August 2022 - Alsa-devel - mailman.alsa-project.org

[PATCH] soundwire: dmi-quirks: add remapping for HP Omen 16-k0005TX
by Bard Liao 23 Aug '22

23 Aug '22

From: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> The DSDT for this device has a number of problems: a) it lists rt711 on link0 and link1, but link1 is disabled b) the rt711 entry on link0 uses the wrong v2 instead of v3 (SDCA) c) the rt1316 amplifier on link3 is not listed. Add a remapping table to work-around these BIOS shenanigans. BugLink: https://github.com/thesofproject/sof/issues/5955 Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan(a)linux.intel.com> Signed-off-by: Bard Liao <yung-chuan.liao(a)linux.intel.com> --- drivers/soundwire/dmi-quirks.c | 27 +++++++++++++++++++++++++++ 1 file changed, 27 insertions(+) diff --git a/drivers/soundwire/dmi-quirks.c b/drivers/soundwire/dmi-quirks.c index 747983743a14..f81cdd83ec26 100644 --- a/drivers/soundwire/dmi-quirks.c +++ b/drivers/soundwire/dmi-quirks.c @@ -55,7 +55,26 @@ static const struct adr_remap dell_sku_0A3E[] = { {} }; +/* + * The HP Omen 16-k0005TX does not expose the correct version of RT711 on link0 + * and does not expose a RT1316 on link3 + */ +static const struct adr_remap hp_omen_16[] = { + /* rt711-sdca on link0 */ + { + 0x000020025d071100ull, + 0x000030025d071101ull + }, + /* rt1316-sdca on link3 */ + { + 0x000120025d071100ull, + 0x000330025d131601ull + }, + {} +}; + static const struct dmi_system_id adr_remap_quirk_table[] = { + /* TGL devices */ { .matches = { DMI_MATCH(DMI_SYS_VENDOR, "HP"), @@ -78,6 +97,14 @@ static const struct dmi_system_id adr_remap_quirk_table[] = { }, .driver_data = (void *)dell_sku_0A3E, }, + /* ADL devices */ + { + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "HP"), + DMI_MATCH(DMI_PRODUCT_NAME, "OMEN by HP Gaming Laptop 16-k0xxx"), + }, + .driver_data = (void *)hp_omen_16, + }, {} }; -- 2.25.1

2 1

[PATCH v3 linux-next] soundwire: intel: Remove unnecessary TODO
by Khalid Masum 23 Aug '22

23 Aug '22

The capabilities enabled for multi-link are required as part of the programming sequences, even when a stream uses a single link we still use the syncArm/syncGo sequences. Therefore the TODO is no longer necessary. Suggested-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> Signed-off-by: Khalid Masum <khalid.masum.92(a)gmail.com> --- Previous discussions: - v2 Link: https://lore.kernel.org/lkml/20220816175954.5259-1-khalid.masum.92@gmail.co… - v1 Link: https://lore.kernel.org/lkml/20220814080416.7531-1-khalid.masum.92@gmail.co… Changes since v2: - Update commit title. Changes since v1: - Remove the todo, leave rest of the code as is. - Update commit message. drivers/soundwire/bus.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/soundwire/bus.c b/drivers/soundwire/bus.c index 8d4000664fa3..18158825a6d8 100644 --- a/drivers/soundwire/bus.c +++ b/drivers/soundwire/bus.c @@ -75,7 +75,6 @@ int sdw_bus_master_add(struct sdw_bus *bus, struct device *parent, /* * Initialize multi_link flag - * TODO: populate this flag by reading property from FW node */ bus->multi_link = false; if (bus->ops->read_prop) { -- 2.37.1

2 1

[PATCH v2] soundwire: qcom: Update error prints to debug prints
by Srinivasa Rao Mandadapu 23 Aug '22

