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

[PATCH v2 0/3] ALSA: Support for non-contiguous and non-coherent page allocations
by Takashi Iwai 17 Oct '21

17 Oct '21

Hi, this is a revised patchset that is created from the latest code. The previous one caused a slight conflict with the recent changes, also some patch descriptions have been updated. *** Here is a patch set to add the support for non-contiguous and non-coherent page allocations to the ALSA core memalloc helpers. Currently no driver is using, but this also gave a nice cleanup for the existing SG-buffer helpers on x86. Takashi === Takashi Iwai (3): ALSA: memalloc: Support for non-contiguous page allocation ALSA: memalloc: Support for non-coherent page allocation ALSA: memalloc: Convert x86 SG-buffer handling with non-contiguous type include/sound/memalloc.h | 61 +++++++++-- sound/core/Makefile | 1 - sound/core/memalloc.c | 207 +++++++++++++++++++++++++++++++++--- sound/core/memalloc_local.h | 1 + sound/core/pcm_lib.c | 3 + sound/core/pcm_local.h | 7 ++ sound/core/pcm_memory.c | 13 ++- sound/core/pcm_native.c | 16 +++ sound/core/sgbuf.c | 201 ---------------------------------- 9 files changed, 282 insertions(+), 228 deletions(-) delete mode 100644 sound/core/sgbuf.c -- 2.26.2

1 3

[PATCH 0/3] ALSA: Support for non-contiguous and non-coherent page allocations
by Takashi Iwai 17 Oct '21

17 Oct '21

Hi, here is a patch set to add the support for non-contiguous and non-coherent page allocations to the ALSA core memalloc helpers. Currently no driver is using, but this also gave a nice cleanup for the existing SG-buffer helpers on x86. Takashi === Takashi Iwai (3): ALSA: memalloc: Support for non-contiguous page allocation ALSA: memalloc: Support for non-coherent page allocation ALSA: memalloc: Convert x86 SG-buffer handling with non-contiguous type include/sound/memalloc.h | 61 +++++++++-- sound/core/Makefile | 1 - sound/core/memalloc.c | 207 +++++++++++++++++++++++++++++++++--- sound/core/memalloc_local.h | 1 + sound/core/pcm_lib.c | 3 + sound/core/pcm_local.h | 7 ++ sound/core/pcm_memory.c | 13 ++- sound/core/pcm_native.c | 16 +++ sound/core/sgbuf.c | 201 ---------------------------------- 9 files changed, 282 insertions(+), 228 deletions(-) delete mode 100644 sound/core/sgbuf.c -- 2.26.2

