September 2020 - Alsa-devel - mailman.alsa-project.org

[PATCH v4] ASoC: tlv320adcx140: Add support for configuring GPIO pin
by Camel Guo 18 Sep '20

18 Sep '20

From: Camel Guo <camelg(a)axis.com> Add support to configure the GPIO pin to the specific configuration. The GPIO pin can be configured as GPO, IRQ, SDOUT2, PDMCLK, MICBASE_EN, GPI, MCLK, SDIN, PDMDIN1, PDMDIN2, PDMDIN3 or PDMDIN4 and the output drive can be configured with various configuration. Signed-off-by: Camel Guo <camelg(a)axis.com> Acked-by: Dan Murphy <dmurphy(a)ti.com> --- v4: - Rebase and fix merge conflict - Add Acked-by from Dan v3: - Add ADCX140_NUM_GPIO_CFGS avoiding using magic number - Remove unneeded check on ret in adcx140_configure_gpio sound/soc/codecs/tlv320adcx140.c | 40 ++++++++++++++++++++++++++++++++ sound/soc/codecs/tlv320adcx140.h | 5 ++++ 2 files changed, 45 insertions(+) diff --git a/sound/soc/codecs/tlv320adcx140.c b/sound/soc/codecs/tlv320adcx140.c index 28dbd7d5e149..53a80246aee1 100644 --- a/sound/soc/codecs/tlv320adcx140.c +++ b/sound/soc/codecs/tlv320adcx140.c @@ -861,6 +861,42 @@ static int adcx140_configure_gpo(struct adcx140_priv *adcx140) } +static int adcx140_configure_gpio(struct adcx140_priv *adcx140) +{ + int gpio_count = 0; + u32 gpio_outputs[ADCX140_NUM_GPIO_CFGS]; + u32 gpio_output_val = 0; + int ret; + + gpio_count = device_property_count_u32(adcx140->dev, + "ti,gpio-config"); + if (gpio_count == 0) + return 0; + + if (gpio_count != ADCX140_NUM_GPIO_CFGS) + return -EINVAL; + + ret = device_property_read_u32_array(adcx140->dev, "ti,gpio-config", + gpio_outputs, gpio_count); + if (ret) + return ret; + + if (gpio_outputs[0] > ADCX140_GPIO_CFG_MAX) { + dev_err(adcx140->dev, "GPIO config out of range\n"); + return -EINVAL; + } + + if (gpio_outputs[1] > ADCX140_GPIO_DRV_MAX) { + dev_err(adcx140->dev, "GPIO drive out of range\n"); + return -EINVAL; + } + + gpio_output_val = gpio_outputs[0] << ADCX140_GPIO_SHIFT + | gpio_outputs[1]; + + return regmap_write(adcx140->regmap, ADCX140_GPIO_CFG0, gpio_output_val); +} + static int adcx140_codec_probe(struct snd_soc_component *component) { struct adcx140_priv *adcx140 = snd_soc_component_get_drvdata(component); @@ -958,6 +994,10 @@ static int adcx140_codec_probe(struct snd_soc_component *component) return ret; } + ret = adcx140_configure_gpio(adcx140); + if (ret) + return ret; + ret = adcx140_configure_gpo(adcx140); if (ret) goto out; diff --git a/sound/soc/codecs/tlv320adcx140.h b/sound/soc/codecs/tlv320adcx140.h index 107bd7927d9c..d7d4e3a88b5c 100644 --- a/sound/soc/codecs/tlv320adcx140.h +++ b/sound/soc/codecs/tlv320adcx140.h @@ -148,4 +148,9 @@ #define ADCX140_TX_FILL BIT(0) +#define ADCX140_NUM_GPIO_CFGS 2 +#define ADCX140_GPIO_SHIFT 4 +#define ADCX140_GPIO_CFG_MAX 15 +#define ADCX140_GPIO_DRV_MAX 5 + #endif /* _TLV320ADCX140_ */ -- 2.20.1

3 3

[PATCH v4 1/2] dt-bindings: tlv320adcx140: Add GPIO config and drive config
by Camel Guo 18 Sep '20

18 Sep '20

From: Camel Guo <camelg(a)axis.com> Add properties for configuring the General Purpose Input Output (GPIO). There are 2 settings for GPIO, configuration and the output drive type. Signed-off-by: Camel Guo <camelg(a)axis.com> Acked-by: Dan Murphy <dmurphy(a)ti.com> --- v4: - Rebase v3: - Fix typo - Add Acked-By from Dan .../bindings/sound/tlv320adcx140.yaml | 44 +++++++++++++++++++ 1 file changed, 44 insertions(+) diff --git a/Documentation/devicetree/bindings/sound/tlv320adcx140.yaml b/Documentation/devicetree/bindings/sound/tlv320adcx140.yaml index e79f8d1891e4..798b366fe6d1 100644 --- a/Documentation/devicetree/bindings/sound/tlv320adcx140.yaml +++ b/Documentation/devicetree/bindings/sound/tlv320adcx140.yaml @@ -140,6 +140,49 @@ patternProperties: 4d - Drive weak low and active high 5d - Drive Hi-Z and active high + ti,gpio-config: + description: | + Defines the configuration and output drive for the General Purpose + Input and Output pin (GPIO1). Its value is a pair, the first value is for + the configuration type and the second value is for the output drive + type. The array is defined as <GPIO1_CFG GPIO1_DRV> + + configuration for the GPIO pin can be one of the following: + 0 - disabled + 1 - GPIO1 is configured as a general-purpose output (GPO) + 2 - (default) GPIO1 is configured as a device interrupt output (IRQ) + 3 - GPIO1 is configured as a secondary ASI output (SDOUT2) + 4 - GPIO1 is configured as a PDM clock output (PDMCLK) + 8 - GPIO1 is configured as an input to control when MICBIAS turns on or + off (MICBIAS_EN) + 9 - GPIO1 is configured as a general-purpose input (GPI) + 10 - GPIO1 is configured as a master clock input (MCLK) + 11 - GPIO1 is configured as an ASI input for daisy-chain (SDIN) + 12 - GPIO1 is configured as a PDM data input for channel 1 and channel 2 + (PDMDIN1) + 13 - GPIO1 is configured as a PDM data input for channel 3 and channel 4 + (PDMDIN2) + 14 - GPIO1 is configured as a PDM data input for channel 5 and channel 6 + (PDMDIN3) + 15 - GPIO1 is configured as a PDM data input for channel 7 and channel 8 + (PDMDIN4) + + output drive type for the GPIO pin can be one of the following: + 0 - Hi-Z output + 1 - Drive active low and active high + 2 - (default) Drive active low and weak high + 3 - Drive active low and Hi-Z + 4 - Drive weak low and active high + 5 - Drive Hi-Z and active high + + allOf: + - $ref: /schemas/types.yaml#/definitions/uint32-array + - minItems: 2 + maxItems: 2 + items: + maximum: 15 + default: [2, 2] + required: - compatible - reg @@ -156,6 +199,7 @@ examples: ti,mic-bias-source = <6>; ti,pdm-edge-select = <0 1 0 1>; ti,gpi-config = <4 5 6 7>; + ti,gpio-config = <10 2>; ti,gpo-config-1 = <0 0>; ti,gpo-config-2 = <0 0>; reset-gpios = <&gpio0 14 GPIO_ACTIVE_HIGH>; -- 2.20.1

1 0

aplay: playing mono audio file alteres pulseaudio default Master left or right volume level
by GitHub issues - edited 18 Sep '20

