Re: [PATCH v2 4/7] ASoC: mediatek: mt7986: add platform driver

From: Maso Huang (黃加竹)
Date: Thu Jul 13 2023 - 03:44:52 EST

Next message: Juergen Gross: "Re: [PATCH] xen: privcmd: Add support for irqfd"
Previous message: Tian, Kevin: "RE: [PATCH v10 2/7] iommu: Move global PASID allocation from SVA to core"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Wed, 2023-07-05 at 16:03 +0800, Maso Huang wrote:
> On Tue, 2023-07-04 at 11:13 +0200, AngeloGioacchino Del Regno wrote:
> >
> > External email : Please do not click links or open attachments
> > until
> > you have verified the sender or the content.
> > Il 26/06/23 04:34, Maso Huang ha scritto:
> > > Add mt7986 platform driver.
> > >
> > > Signed-off-by: Maso Huang <maso.huang@xxxxxxxxxxxx>
> > > ---
> > > sound/soc/mediatek/Kconfig | 10 +
> > > sound/soc/mediatek/Makefile | 1 +
> > > sound/soc/mediatek/mt7986/Makefile | 9 +
> > > sound/soc/mediatek/mt7986/mt7986-afe-pcm.c | 598
> >
> > +++++++++++++++++++++
> > > 4 files changed, 618 insertions(+)
> > > create mode 100644 sound/soc/mediatek/mt7986/Makefile
> > > create mode 100644 sound/soc/mediatek/mt7986/mt7986-afe-pcm.c
> > >
> >
> > ..snip..
> >
> > > diff --git a/sound/soc/mediatek/mt7986/mt7986-afe-pcm.c
> >
> > b/sound/soc/mediatek/mt7986/mt7986-afe-pcm.c
> > > new file mode 100644
> > > index 000000000000..9eef21762e93
> > > --- /dev/null
> > > +++ b/sound/soc/mediatek/mt7986/mt7986-afe-pcm.c
> > > @@ -0,0 +1,598 @@
> > > +// SPDX-License-Identifier: GPL-2.0
> > > +/*
> > > + * MediaTek ALSA SoC AFE platform driver for MT7986
> > > + *
> > > + * Copyright (c) 2021 MediaTek Inc.
> > > + * Author: Vic Wu <vic.wu@xxxxxxxxxxxx>
> > > + * Maso Huang <maso.huang@xxxxxxxxxxxx>
> > > + */
> > > +
> > > +#include <linux/delay.h>
> > > +#include <linux/module.h>
> > > +#include <linux/of.h>
> > > +#include <linux/of_address.h>
> > > +#include <linux/pm_runtime.h>
> > > +
> > > +#include "mt7986-afe-common.h"
> > > +#include "mt7986-afe-clk.h"
> > > +#include "mt7986-reg.h"
> > > +#include "../common/mtk-afe-platform-driver.h"
> > > +#include "../common/mtk-afe-fe-dai.h"
> > > +
> > > +enum {
> > > +MTK_AFE_RATE_8K = 0,
> > > +MTK_AFE_RATE_11K = 1,
> > > +MTK_AFE_RATE_12K = 2,
> > > +MTK_AFE_RATE_16K = 4,
> > > +MTK_AFE_RATE_22K = 5,
> > > +MTK_AFE_RATE_24K = 6,
> > > +MTK_AFE_RATE_32K = 8,
> > > +MTK_AFE_RATE_44K = 9,
> > > +MTK_AFE_RATE_48K = 10,
> > > +MTK_AFE_RATE_88K = 13,
> > > +MTK_AFE_RATE_96K = 14,
> > > +MTK_AFE_RATE_176K = 17,
> > > +MTK_AFE_RATE_192K = 18,
> > > +};
> > > +
> > > +unsigned int mt7986_afe_rate_transform(struct device *dev,
> > > + unsigned int rate)
> >
> > Fits in one line.
> >
>
> OK.
>
> > > +{
> > > +switch (rate) {
> > > +case 8000:
> > > +return MTK_AFE_RATE_8K;
> > > +case 11025:
> > > +return MTK_AFE_RATE_11K;
> > > +case 12000:
> > > +return MTK_AFE_RATE_12K;
> > > +case 16000:
> > > +return MTK_AFE_RATE_16K;
> > > +case 22050:
> > > +return MTK_AFE_RATE_22K;
> > > +case 24000:
> > > +return MTK_AFE_RATE_24K;
> > > +case 32000:
> > > +return MTK_AFE_RATE_32K;
> > > +case 44100:
> > > +return MTK_AFE_RATE_44K;
> > > +case 48000:
> > > +return MTK_AFE_RATE_48K;
> > > +case 88200:
> > > +return MTK_AFE_RATE_88K;
> > > +case 96000:
> > > +return MTK_AFE_RATE_96K;
> > > +case 176400:
> > > +return MTK_AFE_RATE_176K;
> > > +case 192000:
> > > +return MTK_AFE_RATE_192K;
> > > +default:
> > > +dev_warn(dev, "%s(), rate %u invalid, use %d!!!\n",
> >
> > s/use/using/g
> >
>
> OK.
>
> > > + __func__, rate, MTK_AFE_RATE_48K);
> > > +return MTK_AFE_RATE_48K;
> > > +}
> > > +}
> > > +
> >
> > ..snip..
> >
> > > +