1 4

[RFC PATCH] sound: ISA: not for M68K
by Randy Dunlap 17 Oct '21

17 Oct '21

On m68k, compiling drivers under SND_ISA causes build errors: ../sound/core/isadma.c: In function 'snd_dma_program': ../sound/core/isadma.c:33:17: error: implicit declaration of function 'claim_dma_lock' [-Werror=implicit-function-declaration] 33 | flags = claim_dma_lock(); | ^~~~~~~~~~~~~~ ../sound/core/isadma.c:41:9: error: implicit declaration of function 'release_dma_lock' [-Werror=implicit-function-declaration] 41 | release_dma_lock(flags); | ^~~~~~~~~~~~~~~~ ../sound/isa/sb/sb16_main.c: In function 'snd_sb16_playback_prepare': ../sound/isa/sb/sb16_main.c:253:72: error: 'DMA_AUTOINIT' undeclared (first use in this function) 253 | snd_dma_program(dma, runtime->dma_addr, size, DMA_MODE_WRITE | DMA_AUTOINIT); | ^~~~~~~~~~~~ ../sound/isa/sb/sb16_main.c:253:72: note: each undeclared identifier is reported only once for each function it appears in ../sound/isa/sb/sb16_main.c: In function 'snd_sb16_capture_prepare': ../sound/isa/sb/sb16_main.c:322:71: error: 'DMA_AUTOINIT' undeclared (first use in this function) 322 | snd_dma_program(dma, runtime->dma_addr, size, DMA_MODE_READ | DMA_AUTOINIT); | ^~~~~~~~~~~~ and more... Signed-off-by: Randy Dunlap <rdunlap(a)infradead.org> Cc: Jaroslav Kysela <perex(a)perex.cz> Cc: Takashi Iwai <tiwai(a)suse.com> Cc: alsa-devel(a)alsa-project.org Cc: linux-m68k(a)lists.linux-m68k.org Cc: Geert Uytterhoeven <geert(a)linux-m68k.org> --- sound/core/Makefile | 2 ++ sound/isa/Kconfig | 2 +- sound/pci/Kconfig | 1 + 3 files changed, 4 insertions(+), 1 deletion(-) --- linux-next-20211015.orig/sound/isa/Kconfig +++ linux-next-20211015/sound/isa/Kconfig @@ -22,7 +22,7 @@ config SND_SB16_DSP menuconfig SND_ISA bool "ISA sound devices" depends on ISA || COMPILE_TEST - depends on ISA_DMA_API + depends on ISA_DMA_API && !M68K default y help Support for sound devices connected via the ISA bus. --- linux-next-20211015.orig/sound/pci/Kconfig +++ linux-next-20211015/sound/pci/Kconfig @@ -279,6 +279,7 @@ config SND_CS46XX_NEW_DSP config SND_CS5530 tristate "CS5530 Audio" depends on ISA_DMA_API && (X86_32 || COMPILE_TEST) + depends on !M68K select SND_SB16_DSP help Say Y here to include support for audio on Cyrix/NatSemi CS5530 chips. --- linux-next-20211015.orig/sound/core/Makefile +++ linux-next-20211015/sound/core/Makefile @@ -9,7 +9,9 @@ ifneq ($(CONFIG_SND_PROC_FS),) snd-y += info.o snd-$(CONFIG_SND_OSSEMUL) += info_oss.o endif +ifneq ($(CONFIG_M68K),y) snd-$(CONFIG_ISA_DMA_API) += isadma.o +endif snd-$(CONFIG_SND_OSSEMUL) += sound_oss.o snd-$(CONFIG_SND_VMASTER) += vmaster.o snd-$(CONFIG_SND_JACK) += ctljack.o jack.o

2 1

[PATCH] ASoC: dt-bindings: rockchip: i2s-tdm: fix rockchip, cru requirement
by Nicolas Frattaroli 15 Oct '21

15 Oct '21

The behaviour described in the bindings is exactly the opposite of what it should be, so this patch fixes the description and adjusts the if conditions for adding rockchip,cru to required. Fixes: 510f1c133aed ("ASoC: dt-bindings: rockchip: add i2s-tdm bindings") Signed-off-by: Nicolas Frattaroli <frattaroli.nicolas(a)gmail.com> --- .../devicetree/bindings/sound/rockchip,i2s-tdm.yaml | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/Documentation/devicetree/bindings/sound/rockchip,i2s-tdm.yaml b/Documentation/devicetree/bindings/sound/rockchip,i2s-tdm.yaml index ce3e18b50230..93f5221e9381 100644 --- a/Documentation/devicetree/bindings/sound/rockchip,i2s-tdm.yaml +++ b/Documentation/devicetree/bindings/sound/rockchip,i2s-tdm.yaml @@ -86,7 +86,7 @@ properties: $ref: /schemas/types.yaml#/definitions/phandle description: The phandle of the cru. - Required if neither trcm-sync-tx-only nor trcm-sync-rx-only are specified. + Required if either trcm-sync-tx-only or trcm-sync-rx-only are specified. rockchip,grf: $ref: /schemas/types.yaml#/definitions/phandle @@ -147,8 +147,13 @@ required: allOf: - if: properties: - rockchip,trcm-sync-tx-only: false - rockchip,trcm-sync-rx-only: false + rockchip,trcm-sync-tx-only: true + then: + required: + - rockchip,cru + - if: + properties: + rockchip,trcm-sync-rx-only: true then: required: - rockchip,cru -- 2.33.0