18 Sep '20

alsa-project/alsa-utils issue #55 was edited from maxmbed: I am encountering an unexpected behaviour with pulse audio master volume in my application. ### Description If I set for example left channel to 100% and right channel 0%, and play mono audio file (i.e. audio file with 1 channel only), then I would expect to hear only the audio file on left channel. The actual behaviour in the example of above is when playing the audio file, the Master volume right channel will be automatically adjusted to 50%. If play again it increases to 75% and so on. This issue happens also with other R/L volume level, the case of above is just an example. **Important to note the case does not happen with stereo audio file** where audio sounds on left only channel as we can expected. ### Steps to repeat issue: (a) Set mixer Lch to 100%, Rch to 0% ``` $ amixer -D default sset Master 65535,0 Simple mixer control 'Master',0 Capabilities: pvolume pswitch pswitch-joined Playback channels: Front Left - Front Right Limits: Playback 0 - 65536 Mono: Front Left: Playback 65535 [100%] [on] Front Right: Playback 0 [0%] [on] ``` (b) Play mono audio file ``` $ aplay -D default ringtone.wav Playing WAVE 'ringtone.wav' : Signed 16 bit Little Endian, Rate 44100 Hz, Mono ``` ### Request (1) Is it by design expected ? (2) Is there a way to prevent this phenomenon (command option) ? Issue URL : https://github.com/alsa-project/alsa-utils/issues/55 Repository URL: https://github.com/alsa-project/alsa-utils

1 0

[PATCH AUTOSEL 4.4 52/64] ALSA: hda: Fix potential race in unsol event handler
by Sasha Levin 18 Sep '20

18 Sep '20

From: Takashi Iwai <tiwai(a)suse.de> [ Upstream commit c637fa151259c0f74665fde7cba5b7eac1417ae5 ] The unsol event handling code has a loop retrieving the read/write indices and the arrays without locking while the append to the array may happen concurrently. This may lead to some inconsistency. Although there hasn't been any proof of this bad results, it's still safer to protect the racy accesses. This patch adds the spinlock protection around the unsol handling loop for addressing it. Here we take bus->reg_lock as the writer side snd_hdac_bus_queue_event() is also protected by that lock. Link: https://lore.kernel.org/r/20200516062556.30951-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/hda/hdac_bus.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/sound/hda/hdac_bus.c b/sound/hda/hdac_bus.c index 0e81ea89a5965..e3f68a76d90eb 100644 --- a/sound/hda/hdac_bus.c +++ b/sound/hda/hdac_bus.c @@ -155,6 +155,7 @@ static void process_unsol_events(struct work_struct *work) struct hdac_driver *drv; unsigned int rp, caddr, res; + spin_lock_irq(&bus->reg_lock); while (bus->unsol_rp != bus->unsol_wp) { rp = (bus->unsol_rp + 1) % HDA_UNSOL_QUEUE_SIZE; bus->unsol_rp = rp; @@ -166,10 +167,13 @@ static void process_unsol_events(struct work_struct *work) codec = bus->caddr_tbl[caddr & 0x0f]; if (!codec || !codec->dev.driver) continue; + spin_unlock_irq(&bus->reg_lock); drv = drv_to_hdac_driver(codec->dev.driver); if (drv->unsol_event) drv->unsol_event(codec, res); + spin_lock_irq(&bus->reg_lock); } + spin_unlock_irq(&bus->reg_lock); } /** -- 2.25.1

1 0

[PATCH AUTOSEL 4.4 39/64] ALSA: usb-audio: Fix case when USB MIDI interface has more than one extra endpoint descriptor
by Sasha Levin 18 Sep '20

18 Sep '20

From: Andreas Steinmetz <ast(a)domdv.de> [ Upstream commit 5c6cd7021a05a02fcf37f360592d7c18d4d807fb ] The Miditech MIDIFACE 16x16 (USB ID 1290:1749) has more than one extra endpoint descriptor. The first extra descriptor is: 0x06 0x30 0x00 0x00 0x00 0x00 As the code in snd_usbmidi_get_ms_info() looks only at the first extra descriptor to find USB_DT_CS_ENDPOINT the device as such is recognized but there is neither input nor output configured. The patch iterates through the extra descriptors to find the proper one. With this patch the device is correctly configured. Signed-off-by: Andreas Steinmetz <ast(a)domdv.de> Link: https://lore.kernel.org/r/1c3b431a86f69e1d60745b6110cdb93c299f120b.camel@do… Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/usb/midi.c | 29 ++++++++++++++++++++++++----- 1 file changed, 24 insertions(+), 5 deletions(-) diff --git a/sound/usb/midi.c b/sound/usb/midi.c index 5c4a3d6c42341..934540042bc2e 100644 --- a/sound/usb/midi.c +++ b/sound/usb/midi.c @@ -1803,6 +1803,28 @@ static int snd_usbmidi_create_endpoints(struct snd_usb_midi *umidi, return 0; } +static struct usb_ms_endpoint_descriptor *find_usb_ms_endpoint_descriptor( + struct usb_host_endpoint *hostep) +{ + unsigned char *extra = hostep->extra; + int extralen = hostep->extralen; + + while (extralen > 3) { + struct usb_ms_endpoint_descriptor *ms_ep = + (struct usb_ms_endpoint_descriptor *)extra; + + if (ms_ep->bLength > 3 && + ms_ep->bDescriptorType == USB_DT_CS_ENDPOINT && + ms_ep->bDescriptorSubtype == UAC_MS_GENERAL) + return ms_ep; + if (!extra[0]) + break; + extralen -= extra[0]; + extra += extra[0]; + } + return NULL; +} + /* * Returns MIDIStreaming device capabilities. */ @@ -1840,11 +1862,8 @@ static int snd_usbmidi_get_ms_info(struct snd_usb_midi *umidi, ep = get_ep_desc(hostep); if (!usb_endpoint_xfer_bulk(ep) && !usb_endpoint_xfer_int(ep)) continue; - ms_ep = (struct usb_ms_endpoint_descriptor *)hostep->extra; - if (hostep->extralen < 4 || - ms_ep->bLength < 4 || - ms_ep->bDescriptorType != USB_DT_CS_ENDPOINT || - ms_ep->bDescriptorSubtype != UAC_MS_GENERAL) + ms_ep = find_usb_ms_endpoint_descriptor(hostep); + if (!ms_ep) continue; if (usb_endpoint_dir_out(ep)) { if (endpoints[epidx].out_ep) { -- 2.25.1

1 0

[PATCH AUTOSEL 4.4 04/64] ASoC: kirkwood: fix IRQ error handling
by Sasha Levin 18 Sep '20

18 Sep '20

From: Russell King <rmk+kernel(a)armlinux.org.uk> [ Upstream commit 175fc928198236037174e5c5c066fe3c4691903e ] Propagate the error code from request_irq(), rather than returning -EBUSY. Signed-off-by: Russell King <rmk+kernel(a)armlinux.org.uk> Link: https://lore.kernel.org/r/E1iNIqh-0000tW-EZ@rmk-PC.armlinux.org.uk Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/kirkwood/kirkwood-dma.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/kirkwood/kirkwood-dma.c b/sound/soc/kirkwood/kirkwood-dma.c index dbfdfe99c69df..231c7d97333c7 100644 --- a/sound/soc/kirkwood/kirkwood-dma.c +++ b/sound/soc/kirkwood/kirkwood-dma.c @@ -136,7 +136,7 @@ static int kirkwood_dma_open(struct snd_pcm_substream *substream) err = request_irq(priv->irq, kirkwood_dma_irq, IRQF_SHARED, "kirkwood-i2s", priv); if (err) - return -EBUSY; + return err; /* * Enable Error interrupts. We're only ack'ing them but -- 2.25.1

