ Alsa-devel mailing list [email protected] https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Quoted 7 lines · Kai Vehmanen
Hi all,
here's the 4th round for this series that adapts SOF to use snd-hda-codec-hdmi (
patch_hdmi.c) codec driver instead ofhdac_hdmi(soc/codecs/hdac_hdmi.c). The primary goal is to unify the HDMI codec implementation between DSP and non-DSP HDA configurations, offer same interface to user-space and reduce maintenance load for all.
The series looks good to me so
I would recommend that we have a matching change for the Skylake driver and validation that both SOF and cAVS/SST drivers can operate with this mode set. Our goal is still to have coexistence between the two drivers in a single build/distro, e.g. cAVS for SKL/KBL/APL and SOF for newer platforms. This can be done in a follow-up patch but it needs to be done before distros start selecting this common HDMI mode.
The other concern I have is that we have other changes coming for soc-acpi and machine drivers to deal with SoundWire, so it'd be nice to have the changes mirrored between Mark and Takashi trees e.g. with the merge of a fixed branch.
Quoted 48 lines
v4 changes:
- Change order of patches to not break bisect (Pierre's feedback).
- Improve the explanation in commit message formst_no_extra_pcms, patch 1. (Pierre's feedback).
- Fix errors in PCM constraints for HDMI (Pierre's feedback).
- Fix an issue on Ice Lake platforms (patch 3).Feature and testing info:
- Tested on multiple Intel platforms supported by SOF.
- Tested with ALSA console tools as well as with Pulseaudio.
- requires Pulseaudio 12.x or newer, see
https://lists.freedesktop.org/archives/pulseaudio-discuss/2019-August/031358.html
- HDMI, DP, DP-MST with multi-monitor use-scenarios work ok.
- New feature for SOF: ELD /proc fs works just like in DSP-less mode.
- New feature for SOF: jack detection works out-of-the-box with Pulseaudio (no need for card specific UCM for HDMI)Kai Vehmanen (9):
ALSA: hda/hdmi - implementmst_no_extra_pcmsflag
ASoC:hdac_hda: add support for HDMI/DP as a HDA codec
ASoC: Intel: skl-hda-dsp-generic: use snd-hda-codec-hdmi
ASoC: Intel: skl-hda-dsp-generic: fix include guard name
ASoC: SOF: Intel: add support for snd-hda-codec-hdmi
ASoC: Intel: bxt-da7219-max98357a: common hdmi codec support
ASoC: Intel:glk_rt5682_max98357a: common hdmi codec support
ASoC: intel:sof_rt5682: common hdmi codec support
ASoC: Intel:bxt_rt298: common hdmi codec supportinclude/sound/hda_codec.h | 1 + include/sound/soc-acpi.h | 2 + sound/pci/hda/patch_hdmi.c | 19 +++- sound/soc/codecs/hdac_hda.c | 100 +++++++++++++++--- sound/soc/codecs/hdac_hda.h | 12 ++- sound/soc/intel/boards/bxt_da7219_max98357a.c | 11 ++ sound/soc/intel/boards/bxt_rt298.c | 11 ++ sound/soc/intel/boards/glk_rt5682_max98357a.c | 11 ++ sound/soc/intel/boards/hda_dsp_common.h | 93 ++++++++++++++++ sound/soc/intel/boards/skl_hda_dsp_common.c | 10 +- sound/soc/intel/boards/skl_hda_dsp_common.h | 27 ++++- sound/soc/intel/boards/skl_hda_dsp_generic.c | 1 + sound/soc/intel/boards/sof_rt5682.c | 11 ++ sound/soc/sof/intel/Kconfig | 10 ++ sound/soc/sof/intel/hda-codec.c | 19 +++- sound/soc/sof/intel/hda.c | 6 ++ sound/soc/sof/intel/hda.h | 6 +- 17 files changed, 323 insertions(+), 27 deletions(-)
Quoted 2 lines
create mode 100644
sound/soc/intel/boards/hda_dsp_common.h
_______________________________________________
Alsa-devel mailing list
[email protected]
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Quoted 21 lines · Pierre-Louis Bossart
Quoted 7 lines · Kai Vehmanen
Hi all,
here's the 4th round for this series that adapts SOF to use snd-hda-codec-hdmi (
patch_hdmi.c) codec driver instead ofhdac_hdmi(soc/codecs/hdac_hdmi.c). The primary goal is to unify the HDMI codec implementation between DSP and non-DSP HDA configurations, offer same interface to user-space and reduce maintenance load for all.The series looks good to me so
Reviewed-by Pierre-Louis Bossart <[email protected]>I would recommend that we have a matching change for the Skylake driver and validation that both SOF and cAVS/SST drivers can operate with this mode set. Our goal is still to have coexistence between the two drivers in a single build/distro, e.g. cAVS for SKL/KBL/APL and SOF for newer platforms. This can be done in a follow-up patch but it needs to be done before distros start selecting this common HDMI mode.
Quoted 4 lines
The other concern I have is that we have other changes coming for soc-acpi and machine drivers to deal with SoundWire, so it'd be nice to have the changes mirrored between Mark and Takashi trees e.g. with the merge of a fixed branch.