23 Aug '22

Update error prints to debug prints to avoid redundant logging in kernel boot time, as these prints are informative prints in irq handler. Signed-off-by: Srinivasa Rao Mandadapu <quic_srivasam(a)quicinc.com> Reviewed-by: Andrew Halaney <ahalaney(a)redhat.com> --- Changes since v1: -- Remove redundant __func__ argument. -- Replace dev_dbg with dev_dbg_ratelimited. drivers/soundwire/qcom.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c index 9df970e..976ae75 100644 --- a/drivers/soundwire/qcom.c +++ b/drivers/soundwire/qcom.c @@ -573,11 +573,10 @@ static irqreturn_t qcom_swrm_irq_handler(int irq, void *dev_id) break; case SWRM_INTERRUPT_STATUS_NEW_SLAVE_ATTACHED: case SWRM_INTERRUPT_STATUS_CHANGE_ENUM_SLAVE_STATUS: - dev_err_ratelimited(swrm->dev, "%s: SWR new slave attached\n", - __func__); + dev_dbg_ratelimited(swrm->dev, "SWR new slave attached\n"); swrm->reg_read(swrm, SWRM_MCP_SLV_STATUS, &slave_status); if (swrm->slave_status == slave_status) { - dev_err(swrm->dev, "Slave status not changed %x\n", + dev_dbg(swrm->dev, "Slave status not changed %x\n", slave_status); } else { qcom_swrm_get_device_status(swrm); -- 2.7.4

3 2

[PATCH] ASoC: fsl_mqs: Fix supported clock DAI format
by Shengjiu Wang 23 Aug '22

23 Aug '22

The MQS works as codec DAI, not cpu DAI. It is clock consumer, not clock privider. Fixes: 3b14c15a333b ("ASoC: fsl: Update to use set_fmt_new callback") Signed-off-by: Shengjiu Wang <shengjiu.wang(a)nxp.com> --- sound/soc/fsl/fsl_mqs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/fsl/fsl_mqs.c b/sound/soc/fsl/fsl_mqs.c index c1e2f671191b..4922e6795b73 100644 --- a/sound/soc/fsl/fsl_mqs.c +++ b/sound/soc/fsl/fsl_mqs.c @@ -122,7 +122,7 @@ static int fsl_mqs_set_dai_fmt(struct snd_soc_dai *dai, unsigned int fmt) } switch (fmt & SND_SOC_DAIFMT_CLOCK_PROVIDER_MASK) { - case SND_SOC_DAIFMT_BP_FP: + case SND_SOC_DAIFMT_CBC_CFC: break; default: return -EINVAL; -- 2.34.1

2 1

Re: [PATCH] drm/bridge: dw-hdmi-i2s: set insert_pcuv bit if hardware supports it
by Geraldo Nascimento 23 Aug '22

23 Aug '22

On Mon, Aug 22, 2022 at 02:27:26PM +0100, Mark Brown wrote: > On Sat, Aug 20, 2022 at 09:36:53PM -0300, Geraldo Nascimento wrote: > > > + /* > > + * dw-hdmi introduced insert_pcuv bit in version 2.10a. > > + * When set (1'b1), this bit enables the insertion of the PCUV > > + * (Parity, Channel Status, User bit and Validity) bits on the > > + * incoming audio stream (support limited to Linear PCM audio) > > + */ > > + > > + if (hdmi_read(audio, HDMI_DESIGN_ID) >= 0x21) > > + conf2 |= HDMI_AUD_CONF2_INSERT_PCUV; > > So what if we're not handlign linear PCM? Hi Mark, I think we may be able to restrict the fix for L-PCM only by checking byte 0 of iec.status, specifically bit 1. Our define is called IEC958_AES0_NONAUDIO but https://tech.ebu.ch/docs/tech/tech3250.pdf calls it the "Linear PCM identification" bit. There's also a supplement to AES/EBU 3250 in https://tech.ebu.ch/docs/n/n009_1.pdf Let me know if the following is OK. I'll be happy to submit V2 if it is. Thanks, Geraldo Nascimento --- a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-i2s-audio.c +++ b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-i2s-audio.c @@ -42,6 +42,7 @@ static int dw_hdmi_i2s_hw_params(struct device *dev, void *data, struct dw_hdmi *hdmi = audio->hdmi; u8 conf0 = 0; u8 conf1 = 0; + u8 conf2 = 0; u8 inputclkfs = 0; /* it cares I2S only */ @@ -101,6 +102,18 @@ static int dw_hdmi_i2s_hw_params(struct device *dev, void *data, return -EINVAL; } + /* + * dw-hdmi introduced insert_pcuv bit in version 2.10a. + * When set (1'b1), this bit enables the insertion of the PCUV + * (Parity, Channel Status, User bit and Validity) bits on the + * incoming audio stream (support limited to Linear PCM audio) + */ + + if (hdmi_read(audio, HDMI_DESIGN_ID) >= 0x21 && + !(hparms->iec.status[0] & IEC958_AES0_NONAUDIO)) + conf2 = HDMI_AUD_CONF2_INSERT_PCUV; + dw_hdmi_set_sample_rate(hdmi, hparms->sample_rate); dw_hdmi_set_channel_status(hdmi, hparms->iec.status); dw_hdmi_set_channel_count(hdmi, hparms->channels); @@ -109,6 +122,7 @@ static int dw_hdmi_i2s_hw_params(struct device *dev, void *data, hdmi_write(audio, inputclkfs, HDMI_AUD_INPUTCLKFS); hdmi_write(audio, conf0, HDMI_AUD_CONF0); hdmi_write(audio, conf1, HDMI_AUD_CONF1); + hdmi_write(audio, conf2, HDMI_AUD_CONF2); return 0; }

