Linaro-mm-sig

linaro-mm-sig@lists.linaro.org

45 participants
6400 discussions

Re: [PATCH] dma-buf/heaps: Correct the types of fd_flags and heap_flags

by Carlos Llamas

On Thu, Jun 06, 2024 at 02:02:13PM +1200, Barry Song wrote: > From: Barry Song <v-songbaohua(a)oppo.com> > > dma_heap_allocation_data defines the UAPI as follows: > > struct dma_heap_allocation_data { > __u64 len; > __u32 fd; > __u32 fd_flags; > __u64 heap_flags; > }; > > But dma heaps are casting both fd_flags and heap_flags into > unsigned long. This patch makes dma heaps - cma heap and > system heap have consistent types with UAPI. > > Signed-off-by: Barry Song <v-songbaohua(a)oppo.com> > --- Looks good to me, thanks! Reviewed-by: Carlos Llamas <cmllamas(a)google.com>

1 year, 11 months

Re: [v11 3/7] iio: core: Add new DMABUF interface infrastructure

by Paul Cercueil

Le mercredi 19 juin 2024 à 13:56 +0200, Markus Elfring a écrit : > … > > https://lore.kernel.org/linux-iio/219abc43b4fdd4a13b307ed2efaa0e6869e68e3f.… > > > > (and responses below) > > > > It's more nuanced than I remembered. > … > > > > > * Will the desire grow for further collateral evolution according > > > to > > > affected software components? > > > > Not sure what you mean by that. > > Advanced programming interfaces were added a while ago. > > Example: > https://elixir.bootlin.com/linux/v6.10-rc4/source/include/linux/cleanup.h#L8 > > Corresponding attempts for increasing API usage need to adapt to > remaining change reluctance, > don't they? Sure, I guess. But that does not change the fact that I cannot use cleanup.h magic in this patchset, yet, as the required changes would have to be done in a separate one. Cheers, -Paul

1 year, 11 months

[PATCH v11 0/7] iio: new DMABUF based API v11

by Paul Cercueil

Hi Jonathan, Here's the v11 of my patchset that introduces DMABUF support to IIO. It addresses the few points that were raised in the review of the v10. It also adds Nuno as a co-developer. Changelog: - [3/7]: - Document .lock_queue / .unlock_queue buffer callbacks - Add small comment to explain what the spinlock protects - Use guard(mutex) - [4/7]: - Remove useless field "attach" in struct iio_dma_buffer_block - Document "sg_table" and "fence" fields in struct iio_block_state - [6/7]: - "a IIO buffer" -> "an IIO buffer" - Add variable name in IOCTL calls - [7/7]: New patch, to document the DMA changes Cheers, -Paul Paul Cercueil (7): dmaengine: Add API function dmaengine_prep_peripheral_dma_vec() dmaengine: dma-axi-dmac: Implement device_prep_peripheral_dma_vec iio: core: Add new DMABUF interface infrastructure iio: buffer-dma: Enable support for DMABUFs iio: buffer-dmaengine: Support new DMABUF based userspace API Documentation: iio: Document high-speed DMABUF based API Documentation: dmaengine: Document new dma_vec API Documentation/driver-api/dmaengine/client.rst | 9 + .../driver-api/dmaengine/provider.rst | 10 + Documentation/iio/iio_dmabuf_api.rst | 54 +++ Documentation/iio/index.rst | 1 + drivers/dma/dma-axi-dmac.c | 40 ++ drivers/iio/Kconfig | 1 + drivers/iio/buffer/industrialio-buffer-dma.c | 178 ++++++- .../buffer/industrialio-buffer-dmaengine.c | 62 ++- drivers/iio/industrialio-buffer.c | 457 ++++++++++++++++++ include/linux/dmaengine.h | 33 ++ include/linux/iio/buffer-dma.h | 31 ++ include/linux/iio/buffer_impl.h | 33 ++ include/uapi/linux/iio/buffer.h | 22 + 13 files changed, 911 insertions(+), 20 deletions(-) create mode 100644 Documentation/iio/iio_dmabuf_api.rst -- 2.43.0

