February 2023 - Alsa-devel - mailman.alsa-project.org

[PATCH -next] ASoC: soc-ac97: fix wrong pointer pass to PTR_ERR()
by Yang Yingliang 17 Feb '23

17 Feb '23

Fix wrong pointer pass to PTR_ERR() if devm_gpiod_get_index() fails. Fixes: 3ee0d39c50dc ("ASoC: soc-ac97: Convert to agnostic GPIO API") Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- sound/soc/soc-ac97.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/sound/soc/soc-ac97.c b/sound/soc/soc-ac97.c index 40051f513c6a..4e4fe29ade50 100644 --- a/sound/soc/soc-ac97.c +++ b/sound/soc/soc-ac97.c @@ -336,12 +336,12 @@ static int snd_soc_ac97_parse_pinctl(struct device *dev, cfg->sdata_gpio = devm_gpiod_get_index(dev, "ac97", 1, GPIOD_ASIS); if (IS_ERR(cfg->sdata_gpio)) - return dev_err_probe(dev, PTR_ERR(cfg->sync_gpio), "Can't find ac97-sdata gpio\n"); + return dev_err_probe(dev, PTR_ERR(cfg->sdata_gpio), "Can't find ac97-sdata gpio\n"); gpiod_set_consumer_name(cfg->sdata_gpio, "AC97 link sdata"); cfg->reset_gpio = devm_gpiod_get_index(dev, "ac97", 2, GPIOD_ASIS); if (IS_ERR(cfg->reset_gpio)) - return dev_err_probe(dev, PTR_ERR(cfg->sync_gpio), "Can't find ac97-reset gpio\n"); + return dev_err_probe(dev, PTR_ERR(cfg->reset_gpio), "Can't find ac97-reset gpio\n"); gpiod_set_consumer_name(cfg->reset_gpio, "AC97 link reset"); return 0; -- 2.25.1

2 1

[PATCH] ASoC: SMA1303: Change the value for right output
by Kiseok Jo 17 Feb '23

17 Feb '23

This device can output mono, left or right. LR data should be swapped to output right data. Signed-off-by: Kiseok Jo <kiseok.jo(a)irondevice.com> --- sound/soc/codecs/sma1303.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/codecs/sma1303.c b/sound/soc/codecs/sma1303.c index 727c01facf52..fa4b0a60f8a9 100644 --- a/sound/soc/codecs/sma1303.c +++ b/sound/soc/codecs/sma1303.c @@ -569,7 +569,7 @@ static int sma1303_aif_in_event(struct snd_soc_dapm_widget *w, ret += sma1303_regmap_update_bits(sma1303, SMA1303_11_SYSTEM_CTRL2, SMA1303_LR_DATA_SW_MASK, - SMA1303_LR_DATA_SW_NORMAL, + SMA1303_LR_DATA_SW_SWAP, &temp); if (temp == true) change = true; base-commit: ab2932214588224b11102e40ba2e78a6f099b8ef -- 2.20.1

1 0

[PATCH] ASoC: Intel: Skylake: Fix struct definition
by Amadeusz Sławiński 16 Feb '23