1 0

[PATCH AUTOSEL 4.9 74/90] ALSA: hda: Fix potential race in unsol event handler
by Sasha Levin 18 Sep '20

18 Sep '20

From: Takashi Iwai <tiwai(a)suse.de> [ Upstream commit c637fa151259c0f74665fde7cba5b7eac1417ae5 ] The unsol event handling code has a loop retrieving the read/write indices and the arrays without locking while the append to the array may happen concurrently. This may lead to some inconsistency. Although there hasn't been any proof of this bad results, it's still safer to protect the racy accesses. This patch adds the spinlock protection around the unsol handling loop for addressing it. Here we take bus->reg_lock as the writer side snd_hdac_bus_queue_event() is also protected by that lock. Link: https://lore.kernel.org/r/20200516062556.30951-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/hda/hdac_bus.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/sound/hda/hdac_bus.c b/sound/hda/hdac_bus.c index 0e81ea89a5965..e3f68a76d90eb 100644 --- a/sound/hda/hdac_bus.c +++ b/sound/hda/hdac_bus.c @@ -155,6 +155,7 @@ static void process_unsol_events(struct work_struct *work) struct hdac_driver *drv; unsigned int rp, caddr, res; + spin_lock_irq(&bus->reg_lock); while (bus->unsol_rp != bus->unsol_wp) { rp = (bus->unsol_rp + 1) % HDA_UNSOL_QUEUE_SIZE; bus->unsol_rp = rp; @@ -166,10 +167,13 @@ static void process_unsol_events(struct work_struct *work) codec = bus->caddr_tbl[caddr & 0x0f]; if (!codec || !codec->dev.driver) continue; + spin_unlock_irq(&bus->reg_lock); drv = drv_to_hdac_driver(codec->dev.driver); if (drv->unsol_event) drv->unsol_event(codec, res); + spin_lock_irq(&bus->reg_lock); } + spin_unlock_irq(&bus->reg_lock); } /** -- 2.25.1

1 0

[PATCH AUTOSEL 4.9 61/90] ALSA: usb-audio: Fix case when USB MIDI interface has more than one extra endpoint descriptor
by Sasha Levin 18 Sep '20

18 Sep '20

From: Andreas Steinmetz <ast(a)domdv.de> [ Upstream commit 5c6cd7021a05a02fcf37f360592d7c18d4d807fb ] The Miditech MIDIFACE 16x16 (USB ID 1290:1749) has more than one extra endpoint descriptor. The first extra descriptor is: 0x06 0x30 0x00 0x00 0x00 0x00 As the code in snd_usbmidi_get_ms_info() looks only at the first extra descriptor to find USB_DT_CS_ENDPOINT the device as such is recognized but there is neither input nor output configured. The patch iterates through the extra descriptors to find the proper one. With this patch the device is correctly configured. Signed-off-by: Andreas Steinmetz <ast(a)domdv.de> Link: https://lore.kernel.org/r/1c3b431a86f69e1d60745b6110cdb93c299f120b.camel@do… Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/usb/midi.c | 29 ++++++++++++++++++++++++----- 1 file changed, 24 insertions(+), 5 deletions(-) diff --git a/sound/usb/midi.c b/sound/usb/midi.c index 0676e7d485def..b8d4b5b3e54a1 100644 --- a/sound/usb/midi.c +++ b/sound/usb/midi.c @@ -1805,6 +1805,28 @@ static int snd_usbmidi_create_endpoints(struct snd_usb_midi *umidi, return 0; } +static struct usb_ms_endpoint_descriptor *find_usb_ms_endpoint_descriptor( + struct usb_host_endpoint *hostep) +{ + unsigned char *extra = hostep->extra; + int extralen = hostep->extralen; + + while (extralen > 3) { + struct usb_ms_endpoint_descriptor *ms_ep = + (struct usb_ms_endpoint_descriptor *)extra; + + if (ms_ep->bLength > 3 && + ms_ep->bDescriptorType == USB_DT_CS_ENDPOINT && + ms_ep->bDescriptorSubtype == UAC_MS_GENERAL) + return ms_ep; + if (!extra[0]) + break; + extralen -= extra[0]; + extra += extra[0]; + } + return NULL; +} + /* * Returns MIDIStreaming device capabilities. */ @@ -1842,11 +1864,8 @@ static int snd_usbmidi_get_ms_info(struct snd_usb_midi *umidi, ep = get_ep_desc(hostep); if (!usb_endpoint_xfer_bulk(ep) && !usb_endpoint_xfer_int(ep)) continue; - ms_ep = (struct usb_ms_endpoint_descriptor *)hostep->extra; - if (hostep->extralen < 4 || - ms_ep->bLength < 4 || - ms_ep->bDescriptorType != USB_DT_CS_ENDPOINT || - ms_ep->bDescriptorSubtype != UAC_MS_GENERAL) + ms_ep = find_usb_ms_endpoint_descriptor(hostep); + if (!ms_ep) continue; if (usb_endpoint_dir_out(ep)) { if (endpoints[epidx].out_ep) { -- 2.25.1

1 0

[PATCH AUTOSEL 4.9 32/90] ALSA: hda: Clear RIRB status before reading WP
by Sasha Levin 18 Sep '20

18 Sep '20

From: Mohan Kumar <mkumard(a)nvidia.com> [ Upstream commit 6d011d5057ff88ee556c000ac6fe0be23bdfcd72 ] RIRB interrupt status getting cleared after the write pointer is read causes a race condition, where last response(s) into RIRB may remain unserviced by IRQ, eventually causing azx_rirb_get_response to fall back to polling mode. Clearing the RIRB interrupt status ahead of write pointer access ensures that this condition is avoided. Signed-off-by: Mohan Kumar <mkumard(a)nvidia.com> Signed-off-by: Viswanath L <viswanathl(a)nvidia.com> Link: https://lore.kernel.org/r/1580983853-351-1-git-send-email-viswanathl@nvidia… Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/pci/hda/hda_controller.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/sound/pci/hda/hda_controller.c b/sound/pci/hda/hda_controller.c index bd0e4710d15d7..79043b481d7b6 100644 --- a/sound/pci/hda/hda_controller.c +++ b/sound/pci/hda/hda_controller.c @@ -1158,16 +1158,23 @@ irqreturn_t azx_interrupt(int irq, void *dev_id) if (snd_hdac_bus_handle_stream_irq(bus, status, stream_update)) active = true; - /* clear rirb int */ status = azx_readb(chip, RIRBSTS); if (status & RIRB_INT_MASK) { + /* + * Clearing the interrupt status here ensures that no + * interrupt gets masked after the RIRB wp is read in + * snd_hdac_bus_update_rirb. This avoids a possible + * race condition where codec response in RIRB may + * remain unserviced by IRQ, eventually falling back + * to polling mode in azx_rirb_get_response. + */ + azx_writeb(chip, RIRBSTS, RIRB_INT_MASK); active = true; if (status & RIRB_INT_RESPONSE) { if (chip->driver_caps & AZX_DCAPS_CTX_WORKAROUND) udelay(80); snd_hdac_bus_update_rirb(bus); } - azx_writeb(chip, RIRBSTS, RIRB_INT_MASK); } } while (active && ++repeat < 10); -- 2.25.1