1 year, 11 months

Re: [PATCH v11 3/7] iio: core: Add new DMABUF interface infrastructure

by Paul Cercueil

Le mercredi 19 juin 2024 à 13:43 +0200, Markus Elfring a écrit : > … > > +++ b/drivers/iio/industrialio-buffer.c > … > > +static void iio_buffer_dmabuf_release(struct kref *ref) > > +{ > … > > + dma_resv_lock(dmabuf->resv, NULL); > > + dma_buf_unmap_attachment(attach, priv->sgt, priv->dir); > > + dma_resv_unlock(dmabuf->resv); > … > > Under which circumstances will another lock guard become applicable? > https://elixir.bootlin.com/linux/v6.10-rc4/source/include/linux/cleanup.h#L… As soon as "struct dma_resv" gets a DEFINE_GUARD(). -Paul

1 year, 11 months

Re: [v11 3/7] iio: core: Add new DMABUF interface infrastructure

by Paul Cercueil

Le mercredi 19 juin 2024 à 13:13 +0200, Markus Elfring a écrit : > > > Would you dare to transform the remaining goto chain into further > > > applications > > > of scope-based resource management? > > > > We discussed this after v6 or v7, DRM/DMABUF maintainers were not > > keen > > on doing that *just yet*. > > * Would you like to add any links for corresponding development > discussions? https://lore.kernel.org/linux-iio/219abc43b4fdd4a13b307ed2efaa0e6869e68e3f.… (and responses below) It's more nuanced than I remembered. Christian was OK to add cleanup.h support to the DMABUF code as long as the examples were updated as well, but those aren't good candidates as they don't free up the resources in all code paths. > > * Will the desire grow for further collateral evolution according to > affected software components? Not sure what you mean by that. Cheers, -Paul

1 year, 11 months

Re: [v11 3/7] iio: core: Add new DMABUF interface infrastructure

by Paul Cercueil

Hi Markus, Le mercredi 19 juin 2024 à 12:03 +0200, Markus Elfring a écrit : > … > > All errors (new ones prefixed by >>): > > > > > > drivers/iio/industrialio-buffer.c:1715:3: error: cannot jump > > > > from this goto statement to its label > > 1715 | goto err_dmabuf_unmap_attachment; > … > > Would you dare to transform the remaining goto chain into further > applications > of scope-based resource management? We discussed this after v6 or v7, DRM/DMABUF maintainers were not keen on doing that *just yet*. Cheers, -Paul

1 year, 11 months

Re: [PATCH RESEND v5 16/16] arm64: dts: mediatek: add audio support for mt8365-evk

