Message-ID: <ff5616ca0cbace3c304b94246f405b5831de1a41.camel@mediatek.com>
Subject: Re: [PATCH v1 3/3] mmc: mediatek: add support for SDIO eint irq
From:   Axe Yang <axe.yang@mediatek.com>
To:     AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@collabora.com>,
        "Ulf Hansson" <ulf.hansson@linaro.org>,
        Rob Herring <robh+dt@kernel.org>,
        "Chaotian Jing" <chaotian.jing@mediatek.com>,
        Matthias Brugger <matthias.bgg@gmail.com>,
        Adrian Hunter <adrian.hunter@intel.com>
CC:     Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>,
        Satya Tangirala <satyat@google.com>,
        Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        Wolfram Sang <wsa+renesas@sang-engineering.com>,
        Lucas Stach <dev@lynxeye.de>,
        Eric Biggers <ebiggers@google.com>,
        Andrew Jeffery <andrew@aj.id.au>,
        Stephen Boyd <swboyd@chromium.org>,
        Kiwoong Kim <kwmad.kim@samsung.com>,
        Yue Hu <huyue2@yulong.com>, Tian Tao <tiantao6@hisilicon.com>,
        <linux-mmc@vger.kernel.org>, <devicetree@vger.kernel.org>,
        <linux-kernel@vger.kernel.org>,
        <linux-arm-kernel@lists.infradead.org>,
        <linux-mediatek@lists.infradead.org>
Date:   Fri, 7 Jan 2022 15:53:35 +0800
In-Reply-To: <63cd9b04-8ff5-00ff-bbc4-7bc110080e41@collabora.com>
References: <20211227083641.12538-1-axe.yang@mediatek.com>
         <20211227083641.12538-4-axe.yang@mediatek.com>
         <63cd9b04-8ff5-00ff-bbc4-7bc110080e41@collabora.com>
Content-Type: text/plain; charset="UTF-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Tue, 2022-01-04 at 11:44 +0100, AngeloGioacchino Del Regno wrote:
> Il 27/12/21 09:36, Axe Yang ha scritto:
> > Add support for eint irq when MSDC is used as an SDIO host. This
> > feature requires SDIO device support async irq function. With this
> > feature,SDIO host can be awakened by SDIO card in suspend state,
> > without additional pin.
> > 
> > MSDC driver will time-share the SDIO DAT1 pin. During suspend, MSDC
> > turn off clock and switch SDIO DAT1 pin to GPIO mode. And during
> > resume, switch GPIO function back to DAT1 mode then turn on clock.
> > 
> > Some device tree property should be added or modified in msdc node
> > to support SDIO eint irq. Pinctrls named state_dat1 and state_eint
> > are mandatory. And cap-sdio-async-int flag is necessary since this
> > feature depends on asynchronous interrupt:
> > 	&mmcX {
> > 		...
> > 		pinctrl-names = "default", "state_uhs", "state_eint",
> > 						"state_dat1";
> > 		...
> > 		pinctrl-2 = <&mmc2_pins_eint>;
> > 		pinctrl-3 = <&mmc2_pins_dat1>;
> > 		...
> > 		cap-sdio-async-int;
> > 		...
> > 	};
> > 
> > Signed-off-by: Axe Yang <axe.yang@mediatek.com>
> 
> Hello Axe,
> I agree on Andy's review, and I have some more comments...
> 
> > ---
> >   drivers/mmc/host/mtk-sd.c | 113
> > ++++++++++++++++++++++++++++++++++++--
> >   1 file changed, 107 insertions(+), 6 deletions(-)
> > 
> > diff --git a/drivers/mmc/host/mtk-sd.c b/drivers/mmc/host/mtk-sd.c
> > index 4dfc246c5f95..8f23349f2963 100644
> > --- a/drivers/mmc/host/mtk-sd.c
> > +++ b/drivers/mmc/host/mtk-sd.c
> > @@ -1,6 +1,6 @@
> >   // SPDX-License-Identifier: GPL-2.0-only
> >   /*
> > - * Copyright (c) 2014-2015 MediaTek Inc.
> > + * Copyright (c) 2014-2021 MediaTek Inc.
> >    * Author: Chaotian.Jing <chaotian.jing@mediatek.com>
> >    */
> >   
> > @@ -432,9 +432,13 @@ struct msdc_host {
> >   	struct pinctrl *pinctrl;
> >   	struct pinctrl_state *pins_default;
> >   	struct pinctrl_state *pins_uhs;
> > +	struct pinctrl_state *pins_eint;
> > +	struct pinctrl_state *pins_dat1;
> >   	struct delayed_work req_timeout;
> >   	int irq;		/* host interrupt */
> 
> int eint_irq; should be here, under the host interrupt, for the sake
> of perfection.