1 0

[PATCH AUTOSEL 4.9 04/90] ASoC: kirkwood: fix IRQ error handling
by Sasha Levin 18 Sep '20

18 Sep '20

From: Russell King <rmk+kernel(a)armlinux.org.uk> [ Upstream commit 175fc928198236037174e5c5c066fe3c4691903e ] Propagate the error code from request_irq(), rather than returning -EBUSY. Signed-off-by: Russell King <rmk+kernel(a)armlinux.org.uk> Link: https://lore.kernel.org/r/E1iNIqh-0000tW-EZ@rmk-PC.armlinux.org.uk Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/kirkwood/kirkwood-dma.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/kirkwood/kirkwood-dma.c b/sound/soc/kirkwood/kirkwood-dma.c index dafd22e874e99..e655425e4819e 100644 --- a/sound/soc/kirkwood/kirkwood-dma.c +++ b/sound/soc/kirkwood/kirkwood-dma.c @@ -136,7 +136,7 @@ static int kirkwood_dma_open(struct snd_pcm_substream *substream) err = request_irq(priv->irq, kirkwood_dma_irq, IRQF_SHARED, "kirkwood-i2s", priv); if (err) - return -EBUSY; + return err; /* * Enable Error interrupts. We're only ack'ing them but -- 2.25.1

1 0

[PATCH AUTOSEL 4.14 101/127] ALSA: hda: Fix potential race in unsol event handler
by Sasha Levin 18 Sep '20

18 Sep '20

From: Takashi Iwai <tiwai(a)suse.de> [ Upstream commit c637fa151259c0f74665fde7cba5b7eac1417ae5 ] The unsol event handling code has a loop retrieving the read/write indices and the arrays without locking while the append to the array may happen concurrently. This may lead to some inconsistency. Although there hasn't been any proof of this bad results, it's still safer to protect the racy accesses. This patch adds the spinlock protection around the unsol handling loop for addressing it. Here we take bus->reg_lock as the writer side snd_hdac_bus_queue_event() is also protected by that lock. Link: https://lore.kernel.org/r/20200516062556.30951-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/hda/hdac_bus.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/sound/hda/hdac_bus.c b/sound/hda/hdac_bus.c index 714a51721a313..ab9236e4c157e 100644 --- a/sound/hda/hdac_bus.c +++ b/sound/hda/hdac_bus.c @@ -155,6 +155,7 @@ static void process_unsol_events(struct work_struct *work) struct hdac_driver *drv; unsigned int rp, caddr, res; + spin_lock_irq(&bus->reg_lock); while (bus->unsol_rp != bus->unsol_wp) { rp = (bus->unsol_rp + 1) % HDA_UNSOL_QUEUE_SIZE; bus->unsol_rp = rp; @@ -166,10 +167,13 @@ static void process_unsol_events(struct work_struct *work) codec = bus->caddr_tbl[caddr & 0x0f]; if (!codec || !codec->dev.driver) continue; + spin_unlock_irq(&bus->reg_lock); drv = drv_to_hdac_driver(codec->dev.driver); if (drv->unsol_event) drv->unsol_event(codec, res); + spin_lock_irq(&bus->reg_lock); } + spin_unlock_irq(&bus->reg_lock); } /** -- 2.25.1

1 0

[PATCH AUTOSEL 4.14 079/127] ALSA: usb-audio: Fix case when USB MIDI interface has more than one extra endpoint descriptor
by Sasha Levin 18 Sep '20

18 Sep '20

From: Andreas Steinmetz <ast(a)domdv.de> [ Upstream commit 5c6cd7021a05a02fcf37f360592d7c18d4d807fb ] The Miditech MIDIFACE 16x16 (USB ID 1290:1749) has more than one extra endpoint descriptor. The first extra descriptor is: 0x06 0x30 0x00 0x00 0x00 0x00 As the code in snd_usbmidi_get_ms_info() looks only at the first extra descriptor to find USB_DT_CS_ENDPOINT the device as such is recognized but there is neither input nor output configured. The patch iterates through the extra descriptors to find the proper one. With this patch the device is correctly configured. Signed-off-by: Andreas Steinmetz <ast(a)domdv.de> Link: https://lore.kernel.org/r/1c3b431a86f69e1d60745b6110cdb93c299f120b.camel@do… Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/usb/midi.c | 29 ++++++++++++++++++++++++----- 1 file changed, 24 insertions(+), 5 deletions(-) diff --git a/sound/usb/midi.c b/sound/usb/midi.c index 1bfae7a1c32f1..a3d1c0c1b4a67 100644 --- a/sound/usb/midi.c +++ b/sound/usb/midi.c @@ -1805,6 +1805,28 @@ static int snd_usbmidi_create_endpoints(struct snd_usb_midi *umidi, return 0; } +static struct usb_ms_endpoint_descriptor *find_usb_ms_endpoint_descriptor( + struct usb_host_endpoint *hostep) +{ + unsigned char *extra = hostep->extra; + int extralen = hostep->extralen; + + while (extralen > 3) { + struct usb_ms_endpoint_descriptor *ms_ep = + (struct usb_ms_endpoint_descriptor *)extra; + + if (ms_ep->bLength > 3 && + ms_ep->bDescriptorType == USB_DT_CS_ENDPOINT && + ms_ep->bDescriptorSubtype == UAC_MS_GENERAL) + return ms_ep; + if (!extra[0]) + break; + extralen -= extra[0]; + extra += extra[0]; + } + return NULL; +} + /* * Returns MIDIStreaming device capabilities. */ @@ -1842,11 +1864,8 @@ static int snd_usbmidi_get_ms_info(struct snd_usb_midi *umidi, ep = get_ep_desc(hostep); if (!usb_endpoint_xfer_bulk(ep) && !usb_endpoint_xfer_int(ep)) continue; - ms_ep = (struct usb_ms_endpoint_descriptor *)hostep->extra; - if (hostep->extralen < 4 || - ms_ep->bLength < 4 || - ms_ep->bDescriptorType != USB_DT_CS_ENDPOINT || - ms_ep->bDescriptorSubtype != UAC_MS_GENERAL) + ms_ep = find_usb_ms_endpoint_descriptor(hostep); + if (!ms_ep) continue; if (usb_endpoint_dir_out(ep)) { if (endpoints[epidx].out_ep) { -- 2.25.1

1 0

[PATCH AUTOSEL 4.14 041/127] ALSA: hda: Clear RIRB status before reading WP
by Sasha Levin 18 Sep '20

