[PATCH 6.1.y] ALSA: hda: codecs: hdmi: disable keep-alive before audio format change
Sasha Levin <[email protected]>
| Newsgroups | org.kernel.vger.stable |
|---|---|
| Message-ID | <[email protected]> |
From: Kai Vehmanen <[email protected]> [ Upstream commit a3d6d3cedfe87bbd5a677d52b22ac20d28e59cf8 ] When a keep-alive (KAE) silent stream is active on an Intel HDMI/DP codec, opening a real PCM stream reprograms the converter format and the audio infoframe in snd_hda_hdmi_generic_pcm_prepare(). Part of that reprogramming - the converter channel count and the channel mapping in snd_hda_hdmi_setup_audio_infoframe() - is not safe to do while a keep-alive stream is active. This is most visible when switching to a multichannel PCM configuration, where the active channel count actually changes. In that case the newly opened PCM stream plays no sound. Add an optional hdmi_ops .prepare hook, called at the start of the PCM prepare sequence (before the format and infoframe are touched), and implement it for HSW+ to release keep-alive. Keep-alive is then re-enabled as before once the new stream has been set up, in the setup_stream op. Fixes: 15175a4f2bbb ("ALSA: hda/hdmi: add keep-alive support for ADL-P and DG2") Reported-by: Alexander Kaplan <[email protected]> Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/work_items/8412 Tested-by: Alexander Kaplan <[email protected]> Cc: <[email protected]> Signed-off-by: Kai Vehmanen <[email protected]> Link: https://patch.msgid.link/[email protected] Signed-off-by: Takashi Iwai <[email protected]> [ adapted three hunks from the post-6.12 split files (hdmi.c/hdmi_local.h/intelhdmi.c) back into the monolithic sound/pci/hda/patch_hdmi.c, with the prepare hook un-indented one level since 6.12 uses plain mutex_lock() instead of scoped_guard() ] Signed-off-by: Sasha Levin <[email protected]> --- sound/pci/hda/patch_hdmi.c | 48 +++++++++++++++++++++++++++++++++----- 1 file changed, 42 insertions(+), 6 deletions(-) diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index f5712c0aeb504..83ba6dfd28c3b 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -111,6 +111,15 @@ struct hdmi_ops { hda_nid_t pin_nid, int dev_id, u32 stream_tag, int format); + /* + * Optional hook invoked at the beginning of the PCM prepare + * sequence, before the audio infoframe and stream format are + * (re)programmed. Used to disable keep-alive / silent stream so + * that the format change is not done while keep-alive is active. + */ + void (*prepare)(struct hda_codec *codec, + struct hdmi_spec_per_pin *per_pin); + void (*pin_cvt_fixup)(struct hda_codec *codec, struct hdmi_spec_per_pin *per_pin, hda_nid_t cvt_nid); @@ -2138,6 +2147,9 @@ static int generic_hdmi_playback_pcm_prepare(struct hda_pcm_stream *hinfo, per_pin->channels = substream->runtime->channels; per_pin->setup = true; + if (spec->ops.prepare) + spec->ops.prepare(codec, per_pin); + if (get_wcaps(codec, cvt_nid) & AC_WCAP_STRIPE) { stripe = snd_hdac_get_stream_stripe_ctl(&codec->bus->core, substream); @@ -2903,6 +2915,28 @@ static void register_i915_notifier(struct hda_codec *codec) codec->relaxed_resume = 1; } +/* + * prepare ops override for HSW+ + * + * Disable keep-alive before the converter format and audio infoframe are + * reprogrammed by the PCM prepare sequence. Changing the audio format (e.g. + * the channel count when switching to multichannel PCM) while a keep-alive + * stream is active is not safe, so release keep-alive here, early in the + * sequence. It is re-enabled once the new stream has been set up, in + * i915_hsw_setup_stream(). + */ +static void i915_hsw_prepare(struct hda_codec *codec, + struct hdmi_spec_per_pin *per_pin) +{ + struct hdmi_spec *spec = codec->spec; + + if (spec->silent_stream_type == SILENT_STREAM_KAE && per_pin->silent_stream) { + silent_stream_set_kae(codec, per_pin, false); + /* wait for pending transfers in codec to clear */ + usleep_range(100, 200); + } +} + /* setup_stream ops override for HSW+ */ static int i915_hsw_setup_stream(struct hda_codec *codec, hda_nid_t cvt_nid, hda_nid_t pin_nid, int dev_id, u32 stream_tag, @@ -2920,15 +2954,16 @@ static int i915_hsw_setup_stream(struct hda_codec *codec, hda_nid_t cvt_nid, haswell_verify_D0(codec, cvt_nid, pin_nid); - if (spec->silent_stream_type == SILENT_STREAM_KAE && per_pin && per_pin->silent_stream) { - silent_stream_set_kae(codec, per_pin, false); - /* wait for pending transfers in codec to clear */ - usleep_range(100, 200); - } - res = hdmi_setup_stream(codec, cvt_nid, pin_nid, dev_id, stream_tag, format); + /* + * Keep-alive was disabled in i915_hsw_prepare(), re-enable it now. + * The pin lookup above resolves to the same per_pin that prepare + * used (pin_nid comes from that per_pin), so this stays balanced; a + * NULL per_pin only occurs on a lookup failure that also implies no + * active keep-alive stream to restore. + */ if (spec->silent_stream_type == SILENT_STREAM_KAE && per_pin && per_pin->silent_stream) { usleep_range(100, 200); silent_stream_set_kae(codec, per_pin, true); @@ -3104,6 +3139,7 @@ static int intel_hsw_common_init(struct hda_codec *codec, hda_nid_t vendor_nid, codec->depop_delay = 0; codec->auto_runtime_pm = 1; + spec->ops.prepare = i915_hsw_prepare; spec->ops.setup_stream = i915_hsw_setup_stream; spec->ops.pin_cvt_fixup = i915_pin_cvt_fixup; -- 2.53.0