Will fix it in next version.

> 
> >   	struct reset_control *reset;
> > +	int eint_irq;		/* device interrupt */
> > +	int sdio_irq_cnt;	/* irq enable cnt */
> >   
> >   	struct clk *src_clk;	/* msdc source clock */
> >   	struct clk *h_clk;      /* msdc h_clk */
> > @@ -1519,10 +1523,12 @@ static void msdc_enable_sdio_irq(struct
> > mmc_host *mmc, int enb)
> >   	__msdc_enable_sdio_irq(host, enb);
> >   	spin_unlock_irqrestore(&host->lock, flags);
> >   
> > -	if (enb)
> > -		pm_runtime_get_noresume(host->dev);
> > -	else
> > -		pm_runtime_put_noidle(host->dev);
> > +	if (mmc->card && !mmc->card->cccr.enable_async_int) {
> > +		if (enb)
> 
> If you are going to rename `enb` to `enable, this means that you're
> changing the
> function signature... so, please, do that as a separated commit.

Okay, will keep 'enb' in this commit.

> 
> > +			pm_runtime_get_noresume(host->dev);
> > +		else
> > +			pm_runtime_put_noidle(host->dev);
> > +	}
> >   }
> >   
> >   static irqreturn_t msdc_cmdq_irq(struct msdc_host *host, u32
> > intsts)
> > @@ -2380,6 +2386,49 @@ static const struct mmc_host_ops mt_msdc_ops
> > = {
> >   	.hw_reset = msdc_hw_reset,
> >   };
> >   
> > +static irqreturn_t msdc_sdio_eint_irq(int irq, void *dev_id)
> > +{
> > +	unsigned long flags;
> > +	struct msdc_host *host = (struct msdc_host *)dev_id;
> > +	struct mmc_host *mmc = mmc_from_priv(host);
> > +
> > +	spin_lock_irqsave(&host->lock, flags);
> > +	if (likely(host->sdio_irq_cnt > 0)) {
> > +		disable_irq_nosync(host->eint_irq);
> > +		disable_irq_wake(host->eint_irq);
> > +		host->sdio_irq_cnt--;
> > +	}
> > +	spin_unlock_irqrestore(&host->lock, flags);
> > +
> > +	sdio_signal_irq(mmc);
> > +
> > +	return IRQ_HANDLED;
> > +}
> > +
> > +static int msdc_request_dat1_eint_irq(struct msdc_host *host)
> > +{
> > +	struct gpio_desc *desc;
> > +	int ret = 0;
> > +	int irq;
> > +
> > +	desc = devm_gpiod_get_index(host->dev, "eint", 0, GPIOD_IN);
> > +	if (IS_ERR(desc))
> > +		return PTR_ERR(desc);
> > +
> > +	irq = gpiod_to_irq(desc);
> > +	if (irq >= 0) {
> > +		irq_set_status_flags(irq, IRQ_NOAUTOEN);
> > +		ret = devm_request_threaded_irq(host->dev, irq, NULL,
> > msdc_sdio_eint_irq,
> > +						IRQF_TRIGGER_LOW |
> > IRQF_ONESHOT,
> > +						"sdio-eint", host);
> 
> I think you may have misunderstood what Andy tried to say in his
> review: your call
> to the function irq_set_status_flags() is using the right flag -- the
> suggestion
> here is to add the IRQF_NO_AUTOEN flag to the call to
> devm_request_threaded_irq()
> (ex: IRQF_TRIGGER_LOW | IRQF_ONESHOT | IRQF_NO_AUTOEN) and remove the
> now redundant
> call to irq_set_status_flags().

Thank you for explanation. Will fix it in next version.