by AngeloGioacchino Del Regno

Il 14/06/24 09:27, Alexandre Mergnat ha scritto: > Add the sound node which is linked to the MT8365 SoC AFE and > the MT6357 audio codec. > > Update the file header. > > Signed-off-by: Alexandre Mergnat <amergnat(a)baylibre.com> > --- > arch/arm64/boot/dts/mediatek/mt8365-evk.dts | 89 +++++++++++++++++++++++++++++ > 1 file changed, 89 insertions(+) > > diff --git a/arch/arm64/boot/dts/mediatek/mt8365-evk.dts b/arch/arm64/boot/dts/mediatek/mt8365-evk.dts > index 50cbaefa1a99..1d5457f9a4c2 100644 > --- a/arch/arm64/boot/dts/mediatek/mt8365-evk.dts > +++ b/arch/arm64/boot/dts/mediatek/mt8365-evk.dts > @@ -4,6 +4,7 @@ > * Authors: > * Fabien Parent <fparent(a)baylibre.com> > * Bernhard Rosenkränzer <bero(a)baylibre.com> > + * Alexandre Mergnat <amergnat(a)baylibre.com> > */ > > /dts-v1/; > @@ -86,6 +87,28 @@ optee_reserved: optee@43200000 { > reg = <0 0x43200000 0 0x00c00000>; > }; > }; > + > + sound: sound { > + compatible = "mediatek,mt8365-mt6357"; > + pinctrl-names = "default", > + "dmic", > + "miso_off", > + "miso_on", > + "mosi_off", > + "mosi_on"; > + pinctrl-0 = <&aud_default_pins>; > + pinctrl-1 = <&aud_dmic_pins>; > + pinctrl-2 = <&aud_miso_off_pins>; > + pinctrl-3 = <&aud_miso_on_pins>; > + pinctrl-4 = <&aud_mosi_off_pins>; > + pinctrl-5 = <&aud_mosi_on_pins>; > + mediatek,platform = <&afe>; > + }; > +}; > + > +&afe { > + mediatek,dmic-mode = <1>; > + status = "okay"; > }; > > &cpu0 { > @@ -178,9 +201,75 @@ &mt6357_pmic { > interrupts-extended = <&pio 145 IRQ_TYPE_LEVEL_HIGH>; > interrupt-controller; > #interrupt-cells = <2>; > + vaud28-supply = <&mt6357_vaud28_reg>; > + audio-codec { > + mediatek,micbias0-microvolt = <1900000>; > + mediatek,micbias1-microvolt = <1700000>; > + }; > }; > > &pio { > + aud_default_pins: audiodefault-pins { > + pins { clk-dat-pins > + pinmux = <MT8365_PIN_72_CMDAT4__FUNC_I2S3_BCK>, > + <MT8365_PIN_73_CMDAT5__FUNC_I2S3_LRCK>, > + <MT8365_PIN_74_CMDAT6__FUNC_I2S3_MCK>, > + <MT8365_PIN_75_CMDAT7__FUNC_I2S3_DO>; > + }; > + }; > + > + aud_dmic_pins: audiodmic-pins { > + pins { clk-dat-pins > + pinmux = <MT8365_PIN_117_DMIC0_CLK__FUNC_DMIC0_CLK>, > + <MT8365_PIN_118_DMIC0_DAT0__FUNC_DMIC0_DAT0>, > + <MT8365_PIN_119_DMIC0_DAT1__FUNC_DMIC0_DAT1>; > + }; > + }; > + > + aud_miso_off_pins: misooff-pins { > + pins { clk-dat-pins and same for all the others. > + pinmux = <MT8365_PIN_53_AUD_CLK_MISO__FUNC_GPIO53>, > + <MT8365_PIN_54_AUD_SYNC_MISO__FUNC_GPIO54>, > + <MT8365_PIN_55_AUD_DAT_MISO0__FUNC_GPIO55>, > + <MT8365_PIN_56_AUD_DAT_MISO1__FUNC_GPIO56>; > + input-enable; > + bias-pull-down; > + drive-strength = <MTK_DRIVE_2mA>; drive-strength = <2>; > + }; > + }; > + > + aud_miso_on_pins: misoon-pins { > + pins { > + pinmux = <MT8365_PIN_53_AUD_CLK_MISO__FUNC_AUD_CLK_MISO>, > + <MT8365_PIN_54_AUD_SYNC_MISO__FUNC_AUD_SYNC_MISO>, > + <MT8365_PIN_55_AUD_DAT_MISO0__FUNC_AUD_DAT_MISO0>, > + <MT8365_PIN_56_AUD_DAT_MISO1__FUNC_AUD_DAT_MISO1>; > + drive-strength = <MTK_DRIVE_6mA>; = <6>; > + }; > + }; > + > + aud_mosi_off_pins: mosioff-pins { > + pins { > + pinmux = <MT8365_PIN_49_AUD_CLK_MOSI__FUNC_GPIO49>, > + <MT8365_PIN_50_AUD_SYNC_MOSI__FUNC_GPIO50>, > + <MT8365_PIN_51_AUD_DAT_MOSI0__FUNC_GPIO51>, > + <MT8365_PIN_52_AUD_DAT_MOSI1__FUNC_GPIO52>; > + input-enable; > + bias-pull-down; > + drive-strength = <MTK_DRIVE_2mA>; = <2>; > + }; > + }; > + > + aud_mosi_on_pins: mosion-pins { > + pins { > + pinmux = <MT8365_PIN_49_AUD_CLK_MOSI__FUNC_AUD_CLK_MOSI>, > + <MT8365_PIN_50_AUD_SYNC_MOSI__FUNC_AUD_SYNC_MOSI>, > + <MT8365_PIN_51_AUD_DAT_MOSI0__FUNC_AUD_DAT_MOSI0>, > + <MT8365_PIN_52_AUD_DAT_MOSI1__FUNC_AUD_DAT_MOSI1>; > + drive-strength = <MTK_DRIVE_6mA>; = <6>; Cheers, Angelo