2 2

[PATCH v5 00/16] ASoC: Add Audio Graph Card2 support
by Kuninori Morimoto 15 Oct '21

15 Oct '21

Hi Mark We already have Audio-Graph-Card which is Of-Graph base general sound card driver. Basically it supports basic CPU-Codec connection, and is also supporting DPCM connection. Because it was forcibly expanded to DPCM, DT parsing is very limited and very difficult to add new features on it, for example Multi-CPU/Codec support, Codec2Codec support, etc. This patch adds more flexible new Audio-Graph-Card2 driver for it. Audio-Graph-Card and Audio-Graph-Card2 are similar, but don't have full compatibility. The reason why I need Audio-Graph-Card2 instead of updating Audio-Graph-Card is that it is very difficult to keep compatibility. Audio-Graph-Card2 supports Normal/DPCM/Codec2Codec Connection wich Single/Multi DAIs. And it is possible to Customizing. This patch-set adds Audio-Graph-Card2 driver and its custom driver sample, and DT settings sample which can be used for testing. To enable testing/debuging, this patch-set also adds Test-Component driver. We already have Dummy Component and/or Dummy DAI on soc-utils, but 1) we can't use it from DT, 2) it do nothing. Added new Test-Component can be used from DT, and it can indicate called function name. We can use it to trace callback order, understanding ALSA SoC behavior, etc, etc... Sample DT settings of Audio Graph Card2 is using Test-Component as CPU/Codec DAI. You can easily try to use/test it if you added below line to your DT file. Your .config needs to have below CONFIGs to use/test it. It will probe sample Sound Card which has Normal/DPCM/Multi/Codec2Codec connections. #include "../../../../../sound/soc/generic/audio-graph-card2-custom-sample.dtsi" CONFIG_SND_AUDIO_GRAPH_CARD2 CONFIG_SND_AUDIO_GRAPH_CARD2_CUSTOM_SAMPLE CONFIG_SND_TEST_COMPONENT Because Audio Graph Card2 is still under experimental stage, it will indicate such warning when probing, and the DT might be updated/exchanged. It can use Codec2Codec, but it will start automatically when probed, and can't stop it so far. It should be updated. Link: https://lore.kernel.org/r/87k0xszlep.wl-kuninori.morimoto.gx@renesas.com Link: https://lore.kernel.org/r/871r8u4s6q.wl-kuninori.morimoto.gx@renesas.com Link: https://lore.kernel.org/r/87a6mhwyqn.wl-kuninori.morimoto.gx@renesas.com Link: https://lore.kernel.org/r/87tuitusy4.wl-kuninori.morimoto.gx@renesas.com Link: https://lore.kernel.org/r/87a6jn56x0.wl-kuninori.morimoto.gx@renesas.com v1 -> v2 - don't use "port" base for_each loop v2 -> v3 - Rename audio-graph-card2 to rich-graph-card - Rename DSP to DPCM not to confuse - Normal/DPCM/Codec2Codec can use Single/Multi DAIs. - use dpcm/multi/codec2codec node instead of using extra compatible - Sample DTSI patch is separated to Single/Multi. v3 -> v4 - Rename rich-graph-card to audio-graph-card2 - fixup custom sample driver's connection bug - test-component compatible uses "verbose" instead of "vv" v4 -> v5 - tidyup git-log comment at - tidyup Custom Sample comment Kuninori Morimoto (16): ASoC: test-component: add Test Component YAML bindings ASoC: test-component: add Test Component for Sound debug/test ASoC: simple-card-utils: add asoc_graph_is_ports0() ASoC: simple-card-utils: add codec2codec support ASoC: add Audio Graph Card2 driver ASoC: audio-graph-card2: add Multi CPU/Codec support ASoC: audio-graph-card2: add DPCM support ASoC: audio-graph-card2: add Codec2Codec support ASoC: add Audio Graph Card2 Yaml Document ASoC: add Audio Graph Card2 Custom Sample ASoC: audio-graph-card2-custom-sample.dtsi: add Sample DT for Normal (Single) ASoC: audio-graph-card2-custom-sample.dtsi: add Sample DT for Normal (Nulti) ASoC: audio-graph-card2-custom-sample.dtsi: add DPCM sample (Single) ASoC: audio-graph-card2-custom-sample.dtsi: add DPCM sample (Multi) ASoC: audio-graph-card2-custom-sample.dtsi: add Codec2Codec sample (Single) ASoC: audio-graph-card2-custom-sample.dtsi: add Codec2Codec sample (Multi) .../bindings/sound/audio-graph-card2.yaml | 57 + .../bindings/sound/test-component.yaml | 33 + include/sound/graph_card.h | 21 + include/sound/simple_card_utils.h | 4 + sound/soc/generic/Kconfig | 20 + sound/soc/generic/Makefile | 6 + .../generic/audio-graph-card2-custom-sample.c | 183 +++ .../audio-graph-card2-custom-sample.dtsi | 227 +++ sound/soc/generic/audio-graph-card2.c | 1281 +++++++++++++++++ sound/soc/generic/simple-card-utils.c | 46 +- sound/soc/generic/test-component.c | 659 +++++++++ 11 files changed, 2536 insertions(+), 1 deletion(-) create mode 100644 Documentation/devicetree/bindings/sound/audio-graph-card2.yaml create mode 100644 Documentation/devicetree/bindings/sound/test-component.yaml create mode 100644 sound/soc/generic/audio-graph-card2-custom-sample.c create mode 100644 sound/soc/generic/audio-graph-card2-custom-sample.dtsi create mode 100644 sound/soc/generic/audio-graph-card2.c create mode 100644 sound/soc/generic/test-component.c -- 2.25.1

