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

[PATCH 00/14] ASoC: simple-card-utils: prepare for multi support
by Kuninori Morimoto 15 Apr '21

15 Apr '21

Hi Mark I want to add new audio-graph-card2 driver which can support not only DPCM, but also Multi-CPU/Codec, and Codec2Codec. And it is also supporting audio-graph-card2 base custom driver. But before supporting such driver, we need to cleanup existing simple-card / audio-graph, because these and new driver are sharing code. This patch-set are for Multi-CPU/Codec support, and some cleanups. Kuninori Morimoto (14): ASoC: simple-card-utils: enable flexible CPU/Codec/Platform ASoC: simple-card-utils: share dummy DAI and reduce memory ASoC: simple-card-utils: setup dai_props cpu_dai/codec_dai at initial timing ASoC: simple-card-utils: remove li->dais/li->conf ASoC: simple-card-utils: use for_each_prop_xxx() ASoC: simple-card-utils: remove asoc_simple_parse_xxx() ASoC: simple-card-utils: care multi DAI at asoc_simple_clean_reference() ASoC: simple-card-utils: indicate dai_fmt if exist ASoC: simple-card-utils: indicate missing CPU/Codec numbers for debug ASoC: simple-card-utils: add simple_props_to_xxx() macro ASoC: simple-card-utils: multi support at asoc_simple_canonicalize_cpu/platform() ASoC: simple-card-utils: tidyup debug info for clock ASoC: simple-card-utils: tidyup dev_dbg() to use 1 line ASoC: simple-card-utils: tidyup asoc_simple_parse_convert() include/sound/simple_card_utils.h | 116 +++++++---- sound/soc/generic/audio-graph-card.c | 90 ++++----- sound/soc/generic/simple-card-utils.c | 272 +++++++++++++++++--------- sound/soc/generic/simple-card.c | 104 +++++----- 4 files changed, 352 insertions(+), 230 deletions(-) -- 2.25.1

3 20

[RFC PATCH 0/2] ASoC: remove cppchecks warnings on lm49453 and da732x
by Pierre-Louis Bossart 15 Apr '21

15 Apr '21

There are the last two patches in the cleanups, this time I am not sure what the code does and what the proper fix might be. Feedback welcome. Pierre-Louis Bossart (2): ASoC: lm49453: fix useless assignment before return ASoC: da732x: simplify code sound/soc/codecs/da732x.c | 17 ++++++----------- sound/soc/codecs/da732x.h | 12 ++++-------- sound/soc/codecs/lm49453.c | 2 -- 3 files changed, 10 insertions(+), 21 deletions(-) -- 2.25.1

3 6

[PATCH 4.14 30/68] ASoC: sunxi: sun4i-codec: fill ASoC card owner
by Greg Kroah-Hartman 15 Apr '21