18 Sep '20

From: Mohan Kumar <mkumard(a)nvidia.com> [ Upstream commit 6d011d5057ff88ee556c000ac6fe0be23bdfcd72 ] RIRB interrupt status getting cleared after the write pointer is read causes a race condition, where last response(s) into RIRB may remain unserviced by IRQ, eventually causing azx_rirb_get_response to fall back to polling mode. Clearing the RIRB interrupt status ahead of write pointer access ensures that this condition is avoided. Signed-off-by: Mohan Kumar <mkumard(a)nvidia.com> Signed-off-by: Viswanath L <viswanathl(a)nvidia.com> Link: https://lore.kernel.org/r/1580983853-351-1-git-send-email-viswanathl@nvidia… Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/pci/hda/hda_controller.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/sound/pci/hda/hda_controller.c b/sound/pci/hda/hda_controller.c index fa261b27d8588..8198d2e53b7df 100644 --- a/sound/pci/hda/hda_controller.c +++ b/sound/pci/hda/hda_controller.c @@ -1169,16 +1169,23 @@ irqreturn_t azx_interrupt(int irq, void *dev_id) if (snd_hdac_bus_handle_stream_irq(bus, status, stream_update)) active = true; - /* clear rirb int */ status = azx_readb(chip, RIRBSTS); if (status & RIRB_INT_MASK) { + /* + * Clearing the interrupt status here ensures that no + * interrupt gets masked after the RIRB wp is read in + * snd_hdac_bus_update_rirb. This avoids a possible + * race condition where codec response in RIRB may + * remain unserviced by IRQ, eventually falling back + * to polling mode in azx_rirb_get_response. + */ + azx_writeb(chip, RIRBSTS, RIRB_INT_MASK); active = true; if (status & RIRB_INT_RESPONSE) { if (chip->driver_caps & AZX_DCAPS_CTX_WORKAROUND) udelay(80); snd_hdac_bus_update_rirb(bus); } - azx_writeb(chip, RIRBSTS, RIRB_INT_MASK); } } while (active && ++repeat < 10); -- 2.25.1

1 0

[PATCH AUTOSEL 4.14 005/127] ASoC: kirkwood: fix IRQ error handling
by Sasha Levin 18 Sep '20

18 Sep '20

From: Russell King <rmk+kernel(a)armlinux.org.uk> [ Upstream commit 175fc928198236037174e5c5c066fe3c4691903e ] Propagate the error code from request_irq(), rather than returning -EBUSY. Signed-off-by: Russell King <rmk+kernel(a)armlinux.org.uk> Link: https://lore.kernel.org/r/E1iNIqh-0000tW-EZ@rmk-PC.armlinux.org.uk Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/kirkwood/kirkwood-dma.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/kirkwood/kirkwood-dma.c b/sound/soc/kirkwood/kirkwood-dma.c index cf23af159acf4..35ca8e8bb5e52 100644 --- a/sound/soc/kirkwood/kirkwood-dma.c +++ b/sound/soc/kirkwood/kirkwood-dma.c @@ -136,7 +136,7 @@ static int kirkwood_dma_open(struct snd_pcm_substream *substream) err = request_irq(priv->irq, kirkwood_dma_irq, IRQF_SHARED, "kirkwood-i2s", priv); if (err) - return -EBUSY; + return err; /* * Enable Error interrupts. We're only ack'ing them but -- 2.25.1

1 0

[PATCH AUTOSEL 4.19 185/206] ASoC: img-i2s-out: Fix runtime PM imbalance on error
by Sasha Levin 18 Sep '20

18 Sep '20

From: Dinghao Liu <dinghao.liu(a)zju.edu.cn> [ Upstream commit 65bd91dd6957390c42a0491b9622cf31a2cdb140 ] pm_runtime_get_sync() increments the runtime PM usage counter even the call returns an error code. Thus a pairing decrement is needed on the error handling path to keep the counter balanced. Signed-off-by: Dinghao Liu <dinghao.liu(a)zju.edu.cn> Link: https://lore.kernel.org/r/20200529012230.5863-1-dinghao.liu@zju.edu.cn Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/img/img-i2s-out.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/sound/soc/img/img-i2s-out.c b/sound/soc/img/img-i2s-out.c index fc2d1dac63339..798ab579564cb 100644 --- a/sound/soc/img/img-i2s-out.c +++ b/sound/soc/img/img-i2s-out.c @@ -350,8 +350,10 @@ static int img_i2s_out_set_fmt(struct snd_soc_dai *dai, unsigned int fmt) chan_control_mask = IMG_I2S_OUT_CHAN_CTL_CLKT_MASK; ret = pm_runtime_get_sync(i2s->dev); - if (ret < 0) + if (ret < 0) { + pm_runtime_put_noidle(i2s->dev); return ret; + } img_i2s_out_disable(i2s); @@ -491,8 +493,10 @@ static int img_i2s_out_probe(struct platform_device *pdev) goto err_pm_disable; } ret = pm_runtime_get_sync(&pdev->dev); - if (ret < 0) + if (ret < 0) { + pm_runtime_put_noidle(&pdev->dev); goto err_suspend; + } reg = IMG_I2S_OUT_CTL_FRM_SIZE_MASK; img_i2s_out_writel(i2s, reg, IMG_I2S_OUT_CTL); -- 2.25.1

1 0

[PATCH AUTOSEL 4.19 164/206] ALSA: hda: Fix potential race in unsol event handler
by Sasha Levin 18 Sep '20

18 Sep '20

From: Takashi Iwai <tiwai(a)suse.de> [ Upstream commit c637fa151259c0f74665fde7cba5b7eac1417ae5 ] The unsol event handling code has a loop retrieving the read/write indices and the arrays without locking while the append to the array may happen concurrently. This may lead to some inconsistency. Although there hasn't been any proof of this bad results, it's still safer to protect the racy accesses. This patch adds the spinlock protection around the unsol handling loop for addressing it. Here we take bus->reg_lock as the writer side snd_hdac_bus_queue_event() is also protected by that lock. Link: https://lore.kernel.org/r/20200516062556.30951-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/hda/hdac_bus.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/sound/hda/hdac_bus.c b/sound/hda/hdac_bus.c index 714a51721a313..ab9236e4c157e 100644 --- a/sound/hda/hdac_bus.c +++ b/sound/hda/hdac_bus.c @@ -155,6 +155,7 @@ static void process_unsol_events(struct work_struct *work) struct hdac_driver *drv; unsigned int rp, caddr, res; + spin_lock_irq(&bus->reg_lock); while (bus->unsol_rp != bus->unsol_wp) { rp = (bus->unsol_rp + 1) % HDA_UNSOL_QUEUE_SIZE; bus->unsol_rp = rp; @@ -166,10 +167,13 @@ static void process_unsol_events(struct work_struct *work) codec = bus->caddr_tbl[caddr & 0x0f]; if (!codec || !codec->dev.driver) continue; + spin_unlock_irq(&bus->reg_lock); drv = drv_to_hdac_driver(codec->dev.driver); if (drv->unsol_event) drv->unsol_event(codec, res); + spin_lock_irq(&bus->reg_lock); } + spin_unlock_irq(&bus->reg_lock); } /** -- 2.25.1

1 0

[PATCH AUTOSEL 4.19 134/206] ALSA: usb-audio: Fix case when USB MIDI interface has more than one extra endpoint descriptor
by Sasha Levin 18 Sep '20