2 17

[PATCH V2 1/3] ASoc: amd: create platform device for VG machine driver
by Vijendar Mukunda 15 Oct '21

15 Oct '21

Create platform device for Vangogh machine driver. Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda(a)amd.com> --- Changes since v1: - changed machine driver platform device name as "acp5x_mach" sound/soc/amd/vangogh/acp5x.h | 2 +- sound/soc/amd/vangogh/pci-acp5x.c | 3 +++ 2 files changed, 4 insertions(+), 1 deletion(-) diff --git a/sound/soc/amd/vangogh/acp5x.h b/sound/soc/amd/vangogh/acp5x.h index a808635f9740..fe5e1fa98974 100644 --- a/sound/soc/amd/vangogh/acp5x.h +++ b/sound/soc/amd/vangogh/acp5x.h @@ -23,7 +23,7 @@ #define ACP_ERR_INTR_MASK 0x20000000 #define ACP_EXT_INTR_STAT_CLEAR_MASK 0xFFFFFFFF -#define ACP5x_DEVS 3 +#define ACP5x_DEVS 4 #define ACP5x_REG_START 0x1240000 #define ACP5x_REG_END 0x1250200 #define ACP5x_I2STDM_REG_START 0x1242400 diff --git a/sound/soc/amd/vangogh/pci-acp5x.c b/sound/soc/amd/vangogh/pci-acp5x.c index a57b762d9f2e..2b6b9edc36e2 100644 --- a/sound/soc/amd/vangogh/pci-acp5x.c +++ b/sound/soc/amd/vangogh/pci-acp5x.c @@ -213,6 +213,9 @@ static int snd_acp5x_probe(struct pci_dev *pci, pdevinfo[2].num_res = 1; pdevinfo[2].res = &adata->res[2]; + pdevinfo[3].name = "acp5x_mach"; + pdevinfo[3].id = 0; + pdevinfo[3].parent = &pci->dev; for (i = 0; i < ACP5x_DEVS; i++) { adata->pdev[i] = platform_device_register_full(&pdevinfo[i]); -- 2.25.1

2 3

[PATCH 00/16] ASoC: cs42l42: Collection of bugfixes
by Richard Fitzgerald 15 Oct '21