15 Apr '21

From: Bastian Germann <bage(a)linutronix.de> [ Upstream commit 7c0d6e482062eb5c06ecccfab340abc523bdca00 ] card->owner is a required property and since commit 81033c6b584b ("ALSA: core: Warn on empty module") a warning is issued if it is empty. Add it. This fixes following warning observed on Lamobo R1: WARNING: CPU: 1 PID: 190 at sound/core/init.c:207 snd_card_new+0x430/0x480 [snd] Modules linked in: sun4i_codec(E+) sun4i_backend(E+) snd_soc_core(E) ... CPU: 1 PID: 190 Comm: systemd-udevd Tainted: G C E 5.10.0-1-armmp #1 Debian 5.10.4-1 Hardware name: Allwinner sun7i (A20) Family Call trace: (snd_card_new [snd]) (snd_soc_bind_card [snd_soc_core]) (snd_soc_register_card [snd_soc_core]) (sun4i_codec_probe [sun4i_codec]) Fixes: 45fb6b6f2aa3 ("ASoC: sunxi: add support for the on-chip codec on early Allwinner SoCs") Related: commit 3c27ea23ffb4 ("ASoC: qcom: Set card->owner to avoid warnings") Related: commit ec653df2a0cb ("drm/vc4/vc4_hdmi: fill ASoC card owner") Cc: linux-arm-kernel(a)lists.infradead.org Cc: alsa-devel(a)alsa-project.org Signed-off-by: Bastian Germann <bage(a)linutronix.de> Link: https://lore.kernel.org/r/20210331151843.30583-1-bage@linutronix.de Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/sunxi/sun4i-codec.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/sound/soc/sunxi/sun4i-codec.c b/sound/soc/sunxi/sun4i-codec.c index baa9007464ed..700779ca82d0 100644 --- a/sound/soc/sunxi/sun4i-codec.c +++ b/sound/soc/sunxi/sun4i-codec.c @@ -1199,6 +1199,7 @@ static struct snd_soc_card *sun4i_codec_create_card(struct device *dev) return ERR_PTR(-ENOMEM); card->dev = dev; + card->owner = THIS_MODULE; card->name = "sun4i-codec"; card->dapm_widgets = sun4i_codec_card_dapm_widgets; card->num_dapm_widgets = ARRAY_SIZE(sun4i_codec_card_dapm_widgets); @@ -1231,6 +1232,7 @@ static struct snd_soc_card *sun6i_codec_create_card(struct device *dev) return ERR_PTR(-ENOMEM); card->dev = dev; + card->owner = THIS_MODULE; card->name = "A31 Audio Codec"; card->dapm_widgets = sun6i_codec_card_dapm_widgets; card->num_dapm_widgets = ARRAY_SIZE(sun6i_codec_card_dapm_widgets); @@ -1284,6 +1286,7 @@ static struct snd_soc_card *sun8i_a23_codec_create_card(struct device *dev) return ERR_PTR(-ENOMEM); card->dev = dev; + card->owner = THIS_MODULE; card->name = "A23 Audio Codec"; card->dapm_widgets = sun6i_codec_card_dapm_widgets; card->num_dapm_widgets = ARRAY_SIZE(sun6i_codec_card_dapm_widgets); @@ -1322,6 +1325,7 @@ static struct snd_soc_card *sun8i_h3_codec_create_card(struct device *dev) return ERR_PTR(-ENOMEM); card->dev = dev; + card->owner = THIS_MODULE; card->name = "H3 Audio Codec"; card->dapm_widgets = sun6i_codec_card_dapm_widgets; card->num_dapm_widgets = ARRAY_SIZE(sun6i_codec_card_dapm_widgets); @@ -1360,6 +1364,7 @@ static struct snd_soc_card *sun8i_v3s_codec_create_card(struct device *dev) return ERR_PTR(-ENOMEM); card->dev = dev; + card->owner = THIS_MODULE; card->name = "V3s Audio Codec"; card->dapm_widgets = sun6i_codec_card_dapm_widgets; card->num_dapm_widgets = ARRAY_SIZE(sun6i_codec_card_dapm_widgets); -- 2.30.2

1 0

[Problem] A data race in snd_ctl_elem_add()
by Gong, Sishuai 15 Apr '21

15 Apr '21

Hi, We found a data race in sound/core/control.c in linux-5.12-rc3 and we are able to reproduce it under x86. In general, we found when 2 kernel threads are both running snd_ctl_elem_add(), one may read a stale value of card->user_ctl_count, as shown below. Currently, we haven’t found any explicit errors due to this data race, but it looks the reader threads may operate in an inconsistent state, where card->user_ctl_count + 1 is actually bigger than MAX_USER_CONTROLS, so we want to point it out. Thread 1 Thread 2 //snd_ctl_elem_add() //snd_ctl_elem_add() if (card->user_ctl_count + 1 > MAX_USER_CONTROLS) return -ENOMEM; card->user_ctl_count++; unlock: up_write(&card->controls_rwsem); return err; Thanks, Sishuai

2 2

SOF SoundWire: Improve dmic detection
by GitHub pull_request - edited 15 Apr '21

15 Apr '21

alsa-project/alsa-ucm-conf pull request #75 was edited from plbossart: DMIC support seems incomplete for HDaudio (not handling all possibilities for cfg-dmics) and missing completely for SoundWire. This is an untested draft just to get UCM3 feedback from @perexg and tests from @aigilea on an HP Spectre x360 device. Edit: adding @kv2019i @RanderWang @bardliao and @libinyang for reviews. Request URL : https://github.com/alsa-project/alsa-ucm-conf/pull/75 Patch URL : https://github.com/alsa-project/alsa-ucm-conf/pull/75.patch Repository URL: https://github.com/alsa-project/alsa-ucm-conf

1 0

[PATCH] ALSA: control: Fix racy management of user ctl memory size account
by Takashi Iwai 15 Apr '21