1 year, 11 months

Re: [PATCH RESEND v5 15/16] arm64: dts: mediatek: add afe support for mt8365 SoC

by AngeloGioacchino Del Regno

Il 14/06/24 09:27, Alexandre Mergnat ha scritto: > Add audio front end support of MT8365 SoC. > Update the file header. > > Signed-off-by: Alexandre Mergnat <amergnat(a)baylibre.com> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno(a)collabora.com>

1 year, 11 months

Re: [PATCH RESEND v5 04/16] ASoC: mediatek: mt8365: Add common header

by AngeloGioacchino Del Regno

Il 14/06/24 09:27, Alexandre Mergnat ha scritto: > Add header files for register definition and structure. > > Signed-off-by: Alexandre Mergnat <amergnat(a)baylibre.com> > --- > sound/soc/mediatek/mt8365/mt8365-afe-common.h | 491 +++++++++++++ > sound/soc/mediatek/mt8365/mt8365-reg.h | 991 ++++++++++++++++++++++++++ > 2 files changed, 1482 insertions(+) > > diff --git a/sound/soc/mediatek/mt8365/mt8365-afe-common.h b/sound/soc/mediatek/mt8365/mt8365-afe-common.h > new file mode 100644 > index 000000000000..4d8f8c4b19e3 > --- /dev/null > +++ b/sound/soc/mediatek/mt8365/mt8365-afe-common.h > @@ -0,0 +1,491 @@ > +/* SPDX-License-Identifier: GPL-2.0 > + * > + * Mediatek 8365 audio driver common definitions > + * > + * Copyright (c) 2024 MediaTek Inc. > + * Authors: Jia Zeng <jia.zeng(a)mediatek.com> > + * Alexandre Mergnat <amergnat(a)baylibre.com> > + */ > + > +#ifndef _MT8365_AFE_COMMON_H_ > +#define _MT8365_AFE_COMMON_H_ > + > +#define COMMON_CLOCK_FRAMEWORK_API > +#define IDLE_TASK_DRIVER_API > +#define ENABLE_AFE_APLL_TUNER Those three definitions do not exist upstream. Please remove. > + > +#include <linux/clk.h> > +#include <linux/list.h> > +#include <linux/regmap.h> > +#include <sound/soc.h> > +#include <sound/asound.h> > +#include "../common/mtk-base-afe.h" > +#include "mt8365-reg.h" > + > +#define ENUM_TO_STR(enum) #enum Unused definition > + > +#define snd_soc_dai_stream_active_playback(dai) \ > + snd_soc_dai_stream_active(dai, SNDRV_PCM_STREAM_PLAYBACK) > +#define snd_soc_dai_stream_active_capture(dai) \ > + snd_soc_dai_stream_active(dai, SNDRV_PCM_STREAM_CAPTURE) > + Those are used only once and only in mt8365-dai-pcm.c, and I just noticed that. Can you please just remove those two and directly call snd_soc_dai_stream_active() with the right params in function mt8365_dai_pcm1_prepare()? > +enum { > + MT8365_AFE_MEMIF_DL1, > + MT8365_AFE_MEMIF_DL2, > + MT8365_AFE_MEMIF_TDM_OUT, ..snip.. > + > +#ifdef CONFIG_MTK_HIFIXDSP_SUPPORT This configuration option doesn't exist. Please remove the ifdef and the enclosed code entirely, as it's unused. > +struct mt8365_adsp_data { > + /* information adsp supply */ > + bool adsp_on; > + int (*hostless_active)(void); > + /* information afe supply */ > + int (*set_afe_memif)(struct mtk_base_afe *afe, > + int memif_id, > + unsigned int rate, > + unsigned int channels, > + snd_pcm_format_t format); > + int (*set_afe_memif_enable)(struct mtk_base_afe *afe, > + int memif_id, > + unsigned int rate, > + unsigned int period_size, > + int enable); > + void (*get_afe_memif_sram)(struct mtk_base_afe *afe, > + int memif_id, > + unsigned int *paddr, > + unsigned int *size); > + void (*set_afe_init)(struct mtk_base_afe *afe); > + void (*set_afe_uninit)(struct mtk_base_afe *afe); > +}; > +#endif > + > +struct mt8365_afe_private { > + struct clk *clocks[MT8365_CLK_NUM]; > + struct regmap *topckgen; > + struct mt8365_fe_dai_data fe_data[MT8365_AFE_MEMIF_NUM]; > + struct mt8365_be_dai_data be_data[MT8365_AFE_BACKEND_NUM]; > + struct mt8365_control_data ctrl_data; > + struct mt8365_gasrc_data gasrc_data[MT8365_TDM_ASRC_NUM]; > +#ifdef CONFIG_MTK_HIFIXDSP_SUPPORT > + struct mt8365_adsp_data adsp_data; ditto > +#endif > + int afe_on_ref_cnt; > + int top_cg_ref_cnt[MT8365_TOP_CG_NUM]; > + void __iomem *afe_sram_vir_addr; > + unsigned int afe_sram_phy_addr; > + unsigned int afe_sram_size; > + /* locks */ > + spinlock_t afe_ctrl_lock; > + struct mutex afe_clk_mutex; /* Protect & sync APLL TUNER registers access*/ > +#ifdef CONFIG_DEBUG_FS > + struct dentry *debugfs_dentry[MT8365_AFE_DEBUGFS_NUM]; > +#endif > + int apll_tuner_ref_cnt[MT8365_AFE_APLL_NUM]; > + unsigned int tdm_out_mode; > + unsigned int cm2_mux_input; > + > + /* dai */ > + bool dai_on[MT8365_AFE_BACKEND_END]; > + void *dai_priv[MT8365_AFE_BACKEND_END]; > +}; > + .... > +#ifdef CONFIG_MTK_HIFIXDSP_SUPPORT same > +struct mtk_base_afe *mt8365_afe_pcm_get_info(void); > +#endif > + > +int mt8365_dai_i2s_register(struct mtk_base_afe *afe); > +int mt8365_dai_set_priv(struct mtk_base_afe *afe, > + int id, > + int priv_size, > + const void *priv_data); > + Everything else looks good. After applying the proposed cleanups, Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno(a)collabora.com> Cheers, Angelo