15 Oct '21

This patch set contains various bugfixes for the cs42l42 codec driver. Patches marked "Fixes" will apply cleanly to the patch that first introduced the bug. Patches NOT marked "Fixes" will not apply cleanly to the point the bug was first introduced and/or the bug is not having sufficient impact to risk churning older code versions. Richard Fitzgerald (16): ASoC: cs42l42: Don't reconfigure the PLL while it is running ASoC: cs42l42: Always configure both ASP TX channels ASoC: cs42l42: Correct some register default values ASoC: cs42l42: Don't set defaults for volatile registers ASoC: cs42l42: Defer probe if request_threaded_irq() returns EPROBE_DEFER ASoC: cs42l42: Reset GPIO is mandatory ASoC: cs42l42: Correct power-up sequence to match datasheet ASoC: cs42l42: Reset and power-down on driver remove() ASoC: cs42l42: Prevent NULL pointer deref in interrupt handler ASoC: cs42l42: Don't claim to support 192k ASoC: cs42l42: Use PLL for SCLK > 12.288MHz ASoC: cs42l42: Allow time for HP/ADC to power-up after enable ASoC: cs42l42: Set correct SRC MCLK ASoC: cs42l42: Mark OSC_SWITCH_STATUS register volatile ASoC: cs42l42: Fix WARN in remove() if running without an interrupt ASoC: cs42l42: Always enable TS_PLUG and TS_UNPLUG interrupts sound/soc/codecs/cs42l42.c | 282 +++++++++++++++++++++++++-------------------- sound/soc/codecs/cs42l42.h | 9 +- 2 files changed, 160 insertions(+), 131 deletions(-) -- 2.11.0

2 20

[PATCH v4 00/16] ASoC: Add Audio Graph Card2 support
by Kuninori Morimoto 15 Oct '21

