January 2021 - Alsa-devel - mailman.alsa-project.org

Re: Haswell audio no longer working with new Catpt driver
by Lars-Peter Clausen 02 Jan '21

02 Jan '21

Hi, I'm taking this conversation back to the mailinglist. Hope you don't mind. On 12/31/20 2:15 PM, Christian Labisch wrote: > Hi Amadeusz, Greg, and Lars-Peter, > > Do you have a suggestion for a workaround ? > A special boot parameter that I could add ? > Will the drivers work in the next version ? I had a quick look at the changes to the HDA driver between v5.9 and v5.10. And there isn't that much, expect some changes to power management. Curiously enough pretty much your only difference between your v5.9 and v5.10 alsa-info.sh is that a few of the nodes are in D3 state (meaning powered down) in v5.10. One of them is the internal speaker. But this could also be a coincidence, would be interesting to see what the output of alsa-info.sh looks like when the you are trying to play some audio. If it is still powered down then there is a bug. - Lars > > Regards, > Christian > > >> On Thu, 2020-12-31 at 12:20 +0100, Amadeusz Sławiński wrote: >>> On 12/31/2020 11:50 AM, Christian Labisch wrote: >>>> Hi Lars-Peter, >>>> >>>> Thank you, please find attached the requested information from both kernels. >>>> I freshly installed the fedora kernel 5.10.4 to give you the latest results. >>>> >>>> Regards, >>>> Christian >>>> >>>> Christian Labisch >>>> Red Hat Accelerator >>>> clnetbox.blogspot.com >>>> access.redhat.com/community >>>> access.redhat.com/accelerators >>>> >>>> >>>> On Thu, 2020-12-31 at 11:04 +0100, Lars-Peter Clausen wrote: >>>>> On 12/31/20 9:33 AM, Greg Kroah-Hartman wrote: >>>>>> On Wed, Dec 30, 2020 at 07:10:16PM +0100, Christian Labisch wrote: >>>>>>> Update : >>>>>>> >>>>>>> I've just tested the kernel 5.10.4 from ELRepo. >>>>>>> Unfortunately nothing changed - still no sound. >>>>>> Ah, sad. Can you run 'git bisect' between 5.9 and 5.10 to determine the >>>>>> commit that caused the problem? >>>>> The problem is that one driver was replaced with another driver. git >>>>> bisect wont really help to narrow down why the new driver does not work. >>>>> >>>>> Christian can you run the alsa-info.sh[1] script on your system and send >>>>> back the result? >>>>> >>>>> You say sound is not working, can you clarify that a bit? Is no sound >>>>> card registered? Is it registered but output stays silent? >>>>> >>>>> - Lars >>>>> >>>>> [1] https://www.alsa-project.org/wiki/AlsaInfo >>>>> <https://www.alsa-project.org/wiki/AlsaInfo> >>>>> >>>>> >>> Hi, >>> >>> from reading provided files it seems that you use snd_intel_hda driver, >>> it should be possible to git bisect it between 5.9 and 5.10 as it wasn't >>> replaced. >>> >>> Catpt driver is used on machines using DSP. >>> >>> Amadeusz

4 8

[PATCH v2] ALSA: hda/hdmi: fix silent stream for first playback to DP
by Kai Vehmanen 02 Jan '21

02 Jan '21

A problem exists in enabling silent stream when connection type is DisplayPort. Silent stream programming is completed when a new DP receiver is connected, but infoframe transmission does not actually start until PCM is opened for the first time. This can result in audible gap of multiple seconds. This only affects the first PCM open. Fix the issue by properly assigning a converter to the silent stream, and modifying the required stream ID programming sequence. This change only affects Intel display audio codecs. BugLink: https://github.com/thesofproject/linux/issues/2468 Fixes: 951894cf30f4 ("ALSA: hda/hdmi: Add Intel silent stream support") Signed-off-by: Kai Vehmanen <kai.vehmanen(a)linux.intel.com> --- sound/pci/hda/patch_hdmi.c | 98 +++++++++++++++++++++++++++++++++----- 1 file changed, 86 insertions(+), 12 deletions(-) v2 changes: - locking missing in silent_stream_disable() diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index b0068f8ca46d..2ddc27db8c01 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -78,6 +78,7 @@ struct hdmi_spec_per_pin { int pcm_idx; /* which pcm is attached. -1 means no pcm is attached */ int repoll_count; bool setup; /* the stream has been set up by prepare callback */ + bool silent_stream; int channels; /* current number of channels */ bool non_pcm; bool chmap_set; /* channel-map override by ALSA API? */ @@ -979,6 +980,13 @@ static int hdmi_choose_cvt(struct hda_codec *codec, else per_pin = get_pin(spec, pin_idx); + if (per_pin && per_pin->silent_stream) { + cvt_idx = cvt_nid_to_cvt_index(codec, per_pin->cvt_nid); + if (cvt_id) + *cvt_id = cvt_idx; + return 0; + } + /* Dynamically assign converter to stream */ for (cvt_idx = 0; cvt_idx < spec->num_cvts; cvt_idx++) { per_cvt = get_cvt(spec, cvt_idx); @@ -1642,30 +1650,95 @@ static void hdmi_present_sense_via_verbs(struct hdmi_spec_per_pin *per_pin, snd_hda_power_down_pm(codec); } +#define I915_SILENT_RATE 48000 +#define I915_SILENT_CHANNELS 2 +#define I915_SILENT_FORMAT SNDRV_PCM_FORMAT_S16_LE +#define I915_SILENT_FORMAT_BITS 16 +#define I915_SILENT_FMT_MASK 0xf + static void silent_stream_enable(struct hda_codec *codec, - struct hdmi_spec_per_pin *per_pin) + struct hdmi_spec_per_pin *per_pin) { - unsigned int newval, oldval; - - codec_dbg(codec, "hdmi: enabling silent stream for NID %d\n", - per_pin->pin_nid); + struct hdmi_spec *spec = codec->spec; + struct hdmi_spec_per_cvt *per_cvt; + int cvt_idx, pin_idx, err; + unsigned int format; mutex_lock(&per_pin->lock); - if (!per_pin->channels) - per_pin->channels = 2; + if (per_pin->setup) { + codec_dbg(codec, "hdmi: PCM already open, no silent stream\n"); + goto unlock_out; + } - oldval = snd_hda_codec_read(codec, per_pin->pin_nid, 0, - AC_VERB_GET_CONV, 0); - newval = (oldval & 0xF0) | 0xF; - snd_hda_codec_write(codec, per_pin->pin_nid, 0, - AC_VERB_SET_CHANNEL_STREAMID, newval); + pin_idx = pin_id_to_pin_index(codec, per_pin->pin_nid, per_pin->dev_id); + err = hdmi_choose_cvt(codec, pin_idx, &cvt_idx); + if (err) { + codec_err(codec, "hdmi: no free converter to enable silent mode\n"); + goto unlock_out; + } + + per_cvt = get_cvt(spec, cvt_idx); + per_cvt->assigned = 1; + per_pin->cvt_nid = per_cvt->cvt_nid; + per_pin->silent_stream = true; + codec_dbg(codec, "hdmi: enabling silent stream pin-NID=0x%x cvt-NID=0x%x\n", + per_pin->pin_nid, per_cvt->cvt_nid); + + snd_hda_set_dev_select(codec, per_pin->pin_nid, per_pin->dev_id); + snd_hda_codec_write_cache(codec, per_pin->pin_nid, 0, + AC_VERB_SET_CONNECT_SEL, + per_pin->mux_idx); + + /* configure unused pins to choose other converters */ + pin_cvt_fixup(codec, per_pin, 0); + + snd_hdac_sync_audio_rate(&codec->core, per_pin->pin_nid, + per_pin->dev_id, I915_SILENT_RATE); + + /* trigger silent stream generation in hw */ + format = snd_hdac_calc_stream_format(I915_SILENT_RATE, I915_SILENT_CHANNELS, + I915_SILENT_FORMAT, I915_SILENT_FORMAT_BITS, 0); + snd_hda_codec_setup_stream(codec, per_pin->cvt_nid, + I915_SILENT_FMT_MASK, I915_SILENT_FMT_MASK, format); + usleep_range(100, 200); + snd_hda_codec_setup_stream(codec, per_pin->cvt_nid, I915_SILENT_FMT_MASK, 0, format); + + per_pin->channels = I915_SILENT_CHANNELS; hdmi_setup_audio_infoframe(codec, per_pin, per_pin->non_pcm); + unlock_out: mutex_unlock(&per_pin->lock); } +static void silent_stream_disable(struct hda_codec *codec, + struct hdmi_spec_per_pin *per_pin) +{ + struct hdmi_spec *spec = codec->spec; + struct hdmi_spec_per_cvt *per_cvt; + int cvt_idx; + + mutex_lock(&per_pin->lock); + if (!per_pin->silent_stream) + goto unlock_out; + + codec_dbg(codec, "HDMI: disable silent stream on pin-NID=0x%x cvt-NID=0x%x\n", + per_pin->pin_nid, per_pin->cvt_nid); + + cvt_idx = cvt_nid_to_cvt_index(codec, per_pin->cvt_nid); + if (cvt_idx >= 0 && cvt_idx < spec->num_cvts) { + per_cvt = get_cvt(spec, cvt_idx); + per_cvt->assigned = 0; + } + + per_pin->cvt_nid = 0; + per_pin->silent_stream = false; + + unlock_out: + mutex_unlock(&spec->pcm_lock); +} + /* update ELD and jack state via audio component */ static void sync_eld_via_acomp(struct hda_codec *codec, struct hdmi_spec_per_pin *per_pin) @@ -1701,6 +1774,7 @@ static void sync_eld_via_acomp(struct hda_codec *codec, pm_ret); silent_stream_enable(codec, per_pin); } else if (monitor_prev && !monitor_next) { + silent_stream_disable(codec, per_pin); pm_ret = snd_hda_power_down_pm(codec); if (pm_ret < 0) codec_err(codec, -- 2.29.2

4 4

[PATCH] ALSA: hda/hdmi: Fix incorrect mutex unlock in silent_stream_disable()
by Takashi Iwai 01 Jan '21

01 Jan '21

The silent_stream_disable() function introduced by the commit b1a5039759cb ("ALSA: hda/hdmi: fix silent stream for first playback to DP") takes the per_pin->lock mutex, but it unlocks the wrong one, spec->pcm_lock, which causes a deadlock. This patch corrects it. Fixes: b1a5039759cb ("ALSA: hda/hdmi: fix silent stream for first playback to DP") Reported-by: Jan Alexander Steffens (heftig) <heftig(a)archlinux.org> Cc: <stable(a)vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- sound/pci/hda/patch_hdmi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index 1e4a4b83fbf6..74d246a0dc6d 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -1733,7 +1733,7 @@ static void silent_stream_disable(struct hda_codec *codec, per_pin->silent_stream = false; unlock_out: - mutex_unlock(&spec->pcm_lock); + mutex_unlock(&per_pin->lock); } /* update ELD and jack state via audio component */ -- 2.26.2

2 1