2 1

[PATCH] soundwire: qcom: remove duplicate reset control get
by Srinivas Kandagatla 23 Aug '22

23 Aug '22

Looks like adding clock gate flag patch forgot to remove the old code that gets reset control. This causes below crash on platforms that do not need reset. [ 15.653501] reset_control_reset+0x124/0x170 [ 15.653508] qcom_swrm_init+0x50/0x1a0 [ 15.653514] qcom_swrm_probe+0x320/0x668 [ 15.653519] platform_probe+0x68/0xe0 [ 15.653529] really_probe+0xbc/0x2a8 [ 15.653535] __driver_probe_device+0x7c/0xe8 [ 15.653541] driver_probe_device+0x40/0x110 [ 15.653547] __device_attach_driver+0x98/0xd0 [ 15.653553] bus_for_each_drv+0x68/0xd0 [ 15.653559] __device_attach+0xf4/0x188 [ 15.653565] device_initial_probe+0x14/0x20 Fix this by removing old code. Reported-by: Amit Pundir <amit.pundir(a)linaro.org> Fixes: 1fd0d85affe4 ("soundwire: qcom: Add flag for software clock gating check") Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla(a)linaro.org> --- drivers/soundwire/qcom.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c index 9df970eeca45..a43961ad4614 100644 --- a/drivers/soundwire/qcom.c +++ b/drivers/soundwire/qcom.c @@ -1356,10 +1356,6 @@ static int qcom_swrm_probe(struct platform_device *pdev) ctrl->bus.compute_params = &qcom_swrm_compute_params; ctrl->bus.clk_stop_timeout = 300; - ctrl->audio_cgcr = devm_reset_control_get_exclusive(dev, "swr_audio_cgcr"); - if (IS_ERR(ctrl->audio_cgcr)) - dev_err(dev, "Failed to get audio_cgcr reset required for soundwire-v1.6.0\n"); - ret = qcom_swrm_get_port_config(ctrl); if (ret) goto err_clk; -- 2.21.0

3 2

Re: [PATCH v8 1/2] ASoC: sunxi: Add Allwinner H6 Digital MIC driver
by 班涛 23 Aug '22

23 Aug '22

Amadeusz Sławiński <amadeuszx.slawinski(a)linux.intel.com> 于2022年8月12日周五 23:20写道： > On 8/11/2022 4:49 PM, Ban Tao wrote: > > The Allwinner H6 and later SoCs have an DMIC block > > which is capable of capture. > > > > Signed-off-by: Ban Tao <fengzheng923(a)gmail.com> > > --- > > ... > > > + > > +#define SUN50I_DMIC_RATES (SNDRV_PCM_RATE_8000_48000) > > Seeing supported rates in hw_params callback, you probably want to take > a look at SNDRV_PCM_RATE_KNOT? > No, I don't need SNDRV_PCM_RATE_KNOT. Dmic does not support unknown rates. > > > +#define SUN50I_DMIC_FORMATS (SNDRV_PCM_FMTBIT_S16_LE | > SNDRV_PCM_FMTBIT_S24_LE) > > + > > +static struct snd_soc_dai_driver sun50i_dmic_dai = { > > + .capture = { > > + .channels_min = 1, > > + .channels_max = 8, > > + .rates = SUN50I_DMIC_RATES, > > + .formats = SUN50I_DMIC_FORMATS, > > + .sig_bits = 21, > > + }, > > + .probe = sun50i_dmic_soc_dai_probe, > > + .ops = &sun50i_dmic_dai_ops, > > + .name = "dmic", > > +}; > > + > > ... > >