18 Sep '20

From: Andreas Steinmetz <ast(a)domdv.de> [ Upstream commit 5c6cd7021a05a02fcf37f360592d7c18d4d807fb ] The Miditech MIDIFACE 16x16 (USB ID 1290:1749) has more than one extra endpoint descriptor. The first extra descriptor is: 0x06 0x30 0x00 0x00 0x00 0x00 As the code in snd_usbmidi_get_ms_info() looks only at the first extra descriptor to find USB_DT_CS_ENDPOINT the device as such is recognized but there is neither input nor output configured. The patch iterates through the extra descriptors to find the proper one. With this patch the device is correctly configured. Signed-off-by: Andreas Steinmetz <ast(a)domdv.de> Link: https://lore.kernel.org/r/1c3b431a86f69e1d60745b6110cdb93c299f120b.camel@do… Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/usb/midi.c | 29 ++++++++++++++++++++++++----- 1 file changed, 24 insertions(+), 5 deletions(-) diff --git a/sound/usb/midi.c b/sound/usb/midi.c index 28a3ad8b1d74b..137e1e8718d6f 100644 --- a/sound/usb/midi.c +++ b/sound/usb/midi.c @@ -1828,6 +1828,28 @@ static int snd_usbmidi_create_endpoints(struct snd_usb_midi *umidi, return 0; } +static struct usb_ms_endpoint_descriptor *find_usb_ms_endpoint_descriptor( + struct usb_host_endpoint *hostep) +{ + unsigned char *extra = hostep->extra; + int extralen = hostep->extralen; + + while (extralen > 3) { + struct usb_ms_endpoint_descriptor *ms_ep = + (struct usb_ms_endpoint_descriptor *)extra; + + if (ms_ep->bLength > 3 && + ms_ep->bDescriptorType == USB_DT_CS_ENDPOINT && + ms_ep->bDescriptorSubtype == UAC_MS_GENERAL) + return ms_ep; + if (!extra[0]) + break; + extralen -= extra[0]; + extra += extra[0]; + } + return NULL; +} + /* * Returns MIDIStreaming device capabilities. */ @@ -1865,11 +1887,8 @@ static int snd_usbmidi_get_ms_info(struct snd_usb_midi *umidi, ep = get_ep_desc(hostep); if (!usb_endpoint_xfer_bulk(ep) && !usb_endpoint_xfer_int(ep)) continue; - ms_ep = (struct usb_ms_endpoint_descriptor *)hostep->extra; - if (hostep->extralen < 4 || - ms_ep->bLength < 4 || - ms_ep->bDescriptorType != USB_DT_CS_ENDPOINT || - ms_ep->bDescriptorSubtype != UAC_MS_GENERAL) + ms_ep = find_usb_ms_endpoint_descriptor(hostep); + if (!ms_ep) continue; if (usb_endpoint_dir_out(ep)) { if (endpoints[epidx].out_ep) { -- 2.25.1

1 0

[PATCH AUTOSEL 4.19 080/206] ALSA: usb-audio: Don't create a mixer element with bogus volume range
by Sasha Levin 18 Sep '20

18 Sep '20

From: Takashi Iwai <tiwai(a)suse.de> [ Upstream commit e9a0ef0b5ddcbc0d56c65aefc0f18d16e6f71207 ] Some USB-audio descriptors provide a bogus volume range (e.g. volume min and max are identical), which confuses user-space. This patch makes the driver skipping such a control element. BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=206221 Link: https://lore.kernel.org/r/20200214144928.23628-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/usb/mixer.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/sound/usb/mixer.c b/sound/usb/mixer.c index 45bd3d54be54b..451b8ea383c61 100644 --- a/sound/usb/mixer.c +++ b/sound/usb/mixer.c @@ -1699,6 +1699,16 @@ static void __build_feature_ctl(struct usb_mixer_interface *mixer, /* get min/max values */ get_min_max_with_quirks(cval, 0, kctl); + /* skip a bogus volume range */ + if (cval->max <= cval->min) { + usb_audio_dbg(mixer->chip, + "[%d] FU [%s] skipped due to invalid volume\n", + cval->head.id, kctl->id.name); + snd_ctl_free_one(kctl); + return; + } + + if (control == UAC_FU_VOLUME) { check_mapped_dB(map, cval); if (cval->dBmin < cval->dBmax || !cval->initialized) { -- 2.25.1

1 0

[PATCH AUTOSEL 4.19 070/206] ALSA: hda: Clear RIRB status before reading WP
by Sasha Levin 18 Sep '20

18 Sep '20

From: Mohan Kumar <mkumard(a)nvidia.com> [ Upstream commit 6d011d5057ff88ee556c000ac6fe0be23bdfcd72 ] RIRB interrupt status getting cleared after the write pointer is read causes a race condition, where last response(s) into RIRB may remain unserviced by IRQ, eventually causing azx_rirb_get_response to fall back to polling mode. Clearing the RIRB interrupt status ahead of write pointer access ensures that this condition is avoided. Signed-off-by: Mohan Kumar <mkumard(a)nvidia.com> Signed-off-by: Viswanath L <viswanathl(a)nvidia.com> Link: https://lore.kernel.org/r/1580983853-351-1-git-send-email-viswanathl@nvidia… Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/pci/hda/hda_controller.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/sound/pci/hda/hda_controller.c b/sound/pci/hda/hda_controller.c index fa261b27d8588..8198d2e53b7df 100644 --- a/sound/pci/hda/hda_controller.c +++ b/sound/pci/hda/hda_controller.c @@ -1169,16 +1169,23 @@ irqreturn_t azx_interrupt(int irq, void *dev_id) if (snd_hdac_bus_handle_stream_irq(bus, status, stream_update)) active = true; - /* clear rirb int */ status = azx_readb(chip, RIRBSTS); if (status & RIRB_INT_MASK) { + /* + * Clearing the interrupt status here ensures that no + * interrupt gets masked after the RIRB wp is read in + * snd_hdac_bus_update_rirb. This avoids a possible + * race condition where codec response in RIRB may + * remain unserviced by IRQ, eventually falling back + * to polling mode in azx_rirb_get_response. + */ + azx_writeb(chip, RIRBSTS, RIRB_INT_MASK); active = true; if (status & RIRB_INT_RESPONSE) { if (chip->driver_caps & AZX_DCAPS_CTX_WORKAROUND) udelay(80); snd_hdac_bus_update_rirb(bus); } - azx_writeb(chip, RIRBSTS, RIRB_INT_MASK); } } while (active && ++repeat < 10); -- 2.25.1

1 0

[PATCH AUTOSEL 4.19 039/206] ASoC: max98090: remove msleep in PLL unlocked workaround
by Sasha Levin 18 Sep '20

18 Sep '20

From: Tzung-Bi Shih <tzungbi(a)google.com> [ Upstream commit acb874a7c049ec49d8fc66c893170fb42c01bdf7 ] It was observed Baytrail-based chromebooks could cause continuous PLL unlocked when using playback stream and capture stream simultaneously. Specifically, starting a capture stream after started a playback stream. As a result, the audio data could corrupt or turn completely silent. As the datasheet suggested, the maximum PLL lock time should be 7 msec. The workaround resets the codec softly by toggling SHDN off and on if PLL failed to lock for 10 msec. Notably, there is no suggested hold time for SHDN off. On Baytrail-based chromebooks, it would easily happen continuous PLL unlocked if there is a 10 msec delay between SHDN off and on. Removes the msleep(). Signed-off-by: Tzung-Bi Shih <tzungbi(a)google.com> Link: https://lore.kernel.org/r/20191122073114.219945-2-tzungbi@google.com Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/codecs/max98090.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/sound/soc/codecs/max98090.c b/sound/soc/codecs/max98090.c index 89b6e187ac235..a5b0c40ee545f 100644 --- a/sound/soc/codecs/max98090.c +++ b/sound/soc/codecs/max98090.c @@ -2130,10 +2130,16 @@ static void max98090_pll_work(struct max98090_priv *max98090) dev_info_ratelimited(component->dev, "PLL unlocked\n"); + /* + * As the datasheet suggested, the maximum PLL lock time should be + * 7 msec. The workaround resets the codec softly by toggling SHDN + * off and on if PLL failed to lock for 10 msec. Notably, there is + * no suggested hold time for SHDN off. + */ + /* Toggle shutdown OFF then ON */ snd_soc_component_update_bits(component, M98090_REG_DEVICE_SHUTDOWN, M98090_SHDNN_MASK, 0); - msleep(10); snd_soc_component_update_bits(component, M98090_REG_DEVICE_SHUTDOWN, M98090_SHDNN_MASK, M98090_SHDNN_MASK); -- 2.25.1