16 Feb '23

The kernel is globally removing the ambiguous 0-length and 1-element arrays in favor of flexible arrays, so that we can gain both compile-time and run-time array bounds checking[1]. In this instance, struct skl_cpr_cfg contains struct skl_cpr_gtw_cfg, which defined "config_data" as a 1-element array. However, case present in sound/soc/intel/skylake/skl-topology.h is not a simple one as the structure takes part in IPC communication. Apparently original definition missed one field, which while not used by AudioDSP firmware when there is no additional data, is still expected to be part of an IPC message. Currently this works because of how 'config_data' is declared: 'config_data[1]'. Now when one replaces it with a flexible array there would be one field missing. Update struct declaration to fix this. Reported-by: Sasa Ostrouska <casaxa(a)gmail.com> Link: https://lore.kernel.org/all/CALFERdwvq5day_sbDfiUsMSZCQu9HG8-SBpOZDNPeMdZGo… Cc: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> Cc: Liam Girdwood <liam.r.girdwood(a)linux.intel.com> Cc: Peter Ujfalusi <peter.ujfalusi(a)linux.intel.com> Cc: Bard Liao <yung-chuan.liao(a)linux.intel.com> Cc: Ranjani Sridharan <ranjani.sridharan(a)linux.intel.com> Cc: Kai Vehmanen <kai.vehmanen(a)linux.intel.com> Cc: Mark Brown <broonie(a)kernel.org> Cc: Jaroslav Kysela <perex(a)perex.cz> Cc: Takashi Iwai <tiwai(a)suse.com> Cc: "Gustavo A. R. Silva" <gustavoars(a)kernel.org> Cc: alsa-devel(a)alsa-project.org CC: Kees Cook <keescook(a)chromium.org> Reviewed-by: Cezary Rojewski <cezary.rojewski(a)intel.com> Signed-off-by: Amadeusz Sławiński <amadeuszx.slawinski(a)linux.intel.com> --- sound/soc/intel/skylake/skl-messages.c | 2 +- sound/soc/intel/skylake/skl-topology.h | 5 ++++- 2 files changed, 5 insertions(+), 2 deletions(-) diff --git a/sound/soc/intel/skylake/skl-messages.c b/sound/soc/intel/skylake/skl-messages.c index 5ab0917a2b3d..d31509298a0a 100644 --- a/sound/soc/intel/skylake/skl-messages.c +++ b/sound/soc/intel/skylake/skl-messages.c @@ -549,7 +549,7 @@ static void skl_copy_copier_caps(struct skl_module_cfg *mconfig, if (mconfig->formats_config[SKL_PARAM_INIT].caps_size == 0) return; - memcpy(cpr_mconfig->gtw_cfg.config_data, + memcpy(&cpr_mconfig->gtw_cfg.config_data, mconfig->formats_config[SKL_PARAM_INIT].caps, mconfig->formats_config[SKL_PARAM_INIT].caps_size); diff --git a/sound/soc/intel/skylake/skl-topology.h b/sound/soc/intel/skylake/skl-topology.h index 6db0fd7bad49..30a0977af943 100644 --- a/sound/soc/intel/skylake/skl-topology.h +++ b/sound/soc/intel/skylake/skl-topology.h @@ -115,7 +115,10 @@ struct skl_cpr_gtw_cfg { u32 dma_buffer_size; u32 config_length; /* not mandatory; required only for DMIC/I2S */ - u32 config_data[1]; + struct { + u32 gtw_attrs; + u32 data[]; + } config_data; } __packed; struct skl_dma_control { -- 2.34.1

4 3

[PATCH 14/17] dt-bindings: sound: apple,mca: Add t8112-mca compatible
by Janne Grunau 16 Feb '23

16 Feb '23

The block found on Apple's M2 SoC is compatible with the existing driver so add its per-SoC compatible. Signed-off-by: Janne Grunau <j(a)jannau.net> --- This trivial dt-bindings update should be merged through the asahi-soc tree to ensure validation of the Apple M2 (t8112) devicetrees in this series. --- Documentation/devicetree/bindings/sound/apple,mca.yaml | 1 + 1 file changed, 1 insertion(+) diff --git a/Documentation/devicetree/bindings/sound/apple,mca.yaml b/Documentation/devicetree/bindings/sound/apple,mca.yaml index 40e3a202f443..5c6ec08c7d24 100644 --- a/Documentation/devicetree/bindings/sound/apple,mca.yaml +++ b/Documentation/devicetree/bindings/sound/apple,mca.yaml @@ -23,6 +23,7 @@ properties: - enum: - apple,t6000-mca - apple,t8103-mca + - apple,t8112-mca - const: apple,mca reg: -- 2.39.1