1 0

[PATCH] ALSA: hda: intel-nhlt: Correct the handling of fmt_config flexible array
by Peter Ujfalusi 23 Aug '22

23 Aug '22

The struct nhlt_format's fmt_config is a flexible array, it must not be used as normal array. When moving to the next nhlt_fmt_cfg we need to take into account the data behind the ->config.caps (indicated by ->config.size). Fixes: a864e8f159b13 ("ALSA: hda: intel-nhlt: verify config type") Signed-off-by: Peter Ujfalusi <peter.ujfalusi(a)linux.intel.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> Reviewed-by: Jaska Uimonen <jaska.uimonen(a)linux.intel.com> --- sound/hda/intel-nhlt.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/sound/hda/intel-nhlt.c b/sound/hda/intel-nhlt.c index 9db5ccd9aa2d..13bb0ccfb36c 100644 --- a/sound/hda/intel-nhlt.c +++ b/sound/hda/intel-nhlt.c @@ -55,16 +55,22 @@ int intel_nhlt_get_dmic_geo(struct device *dev, struct nhlt_acpi_table *nhlt) /* find max number of channels based on format_configuration */ if (fmt_configs->fmt_count) { + struct nhlt_fmt_cfg *fmt_cfg = fmt_configs->fmt_config; + dev_dbg(dev, "found %d format definitions\n", fmt_configs->fmt_count); for (i = 0; i < fmt_configs->fmt_count; i++) { struct wav_fmt_ext *fmt_ext; - fmt_ext = &fmt_configs->fmt_config[i].fmt_ext; + fmt_ext = &fmt_cfg->fmt_ext; if (fmt_ext->fmt.channels > max_ch) max_ch = fmt_ext->fmt.channels; + + /* Move to the next nhlt_fmt_cfg */ + fmt_cfg = (struct nhlt_fmt_cfg *)(fmt_cfg->config.caps + + fmt_cfg->config.size); } dev_dbg(dev, "max channels found %d\n", max_ch); } else { -- 2.37.2

2 1

data-race in snd_seq_oss_midi_check_exit_port / snd_seq_oss_midi_setup
by Abhishek Shah 23 Aug '22