> > > +static bool mt7986_is_volatile_reg(struct device *dev, unsigned
> >
> > int reg)
> > > +{
> > > +/* these auto-gen reg has read-only bit, so put it as volatile
> > > */
> > > +/* volatile reg cannot be cached, so cannot be set when power
> > > off
> >
> > */
> >
> > Fix the comment format please; for multi-line comments, you want:
> >
> > /*
> > * Those auto-gen regs are read-only, so put it a volatile because
> > * volatile registers cannot be cached, which means that they
> > cannot
> > * be set when power is off
> > */
> >
>
> OK.
>
> >
> > > +switch (reg) {
> > > +case AFE_DL0_CUR_MSB:
> > > +case AFE_DL0_CUR:
> > > +case AFE_DL0_RCH_MON:
> > > +case AFE_DL0_LCH_MON:
> > > +case AFE_VUL0_CUR_MSB:
> > > +case AFE_VUL0_CUR:
> > > +case AFE_IRQ_MCU_STATUS:
> > > +case AFE_MEMIF_RD_MON:
> > > +case AFE_MEMIF_WR_MON:
> > > +return true;
> > > +default:
> > > +return false;
> > > +};
> > > +}
> > > +
> >
> > ..snip..
> >
> > > +
> > > +static int mt7986_afe_pcm_dev_probe(struct platform_device
> > > *pdev)
> > > +{
> > > +struct mtk_base_afe *afe;
> > > +struct mt7986_afe_private *afe_priv;
> > > +struct device *dev;
> > > +int i, irq_id, ret;
> > > +
> > > +afe = devm_kzalloc(&pdev->dev, sizeof(*afe), GFP_KERNEL);
> > > +if (!afe)
> > > +return -ENOMEM;
> > > +platform_set_drvdata(pdev, afe);
> > > +
> > > +afe->platform_priv = devm_kzalloc(&pdev->dev, sizeof(*afe_priv),
> > > + GFP_KERNEL);
> > > +if (!afe->platform_priv)
> > > +return -ENOMEM;
> > > +
> > > +afe_priv = afe->platform_priv;
> > > +afe->dev = &pdev->dev;
> > > +dev = afe->dev;
> > > +
> > > +afe->base_addr = devm_platform_ioremap_resource(pdev, 0);
> > > +if (IS_ERR(afe->base_addr))
> > > +return PTR_ERR(afe->base_addr);
> > > +
> > > +/* initial audio related clock */
> > > +ret = mt7986_init_clock(afe);
> >
> > As said in the review for patch [2/7], function mt7986_init_clock()
> > must be moved
> > in this file instead.
> >
>
> OK.
>
> > > +if (ret)
> > > +return dev_err_probe(dev, ret, "Cannot initialize clocks\n");
> > > +
> > > +ret = devm_pm_runtime_enable(dev);
> > > +if (ret)
> > > +return ret;
> > > +
> > > +/* enable clock for regcache get default value from hw */
> > > +afe_priv->pm_runtime_bypass_reg_ctl = true;
> > > +pm_runtime_get_sync(&pdev->dev);
> > > +
> > > +afe->regmap = devm_regmap_init_mmio(&pdev->dev, afe->base_addr,
> > > + &mt7986_afe_regmap_config);
> >
> > pm_runtime_put_sync(&pdev->dev);
> > if (IS_ERR(afe->regmap))
> > return PTR_ERR(afe->regmap);
> >
> > afe_prov->pm_runtime_bypass_reg_ctl = false;
> >
> > that's better :-)
> >
>
> OK.
>
> > > +if (IS_ERR(afe->regmap)) {
> > > +ret = PTR_ERR(afe->regmap);
> > > +goto err_pm_disable;
> > > +}
> > > +
> > > +pm_runtime_put_sync(&pdev->dev);
> > > +afe_priv->pm_runtime_bypass_reg_ctl = false;
> > > +
> > > +/* init memif */
> > > +afe->memif_size = MT7986_MEMIF_NUM;
> > > +afe->memif = devm_kcalloc(dev, afe->memif_size, sizeof(*afe-
> > > memif),
> > > + GFP_KERNEL);
> > > +if (!afe->memif)
> > > +goto err_pm_disable;
> > > +
> > > +for (i = 0; i < afe->memif_size; i++) {
> > > +afe->memif[i].data = &memif_data[i];
> > > +afe->memif[i].irq_usage = -1;
> > > +}
> > > +
> > > +mutex_init(&afe->irq_alloc_lock);
> > > +
> > > +/* irq initialize */
> > > +afe->irqs_size = MT7986_IRQ_NUM;
> > > +afe->irqs = devm_kcalloc(dev, afe->irqs_size, sizeof(*afe-
> > > >irqs),
> > > + GFP_KERNEL);
> > > +if (!afe->irqs)
> > > +goto err_pm_disable;
> > > +
> > > +for (i = 0; i < afe->irqs_size; i++)
> > > +afe->irqs[i].irq_data = &irq_data[i];
> > > +
> > > +/* request irq */
> > > +irq_id = platform_get_irq(pdev, 0);