3 4

[PATCH 0/9] Refactor Vangogh acp5x machine driver
by Lucas Tanure 16 Feb '23

16 Feb '23

Provide small fixes and refactor the code for easier insertion of a new platform using the same acp5x machine driver. Lucas Tanure (9): ASoC: amd: vangogh: Remove unnecessary init function ASoC: amd: vangogh: Update code indentation ASoC: amd: vangogh: use sizeof of variable instead of struct type ASoC: amd: vangogh: remove unnecessarily included headers ASoC: amd: vangogh: use for_each_rtd_components instead of for ASoC: amd: vangogh: Check Bit Clock rate before snd_soc_dai_set_pll ASoC: amd: vangogh: Move nau8821 and CPU side code up for future platform ASoC: amd: vangogh: Centralize strings definition ASoC: amd: vangogh: Include cs35l41 in structs names sound/soc/amd/vangogh/acp5x-mach.c | 313 +++++++++++++---------------- 1 file changed, 141 insertions(+), 172 deletions(-) -- 2.39.2

2 11

[PATCH] pcm_rate: fix hw_ptr boundary wrapping issue
by Alan Young 16 Feb '23

16 Feb '23

Wrap the hw_ptr using the total position of the slave hw_ptr, including boundary wraps. Otherwise, small errors can creep in due to residuals (when boundary is not a multiple of period size) and which can accumulate. Signed-off-by: Alan Young <consult.awy(a)gmail.com> Fixes: 7570e5d7 ("pcm: rate: fix the hw_ptr update until the boundary available") --- src/pcm/pcm_rate.c | 54 ++++++++++++++++++++++++++++++---------------- 1 file changed, 36 insertions(+), 18 deletions(-) diff --git a/src/pcm/pcm_rate.c b/src/pcm/pcm_rate.c index c8076859..a29fc5a9 100644 --- a/src/pcm/pcm_rate.c +++ b/src/pcm/pcm_rate.c @@ -52,6 +52,7 @@ struct _snd_pcm_rate { snd_pcm_generic_t gen; snd_pcm_uframes_t appl_ptr, hw_ptr, last_slave_hw_ptr; snd_pcm_uframes_t last_commit_ptr; + u_int64_t slave_hw_ptr_wrap; snd_pcm_uframes_t orig_avail_min; snd_pcm_sw_params_t sw_params; snd_pcm_format_t sformat; @@ -562,6 +563,8 @@ static int snd_pcm_rate_init(snd_pcm_t *pcm) if (rate->ops.reset) rate->ops.reset(rate->obj); rate->last_commit_ptr = 0; + rate->last_slave_hw_ptr = 0; + rate->slave_hw_ptr_wrap = 0; rate->start_pending = 0; return 0; } @@ -638,33 +641,48 @@ snd_pcm_rate_read_areas1(snd_pcm_t *pcm, static inline void snd_pcm_rate_sync_hwptr0(snd_pcm_t *pcm, snd_pcm_uframes_t slave_hw_ptr) { snd_pcm_rate_t *rate; - snd_pcm_sframes_t slave_hw_ptr_diff; - snd_pcm_sframes_t last_slave_hw_ptr_frac; + snd_pcm_uframes_t new_hw_ptr; + snd_pcm_uframes_t slave_residual; if (pcm->stream != SND_PCM_STREAM_PLAYBACK) return; rate = pcm->private_data; - slave_hw_ptr_diff = pcm_frame_diff(slave_hw_ptr, rate->last_slave_hw_ptr, rate->gen.slave->boundary); - if (slave_hw_ptr_diff == 0) + if (slave_hw_ptr == rate->last_slave_hw_ptr) return; - last_slave_hw_ptr_frac = rate->last_slave_hw_ptr % rate->gen.slave->period_size; - /* While handling fraction part fo slave period, rounded value will be - * introduced by input_frames(). - * To eliminate rounding issue on rate->hw_ptr, subtract last rounded - * value from rate->hw_ptr and add new rounded value of present - * slave_hw_ptr fraction part to rate->hw_ptr. Hence, - * rate->hw_ptr += [ (no. of updated slave periods * pcm rate period size) - - * fractional part of last_slave_hw_ptr rounded value + - * fractional part of updated slave hw ptr's rounded value ] + + /* + * Our hw_ptr may wrap at a different time to that of the slave as the + * number of period / boundary may differ. It is essential that our hw_ptr + * wraps at the correct point, along with our appl_ptr. Otherwise the + * avail calculation will be incorrect. + * + * Track the number of slave hw_ptr boundary wraps and use that to calculate + * our new_hw_ptr from the total number of samples converted. */ - rate->hw_ptr += ( - (((last_slave_hw_ptr_frac + slave_hw_ptr_diff) / rate->gen.slave->period_size) * pcm->period_size) - - rate->ops.input_frames(rate->obj, last_slave_hw_ptr_frac) + - rate->ops.input_frames(rate->obj, (last_slave_hw_ptr_frac + slave_hw_ptr_diff) % rate->gen.slave->period_size)); + if (slave_hw_ptr < rate->last_slave_hw_ptr) { + rate->slave_hw_ptr_wrap += rate->gen.slave->boundary; + } rate->last_slave_hw_ptr = slave_hw_ptr; - rate->hw_ptr %= pcm->boundary; + /* + * Calculate rate->hw_ptr using total number of frames consumed by slave hw_ptr, + * including any boundary wraps. + */ + if (rate->slave_hw_ptr_wrap) { + /* + * Restrict explicit 64-bit calculations to case where rate->slave_hw_ptr_wrap + * is non-zero, which will only happen in 32-bit environments. + */ + u_int64_t wrapped_slave_hw_ptr = slave_hw_ptr + rate->slave_hw_ptr_wrap; + new_hw_ptr = ((wrapped_slave_hw_ptr / rate->gen.slave->period_size) * pcm->period_size) % pcm->boundary; + slave_residual = wrapped_slave_hw_ptr % rate->gen.slave->period_size; + } else { + new_hw_ptr = (slave_hw_ptr / rate->gen.slave->period_size) * pcm->period_size; + slave_residual = slave_hw_ptr % rate->gen.slave->period_size; + } + new_hw_ptr += rate->ops.input_frames(rate->obj, slave_residual); + rate->hw_ptr = new_hw_ptr % pcm->boundary; } static inline void snd_pcm_rate_sync_hwptr(snd_pcm_t *pcm) -- 2.25.1