23 Aug '22

Hi all, We found a race involving the *max_midi_devs* variable. We see an interleaving where the following check here <https://elixir.bootlin.com/linux/v5.18-rc5/source/sound/core/seq/oss/seq_os…> passes before the *snd_seq_oss_midi_check_exit_port*() finishes, but this check should not pass if *max_midi_devs* will become zero, but we are not sure of its implications in terms of security impact. Please let us know what you think. Thanks! *-------------------Report---------------------* *write* to 0xffffffff88382f80 of 4 bytes by task 6541 on cpu 0: snd_seq_oss_midi_check_exit_port+0x1a6/0x270 sound/core/seq/oss/seq_oss_midi.c:237 receive_announce+0x193/0x1b0 sound/core/seq/oss/seq_oss_init.c:143 snd_seq_deliver_single_event+0x30d/0x4e0 sound/core/seq/seq_clientmgr.c:640 deliver_to_subscribers sound/core/seq/seq_clientmgr.c:695 [inline] snd_seq_deliver_event+0x38c/0x490 sound/core/seq/seq_clientmgr.c:830 snd_seq_kernel_client_dispatch+0x189/0x1a0 sound/core/seq/seq_clientmgr.c:2339 snd_seq_system_broadcast+0x98/0xd0 sound/core/seq/seq_system.c:86 snd_seq_ioctl_delete_port+0x9a/0xc0 sound/core/seq/seq_clientmgr.c:1356 snd_seq_ioctl+0x198/0x2d0 sound/core/seq/seq_clientmgr.c:2173 vfs_ioctl fs/ioctl.c:51 [inline] __do_sys_ioctl fs/ioctl.c:870 [inline] __se_sys_ioctl+0xe1/0x150 fs/ioctl.c:856 __x64_sys_ioctl+0x43/0x50 fs/ioctl.c:856 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x44/0xae *read* to 0xffffffff88382f80 of 4 bytes by task 6542 on cpu 1: snd_seq_oss_midi_setup+0x1b/0x40 sound/core/seq/oss/seq_oss_midi.c:273 snd_seq_oss_open+0x364/0x900 sound/core/seq/oss/seq_oss_init.c:198 odev_open+0x55/0x70 sound/core/seq/oss/seq_oss.c:128 soundcore_open+0x315/0x3a0 sound/sound_core.c:593 chrdev_open+0x373/0x3f0 fs/char_dev.c:414 do_dentry_open+0x543/0x8f0 fs/open.c:824 vfs_open+0x47/0x50 fs/open.c:958 do_open fs/namei.c:3476 [inline] path_openat+0x1906/0x1dc0 fs/namei.c:3609 do_filp_open+0xef/0x200 fs/namei.c:3636 do_sys_openat2+0xa5/0x2a0 fs/open.c:1213 do_sys_open fs/open.c:1229 [inline] __do_sys_openat fs/open.c:1245 [inline] __se_sys_openat fs/open.c:1240 [inline] __x64_sys_openat+0xf0/0x120 fs/open.c:1240 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x44/0xae Reported by Kernel Concurrency Sanitizer on: CPU: 1 PID: 6542 Comm: syz-executor2-n Not tainted 5.18.0-rc5+ #107 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014 *Reproducing Inputs* Input CPU 0: r0 = openat$sndseq(0xffffffffffffff9c, &(0x7f0000000040)='/dev/snd/seq\x00', 0x0) ioctl$SNDRV_SEQ_IOCTL_CREATE_PORT(r0, 0xc0a85320, &(0x7f0000000240)={{0x80}, 'port1\x00', 0x10}) ioctl$SNDRV_SEQ_IOCTL_SET_CLIENT_POOL(r0, 0x40a85321, &(0x7f0000000100)={0x80}) Input CPU 1: r0 = openat$sequencer2(0xffffff9c, &(0x7f0000000000)='/dev/sequencer2\x00', 0x0, 0x0) ioctl$SNDCTL_SYNTH_INFO(r0, 0xc08c5102, &(0x7f0000000200)={"02961a3ce6d4828f8b5559726313251b55fa11d8d65406f1f33c9af8e3f8", 0xffffffff})

3 4

[PATCH 0/5] ALSA: Drop hackish GFP giveaway for CONTINUOUS pages
by Takashi Iwai 23 Aug '22

23 Aug '22

Hi, this is a series of cleanup patches for dropping the current hackish way of passing the GFP_* flags for CONTINOUS and VMALLOC memory allocations. There are only three users for this legacy feature, and all of them seem superfluous. And, if any driver requires the memory restriction in future, it can now pass the proper device pointer for specifying the DMA mask. Takashi === Takashi Iwai (5): ALSA: vx: Drop superfluous GFP setup ALSA: pdaudiocf: Drop superfluous GFP setup ASoC: Intel: sst: Switch to standard device pages ALSA: memalloc: Drop special handling of GFP for CONTINUOUS allocation ALSA: doc: Drop snd_dma_continuous_data() usages .../kernel-api/writing-an-alsa-driver.rst | 21 ++-- include/sound/memalloc.h | 3 - sound/core/memalloc.c | 113 ++++++++---------- sound/drivers/vx/vx_pcm.c | 3 +- sound/pcmcia/pdaudiocf/pdaudiocf_pcm.c | 3 +- sound/soc/intel/atom/sst-mfld-platform-pcm.c | 7 +- 6 files changed, 61 insertions(+), 89 deletions(-) -- 2.35.3

2 6