15 Apr '21

We've got a report about the possible race in the user control element counts (card->user_ctl_count), and it was confirmed that the race wasn't serious in the old code up to 5.12. There, the value modification itself was exclusive and protected via a write semaphore, hence it's at most concurrent reads and evaluations before the increment. Since it's only about the soft-limit to avoid the exhausting memory usage, one-off isn't a big problem at all. Meanwhile, the relevant code has been largely modified recently, and now card->user_ctl_count was replaced with card->user_ctl_alloc_size, and a few more places were added to access this field. And, in this new code, it turned out to be more serious: the modifications are scattered in various places, and a few of them are without protection. It implies that it may lead to an inconsistent value by racy accesses. For addressing it, this patch extends the range covered by the card->controls_rwsem write lock at snd_ctl_elem_add() so that the all code paths that modify and refer to card->user_ctl_alloc_size are protected by the rwsem properly. The patch adds also comments in a couple of functions to indicate that they are under the rwsem lock. Fixes: 66c6d1ef86ff ("ALSA: control: Add memory consumption limit to user controls") Link: https://lore.kernel.org/r/FEEBF384-44BE-42CF-8FB3-93470933F64F@purdue.edu Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- sound/core/control.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) diff --git a/sound/core/control.c b/sound/core/control.c index a076c08c21b6..498e3701514a 100644 --- a/sound/core/control.c +++ b/sound/core/control.c @@ -1337,6 +1337,7 @@ static int snd_ctl_elem_user_put(struct snd_kcontrol *kcontrol, return change; } +/* called in controls_rwsem write lock */ static int replace_user_tlv(struct snd_kcontrol *kctl, unsigned int __user *buf, unsigned int size) { @@ -1414,6 +1415,7 @@ static int snd_ctl_elem_user_tlv(struct snd_kcontrol *kctl, int op_flag, return read_user_tlv(kctl, buf, size); } +/* called in controls_rwsem write lock */ static int snd_ctl_elem_init_enum_names(struct user_element *ue) { char *names, *p; @@ -1529,8 +1531,11 @@ static int snd_ctl_elem_add(struct snd_ctl_file *file, private_size = value_sizes[info->type] * info->count; alloc_size = compute_user_elem_size(private_size, count); - if (check_user_elem_overflow(card, alloc_size)) - return -ENOMEM; + down_write(&card->controls_rwsem); + if (check_user_elem_overflow(card, alloc_size)) { + err = -ENOMEM; + goto unlock; + } /* * Keep memory object for this userspace control. After passing this @@ -1540,12 +1545,13 @@ static int snd_ctl_elem_add(struct snd_ctl_file *file, */ err = snd_ctl_new(&kctl, count, access, file); if (err < 0) - return err; + goto unlock; memcpy(&kctl->id, &info->id, sizeof(kctl->id)); ue = kzalloc(alloc_size, GFP_KERNEL); if (!ue) { kfree(kctl); - return -ENOMEM; + err = -ENOMEM; + goto unlock; } kctl->private_data = ue; kctl->private_free = snd_ctl_elem_user_free; @@ -1563,7 +1569,7 @@ static int snd_ctl_elem_add(struct snd_ctl_file *file, err = snd_ctl_elem_init_enum_names(ue); if (err < 0) { snd_ctl_free_one(kctl); - return err; + goto unlock; } } @@ -1580,7 +1586,6 @@ static int snd_ctl_elem_add(struct snd_ctl_file *file, kctl->tlv.c = snd_ctl_elem_user_tlv; /* This function manage to free the instance on failure. */ - down_write(&card->controls_rwsem); err = __snd_ctl_add_replace(card, kctl, CTL_ADD_EXCLUSIVE); if (err < 0) { snd_ctl_free_one(kctl); -- 2.26.2

1 0

[PATCH] ASoC: Intel: Boards: cml_da7219_max98390: set TDM format configuration
by mac.chiang＠intel.com 15 Apr '21