> 
> 
> > +	} else {
> > +		ret = irq;
> > +	}
> > +
> > +	host->eint_irq = irq;
> > +	return ret;
> > +}
> > +
> >   static const struct cqhci_host_ops msdc_cmdq_ops = {
> >   	.enable         = msdc_cqe_enable,
> >   	.disable        = msdc_cqe_disable,
> > @@ -2534,6 +2583,19 @@ static int msdc_drv_probe(struct
> > platform_device *pdev)
> >   		goto host_free;
> >   	}
> >   
> > +	/* Support for SDIO eint irq */
> > +	host->pins_eint = pinctrl_lookup_state(host->pinctrl,
> > "state_eint");
> > +	if (IS_ERR(host->pins_eint)) {
> > +		dev_dbg(&pdev->dev, "Cannot find pinctrl eint!\n");
> > +	} else {
> > +		host->pins_dat1 = pinctrl_lookup_state(host->pinctrl,
> > "state_dat1");
> > +		if (IS_ERR(host->pins_dat1)) {
> > +			ret = PTR_ERR(host->pins_dat1);
> > +			dev_err(&pdev->dev, "Cannot find pinctrl
> > dat1!\n");
> > +			goto host_free;
> > +		}
> > +	}
> > +
> >   	msdc_of_property_parse(pdev, host);
> >   
> >   	host->dev = &pdev->dev;
> > @@ -2621,6 +2683,16 @@ static int msdc_drv_probe(struct
> > platform_device *pdev)
> >   	if (ret)
> >   		goto release;
> >   
> > +	if (!IS_ERR(host->pins_eint) && !IS_ERR(host->pins_dat1)) {
> > +		ret = msdc_request_dat1_eint_irq(host);
> > +		if (ret) {
> > +			dev_err(host->dev, "Failed to register data1
> > eint irq!\n");
> > +			goto release;
> > +		}
> > +
> > +		pinctrl_select_state(host->pinctrl, host->pins_dat1);
> > +	}
> > +
> >   	pm_runtime_set_active(host->dev);
> >   	pm_runtime_set_autosuspend_delay(host->dev,
> > MTK_MMC_AUTOSUSPEND_DELAY);
> >   	pm_runtime_use_autosuspend(host->dev);
> > @@ -2740,21 +2812,50 @@ static void msdc_restore_reg(struct
> > msdc_host *host)
> >   
> >   static int __maybe_unused msdc_runtime_suspend(struct device
> > *dev)
> >   {
> > +	unsigned long flags;
> >   	struct mmc_host *mmc = dev_get_drvdata(dev);
> >   	struct msdc_host *host = mmc_priv(mmc);
> >   
> >   	msdc_save_reg(host);
> > +
> > +	if (!IS_ERR(host->pins_eint)) {
> > +		disable_irq(host->irq);
> > +		pinctrl_select_state(host->pinctrl, host->pins_eint);
> > +		spin_lock_irqsave(&host->lock, flags);
> > +		if (host->sdio_irq_cnt == 0) {
> > +			enable_irq(host->eint_irq);
> > +			enable_irq_wake(host->eint_irq);
> > +			host->sdio_irq_cnt++;
> > +		}
> > +		sdr_clr_bits(host->base + SDC_CFG, SDC_CFG_SDIOIDE);
> > +		spin_unlock_irqrestore(&host->lock, flags);
> > +	}
> >   	msdc_gate_clock(host);
> >   	return 0;
> >   }
> >   
> >   static int __maybe_unused msdc_runtime_resume(struct device *dev)
> >   {
> > +	unsigned long flags;
> >   	struct mmc_host *mmc = dev_get_drvdata(dev);
> >   	struct msdc_host *host = mmc_priv(mmc);
> >   
> >   	msdc_ungate_clock(host);
> >   	msdc_restore_reg(host);
> > +	if (!IS_ERR(host->pins_eint)) {
> > +		spin_lock_irqsave(&host->lock, flags);
> > +		if (host->sdio_irq_cnt > 0) {
> > +			disable_irq_nosync(host->eint_irq);
> > +			disable_irq_wake(host->eint_irq);
> > +			host->sdio_irq_cnt--;
> > +			sdr_set_bits(host->base + SDC_CFG,
> > SDC_CFG_SDIOIDE);
> > +		} else {
> > +			sdr_clr_bits(host->base + MSDC_INTEN,
> > MSDC_INTEN_SDIOIRQ);
> > +		}
> > +		spin_unlock_irqrestore(&host->lock, flags);
> > +		pinctrl_select_state(host->pinctrl, host->pins_uhs);
> > +		enable_irq(host->irq);
> > +	}
> >   	return 0;
> >   }
> >   
> > @@ -2778,7 +2879,7 @@ static int __maybe_unused msdc_resume(struct
> > device *dev)
> >   }
> >   
> >   static const struct dev_pm_ops msdc_dev_pm_ops = {
> > -	SET_SYSTEM_SLEEP_PM_OPS(msdc_suspend, msdc_resume)
> > +	SET_NOIRQ_SYSTEM_SLEEP_PM_OPS(msdc_suspend, msdc_resume)
> 
> Just a nitpick: msdc_{suspend,resume} are now noirq, so you should
> change the
> function names to msdc_suspend_noirq, msdc_resume_noirq.

Will fix it in next version. Thank you for your suggestion.

> 
> >   	SET_RUNTIME_PM_OPS(msdc_runtime_suspend, msdc_runtime_resume,
> > NULL)
> >   };
> >   
> > 
> 
> Regards,
> - Angelo
>