1 0

[PATCH AUTOSEL 4.19 010/206] ASoC: kirkwood: fix IRQ error handling
by Sasha Levin 18 Sep '20

18 Sep '20

From: Russell King <rmk+kernel(a)armlinux.org.uk> [ Upstream commit 175fc928198236037174e5c5c066fe3c4691903e ] Propagate the error code from request_irq(), rather than returning -EBUSY. Signed-off-by: Russell King <rmk+kernel(a)armlinux.org.uk> Link: https://lore.kernel.org/r/E1iNIqh-0000tW-EZ@rmk-PC.armlinux.org.uk Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/kirkwood/kirkwood-dma.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/kirkwood/kirkwood-dma.c b/sound/soc/kirkwood/kirkwood-dma.c index c6a58520d377a..255cc45905b81 100644 --- a/sound/soc/kirkwood/kirkwood-dma.c +++ b/sound/soc/kirkwood/kirkwood-dma.c @@ -136,7 +136,7 @@ static int kirkwood_dma_open(struct snd_pcm_substream *substream) err = request_irq(priv->irq, kirkwood_dma_irq, IRQF_SHARED, "kirkwood-i2s", priv); if (err) - return -EBUSY; + return err; /* * Enable Error interrupts. We're only ack'ing them but -- 2.25.1

1 0

[PATCH AUTOSEL 5.4 328/330] ALSA: hda: Workaround for spurious wakeups on some Intel platforms
by Sasha Levin 18 Sep '20

18 Sep '20

From: Takashi Iwai <tiwai(a)suse.de> [ Upstream commit a6630529aecb5a3e84370c376ed658e892e6261e ] We've received a regression report on Intel HD-audio controller that wakes up immediately after S3 suspend. The bisection leads to the commit c4c8dd6ef807 ("ALSA: hda: Skip controller resume if not needed"). This commit replaces the system-suspend to use pm_runtime_force_suspend() instead of the direct call of __azx_runtime_suspend(). However, by some really mysterious reason, pm_runtime_force_suspend() causes a spurious wakeup (although it calls the same __azx_runtime_suspend() internally). As an ugly workaround for now, revert the behavior to call __azx_runtime_suspend() and __azx_runtime_resume() for those old Intel platforms that may exhibit such a problem, while keeping the new standard pm_runtime_force_suspend() and pm_runtime_force_resume() pair for the remaining chips. Fixes: c4c8dd6ef807 ("ALSA: hda: Skip controller resume if not needed") BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=208649 Cc: <stable(a)vger.kernel.org> Link: https://lore.kernel.org/r/20200727164443.4233-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/pci/hda/hda_controller.h | 2 +- sound/pci/hda/hda_intel.c | 17 ++++++++++++++--- 2 files changed, 15 insertions(+), 4 deletions(-) diff --git a/sound/pci/hda/hda_controller.h b/sound/pci/hda/hda_controller.h index 82e26442724ba..a356fb0e57738 100644 --- a/sound/pci/hda/hda_controller.h +++ b/sound/pci/hda/hda_controller.h @@ -41,7 +41,7 @@ /* 24 unused */ #define AZX_DCAPS_COUNT_LPIB_DELAY (1 << 25) /* Take LPIB as delay */ #define AZX_DCAPS_PM_RUNTIME (1 << 26) /* runtime PM support */ -/* 27 unused */ +#define AZX_DCAPS_SUSPEND_SPURIOUS_WAKEUP (1 << 27) /* Workaround for spurious wakeups after suspend */ #define AZX_DCAPS_CORBRP_SELF_CLEAR (1 << 28) /* CORBRP clears itself after reset */ #define AZX_DCAPS_NO_MSI64 (1 << 29) /* Stick to 32-bit MSIs */ #define AZX_DCAPS_SEPARATE_STREAM_TAG (1 << 30) /* capture and playback use separate stream tag */ diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c index 754e4d1a86b57..590ea262f2e20 100644 --- a/sound/pci/hda/hda_intel.c +++ b/sound/pci/hda/hda_intel.c @@ -295,7 +295,8 @@ enum { /* PCH for HSW/BDW; with runtime PM */ /* no i915 binding for this as HSW/BDW has another controller for HDMI */ #define AZX_DCAPS_INTEL_PCH \ - (AZX_DCAPS_INTEL_PCH_BASE | AZX_DCAPS_PM_RUNTIME) + (AZX_DCAPS_INTEL_PCH_BASE | AZX_DCAPS_PM_RUNTIME |\ + AZX_DCAPS_SUSPEND_SPURIOUS_WAKEUP) /* HSW HDMI */ #define AZX_DCAPS_INTEL_HASWELL \ @@ -1026,7 +1027,14 @@ static int azx_suspend(struct device *dev) chip = card->private_data; bus = azx_bus(chip); snd_power_change_state(card, SNDRV_CTL_POWER_D3hot); - pm_runtime_force_suspend(dev); + /* An ugly workaround: direct call of __azx_runtime_suspend() and + * __azx_runtime_resume() for old Intel platforms that suffer from + * spurious wakeups after S3 suspend + */ + if (chip->driver_caps & AZX_DCAPS_SUSPEND_SPURIOUS_WAKEUP) + __azx_runtime_suspend(chip); + else + pm_runtime_force_suspend(dev); if (bus->irq >= 0) { free_irq(bus->irq, chip); bus->irq = -1; @@ -1054,7 +1062,10 @@ static int azx_resume(struct device *dev) if (azx_acquire_irq(chip, 1) < 0) return -EIO; - pm_runtime_force_resume(dev); + if (chip->driver_caps & AZX_DCAPS_SUSPEND_SPURIOUS_WAKEUP) + __azx_runtime_resume(chip, false); + else + pm_runtime_force_resume(dev); snd_power_change_state(card, SNDRV_CTL_POWER_D0); trace_azx_resume(chip); -- 2.25.1

1 0

[PATCH AUTOSEL 5.4 327/330] ALSA: hda: Always use jackpoll helper for jack update after resume
by Sasha Levin 18 Sep '20