15 Apr '21

From: Mac Chiang <mac.chiang(a)intel.com> add tx/rx 2 channels, 32bits support Signed-off-by: Steve Lee <SteveS.Lee(a)maximintegrated.com> Signed-off-by: Mac Chiang <mac.chiang(a)intel.com> --- sound/soc/intel/boards/bxt_da7219_max98357a.c | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) diff --git a/sound/soc/intel/boards/bxt_da7219_max98357a.c b/sound/soc/intel/boards/bxt_da7219_max98357a.c index 9ffef39..515379cd 100644 --- a/sound/soc/intel/boards/bxt_da7219_max98357a.c +++ b/sound/soc/intel/boards/bxt_da7219_max98357a.c @@ -253,6 +253,30 @@ static int broxton_da7219_codec_init(struct snd_soc_pcm_runtime *rtd) return ret; } +static int max98390_hw_params(struct snd_pcm_substream *substream, + struct snd_pcm_hw_params *params) +{ + struct snd_soc_pcm_runtime *rtd = substream->private_data; + struct snd_soc_dai *codec_dai; + int j; + + for_each_rtd_codec_dais(rtd, j, codec_dai) { + if (!strcmp(codec_dai->component->name, MAX98390_DEV0_NAME)) { + /* DEV0 tdm slot configuration */ + snd_soc_dai_set_tdm_slot(codec_dai, 0x1, 3, 2, 32); + } + if (!strcmp(codec_dai->component->name, MAX98390_DEV1_NAME)) { + /* DEV1 tdm slot configuration */ + snd_soc_dai_set_tdm_slot(codec_dai, 0x2, 3, 2, 32); + } + } + return 0; +} + +struct snd_soc_ops max_98390_ops = { + .hw_params = max98390_hw_params, +}; + static int broxton_hdmi_init(struct snd_soc_pcm_runtime *rtd) { struct bxt_card_private *ctx = snd_soc_card_get_drvdata(rtd->card); @@ -813,6 +837,7 @@ static int broxton_audio_probe(struct platform_device *pdev) if (ctx->spkamp == SPKAMP_MAX98390) { broxton_dais[i].codecs = max98390_codec; broxton_dais[i].num_codecs = ARRAY_SIZE(max98390_codec); + broxton_dais[i].ops = &max_98390_ops; broxton_dais[i].dpcm_capture = 1; } } -- 2.7.4

2 1

Re: [2/2]ALSA: hda: add Zhaoxin HDAC non-snoop support: patch a design limitation
by Takashi Iwai 15 Apr '21

15 Apr '21

On Wed, 14 Apr 2021 14:00:23 +0200, Hans Hu(SH-RD) wrote: > > Hi Takashi, > > ZHAOXIN HDAC controller has one design limitation when it works in non-snoop mode. This design limitation is: hardware can't guarantee that the write CORB cycle always completes first before the write CORBWP cycle. Here is the error scene: > int snd_hdac_bus_send_cmd(struct hdac_bus *bus, unsigned int val) > { > ... > bus->corb.buf[wp] = cpu_to_le32(val); // cycle_1: write value to CORB > snd_hdac_chip_writew(bus, CORBWP, wp); // cycle_2: write wp to CORBWP > ... > } > Normally, after cycle_2, DMA engine will start working and get data from CORB and sent it out. But if cycle_1 haven’t finished yet at this time(limitation occurs), which means the value haven’t been written into CORB, then DMA engine will get unexpected value, then error occurred. (feels like one kind of CORB underrun). > > If we add one read CORB cycle between cycle_1 and cycle_2, cycle_1 and cycle_2 operations will be synchronized, this limitation will be fixed. We have written a draft patch based on this situation and hope to be accepted, the patch example is as follows: > diff --git a/include/sound/hdaudio.h b/include/sound/hdaudio.h > index 22af68b..c338db00 100644 > --- a/include/sound/hdaudio.h > +++ b/include/sound/hdaudio.h > @@ -339,6 +339,7 @@ struct hdac_bus { > bool sync_write:1; /* sync after verb write */ > bool use_posbuf:1; /* use position buffer */ > bool snoop:1; /* enable snooping */ > + bool no_snoop_corb_sync:1; > bool align_bdle_4k:1; /* BDLE align 4K boundary */ > bool reverse_assign:1; /* assign devices in reverse order */ > bool corbrp_self_clear:1; /* CORBRP clears itself after reset */ > diff --git a/sound/hda/hdac_controller.c b/sound/hda/hdac_controller.c > index 062da7a..6c90cdd 100644 > --- a/sound/hda/hdac_controller.c > +++ b/sound/hda/hdac_controller.c > @@ -167,6 +167,8 @@ int snd_hdac_bus_send_cmd(struct hdac_bus *bus, unsigned int val) > > bus->rirb.cmds[addr]++; > bus->corb.buf[wp] = cpu_to_le32(val); > + if (bus->no_snoop_corb_sync) > + val = bus->corb.buf[wp]; > snd_hdac_chip_writew(bus, CORBWP, wp); Just wondering whether using WRITE_ONCE() macro would be enough? e.g. WRITE_ONCE(bus->corb.buf[wp], cpu_to_le32(val)); Also, no_snoop_corb_sync is a bit confusing. What you do is rather sync of the written value, so it can be taken as if other way round. Maybe no_coherent_corb_write or such would be better understandable. thanks, Takashi

