Re: [PATCH 2/2] mailbox: Add support for QTI CPUCP mailbox controller

[Bjorn Andersson <andersson@kernel.org>]

On Wed, Sep 14, 2022 at 11:33:06AM +0530, Sibi Sankar wrote:
> Add support for CPUSS Control Processor (CPUCP) mailbox controller,
> this driver enables communication between AP and CPUCP by acting as
> a doorbell between them.
> 
> Signed-off-by: Gaurav Kohli <gkohli@codeaurora.org>
> [sibis: moved to mailbox and misc. cleanups]
> Signed-off-by: Sibi Sankar <quic_sibis@quicinc.com>
> ---
>  drivers/mailbox/Kconfig           |   9 ++
>  drivers/mailbox/Makefile          |   2 +
>  drivers/mailbox/qcom-cpucp-mbox.c | 176 ++++++++++++++++++++++++++++++++++++++
>  3 files changed, 187 insertions(+)
>  create mode 100644 drivers/mailbox/qcom-cpucp-mbox.c
> 
> diff --git a/drivers/mailbox/Kconfig b/drivers/mailbox/Kconfig
> index 05d6fae800e3..7766e0ad2f12 100644
> --- a/drivers/mailbox/Kconfig
> +++ b/drivers/mailbox/Kconfig
> @@ -284,6 +284,15 @@ config SPRD_MBOX
>  	  to send message between application processors and MCU. Say Y here if
>  	  you want to build the Spreatrum mailbox controller driver.
>  
> +config QCOM_CPUCP_MBOX
> +	tristate "Qualcomm Technologies, Inc. CPUCP mailbox driver"
> +	depends on ARCH_QCOM || COMPILE_TEST
> +	help
> +	  Qualcomm Technologies, Inc. CPUSS Control Processor (CPUCP) mailbox
> +	  controller driver enables communication between AP and CPUCP by
> +	  acting as a doorbell between them. Say Y here if you want to build
> +	  this driver.

What will consume this interface?

Given that there's a single channel, is there any benefit of separating
the interrupt handling out of the single client driver?

> +
>  config QCOM_IPCC
>  	tristate "Qualcomm Technologies, Inc. IPCC driver"
>  	depends on ARCH_QCOM || COMPILE_TEST
> diff --git a/drivers/mailbox/Makefile b/drivers/mailbox/Makefile
> index fc9376117111..195b7e40541f 100644
> --- a/drivers/mailbox/Makefile
> +++ b/drivers/mailbox/Makefile
> @@ -59,6 +59,8 @@ obj-$(CONFIG_SUN6I_MSGBOX)	+= sun6i-msgbox.o
>  
>  obj-$(CONFIG_SPRD_MBOX)		+= sprd-mailbox.o
>  
> +obj-$(CONFIG_QCOM_CPUCP_MBOX)	+= qcom-cpucp-mbox.o
> +
>  obj-$(CONFIG_QCOM_IPCC)		+= qcom-ipcc.o
>  
>  obj-$(CONFIG_APPLE_MAILBOX)	+= apple-mailbox.o
> diff --git a/drivers/mailbox/qcom-cpucp-mbox.c b/drivers/mailbox/qcom-cpucp-mbox.c
> new file mode 100644
> index 000000000000..063bb2d80f3e
> --- /dev/null
> +++ b/drivers/mailbox/qcom-cpucp-mbox.c
> @@ -0,0 +1,176 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Copyright (c) 2022 Qualcomm Innovation Center, Inc. All rights reserved.
> + */
> +
> +#include <linux/interrupt.h>
> +#include <linux/irq.h>
> +#include <linux/irqdomain.h>
> +#include <linux/mailbox_controller.h>
> +#include <linux/module.h>
> +#include <linux/platform_device.h>
> +
> +/* CPUCP Register offsets */
> +#define CPUCP_INTR_CLEAR_REG		0x8
> +#define CPUCP_INTR_STATUS_REG		0xC
> +#define CPUCP_SEND_IRQ_REG		0xC
> +
> +#define CPUCP_IRQ_CLEAR			BIT(3)
> +#define CPUCP_IRQ_STATUS_ASSERTED	BIT(3)
> +#define CPUCP_SEND_IRQ			BIT(28)
> +
> +/**
> + * struct qcom_cpucp_mbox - Holder for the mailbox driver
> + * @mbox:			The mailbox controller
> + * @chan:			The mailbox channel
> + * @tx_base:			Base address of the CPUCP tx registers
> + * @rx_base:			Base address of the CPUCP rx registers
> + * @dev:			Device associated with this instance
> + * @lock:			Lock protecting private
> + * @irq:			CPUCP to AP irq
> + */
> +struct qcom_cpucp_mbox {
> +	struct mbox_controller mbox;
> +	struct mbox_chan chan;
> +	void __iomem *tx_base;
> +	void __iomem *rx_base;
> +	struct device *dev;
> +	int irq;
> +
> +	/* control access to the chan private data */
> +	spinlock_t lock;
> +};
> +
> +static inline struct qcom_cpucp_mbox *to_qcom_cpucp_mbox(struct mbox_controller *mbox)
> +{
> +	return container_of(mbox, struct qcom_cpucp_mbox, mbox);
> +}
> +
> +static irqreturn_t qcom_cpucp_mbox_irq_fn(int irq, void *data)
> +{
> +	struct qcom_cpucp_mbox *cpucp = data;
> +	unsigned long flags;
> +	u32 val;
> +
> +	val = readl(cpucp->rx_base + CPUCP_INTR_STATUS_REG);
> +	if (val & CPUCP_IRQ_STATUS_ASSERTED) {
> +		writel(CPUCP_IRQ_CLEAR, cpucp->rx_base + CPUCP_INTR_CLEAR_REG);
> +
> +		spin_lock_irqsave(&cpucp->lock, flags);
> +		if (cpucp->chan.con_priv)
> +			mbox_chan_received_data(&cpucp->chan, NULL);