2 3

[PATCH] ASoC: rsnd: adg: Fix BRG typos
by Geert Uytterhoeven 16 Feb '23

16 Feb '23

"BRG" stands for "Baud Rate Generator", but is frequently misspelled as "RBG". Signed-off-by: Geert Uytterhoeven <geert+renesas(a)glider.be> --- sound/soc/sh/rcar/adg.c | 64 ++++++++++++++++++++--------------------- 1 file changed, 32 insertions(+), 32 deletions(-) diff --git a/sound/soc/sh/rcar/adg.c b/sound/soc/sh/rcar/adg.c index 7bc4421835d7261f..0b8926600d900a68 100644 --- a/sound/soc/sh/rcar/adg.c +++ b/sound/soc/sh/rcar/adg.c @@ -39,10 +39,10 @@ struct rsnd_adg { int clkin_size; int clkout_size; u32 ckr; - u32 rbga; - u32 rbgb; + u32 brga; + u32 brgb; - int rbg_rate[ADG_HZ_SIZE]; /* RBGA / RBGB */ + int brg_rate[ADG_HZ_SIZE]; /* BRGA / BRGB */ }; #define for_each_rsnd_clkin(pos, adg, i) \ @@ -75,7 +75,7 @@ static const char * const clkout_name_gen2[] = { [CLKOUT3] = "audio_clkout3", }; -static u32 rsnd_adg_calculate_rbgx(unsigned long div) +static u32 rsnd_adg_calculate_brgx(unsigned long div) { int i; @@ -131,8 +131,8 @@ static void __rsnd_adg_get_timesel_ratio(struct rsnd_priv *priv, adg->clkin_rate[CLKA], /* 0000: CLKA */ adg->clkin_rate[CLKB], /* 0001: CLKB */ adg->clkin_rate[CLKC], /* 0010: CLKC */ - adg->rbg_rate[ADG_HZ_441], /* 0011: RBGA */ - adg->rbg_rate[ADG_HZ_48], /* 0100: RBGB */ + adg->brg_rate[ADG_HZ_441], /* 0011: BRGA */ + adg->brg_rate[ADG_HZ_48], /* 0100: BRGB */ }; min = ~0; @@ -323,10 +323,10 @@ int rsnd_adg_clk_query(struct rsnd_priv *priv, unsigned int rate) /* * find divided clock from BRGA/BRGB */ - if (rate == adg->rbg_rate[ADG_HZ_441]) + if (rate == adg->brg_rate[ADG_HZ_441]) return 0x10; - if (rate == adg->rbg_rate[ADG_HZ_48]) + if (rate == adg->brg_rate[ADG_HZ_48]) return 0x20; return -EIO; @@ -358,13 +358,13 @@ int rsnd_adg_ssi_clk_try_start(struct rsnd_mod *ssi_mod, unsigned int rate) ckr = 0x80000000; /* BRGB output = 48kHz */ rsnd_mod_bset(adg_mod, BRGCKR, 0x80770000, adg->ckr | ckr); - rsnd_mod_write(adg_mod, BRRA, adg->rbga); - rsnd_mod_write(adg_mod, BRRB, adg->rbgb); + rsnd_mod_write(adg_mod, BRRA, adg->brga); + rsnd_mod_write(adg_mod, BRRB, adg->brgb); dev_dbg(dev, "CLKOUT is based on BRG%c (= %dHz)\n", (ckr) ? 