15 Oct '21

Hi Mark We already have Audio-Graph-Card which is Of-Graph base general sound card driver. Basically it supports basic CPU-Codec connection, and is also supporting DPCM connection. Because it was forcibly expanded to DPCM, DT parsing is very limited and very difficult to add new features on it, for example Multi-CPU/Codec support, Codec2Codec support, etc. This patch adds more flexible new Audio-Graph-Card2 driver for it. Audio-Graph-Card and Audio-Graph-Card2 are similar, but don't have full compatibility. The reason why I need Audio-Graph-Card2 instead of updating Audio-Graph-Card is that it is very difficult to keep compatibility. Audio-Graph-Card2 supports Normal/DPCM/Codec2Codec Connection wich Single/Multi DAIs. And it is possible to Customizing. This patch-set adds Audio-Graph-Card2 driver and its custom driver sample, and DT settings sample which can be used for testing. To enable testing/debuging, this patch-set also adds Test-Component driver. We already have Dummy Component and/or Dummy DAI on soc-utils, but 1) we can't use it from DT, 2) it do nothing. Added new Test-Component can be used from DT, and it can indicate called function name. We can use it to trace callback order, understanding ALSA SoC behavior, etc, etc... Sample DT settings of Audio Graph Card2 is using Test-Component as CPU/Codec DAI. You can easily try to use/test it if you added below line to your DT file. Your .config needs to have below CONFIGs to use/test it. It will probe sample Sound Card which has Normal/DPCM/Multi/Codec2Codec connections. #include "../../../../../sound/soc/generic/audio-graph-card2-custom-sample.dtsi" CONFIG_SND_AUDIO_GRAPH_CARD2 CONFIG_SND_AUDIO_GRAPH_CARD2_CUSTOM_SAMPLE CONFIG_SND_TEST_COMPONENT Because Audio Graph Card2 is still under experimental stage, it will indicate such warning when probing, and the DT might be updated/exchanged. It can use Codec2Codec, but it will start automatically when probed, and can't stop it so far. It should be updated. Link: https://lore.kernel.org/r/87k0xszlep.wl-kuninori.morimoto.gx@renesas.com Link: https://lore.kernel.org/r/871r8u4s6q.wl-kuninori.morimoto.gx@renesas.com Link: https://lore.kernel.org/r/87a6mhwyqn.wl-kuninori.morimoto.gx@renesas.com Link: https://lore.kernel.org/r/87tuitusy4.wl-kuninori.morimoto.gx@renesas.com v1 -> v2 - don't use "port" base for_each loop v2 -> v3 - Rename audio-graph-card2 to rich-graph-card - Rename DSP to DPCM not to confuse - Normal/DPCM/Codec2Codec can use Single/Multi DAIs. - use dpcm/multi/codec2codec node instead of using extra compatible - Sample DTSI patch is separated to Single/Multi. v3 -> v4 - Rename rich-graph-card to audio-graph-card2 - fixup custom sample driver's connection bug - test-component compatible uses "verbose" instead of "vv" Kuninori Morimoto (16): ASoC: test-component: add Test Component YAML bindings ASoC: test-component: add Test Component for Sound debug/test ASoC: simple-card-utils: add asoc_graph_is_ports0() ASoC: simple-card-utils: add codec2codec support ASoC: add Audio Graph Card2 driver ASoC: audio-graph-card2: add Multi CPU/Codec support ASoC: audio-graph-card2: add DPCM support ASoC: audio-graph-card2: add Codec2Codec support ASoC: add Audio Graph Card2 Yaml Document ASoC: add Audio Graph Card2 Custom Sample ASoC: audio-graph-card2-custom-sample.dtsi: add Sample DT for Normal (Single) ASoC: audio-graph-card2-custom-sample.dtsi: add Sample DT for Normal (Nulti) ASoC: audio-graph-card2-custom-sample.dtsi: add DPCM sample (Single) ASoC: audio-graph-card2-custom-sample.dtsi: add DPCM sample (Multi) ASoC: audio-graph-card2-custom-sample.dtsi: add Codec2Codec sample (Single) ASoC: audio-graph-card2-custom-sample.dtsi: add Codec2Codec sample (Multi) .../bindings/sound/audio-graph-card2.yaml | 57 + .../bindings/sound/test-component.yaml | 33 + include/sound/graph_card.h | 21 + include/sound/simple_card_utils.h | 4 + sound/soc/generic/Kconfig | 20 + sound/soc/generic/Makefile | 6 + .../generic/audio-graph-card2-custom-sample.c | 174 +++ .../audio-graph-card2-custom-sample.dtsi | 227 +++ sound/soc/generic/audio-graph-card2.c | 1281 +++++++++++++++++ sound/soc/generic/simple-card-utils.c | 46 +- sound/soc/generic/test-component.c | 659 +++++++++ 11 files changed, 2527 insertions(+), 1 deletion(-) create mode 100644 Documentation/devicetree/bindings/sound/audio-graph-card2.yaml create mode 100644 Documentation/devicetree/bindings/sound/test-component.yaml create mode 100644 sound/soc/generic/audio-graph-card2-custom-sample.c create mode 100644 sound/soc/generic/audio-graph-card2-custom-sample.dtsi create mode 100644 sound/soc/generic/audio-graph-card2.c create mode 100644 sound/soc/generic/test-component.c -- 2.25.1

2 20

[PATCH] ASoC: soc-component: improve error reporting for register access
by Srinivas Kandagatla 15 Oct '21

