At Thu, 21 Mar 2013 11:57:08 +0100,
David Henningsson wrote:
>
> On 03/21/2013 11:47 AM, David Henningsson wrote:
> > On 03/21/2013 11:44 AM, Takashi Iwai wrote:
> >> At Thu, 21 Mar 2013 11:32:57 +0100,
> >> David Henningsson wrote:
> >>>
> >>> Headset mic jacks, i e TRRS style jacks with Headphone Left,
> >>> Headphone Right, Mic and GND signals, are becoming increasingly
> >>> common and are now being shipped by several manufacturers.
> >>>
> >>> Unfortunately, the HDA specification does not give us any hint
> >>> of whether a Mic pin belongs to such a jack or not, but it would
> >>> still be helpful for the user to know (especially if there is one
> >>> TRS Mic jack and one TRRS headset jack).
> >>>
> >>> This new fixup causes the first (non-dock, non-internal) mic to
> >>> be a headset mic jack. The algorithm can be later refined if needed.
> >>>
> >>> Signed-off-by: David Henningsson <david.henningsson@xxxxxxxxxxxxx>
> >>> ---
> >>> sound/pci/hda/hda_auto_parser.c | 24 ++++++++++++++++++++++--
> >>> sound/pci/hda/hda_auto_parser.h | 2 ++
> >>> sound/pci/hda/patch_sigmatel.c | 6 +++++-
> >>> 3 files changed, 29 insertions(+), 3 deletions(-)
> >>>
> >>> diff --git a/sound/pci/hda/hda_auto_parser.c
> >>> b/sound/pci/hda/hda_auto_parser.c
> >>> index a3ea76a..6b173b3 100644
> >>> --- a/sound/pci/hda/hda_auto_parser.c
> >>> +++ b/sound/pci/hda/hda_auto_parser.c
> >>> @@ -260,6 +260,22 @@ int snd_hda_parse_pin_defcfg(struct hda_codec
> >>> *codec,
> >>> }
> >>> }
> >>>
> >>> + /* Take first mic to be a headset mic pin */
> >>> + if (cond_flags & HDA_PINCFG_HEADSET_MIC) {
> >>> + for (i = 0; i < cfg->num_inputs; i++) {
> >>> + int attr;
> >>> + unsigned int def_conf;
> >>> + if (cfg->inputs[i].type != AUTO_PIN_MIC)
> >>> + continue;
> >>> + def_conf = snd_hda_codec_get_pincfg(codec,
> >>> cfg->inputs[i].pin);
> >>> + attr = snd_hda_get_input_pin_attr(def_conf);
> >>> + if (attr <= INPUT_PIN_ATTR_DOCK)
> >>> + continue;
> >>> + cfg->inputs[i].is_headset_mic = 1;
> >>> + break;
> >>> + }
> >>> + }
> >>> +
> >>> /* FIX-UP:
> >>> * If no line-out is defined but multiple HPs are found,
> >>> * some of them might be the real line-outs.