2 2

[PATCH] ALSA: usb-audio: Explicitly set up the clock selector
by Takashi Iwai 14 Apr '21

14 Apr '21

In the current code, we have some assumption that the audio clock selector has been set up implicitly and don't want to touch it unless it's really needed for the fallback autoclock setup. This works for most devices but some seem having a problem. Partially this was covered for the devices with a single connector at the initialization phase (commit 086b957cc17f "ALSA: usb-audio: Skip the clock selector inquiry for single connections"), but also there are cases where the wrong clock set up is kept silently. The latter seems to be the cause of the noises on Behringer devices. In this patch, we explicitly set up the audio clock selector whenever the appropriate node is found. Reported-by: Geraldo Nascimento <geraldogabriel(a)gmail.com> BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=199327 Link: https://lore.kernel.org/r/CAEsQvcvF7LnO8PxyyCxuRCx=7jNeSCvFAd-+dE0g_rd1rOxx… Cc: <stable(a)vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- sound/usb/clock.c | 18 ++++++++++++++---- 1 file changed, 14 insertions(+), 4 deletions(-) diff --git a/sound/usb/clock.c b/sound/usb/clock.c index a746802d0ac3..17bbde73d4d1 100644 --- a/sound/usb/clock.c +++ b/sound/usb/clock.c @@ -296,7 +296,7 @@ static int __uac_clock_find_source(struct snd_usb_audio *chip, selector = snd_usb_find_clock_selector(chip->ctrl_intf, entity_id); if (selector) { - int ret, i, cur; + int ret, i, cur, err; if (selector->bNrInPins == 1) { ret = 1; @@ -324,13 +324,17 @@ static int __uac_clock_find_source(struct snd_usb_audio *chip, ret = __uac_clock_find_source(chip, fmt, selector->baCSourceID[ret - 1], visited, validate); + if (ret > 0) { + err = uac_clock_selector_set_val(chip, entity_id, cur); + if (err < 0) + return err; + } + if (!validate || ret > 0 || !chip->autoclock) return ret; /* The current clock source is invalid, try others. */ for (i = 1; i <= selector->bNrInPins; i++) { - int err; - if (i == cur) continue; @@ -396,7 +400,7 @@ static int __uac3_clock_find_source(struct snd_usb_audio *chip, selector = snd_usb_find_clock_selector_v3(chip->ctrl_intf, entity_id); if (selector) { - int ret, i, cur; + int ret, i, cur, err; /* the entity ID we are looking for is a selector. * find out what it currently selects */ @@ -418,6 +422,12 @@ static int __uac3_clock_find_source(struct snd_usb_audio *chip, ret = __uac3_clock_find_source(chip, fmt, selector->baCSourceID[ret - 1], visited, validate); + if (ret > 0) { + err = uac_clock_selector_set_val(chip, entity_id, cur); + if (err < 0) + return err; + } + if (!validate || ret > 0 || !chip->autoclock) return ret; -- 2.26.2

2 1

[PATCH v2 -next 1/2] ASoC: ak5558: correct reset polarity
by Shengjiu Wang 14 Apr '21

14 Apr '21

Reset (aka power off) happens when the reset gpio is made active. The reset gpio is GPIO_ACTIVE_LOW Fixes: 920884777480 ("ASoC: ak5558: Add support for AK5558 ADC driver") Signed-off-by: Shengjiu Wang <shengjiu.wang(a)nxp.com> --- changes in v2: - split the patch to two patches. sound/soc/codecs/ak5558.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/sound/soc/codecs/ak5558.c b/sound/soc/codecs/ak5558.c index 8e4dca753f0b..f24d91b728dd 100644 --- a/sound/soc/codecs/ak5558.c +++ b/sound/soc/codecs/ak5558.c @@ -323,7 +323,7 @@ static void ak5558_power_off(struct ak5558_priv *ak5558) if (!ak5558->reset_gpiod) return; - gpiod_set_value_cansleep(ak5558->reset_gpiod, 0); + gpiod_set_value_cansleep(ak5558->reset_gpiod, 1); usleep_range(1000, 2000); } @@ -332,7 +332,7 @@ static void ak5558_power_on(struct ak5558_priv *ak5558) if (!ak5558->reset_gpiod) return; - gpiod_set_value_cansleep(ak5558->reset_gpiod, 1); + gpiod_set_value_cansleep(ak5558->reset_gpiod, 0); usleep_range(1000, 2000); } -- 2.27.0

2 2