15 Oct '21

Currently errors on register read/write/update are reported with an error code and the corresponding function but does not provide any details on the which register number did it actually fail. register number can give better clue and it should be easy to locate the code and fix. Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla(a)linaro.org> --- Personally I found this patch very useful while developing and debugging. Ex: below error "ASoC: error at soc_component_read_no_lock on soc@0:codec: -16" did not provide much info except that it failed to read, but after this patch error message looks like: "ASoC: error at soc_component_read_no_lock on soc@0:codec for register: [0x00003125] -16" which was easy to locate and fix. sound/soc/soc-component.c | 26 ++++++++++++++++---------- 1 file changed, 16 insertions(+), 10 deletions(-) diff --git a/sound/soc/soc-component.c b/sound/soc/soc-component.c index a08a897c5230..c76ff9c59dfb 100644 --- a/sound/soc/soc-component.c +++ b/sound/soc/soc-component.c @@ -13,9 +13,10 @@ #include <sound/soc.h> #include <linux/bitops.h> -#define soc_component_ret(dai, ret) _soc_component_ret(dai, __func__, ret) +#define soc_component_ret(dai, ret) _soc_component_ret(dai, __func__, ret, -1) +#define soc_component_ret_reg_rw(dai, ret, reg) _soc_component_ret(dai, __func__, ret, reg) static inline int _soc_component_ret(struct snd_soc_component *component, - const char *func, int ret) + const char *func, int ret, int reg) { /* Positive/Zero values are not errors */ if (ret >= 0) @@ -27,9 +28,14 @@ static inline int _soc_component_ret(struct snd_soc_component *component, case -ENOTSUPP: break; default: - dev_err(component->dev, - "ASoC: error at %s on %s: %d\n", - func, component->name, ret); + if (reg == -1) + dev_err(component->dev, + "ASoC: error at %s on %s: %d\n", + func, component->name, ret); + else + dev_err(component->dev, + "ASoC: error at %s on %s for register: [0x%08x] %d\n", + func, component->name, reg, ret); } return ret; @@ -687,7 +693,7 @@ static unsigned int soc_component_read_no_lock( ret = -EIO; if (ret < 0) - return soc_component_ret(component, ret); + return soc_component_ret_reg_rw(component, ret, reg); return val; } @@ -723,7 +729,7 @@ static int soc_component_write_no_lock( else if (component->driver->write) ret = component->driver->write(component, reg, val); - return soc_component_ret(component, ret); + return soc_component_ret_reg_rw(component, ret, reg); } /** @@ -765,7 +771,7 @@ static int snd_soc_component_update_bits_legacy( mutex_unlock(&component->io_mutex); - return soc_component_ret(component, ret); + return soc_component_ret_reg_rw(component, ret, reg); } /** @@ -793,7 +799,7 @@ int snd_soc_component_update_bits(struct snd_soc_component *component, mask, val, &change); if (ret < 0) - return soc_component_ret(component, ret); + return soc_component_ret_reg_rw(component, ret, reg); return change; } EXPORT_SYMBOL_GPL(snd_soc_component_update_bits); @@ -829,7 +835,7 @@ int snd_soc_component_update_bits_async(struct snd_soc_component *component, mask, val, &change); if (ret < 0) - return soc_component_ret(component, ret); + return soc_component_ret_reg_rw(component, ret, reg); return change; } EXPORT_SYMBOL_GPL(snd_soc_component_update_bits_async); -- 2.21.0

3 4

[RFC PATCH v3 00/13] ASoC : soc-pcm: fix trigger race conditions with shared BE
by Pierre-Louis Bossart 15 Oct '21