'B' : 'A', - (ckr) ? adg->rbg_rate[ADG_HZ_48] : - adg->rbg_rate[ADG_HZ_441]); + (ckr) ? adg->brg_rate[ADG_HZ_48] : + adg->brg_rate[ADG_HZ_441]); return 0; } @@ -484,7 +484,7 @@ static int rsnd_adg_get_clkout(struct rsnd_priv *priv) struct device *dev = rsnd_priv_to_dev(priv); struct device_node *np = dev->of_node; struct property *prop; - u32 ckr, rbgx, rbga, rbgb; + u32 ckr, brgx, brga, brgb; u32 rate, div; u32 req_rate[ADG_HZ_SIZE] = {}; uint32_t count = 0; @@ -501,8 +501,8 @@ static int rsnd_adg_get_clkout(struct rsnd_priv *priv) }; ckr = 0; - rbga = 2; /* default 1/6 */ - rbgb = 2; /* default 1/6 */ + brga = 2; /* default 1/6 */ + brgb = 2; /* default 1/6 */ /* * ADG supports BRRA/BRRB output only @@ -543,30 +543,30 @@ static int rsnd_adg_get_clkout(struct rsnd_priv *priv) if (0 == rate) /* not used */ continue; - /* RBGA */ - if (!adg->rbg_rate[ADG_HZ_441] && (0 == rate % 44100)) { + /* BRGA */ + if (!adg->brg_rate[ADG_HZ_441] && (0 == rate % 44100)) { div = 6; if (req_Hz[ADG_HZ_441]) div = rate / req_Hz[ADG_HZ_441]; - rbgx = rsnd_adg_calculate_rbgx(div); - if (BRRx_MASK(rbgx) == rbgx) { - rbga = rbgx; - adg->rbg_rate[ADG_HZ_441] = rate / div; + brgx = rsnd_adg_calculate_brgx(div); + if (BRRx_MASK(brgx) == brgx) { + brga = brgx; + adg->brg_rate[ADG_HZ_441] = rate / div; ckr |= brg_table[i] << 20; if (req_Hz[ADG_HZ_441]) parent_clk_name = __clk_get_name(clk); } } - /* RBGB */ - if (!adg->rbg_rate[ADG_HZ_48] && (0 == rate % 48000)) { + /* BRGB */ + if (!adg->brg_rate[ADG_HZ_48] && (0 == rate % 48000)) { div = 6; if (req_Hz[ADG_HZ_48]) div = rate / req_Hz[ADG_HZ_48]; - rbgx = rsnd_adg_calculate_rbgx(div); - if (BRRx_MASK(rbgx) == rbgx) { - rbgb = rbgx; - adg->rbg_rate[ADG_HZ_48] = rate / div; + brgx = rsnd_adg_calculate_brgx(div); + if (BRRx_MASK(brgx) == brgx) { + brgb = brgx; + adg->brg_rate[ADG_HZ_48] = rate / div; ckr |= brg_table[i] << 16; if (req_Hz[ADG_HZ_48]) parent_clk_name = __clk_get_name(clk); @@ -620,8 +620,8 @@ static int rsnd_adg_get_clkout(struct rsnd_priv *priv) rsnd_adg_get_clkout_end: adg->ckr = ckr; - adg->rbga = rbga; - adg->rbgb = rbgb; + adg->brga = brga; + adg->brgb = brgb; return 0; @@ -663,9 +663,9 @@ void rsnd_adg_clk_dbg_info(struct rsnd_priv *priv, struct seq_file *m) __clk_get_name(clk), clk, clk_get_rate(clk)); dbg_msg(dev, m, "BRGCKR = 0x%08x, BRRA/BRRB = 0x%x/0x%x\n", - adg->ckr, adg->rbga, adg->rbgb); - dbg_msg(dev, m, "BRGA (for 44100 base) = %d\n", adg->rbg_rate[ADG_HZ_441]); - dbg_msg(dev, m, "BRGB (for 48000 base) = %d\n", adg->rbg_rate[ADG_HZ_48]); + adg->ckr, adg->brga, adg->brgb); + dbg_msg(dev, m, "BRGA (for 44100 base) = %d\n", adg->brg_rate[ADG_HZ_441]); + dbg_msg(dev, m, "BRGB (for 48000 base) = %d\n", adg->brg_rate[ADG_HZ_48]); /* * Actual CLKOUT will be exchanged in rsnd_adg_ssi_clk_try_start() -- 2.34.1