> > > +if (irq_id < 0) {
> > > +dev_err(dev, "%pOFn no irq found\n", dev->of_node);
> > > +ret = irq_id;
> > > +goto err_pm_disable;
> >
> > You can just return... and please use dev_err_probe().
> >
>
> OK.
>
> > > +}
> > > +ret = devm_request_irq(dev, irq_id, mt7986_afe_irq_handler,
> > > + IRQF_TRIGGER_NONE, "asys-isr", (void *)afe);
> > > +if (ret) {
> > > +dev_err(dev, "could not request_irq for asys-isr\n");
> > > +goto err_pm_disable;
> > > +}
> > > +
> > > +/* init sub_dais */
> > > +INIT_LIST_HEAD(&afe->sub_dais);
> > > +
> > > +for (i = 0; i < ARRAY_SIZE(dai_register_cbs); i++) {
> > > +ret = dai_register_cbs[i](afe);
> > > +if (ret) {
> > > +dev_err(afe->dev, "dai register i %d fail, ret %d\n",
> > > + i, ret);
> > > +goto err_pm_disable;
> > > +}
> > > +}
> > > +
> > > +/* init dai_driver and component_driver */
> > > +ret = mtk_afe_combine_sub_dai(afe);
> > > +if (ret) {
> > > +dev_err(afe->dev, "mtk_afe_combine_sub_dai fail, ret %d\n",
> > > + ret);
> > > +goto err_pm_disable;
> > > +}
> > > +
> > > +afe->mtk_afe_hardware = &mt7986_afe_hardware;
> > > +afe->memif_fs = mt7986_memif_fs;
> > > +afe->irq_fs = mt7986_irq_fs;
> > > +
> > > +afe->runtime_resume = mt7986_afe_runtime_resume;
> > > +afe->runtime_suspend = mt7986_afe_runtime_suspend;
> > > +
> > > +/* register component */
> > > +ret = devm_snd_soc_register_component(&pdev->dev,
> > > + &mt7986_afe_component,
> > > + NULL, 0);
> > > +if (ret) {
> > > +dev_warn(dev, "err_platform\n");
> >
> > That shall not be a dev_warn, but a dev_err() instead, but then,
> > being this
> > a probe function and not needing that jump, it's dev_err_probe().
> >
>
> OK.
>
> > > +goto err_pm_disable;
> > > +}
> > > +
> > > +ret = devm_snd_soc_register_component(afe->dev,
> > > + &mt7986_afe_pcm_dai_component,
> > > + afe->dai_drivers,
> > > + afe->num_dai_drivers);
> > > +if (ret) {
> > > +dev_warn(dev, "err_dai_component\n");
> >
> > Same here.
> >
>
> OK.
>
> > > +goto err_pm_disable;
> > > +}
> > > +
> > > +return ret;
> > > +
> > > +err_pm_disable:
> >
> > You don't need this label because you're now using
> > devm_pm_runtime_enable() and,
> > following my advice from some lines up, pm_runtime_put_sync() is
> > called at the
> > right time already.
> >
> > This means that you can simplify this function a lot with directly
> > returning the
> > error in error paths (including instances of dev_err_probe() ).
> >
>
> OK.
>
> > > +pm_runtime_put_sync(&pdev->dev);
> > > +pm_runtime_disable(&pdev->dev);
> > > +return ret;
> > > +}
> > > +
> > > +static void mt7986_afe_pcm_dev_remove(struct platform_device
> >
> > *pdev)
> > > +{
> > > +pm_runtime_disable(&pdev->dev);
> > > +if (!pm_runtime_status_suspended(&pdev->dev))
> > > +mt7986_afe_runtime_suspend(&pdev->dev);
> > > +}

Hi Angelo,

Since I'm using devm_pm_runtime_enable right now, shoud I remove
mt7986_afe_pcm_dev_remove as well?
It seems no more needed.

Best regards,
Maso

> > > +
> > > +static const struct of_device_id mt7986_afe_pcm_dt_match[] = {
> > > +{ .compatible = "mediatek,mt7986-afe", },
> >
> > You don't need the comma here ^^
> > so this is just
> >
> > { .compatible = "mediatek,mt7986-afe" },
> >
>
> OK.
>
> > Everything else looks fine.
> >
> > Regards,
> > Angelo
>
> Hi Angelo,
>
> Thanks for your review.
> I'll refine all of them in v3 patch.
>
> Best regards,
> Maso
>

Next message: Juergen Gross: "Re: [PATCH] xen: privcmd: Add support for irqfd"
Previous message: Tian, Kevin: "RE: [PATCH v10 2/7] iommu: Move global PASID allocation from SVA to core"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]