18 Sep '20

From: Takashi Iwai <tiwai(a)suse.de> [ Upstream commit 8d6762af302d69f76fa788a277a56a9d9cd275d5 ] HD-audio codec driver applies a tricky procedure to forcibly perform the runtime resume by mimicking the usage count even if the device has been runtime-suspended beforehand. This was needed to assure to trigger the jack detection update after the system resume. And recently we also applied the similar logic to the HD-audio controller side. However this seems leading to some inconsistency, and eventually PCI controller gets screwed up. This patch is an attempt to fix and clean up those behavior: instead of the tricky runtime resume procedure, the existing jackpoll work is scheduled when such a forced codec resume is required. The jackpoll work will power up the codec, and this alone should suffice for the jack status update in usual cases. If the extra polling is requested (by checking codec->jackpoll_interval), the manual update is invoked after that, and the codec is powered down again. Also, we filter the spurious wake up of the codec from the controller runtime resume by checking codec->relaxed_resume flag. If this flag is set, basically we don't need to wake up explicitly, but it's supposed to be done via the audio component notifier. Fixes: c4c8dd6ef807 ("ALSA: hda: Skip controller resume if not needed") Link: https://lore.kernel.org/r/20200422203744.26299-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/pci/hda/hda_codec.c | 28 +++++++++++++++++----------- sound/pci/hda/hda_intel.c | 17 ++--------------- 2 files changed, 19 insertions(+), 26 deletions(-) diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index 12da263fb02ba..6da296def283e 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -641,8 +641,18 @@ static void hda_jackpoll_work(struct work_struct *work) struct hda_codec *codec = container_of(work, struct hda_codec, jackpoll_work.work); - snd_hda_jack_set_dirty_all(codec); - snd_hda_jack_poll_all(codec); + /* for non-polling trigger: we need nothing if already powered on */ + if (!codec->jackpoll_interval && snd_hdac_is_power_on(&codec->core)) + return; + + /* the power-up/down sequence triggers the runtime resume */ + snd_hda_power_up_pm(codec); + /* update jacks manually if polling is required, too */ + if (codec->jackpoll_interval) { + snd_hda_jack_set_dirty_all(codec); + snd_hda_jack_poll_all(codec); + } + snd_hda_power_down_pm(codec); if (!codec->jackpoll_interval) return; @@ -2958,18 +2968,14 @@ static int hda_codec_runtime_resume(struct device *dev) static int hda_codec_force_resume(struct device *dev) { struct hda_codec *codec = dev_to_hda_codec(dev); - bool forced_resume = hda_codec_need_resume(codec); int ret; - /* The get/put pair below enforces the runtime resume even if the - * device hasn't been used at suspend time. This trick is needed to - * update the jack state change during the sleep. - */ - if (forced_resume) - pm_runtime_get_noresume(dev); ret = pm_runtime_force_resume(dev); - if (forced_resume) - pm_runtime_put(dev); + /* schedule jackpoll work for jack detection update */ + if (codec->jackpoll_interval || + (pm_runtime_suspended(dev) && hda_codec_need_resume(codec))) + schedule_delayed_work(&codec->jackpoll_work, + codec->jackpoll_interval); return ret; } diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c index a6e8aaa091c7d..754e4d1a86b57 100644 --- a/sound/pci/hda/hda_intel.c +++ b/sound/pci/hda/hda_intel.c @@ -1002,7 +1002,8 @@ static void __azx_runtime_resume(struct azx *chip, bool from_rt) if (status && from_rt) { list_for_each_codec(codec, &chip->bus) - if (status & (1 << codec->addr)) + if (!codec->relaxed_resume && + (status & (1 << codec->addr))) schedule_delayed_work(&codec->jackpoll_work, codec->jackpoll_interval); } @@ -1041,9 +1042,7 @@ static int azx_suspend(struct device *dev) static int azx_resume(struct device *dev) { struct snd_card *card = dev_get_drvdata(dev); - struct hda_codec *codec; struct azx *chip; - bool forced_resume = false; if (!azx_is_pm_ready(card)) return 0; @@ -1055,19 +1054,7 @@ static int azx_resume(struct device *dev) if (azx_acquire_irq(chip, 1) < 0) return -EIO; - /* check for the forced resume */ - list_for_each_codec(codec, &chip->bus) { - if (hda_codec_need_resume(codec)) { - forced_resume = true; - break; - } - } - - if (forced_resume) - pm_runtime_get_noresume(dev); pm_runtime_force_resume(dev); - if (forced_resume) - pm_runtime_put(dev); snd_power_change_state(card, SNDRV_CTL_POWER_D0); trace_azx_resume(chip); -- 2.25.1

1 0

[PATCH AUTOSEL 5.4 299/330] ASoC: img-i2s-out: Fix runtime PM imbalance on error
by Sasha Levin 18 Sep '20

18 Sep '20

From: Dinghao Liu <dinghao.liu(a)zju.edu.cn> [ Upstream commit 65bd91dd6957390c42a0491b9622cf31a2cdb140 ] pm_runtime_get_sync() increments the runtime PM usage counter even the call returns an error code. Thus a pairing decrement is needed on the error handling path to keep the counter balanced. Signed-off-by: Dinghao Liu <dinghao.liu(a)zju.edu.cn> Link: https://lore.kernel.org/r/20200529012230.5863-1-dinghao.liu@zju.edu.cn Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/img/img-i2s-out.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/sound/soc/img/img-i2s-out.c b/sound/soc/img/img-i2s-out.c index 4b18534096336..9c4212f2f7269 100644 --- a/sound/soc/img/img-i2s-out.c +++ b/sound/soc/img/img-i2s-out.c @@ -347,8 +347,10 @@ static int img_i2s_out_set_fmt(struct snd_soc_dai *dai, unsigned int fmt) chan_control_mask = IMG_I2S_OUT_CHAN_CTL_CLKT_MASK; ret = pm_runtime_get_sync(i2s->dev); - if (ret < 0) + if (ret < 0) { + pm_runtime_put_noidle(i2s->dev); return ret; + } img_i2s_out_disable(i2s); @@ -488,8 +490,10 @@ static int img_i2s_out_probe(struct platform_device *pdev) goto err_pm_disable; } ret = pm_runtime_get_sync(&pdev->dev); - if (ret < 0) + if (ret < 0) { + pm_runtime_put_noidle(&pdev->dev); goto err_suspend; + } reg = IMG_I2S_OUT_CTL_FRM_SIZE_MASK; img_i2s_out_writel(i2s, reg, IMG_I2S_OUT_CTL); -- 2.25.1

1 0

[PATCH AUTOSEL 5.4 267/330] ALSA: hda: Fix potential race in unsol event handler
by Sasha Levin 18 Sep '20

18 Sep '20

From: Takashi Iwai <tiwai(a)suse.de> [ Upstream commit c637fa151259c0f74665fde7cba5b7eac1417ae5 ] The unsol event handling code has a loop retrieving the read/write indices and the arrays without locking while the append to the array may happen concurrently. This may lead to some inconsistency. Although there hasn't been any proof of this bad results, it's still safer to protect the racy accesses. This patch adds the spinlock protection around the unsol handling loop for addressing it. Here we take bus->reg_lock as the writer side snd_hdac_bus_queue_event() is also protected by that lock. Link: https://lore.kernel.org/r/20200516062556.30951-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/hda/hdac_bus.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/sound/hda/hdac_bus.c b/sound/hda/hdac_bus.c index 8f19876244ebe..53be2cac98e7c 100644 --- a/sound/hda/hdac_bus.c +++ b/sound/hda/hdac_bus.c @@ -158,6 +158,7 @@ static void snd_hdac_bus_process_unsol_events(struct work_struct *work) struct hdac_driver *drv; unsigned int rp, caddr, res; + spin_lock_irq(&bus->reg_lock); while (bus->unsol_rp != bus->unsol_wp) { rp = (bus->unsol_rp + 1) % HDA_UNSOL_QUEUE_SIZE; bus->unsol_rp = rp; @@ -169,10 +170,13 @@ static void snd_hdac_bus_process_unsol_events(struct work_struct *work) codec = bus->caddr_tbl[caddr & 0x0f]; if (!codec || !codec->dev.driver) continue; + spin_unlock_irq(&bus->reg_lock); drv = drv_to_hdac_driver(codec->dev.driver); if (drv->unsol_event) drv->unsol_event(codec, res); + spin_lock_irq(&bus->reg_lock); } + spin_unlock_irq(&bus->reg_lock); } /** -- 2.25.1

1 0