3 2

[PATCH 1/2] ALSA: pcm: generate default audio timestamp
by Alan Young 16 Feb '23

16 Feb '23

If there is no get_time_info() available or the requested audio timestamp type is AUDIO_TSTAMP_TYPE_DEFAULT then need to set audio_tstamp_report.actual_type = AUDIO_TSTAMP_TYPE_DEFAULT (in the else clause near the start of snd_pcm_update_hw_ptr0()) in order for update_audio_tstamp() to generate the audio timestamp from the number of frames. Fixes: 3179f6200188 ("ALSA: core: add .get_time_info") Signed-off-by: consult.awy(a)gmail.com Reviewed-by: Jaroslav Kysela <perex(a)perex.cz> --- sound/core/pcm_lib.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/sound/core/pcm_lib.c b/sound/core/pcm_lib.c index 8b6aeb8a78f7..943f5396fc60 100644 --- a/sound/core/pcm_lib.c +++ b/sound/core/pcm_lib.c @@ -289,8 +289,10 @@ static int snd_pcm_update_hw_ptr0(struct snd_pcm_substream *substream, /* re-test in case tstamp type is not supported in hardware and was demoted to DEFAULT */ if (runtime->audio_tstamp_report.actual_type == SNDRV_PCM_AUDIO_TSTAMP_TYPE_DEFAULT) snd_pcm_gettime(runtime, &curr_tstamp); - } else + } else { + runtime->audio_tstamp_report.actual_type = SNDRV_PCM_AUDIO_TSTAMP_TYPE_DEFAULT; snd_pcm_gettime(runtime, &curr_tstamp); + } } if (pos == SNDRV_PCM_POS_XRUN) { -- 2.31.1

1 0

[PATCH 2/2] ALSA: Change calculation of audio time from frames
by Alan Young 16 Feb '23