15 Oct '21

We've been adding a 'deep buffer' PCM device to several SOF topologies in order to reduce power consumption. The typical use-case would be music playback over a headset: this additional PCM device provides more buffering and longer latencies, leaving the rest of the system sleep for longer periods. Notifications and 'regular' low-latency audio playback would still use the 'normal' PCM device and be mixed with the 'deep buffer' before rendering on the headphone endpoint. The tentative direction would be to expose this alternate device to PulseAudio/PipeWire/CRAS via the UCM SectionModifier definitions. That seemed a straightforward topology change until our automated validation stress tests started reporting issues on SoundWire platforms, when e.g. two START triggers might be send and conversely the STOP trigger is never sent. The SoundWire stream state management flagged inconsistent states when the two 'normal' and 'deep buffer' devices are used concurrently with rapid play/stop/pause monkey testing. Looking at the soc-pcm.c code, it seems that the BE state management needs a lot of love. a) there is no consistent protection for the BE state. In some parts of the code, the state updates are protected by a spinlock but in the trigger they are not. When we open/play/close the two PCM devices in stress tests, we end-up testing a state that is being modified. That can't be good. b) there is a conceptual deadlock: on stop we check the FE states to see if a shared BE can be stopped, but since we trigger the BE first the FE states have not been modified yet, so the TRIGGER_STOP is never sent. This patchset suggests the removal of the dedicated 'dpcm_lock', and the use of the FE PCM lock before walking through the BE list, a mutual exclusion between triggers using the BE PCM lock, and the use of a refcount to decide when to trigger the BE. With these patches I am able to run our entire validation suite without any issues with this new 'deep buffer' topology, and no regressions on existing solutions [1] One might ask 'how come we didn't see this earlier'? The answer is probably that the .trigger callbacks in most implementations seems to perform DAPM operations, and sending the triggers multiple times is not an issue. In the case of SoundWire, we do use the .trigger callback to reconfigure the bus using the 'bank switch' mechanism. It could be acceptable to tolerate a trigger multiple times, but the deadlock on stop cannot be fixed at the SoundWire layer alone. I chose to send this patchset as an RFCv3 to gather more feedback and make use others know about DPCM issues. We're going to spend more time on this but if others can provide feedback/test results it would be greatly appreciated. The change in the locking model could be problematic on other platforms so we do want more time to comment/test before even considering a merge. Opens: 1) is this the right solution? The DPCM code is far from simple, has notions such as SND_SOC_DPCM_UPDATE_NO and 'trigger_pending' that I have no background on. It's not clear if these cases are still needed with the locking changes. 2) There are other reports of kernel oopses [2] that seem related to the lack of protection. I'd be good to confirm if this patchset solve these problems as well. [1] https://github.com/thesofproject/linux/pull/3146 [2] https://lore.kernel.org/alsa-devel/002f01d7b4f5$c030f4a0$4092dde0$@samsung.… changes since RFCv2: Removal of dpcm_lock to use FE PCM locks (credits to Takashi Iwai for the suggestion). The FE PCM lock is now used before each use of for_each_dpcm_be() - with the exception of the trigger where the lock is already taken. This change is also applied in drivers which make use of this loop (compress, SH, FSL). Addition of BE PCM lock to deal with mutual exclusion between triggers for the same BE. Alignment of the BE atomicity on the FE on connections, this is required to avoid sleeping in atomic context. Additional cleanups (indentation, static functions) changes since RFC v1: Removed unused function Removed exported symbols only used in soc-pcm.c, used static instead Use a mutex instead of a spinlock Protect all for_each_dpcm_be() loops Fix bugs introduced in the refcount Pierre-Louis Bossart (13): ASoC: soc-pcm: remove snd_soc_dpcm_fe_can_update() ASoC: soc-pcm: don't export local functions, use static ASoC: soc-pcm: use proper indentation on 'continue' ASoC: soc-pcm: introduce snd_soc_dpcm_fe_lock_irq/unlock_irq() ASoC: soc-pcm: align BE 'atomicity' with that of the FE ASoC: soc-pcm: remove dpcm spin_lock, use PCM stream lock ASoC: soc-pcm: protect for_each_dpcm_be() loops ASoC: soc-compress: protect for_each_dpcm_be() loops ASoC: sh: rcar: protect for_each_dpcm_be() loops ASoC: fsl: asrc_dma: protect for_each_dpcm_be() loops ASoC: soc-pcm: serialize BE triggers ASoC: soc-pcm: test refcount before triggering ASoC: soc-pcm: fix BE handling of PAUSE_RELEASE include/sound/soc-dpcm.h | 20 +-- include/sound/soc.h | 2 - sound/soc/fsl/fsl_asrc_dma.c | 2 + sound/soc/sh/rcar/core.c | 2 + sound/soc/soc-compress.c | 4 + sound/soc/soc-core.c | 1 - sound/soc/soc-pcm.c | 253 ++++++++++++++++++++++++----------- 7 files changed, 187 insertions(+), 97 deletions(-) -- 2.25.1

3 25