> >>> @@ -388,6 +404,7 @@ EXPORT_SYMBOL_HDA(snd_hda_get_input_pin_attr);
> >>> */
> >>>
> >>> static const char *hda_get_input_pin_label(struct hda_codec *codec,
> >>> + const struct auto_pin_cfg_item *item,
> >>> hda_nid_t pin, bool check_location)
> >>> {
> >>> unsigned int def_conf;
> >>> @@ -400,6 +417,8 @@ static const char *hda_get_input_pin_label(struct
> >>> hda_codec *codec,
> >>>
> >>> switch (get_defcfg_device(def_conf)) {
> >>> case AC_JACK_MIC_IN:
> >>> + if (item && item->is_headset_mic)
> >>> + return "Headset Mic";
> >>> if (!check_location)
> >>> return "Mic";
> >>> attr = snd_hda_get_input_pin_attr(def_conf);
> >>> @@ -480,7 +499,8 @@ const char *hda_get_autocfg_input_label(struct
> >>> hda_codec *codec,
> >>> has_multiple_pins = 1;
> >>> if (has_multiple_pins && type == AUTO_PIN_MIC)
> >>> has_multiple_pins &= check_mic_location_need(codec, cfg,
> >>> input);
> >>> - return hda_get_input_pin_label(codec, cfg->inputs[input].pin,
> >>> + return hda_get_input_pin_label(codec, &cfg->inputs[input],
> >>> + cfg->inputs[input].pin,
> >>> has_multiple_pins);
> >>> }
> >>> EXPORT_SYMBOL_HDA(hda_get_autocfg_input_label);
> >>> @@ -649,7 +669,7 @@ int snd_hda_get_pin_label(struct hda_codec
> >>> *codec, hda_nid_t nid,
> >>> }
> >>> }
> >>> if (!name)
> >>> - name = hda_get_input_pin_label(codec, nid, true);
> >>> + name = hda_get_input_pin_label(codec, NULL, nid, true);
> >>> break;
> >>> }
> >>> if (!name)
> >>> diff --git a/sound/pci/hda/hda_auto_parser.h
> >>> b/sound/pci/hda/hda_auto_parser.h
> >>> index f748071..98a72d0 100644
> >>> --- a/sound/pci/hda/hda_auto_parser.h
> >>> +++ b/sound/pci/hda/hda_auto_parser.h
> >>> @@ -36,6 +36,7 @@ enum {
> >>> struct auto_pin_cfg_item {
> >>> hda_nid_t pin;
> >>> int type;
> >>> + int is_headset_mic:1;
> >>
> >> Use unsigned int for bit-fields. Otherwise you'll hit a mess of
> >> signedness.
> >
> > Ok, will fix.
> >
> >>
> >>
> >>> };
> >>>
> >>> struct auto_pin_cfg;
> >>> @@ -80,6 +81,7 @@ struct auto_pin_cfg {
> >>> /* bit-flags for snd_hda_parse_pin_def_config() behavior */
> >>> #define HDA_PINCFG_NO_HP_FIXUP (1 << 0) /* no HP-split */
> >>> #define HDA_PINCFG_NO_LO_FIXUP (1 << 1) /* don't take other outs
> >>> as LO */
> >>> +#define HDA_PINCFG_HEADSET_MIC (1 << 2) /* Take first mic as
> >>> headset mic */
> >>>
> >>> int snd_hda_parse_pin_defcfg(struct hda_codec *codec,
> >>> struct auto_pin_cfg *cfg,
> >>> diff --git a/sound/pci/hda/patch_sigmatel.c
> >>> b/sound/pci/hda/patch_sigmatel.c
> >>> index 3be877b..1d9d642 100644
> >>> --- a/sound/pci/hda/patch_sigmatel.c
> >>> +++ b/sound/pci/hda/patch_sigmatel.c
> >>> @@ -3528,8 +3528,12 @@ static int stac_parse_auto_config(struct
> >>> hda_codec *codec)
> >>> {
> >>> struct sigmatel_spec *spec = codec->spec;
> >>> int err;
> >>> + int flags = 0;
> >>>
> >>> - err = snd_hda_parse_pin_defcfg(codec, &spec->gen.autocfg, NULL, 0);
> >>> + if (spec->headset_jack)
> >>> + flags |= HDA_PINCFG_HEADSET_MIC;
> >>> +
> >>> + err = snd_hda_parse_pin_defcfg(codec, &spec->gen.autocfg, NULL,
> >>> flags);
> >>> if (err < 0)
> >>> return err;
> >>
> >>
> >> The definition of spec->headset_jack is missing?
> >
> > It was added in sigmatel_spec in Oct 2012. If you prefer, I could send
> > the changes to patch_sigmatel in a separate patch?
>
> Btw; this change for patch_sigmatel is just a first example. I intend to
> add more headset mics later, and this fixup also makes it easier for
> others to submit such patches.
>
> But if I get this name into the kernel now I can also add it to
> PulseAudio before the 4.0 freeze, which is quite soon.
It's rather a new feature, so I'll push it to 3.10.
Takashi
_______________________________________________
Alsa-devel mailing list
Alsa-devel@xxxxxxxxxxxxxxxx
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