16 Feb '23

Introduce snd_pcm_lib_frames_to_timespec64() to replace the calculation audio_nsecs = div_u64(audio_frames * 1000000000LL, runtime->rate); which overflows after a playback duration of 4.8 days @ 44100 HZ or 0.56 days @ 384000 Hz. The implementation of snd_pcm_lib_frames_to_timespec64() extends these limits, to 1.5 million years (@ 384000 Hz), or the 68 year limit of struct timespec on 32-bit systems. Fixes: 3179f6200188 ("ALSA: core: add .get_time_info") Fixes: 4eeaaeaea1ce ("ALSA: core: add hooks for audio timestamps") Signed-off-by: consult.awy(a)gmail.com Suggested-by: Jaroslav Kysela <perex(a)perex.cz> Reviewed-by: Jaroslav Kysela <perex(a)perex.cz> --- sound/core/pcm_lib.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/sound/core/pcm_lib.c b/sound/core/pcm_lib.c index 943f5396fc60..e0a6556d33ee 100644 --- a/sound/core/pcm_lib.c +++ b/sound/core/pcm_lib.c @@ -205,12 +205,19 @@ int snd_pcm_update_state(struct snd_pcm_substream *substream, return 0; } +static inline void snd_pcm_lib_frames_to_timespec64(u64 frames, unsigned int rate, struct timespec64 *audio_tstamp) +{ + u32 remainder; + audio_tstamp->tv_sec = div_u64_rem(frames, rate, &remainder); + audio_tstamp->tv_nsec = div_u64(mul_u32_u32(remainder, NSEC_PER_SEC), rate); +} + static void update_audio_tstamp(struct snd_pcm_substream *substream, struct timespec64 *curr_tstamp, struct timespec64 *audio_tstamp) { struct snd_pcm_runtime *runtime = substream->runtime; - u64 audio_frames, audio_nsecs; + u64 audio_frames; struct timespec64 driver_tstamp; if (runtime->tstamp_mode != SNDRV_PCM_TSTAMP_ENABLE) @@ -233,9 +240,7 @@ static void update_audio_tstamp(struct snd_pcm_substream *substream, else audio_frames += runtime->delay; } - audio_nsecs = div_u64(audio_frames * 1000000000LL, - runtime->rate); - *audio_tstamp = ns_to_timespec64(audio_nsecs); + snd_pcm_lib_frames_to_timespec64(audio_frames, runtime->rate, audio_tstamp); } if (runtime->status->audio_tstamp.tv_sec != audio_tstamp->tv_sec || -- 2.31.1

1 0

[PATCH 1/2] ALSA: pcm: generate default audio timestamp
by Alan Young 16 Feb '23

16 Feb '23

If there is no get_time_info() available or the requested audio timestamp type is AUDIO_TSTAMP_TYPE_DEFAULT then need to set audio_tstamp_report.actual_type = AUDIO_TSTAMP_TYPE_DEFAULT (in the else clause near the start of snd_pcm_update_hw_ptr0()) in order for update_audio_tstamp() to generate the audio timestamp from the number of frames. Fixes: 3179f6200188 ("ALSA: core: add .get_time_info") Signed-off-by: consult.awy(a)gmail.com --- sound/core/pcm_lib.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/sound/core/pcm_lib.c b/sound/core/pcm_lib.c index 8b6aeb8a78f7..943f5396fc60 100644 --- a/sound/core/pcm_lib.c +++ b/sound/core/pcm_lib.c @@ -289,8 +289,10 @@ static int snd_pcm_update_hw_ptr0(struct snd_pcm_substream *substream, /* re-test in case tstamp type is not supported in hardware and was demoted to DEFAULT */ if (runtime->audio_tstamp_report.actual_type == SNDRV_PCM_AUDIO_TSTAMP_TYPE_DEFAULT) snd_pcm_gettime(runtime, &curr_tstamp); - } else + } else { + runtime->audio_tstamp_report.actual_type = SNDRV_PCM_AUDIO_TSTAMP_TYPE_DEFAULT; snd_pcm_gettime(runtime, &curr_tstamp); + } } if (pos == SNDRV_PCM_POS_XRUN) { -- 2.31.1

2 1