1 year, 11 months

Re: [PATCH RESEND v5 13/16] ASoC: mediatek: Add MT8365 support

by AngeloGioacchino Del Regno

Il 14/06/24 09:27, Alexandre Mergnat ha scritto: > - Add specific config to enable: > - MT8365 sound support > - MT6357 audio codec support > - Add the mt8365 directory and all drivers under it. > > Signed-off-by: Alexandre Mergnat <amergnat(a)baylibre.com> > --- > sound/soc/mediatek/Kconfig | 20 ++++++++++++++++++++ > sound/soc/mediatek/Makefile | 1 + > sound/soc/mediatek/mt8365/Makefile | 15 +++++++++++++++ > 3 files changed, 36 insertions(+) > > diff --git a/sound/soc/mediatek/Kconfig b/sound/soc/mediatek/Kconfig > index 5a8476e1ecca..c1c86ecc0601 100644 > --- a/sound/soc/mediatek/Kconfig > +++ b/sound/soc/mediatek/Kconfig > @@ -298,3 +298,23 @@ config SND_SOC_MT8195_MT6359 > boards with the MT6359 and other I2S audio codecs. > Select Y if you have such device. > If unsure select "N". > + > +config SND_SOC_MT8365 > + tristate "ASoC support for Mediatek MT8365 chip" > + depends on ARCH_MEDIATEK > + select SND_SOC_MEDIATEK > + help > + This adds ASoC platform driver support for Mediatek MT8365 chip > + that can be used with other codecs. > + Select Y if you have such device. > + If unsure select "N". > + > +config SND_SOC_MT8365_MT6357 > + tristate "ASoC Audio driver for MT8365 with MT6357 codec" > + depends on SND_SOC_MT8365 && MTK_PMIC_WRAP > + select SND_SOC_MT6357 > + help > + This adds support for ASoC machine driver for Mediatek MT8365 s/Mediatek/MediaTek/g (here and everywhere else in other commits as well) > + boards with the MT6357 codec. MT6357 PMIC codec after which Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno(a)collabora.com>

1 year, 11 months

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

Linaro-mm-sig