In general it's already a bit too late to merge for 5.4, so the whole series will be after that, at most, hence we still have enough time :)
Takashi
_______________________________________________
Alsa-devel mailing list
[email protected]
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Quoted 23 lines · Takashi Iwai
Quoted 21 lines · Pierre-Louis Bossart
Quoted 7 lines · Kai Vehmanen
Hi all,
here's the 4th round for this series that adapts SOF to use snd-hda-codec-hdmi (
patch_hdmi.c) codec driver instead ofhdac_hdmi(soc/codecs/hdac_hdmi.c). The primary goal is to unify the HDMI codec implementation between DSP and non-DSP HDA configurations, offer same interface to user-space and reduce maintenance load for all.The series looks good to me so
Reviewed-by Pierre-Louis Bossart <[email protected]>I would recommend that we have a matching change for the Skylake driver and validation that both SOF and cAVS/SST drivers can operate with this mode set. Our goal is still to have coexistence between the two drivers in a single build/distro, e.g. cAVS for SKL/KBL/APL and SOF for newer platforms. This can be done in a follow-up patch but it needs to be done before distros start selecting this common HDMI mode.
Quoted 7 lines
Quoted 4 lines
The other concern I have is that we have other changes coming for soc-acpi and machine drivers to deal with SoundWire, so it'd be nice to have the changes mirrored between Mark and Takashi trees e.g. with the merge of a fixed branch.
In general it's already a bit too late to merge for 5.4, so the whole series will be after that, at most, hence we still have enough time :)
yes indeed, this was just to let you know we'll need extra work to synchronize between you/Mark/Vinod, not to rush things through.
_______________________________________________
Alsa-devel mailing list
[email protected]
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerHi,
Quoted 2 lines · Takashi Iwai
In general it's already a bit too late to merge for 5.4, so the whole series will be after that, at most, hence we still have enough time :)
actually, Takashi, please hold merging this series, for a few days at least. I got a report about a race between powering up display and the codec probe with this patchset, which I'm debugging now. I'll follow-up to this thread once I have rootcaused this.
Br, Kai
_______________________________________________
Alsa-devel mailing list
[email protected]
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Kai Vehmanen
maintainerHi,
Quoted 2 lines · Pierre-Louis Bossart
The series looks good to me so
Reviewed-by Pierre-Louis Bossart <[email protected]>
thank you Pierre and Takashi for the reviews!
Quoted 6 lines
I would recommend that we have a matching change for the Skylake driver and validation that both SOF and cAVS/SST drivers can operate with this mode set. Our goal is still to have coexistence between the two drivers in a single build/distro, e.g. cAVS for SKL/KBL/APL and SOF for newer platforms. This can be done in a follow-up patch but it needs to be done before distros start selecting this common HDMI mode.
The current patchset actually does allow that. You can select the common HDMI codec in kernel config, build both SOF and SST drivers and based on runtime selection of the platform, either SOF with patch_hdmi.c will be used or SST with hdac-hdmi. This is achieved by setting the common-hdmi-codec mach-params flag in sof/intel/hda.c, so with SST drivers, this will never be set.
To change SST to use patch_hdmi.c as well, is a bigger effort. There are more (and much older) platforms impacted by the alsa mixer name changes. I'm not sure whether this is worth the trouble. But open to ideas here.
Br, Kai
_______________________________________________
Alsa-devel mailing list
[email protected]
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Quoted 23 lines · Kai Vehmanen
Hi,
Quoted 2 lines · Pierre-Louis Bossart
The series looks good to me so
Reviewed-by Pierre-Louis Bossart <[email protected]>thank you Pierre and Takashi for the reviews!
Quoted 6 lines
I would recommend that we have a matching change for the Skylake driver and validation that both SOF and cAVS/SST drivers can operate with this mode set. Our goal is still to have coexistence between the two drivers in a single build/distro, e.g. cAVS for SKL/KBL/APL and SOF for newer platforms. This can be done in a follow-up patch but it needs to be done before distros start selecting this common HDMI mode.
The current patchset actually does allow that. You can select the common HDMI codec in kernel config, build both SOF and SST drivers and based on runtime selection of the platform, either SOF with
patch_hdmi.cwill be used or SST with hdac-hdmi. This is achieved by setting the common-hdmi-codec mach-params flag insof/intel/hda.c, so with SST drivers, this will never be set.
Quoted 4 lines
To change SST to use
patch_hdmi.cas well, is a bigger effort. There are more (and much older) platforms impacted by the alsa mixer name changes. I'm not sure whether this is worth the trouble. But open to ideas here.
If the i915 interface was set in stone yes we could probably leave the Skylake driver alone. But there are changes from time to time and bug fixes that will require Intel to work on two parts of the tree, so really wondering if we shouldn't just move to this common mode across the board and remove hdac_hdmi.c. it doesn't need to be done at once but it's not that crazy to aim for the next kernel release.
_______________________________________________
Alsa-devel mailing list
[email protected]
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel