Thank you for your contribution! Sashiko AI review found 5 potential issues to consider: - [Medium] drm/bridge: dw-dp: unselected audio clock disabled before mux switch - [Medium] drm/bridge: dw-dp: unadvertised sample_width limit breaks 32-bit I2S formats - [Medium] drm/bridge: dw-dp: checking params->cea.channels breaks non-PCM audio - [Medium] drm/bridge: dw-dp: missing drm_connector_hdmi_audio_plugged_notify() calls - [Low] drm/bridge: dw-dp: invalid hdmi_audio_dai_port in dw_dp_probe() --
commit 20fac3cf1ff3b933dbf4d3b4b4b914b2c8a55eb1 Author: Sebastian Reichel <[email protected]> drm/bridge: synopsys: dw-dp: Add audio support This commit adds DisplayPort audio support to the Synopsys DesignWare DisplayPort bridge driver using the DRM bridge audio callbacks. > diff --git a/drivers/gpu/drm/bridge/synopsys/dw-dp.c > b/drivers/gpu/drm/bridge/synopsys/dw-dp.c > index 3c10cdb9ca28b..bf3a53463394f 100644 > --- a/drivers/gpu/drm/bridge/synopsys/dw-dp.c > +++ b/drivers/gpu/drm/bridge/synopsys/dw-dp.c [ ... ] > @@ -1696,6 +1727,261 @@ static int dw_dp_link_enable(struct dw_dp *dp) > return ret; > } [ ... ] > +static int __dw_dp_audio_enable(struct dw_dp *dp) > +{ > + u8 audio_data_in_en; > + > + switch (dp->audio_channels) { > + case 1: > + case 2: > + audio_data_in_en = AUDIO_DATA_IN_EN_CHANNEL12; > + break; > + case 8: > + audio_data_in_en = AUDIO_DATA_IN_EN_CHANNEL12 | > + AUDIO_DATA_IN_EN_CHANNEL34 | > + AUDIO_DATA_IN_EN_CHANNEL56 | > + AUDIO_DATA_IN_EN_CHANNEL78; > + break; > + default: > + return -EINVAL; > + } > + > + regmap_update_bits(dp->regmap, DW_DP_AUD_CONFIG1, > + AUDIO_DATA_IN_EN | NUM_CHANNELS | AUDIO_DATA_WIDTH | > + AUDIO_INF_SELECT | HBR_MODE_ENABLE | AUDIO_MUTE, > + FIELD_PREP(AUDIO_DATA_IN_EN, audio_data_in_en) | > + FIELD_PREP(NUM_CHANNELS, dp->audio_channels - 1) | > + FIELD_PREP(AUDIO_DATA_WIDTH, dp->audio_sample_width) > | > + FIELD_PREP(AUDIO_INF_SELECT, dp->audio_interface) | > + FIELD_PREP(HBR_MODE_ENABLE, 0) | > + FIELD_PREP(AUDIO_MUTE, dp->audio_muted)); > + > + /* Wait for inf switch */ > + usleep_range(20, 40); [ ... ] > +static int dw_dp_audio_prepare(struct drm_bridge *bridge, > + struct drm_connector *connector, > + struct hdmi_codec_daifmt *daifmt, > + struct hdmi_codec_params *params) > +{ > + struct dw_dp *dp = bridge_to_dp(bridge); > + u8 supported_audio_interfaces; > + enum dw_dp_audio_interface audio_interface; > + u32 cfg1; > + int ret; > + > + guard(mutex)(&dp->audio_lock); > + > + /* > + * prepare might be called multiple times, so release the clocks > + * from previous calls to keep the calls in balance. > + */ > + if (dp->audio_interface != DW_DP_AUDIO_UNUSED) > + __dw_dp_audio_disable(dp); > + > + /* The hardware is limited to 1,2 or 8 channels */ > + switch (params->cea.channels) { > + case 1: > + case 2: > + case 8: > + break; > + default: > + dev_err(dp->dev, "invalid audio channels %d\n", > params->cea.channels); > + return -EINVAL; > + } [Severity: Medium] Should dw_dp_audio_prepare() check and store params->channels instead of params->cea.channels? In hdmi_codec_fill_codec_params(), params->cea.channels is set to 0 when IEC958_AES0_NONAUDIO is set in the IEC958 status bits (meaning "refer to stream header" in the CTA-861 Audio InfoFrame), while params->channels holds the actual stream channel count. Using params->cea.channels causes dw_dp_audio_prepare() to reject non-PCM streams with -EINVAL, and would also program NUM_CHANNELS and AUDIO_DATA_IN_EN in DW_DP_AUD_CONFIG1 with the InfoFrame field rather than the stream channel count. > + > + if (params->sample_width < 16 || params->sample_width > 24) { > + dev_err(dp->dev, "invalid data sample width %d\n", > params->sample_width); > + return -EINVAL; > + } [Severity: Medium] Does rejecting sample_width > 24 here break 32-bit formats advertised on the I2S DAI? In dw_dp_probe(), bridge->hdmi_audio_i2s_formats is left unset, so hdmi_codec_probe() falls back to its default I2S_FORMATS mask, which includes SNDRV_PCM_FMTBIT_S32_LE and SNDRV_PCM_FMTBIT_IEC958_SUBFRAME_LE (used by ALSA's iec958 plugin to insert PCUV bits in I2S mode). hdmi_codec_prepare() sets params->sample_width to snd_pcm_format_width(runtime->format), which is 32 for both formats, causing dw_dp_audio_prepare() to fail with -EINVAL after hw_params succeeds. Should bridge->hdmi_audio_i2s_formats be restricted to 16-bit through 24-bit formats? > + > + switch (daifmt->fmt) { > + case HDMI_SPDIF: > + audio_interface = DW_DP_AUDIO_SPDIF; > + break; > + case HDMI_I2S: > + /* > + * It is recommended to use SPDIF instead of I2S, since I2S > mode requires > + * manually inserting PCUV control bits from userspace and this > is done > + * automatically in hardware for SPDIF mode. > + */ > + audio_interface = DW_DP_AUDIO_I2S; > + break; > + default: > + dev_err(dp->dev, "invalid DAI format %d\n", daifmt->fmt); > + return -EINVAL; > + } > + > + regmap_read(dp->regmap, DW_DP_CONFIG_REG1, &cfg1); > + supported_audio_interfaces = FIELD_GET(AUDIO_SELECT, cfg1); > + > + if (supported_audio_interfaces != DW_DP_AUDIO_I2S_AND_SPDIF && > + supported_audio_interfaces != audio_interface) { > + dev_err(dp->dev, "unsupported DAI %d\n", daifmt->fmt); > + return -EINVAL; > + } > + > + ret = clk_prepare_enable(dp->spdif_clk); > + if (ret) > + return ret; > + > + ret = clk_prepare_enable(dp->i2s_clk); > + if (ret) { > + clk_disable_unprepare(dp->spdif_clk); > + return ret; > + } > + > + if (audio_interface == DW_DP_AUDIO_I2S) > + clk_disable_unprepare(dp->spdif_clk); > + else if (audio_interface == DW_DP_AUDIO_SPDIF) > + clk_disable_unprepare(dp->i2s_clk); > + > + dp->audio_channels = params->cea.channels; > + dp->audio_channel_allocation = params->cea.channel_allocation; > + dp->audio_sample_width = params->sample_width; > + dp->audio_interface = audio_interface; > + > + ret = __dw_dp_audio_enable(dp); [Severity: Medium] Does disabling the unselected audio clock before calling __dw_dp_audio_enable() defeat keeping both clocks running across the AUDIO_INF_SELECT mux switch? Both dp->spdif_clk and dp->i2s_clk are enabled here and the unselected clock is immediately disabled before any register access occurs. The actual AUDIO_INF_SELECT update in DW_DP_AUD_CONFIG1 and the 20-40 us interface switch delay happen inside __dw_dp_audio_enable(), which is also called from dw_dp_bridge_atomic_enable() after dw_dp_reset() resets AUDIO_INF_SELECT to 0 (I2S). If the hardware mux requires both clocks to be active while switching AUDIO_INF_SELECT, both clocks should be enabled around the DW_DP_AUD_CONFIG1 update and usleep_range(20, 40) inside __dw_dp_audio_enable(). [ ... ] > @@ -1734,12 +2020,22 @@ static void dw_dp_bridge_atomic_enable(struct > drm_bridge *bridge, > dev_err(dp->dev, "failed to enable video: %d\n", ret); > return; > } > + > + scoped_guard(mutex, &dp->audio_lock) { > + if (dp->audio_interface != DW_DP_AUDIO_UNUSED) { > + ret = __dw_dp_audio_enable(dp); > + if (ret < 0) > + dev_err(dp->dev, "failed to restore audio: > %d\n", ret); > + } > + } > } [Severity: Medium] Should the driver call drm_connector_hdmi_audio_plugged_notify() when the DisplayPort link is enabled, disabled, or hotplugged? The patch includes drm/display/drm_hdmi_audio_helper.h and sets DRM_BRIDGE_OP_DP_AUDIO, but never calls drm_connector_hdmi_audio_plugged_notify(). Because drm_bridge_connector only calls drm_atomic_helper_connector_hdmi_hotplug() automatically when DRM_BRIDGE_OP_HDMI is set, hdmi-codec's plugged_cb() is never invoked for this bridge. As a result, ASoC jack detection (SND_JACK_AVOUT) is never signaled and the ALSA ELD control is not updated on hotplug before a PCM stream is opened. [ ... ] > @@ -2368,12 +2673,21 @@ int dw_dp_probe(struct dw_dp *dp) > return ret; > > dp->vsc_sdp_nr = -1; > + dp->audio_interface = DW_DP_AUDIO_UNUSED; > + dp->audio_sdp_nr = -1; > > bridge = &dp->bridge; > bridge->of_node = dev->of_node; > - bridge->ops = DRM_BRIDGE_OP_DETECT | DRM_BRIDGE_OP_EDID | > DRM_BRIDGE_OP_HPD; > + bridge->ops = DRM_BRIDGE_OP_DP_AUDIO | > + DRM_BRIDGE_OP_DETECT | > + DRM_BRIDGE_OP_EDID | > + DRM_BRIDGE_OP_HPD; > bridge->type = DRM_MODE_CONNECTOR_DisplayPort; > bridge->ycbcr_420_allowed = true; > + bridge->hdmi_audio_dev = dev; > + bridge->hdmi_audio_max_i2s_playback_channels = 8; > + bridge->hdmi_audio_dai_port = 1; > + bridge->hdmi_audio_spdif_playback = true; [Severity: Low] Should bridge->hdmi_audio_dai_port be set to -1 instead of 1? In rockchip,dw-dp.yaml (and in dw_dp_probe() via of_drm_get_bridge_by_endpoint(dev->of_node, 1, 0)), OF-graph port 1 is the DisplayPort video output port, while audio DAIs are referenced via #sound-dai-cells = <1> rather than an OF-graph audio port. Per struct drm_bridge documentation, hdmi_audio_dai_port should be -1 when an OF-graph audio port is not used, otherwise drm_connector_hdmi_audio_get_dai_id() will match the video output port. -- Sashiko AI review ยท https://sashiko.dev/#/patchset/20261114-synopsys-dw-dp-improvements-v12-0-d42d3dc5b...@collabora.com?part=22
