Message ID | 1451524942-17288-5-git-send-email-libin.yang@linux.intel.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Thu, 31 Dec 2015 02:22:22 +0100, libin.yang@linux.intel.com wrote: > > From: Libin Yang <libin.yang@linux.intel.com> > > This patch adds the support for monitor hotplug of dynamic pcm assignment. > > 1. unsol_event enabling > - For codec_has_acomp, unsol_event is disabled. > - For !codec_has_acomp && !dyn_pcm_assign, use the hda_jack helper to > enable unsol_event > - For !codec_has_acomp && dyn_pcm_assign, enable unsol_event with verb > directly > > 2. unsol_event handling > - For !dyn_pcm_assign, use hda_jack helper to report the event > - For dyn_pcm_assign, use snd_jack_report() directly I guess we can reduce lots of codes if ignoring the case with dyn_pcm_assign but without audio component? If so, for simplicity, we can limit dyn_pcm_assign only tied with audio component. Just add a WARN_ON() in such a case. Although the functionality is somehow unrelated, the actual testing and coding is more targeted with Intel chips that are only with audio component. Takashi
Hi Takashi, > -----Original Message----- > From: Takashi Iwai [mailto:tiwai@suse.de] > Sent: Thursday, January 07, 2016 10:18 PM > To: libin.yang@linux.intel.com > Cc: alsa-devel@alsa-project.org; Lin, Mengdong; Yang, Libin > Subject: Re: [alsa-devel] [PATCH 4/4] ALSA: hda - hdmi monitor hotplug > support for dynamic pcm assignment > > On Thu, 31 Dec 2015 02:22:22 +0100, > libin.yang@linux.intel.com wrote: > > > > From: Libin Yang <libin.yang@linux.intel.com> > > > > This patch adds the support for monitor hotplug of dynamic pcm > assignment. > > > > 1. unsol_event enabling > > - For codec_has_acomp, unsol_event is disabled. > > - For !codec_has_acomp && !dyn_pcm_assign, use the hda_jack > helper to > > enable unsol_event > > - For !codec_has_acomp && dyn_pcm_assign, enable unsol_event > with verb > > directly > > > > 2. unsol_event handling > > - For !dyn_pcm_assign, use hda_jack helper to report the event > > - For dyn_pcm_assign, use snd_jack_report() directly > > I guess we can reduce lots of codes if ignoring the case with > dyn_pcm_assign but without audio component? If so, for simplicity, we > can limit dyn_pcm_assign only tied with audio component. Just add a > WARN_ON() in such a case. Although the functionality is somehow > unrelated, the actual testing and coding is more targeted with Intel > chips that are only with audio component. My initial idea is to support: 1. dyn_pcm_assign && acomp: Intel case 2. dyn_pcm_assign && !acomp: other vendors MST case 3. !dyn_pcm_assign && acomp: Intel non-mst case 4. !dyn_pcm_assign && !acomp: other vendors current case Do you mean we can ignore other vendors' case so far? If so, the code will be simpler. > > > Takashi
On Fri, 08 Jan 2016 06:25:37 +0100, Yang, Libin wrote: > > Hi Takashi, > > > -----Original Message----- > > From: Takashi Iwai [mailto:tiwai@suse.de] > > Sent: Thursday, January 07, 2016 10:18 PM > > To: libin.yang@linux.intel.com > > Cc: alsa-devel@alsa-project.org; Lin, Mengdong; Yang, Libin > > Subject: Re: [alsa-devel] [PATCH 4/4] ALSA: hda - hdmi monitor hotplug > > support for dynamic pcm assignment > > > > On Thu, 31 Dec 2015 02:22:22 +0100, > > libin.yang@linux.intel.com wrote: > > > > > > From: Libin Yang <libin.yang@linux.intel.com> > > > > > > This patch adds the support for monitor hotplug of dynamic pcm > > assignment. > > > > > > 1. unsol_event enabling > > > - For codec_has_acomp, unsol_event is disabled. > > > - For !codec_has_acomp && !dyn_pcm_assign, use the hda_jack > > helper to > > > enable unsol_event > > > - For !codec_has_acomp && dyn_pcm_assign, enable unsol_event > > with verb > > > directly > > > > > > 2. unsol_event handling > > > - For !dyn_pcm_assign, use hda_jack helper to report the event > > > - For dyn_pcm_assign, use snd_jack_report() directly > > > > I guess we can reduce lots of codes if ignoring the case with > > dyn_pcm_assign but without audio component? If so, for simplicity, we > > can limit dyn_pcm_assign only tied with audio component. Just add a > > WARN_ON() in such a case. Although the functionality is somehow > > unrelated, the actual testing and coding is more targeted with Intel > > chips that are only with audio component. > > My initial idea is to support: > 1. dyn_pcm_assign && acomp: Intel case > 2. dyn_pcm_assign && !acomp: other vendors MST case > 3. !dyn_pcm_assign && acomp: Intel non-mst case > 4. !dyn_pcm_assign && !acomp: other vendors current case > > Do you mean we can ignore other vendors' case so far? If so, the > code will be simpler. Yes. And give WARN_ON(dyn_pcm_assign && !acomp); Takashi
> -----Original Message----- > From: Takashi Iwai [mailto:tiwai@suse.de] > Sent: Friday, January 08, 2016 3:45 PM > To: Yang, Libin > Cc: libin.yang@linux.intel.com; alsa-devel@alsa-project.org; Lin, > Mengdong > Subject: Re: [alsa-devel] [PATCH 4/4] ALSA: hda - hdmi monitor hotplug > support for dynamic pcm assignment > > On Fri, 08 Jan 2016 06:25:37 +0100, > Yang, Libin wrote: > > > > Hi Takashi, > > > > > -----Original Message----- > > > From: Takashi Iwai [mailto:tiwai@suse.de] > > > Sent: Thursday, January 07, 2016 10:18 PM > > > To: libin.yang@linux.intel.com > > > Cc: alsa-devel@alsa-project.org; Lin, Mengdong; Yang, Libin > > > Subject: Re: [alsa-devel] [PATCH 4/4] ALSA: hda - hdmi monitor > hotplug > > > support for dynamic pcm assignment > > > > > > On Thu, 31 Dec 2015 02:22:22 +0100, > > > libin.yang@linux.intel.com wrote: > > > > > > > > From: Libin Yang <libin.yang@linux.intel.com> > > > > > > > > This patch adds the support for monitor hotplug of dynamic pcm > > > assignment. > > > > > > > > 1. unsol_event enabling > > > > - For codec_has_acomp, unsol_event is disabled. > > > > - For !codec_has_acomp && !dyn_pcm_assign, use the hda_jack > > > helper to > > > > enable unsol_event > > > > - For !codec_has_acomp && dyn_pcm_assign, enable unsol_event > > > with verb > > > > directly > > > > > > > > 2. unsol_event handling > > > > - For !dyn_pcm_assign, use hda_jack helper to report the event > > > > - For dyn_pcm_assign, use snd_jack_report() directly > > > > > > I guess we can reduce lots of codes if ignoring the case with > > > dyn_pcm_assign but without audio component? If so, for simplicity, > we > > > can limit dyn_pcm_assign only tied with audio component. Just add a > > > WARN_ON() in such a case. Although the functionality is somehow > > > unrelated, the actual testing and coding is more targeted with Intel > > > chips that are only with audio component. > > > > My initial idea is to support: > > 1. dyn_pcm_assign && acomp: Intel case > > 2. dyn_pcm_assign && !acomp: other vendors MST case > > 3. !dyn_pcm_assign && acomp: Intel non-mst case > > 4. !dyn_pcm_assign && !acomp: other vendors current case > > > > Do you mean we can ignore other vendors' case so far? If so, the > > code will be simpler. > > Yes. And give WARN_ON(dyn_pcm_assign && !acomp); Get it. I will follow it. Regards, Libin > > > Takashi
diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index 356ae04..293b0e3 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -87,6 +87,7 @@ struct hdmi_spec_per_pin { struct snd_kcontrol *eld_ctl; struct hdmi_pcm *pcm; /* pointer to spec->pcm_rec[n] dynamically*/ int pcm_idx; /* which pcm is attached. -1 means no pcm is attached */ + int unsol_tag; int repoll_count; bool setup; /* the stream has been set up by prepare callback */ int channels; /* current number of channels */ @@ -1229,23 +1230,44 @@ static void jack_callback(struct hda_codec *codec, check_presence_and_report(codec, jack->tbl->nid); } +static hda_nid_t hdmi_get_nid_from_tag(struct hda_codec *codec, int tag) +{ + int pin_idx; + struct hdmi_spec *spec = codec->spec; + + for (pin_idx = 0; pin_idx < spec->num_pins; pin_idx++) { + struct hdmi_spec_per_pin *per_pin = get_pin(spec, pin_idx); + + if (per_pin->unsol_tag == tag) + return per_pin->pin_nid; + } + /* this will never happen, tag is checked before */ + return 0; +} + static void hdmi_intrinsic_event(struct hda_codec *codec, unsigned int res) { int tag = res >> AC_UNSOL_RES_TAG_SHIFT; struct hda_jack_tbl *jack; + struct hdmi_spec *spec = codec->spec; int dev_entry = (res & AC_UNSOL_RES_DE) >> AC_UNSOL_RES_DE_SHIFT; + hda_nid_t nid; - jack = snd_hda_jack_tbl_get_from_tag(codec, tag); - if (!jack) - return; - jack->jack_dirty = 1; + if (!spec->dyn_pcm_assign) { + jack = snd_hda_jack_tbl_get_from_tag(codec, tag); + if (!jack) + return; + jack->jack_dirty = 1; + nid = jack->nid; + } else + nid = hdmi_get_nid_from_tag(codec, tag); codec_dbg(codec, "HDMI hot plug event: Codec=%d Pin=%d Device=%d Inactive=%d Presence_Detect=%d ELD_Valid=%d\n", - codec->addr, jack->nid, dev_entry, !!(res & AC_UNSOL_RES_IA), + codec->addr, nid, dev_entry, !!(res & AC_UNSOL_RES_IA), !!(res & AC_UNSOL_RES_PD), !!(res & AC_UNSOL_RES_ELDV)); - check_presence_and_report(codec, jack->nid); + check_presence_and_report(codec, nid); } static void hdmi_non_intrinsic_event(struct hda_codec *codec, unsigned int res) @@ -1270,13 +1292,30 @@ static void hdmi_non_intrinsic_event(struct hda_codec *codec, unsigned int res) ; } +static bool hdmi_unsol_tag_is_valid(struct hda_codec *codec, int tag) +{ + struct hdmi_spec *spec = codec->spec; + int i; + + if (!spec->dyn_pcm_assign) { + if (snd_hda_jack_tbl_get_from_tag(codec, tag)) + return true; + } else { + for (i = 0; i < spec->num_pins; i++) { + if (get_pin(spec, i)->unsol_tag == tag) + return true; + } + } + + return false; +} static void hdmi_unsol_event(struct hda_codec *codec, unsigned int res) { int tag = res >> AC_UNSOL_RES_TAG_SHIFT; int subtag = (res & AC_UNSOL_RES_SUBTAG) >> AC_UNSOL_RES_SUBTAG_SHIFT; - if (!snd_hda_jack_tbl_get_from_tag(codec, tag)) { + if (!hdmi_unsol_tag_is_valid(codec, tag)) { codec_dbg(codec, "Unexpected HDMI event tag 0x%x\n", tag); return; } @@ -1875,7 +1914,8 @@ static void update_eld(struct hda_codec *codec, static bool hdmi_present_sense_via_verbs(struct hdmi_spec_per_pin *per_pin, int repoll) { - struct hda_jack_tbl *jack; + struct hda_jack_tbl *jack = NULL; + struct snd_jack *sjack = NULL; struct hda_codec *codec = per_pin->codec; struct hdmi_spec *spec = codec->spec; struct hdmi_eld *eld = &spec->temp_eld; @@ -1890,7 +1930,7 @@ static bool hdmi_present_sense_via_verbs(struct hdmi_spec_per_pin *per_pin, * the unsolicited response to avoid custom WARs. */ int present; - bool ret; + bool ret = false; bool do_repoll = false; snd_hda_power_up_pm(codec); @@ -1920,14 +1960,26 @@ static bool hdmi_present_sense_via_verbs(struct hdmi_spec_per_pin *per_pin, do_repoll = true; } + if (spec->dyn_pcm_assign && per_pin->pcm_idx >= 0) { + sjack = spec->pcm_rec[per_pin->pcm_idx].jack; + snd_jack_report(sjack, + pin_eld->monitor_present ? SND_JACK_AVOUT : 0); + } + if (do_repoll) schedule_delayed_work(&per_pin->work, msecs_to_jiffies(300)); else update_eld(codec, per_pin, eld); - ret = !repoll || !pin_eld->monitor_present || pin_eld->eld_valid; - - jack = snd_hda_jack_tbl_get(codec, pin_nid); + if (!spec->dyn_pcm_assign) { + ret = !repoll || !pin_eld->monitor_present || + pin_eld->eld_valid; + jack = snd_hda_jack_tbl_get(codec, pin_nid); + } else if (per_pin->pcm_idx >= 0 && !sjack) { + sjack = spec->pcm_rec[per_pin->pcm_idx].jack; + snd_jack_report(sjack, + pin_eld->monitor_present ? SND_JACK_AVOUT : 0); + } if (jack) jack->block_report = !ret; @@ -2511,7 +2563,7 @@ static int generic_hdmi_build_jack(struct hda_codec *codec, int pcm_idx) { char hdmi_str[32] = "HDMI/DP"; struct hdmi_spec *spec = codec->spec; - struct hdmi_spec_per_pin *per_pin; + struct hdmi_spec_per_pin *per_pin = get_pin(spec, pcm_idx); struct hda_jack_tbl *jack; int pcmdev = get_pcm_rec(spec, pcm_idx).pcm->device; bool phantom_jack; @@ -2526,7 +2578,7 @@ static int generic_hdmi_build_jack(struct hda_codec *codec, int pcm_idx) * This means pcms and pins are statically mapped. * And pcm_idx is pin_idx. */ - per_pin = get_pin(spec, pcm_idx); + phantom_jack = !is_jack_detectable(codec, per_pin->pin_nid); if (phantom_jack) strncat(hdmi_str, " Phantom", @@ -2540,10 +2592,17 @@ static int generic_hdmi_build_jack(struct hda_codec *codec, int pcm_idx) return 0; /* statically bind jack */ spec->pcm_rec[pcm_idx].jack = jack->jack; + per_pin->unsol_tag = jack->tag; return 0; } - return add_hdmi_jack_kctl(codec, spec, pcm_idx, hdmi_str); + per_pin->unsol_tag = per_pin->pin_nid_idx + 1; + ret = add_hdmi_jack_kctl(codec, spec, pcm_idx, hdmi_str); + if (!codec_has_acomp(codec)) + snd_hda_codec_write_cache(codec, per_pin->pin_nid, 0, + AC_VERB_SET_UNSOLICITED_ENABLE, + AC_USRSP_EN | per_pin->unsol_tag); + return ret; } static int generic_hdmi_build_controls(struct hda_codec *codec) @@ -2636,7 +2695,13 @@ static int generic_hdmi_init(struct hda_codec *codec) hda_nid_t pin_nid = per_pin->pin_nid; hdmi_init_pin(codec, pin_nid); - if (!codec_has_acomp(codec)) + /* if codec_has_acomp, will not enable unsol event + * if !codec_has_acomp && ! dyn_pcm_assign, + * will use hda_jack + * if !codec_has_acomp && dyn_pcm_assign, + * enable unsol_event when building jack + */ + if (!codec_has_acomp(codec) && !spec->dyn_pcm_assign) snd_hda_jack_detect_enable_callback(codec, pin_nid, codec->jackpoll_interval > 0 ? jack_callback : NULL);