Afaict this will deliver a data-less message to the mailbox receiver to
communicate an incoming interrupt.

I would prefer that you represent this as an irq_chip, like we've done
for similar designs previously.

> +		spin_unlock_irqrestore(&cpucp->lock, flags);
> +	}
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static int qcom_cpucp_mbox_send_data(struct mbox_chan *chan, void *data)
> +{
> +	struct qcom_cpucp_mbox *cpucp = to_qcom_cpucp_mbox(chan->mbox);
> +
> +	writel(CPUCP_SEND_IRQ, cpucp->tx_base + CPUCP_SEND_IRQ_REG);
> +
> +	return 0;
> +}
> +
> +static void qcom_cpucp_mbox_shutdown(struct mbox_chan *chan)
> +{
> +	struct qcom_cpucp_mbox *cpucp = to_qcom_cpucp_mbox(chan->mbox);
> +	unsigned long flags;
> +
> +	spin_lock_irqsave(&cpucp->lock, flags);
> +	chan->con_priv = NULL;
> +	spin_unlock_irqrestore(&cpucp->lock, flags);
> +}
> +
> +static const struct mbox_chan_ops cpucp_mbox_chan_ops = {
> +	.send_data = qcom_cpucp_mbox_send_data,
> +	.shutdown = qcom_cpucp_mbox_shutdown
> +};
> +
> +static struct mbox_chan *qcom_cpucp_mbox_xlate(struct mbox_controller *mbox,
> +					       const struct of_phandle_args *ph)
> +{
> +	struct qcom_cpucp_mbox *cpucp = to_qcom_cpucp_mbox(mbox);
> +
> +	if (ph->args_count != 0)
> +		return ERR_PTR(-EINVAL);
> +
> +	if (mbox->chans[0].con_priv)
> +		return ERR_PTR(-EBUSY);
> +
> +	mbox->chans[0].con_priv = cpucp;
> +
> +	return &mbox->chans[0];
> +}
> +
> +static int qcom_cpucp_mbox_probe(struct platform_device *pdev)
> +{
> +	struct qcom_cpucp_mbox *cpucp;
> +	struct mbox_controller *mbox;
> +	int ret;
> +
> +	cpucp = devm_kzalloc(&pdev->dev, sizeof(*cpucp), GFP_KERNEL);
> +	if (!cpucp)
> +		return -ENOMEM;
> +
> +	spin_lock_init(&cpucp->lock);
> +	cpucp->dev = &pdev->dev;
> +
> +	cpucp->tx_base = devm_platform_ioremap_resource(pdev, 0);
> +	if (IS_ERR(cpucp->tx_base))
> +		return PTR_ERR(cpucp->tx_base);
> +
> +	cpucp->rx_base = devm_platform_ioremap_resource(pdev, 1);
> +	if (IS_ERR(cpucp->rx_base))
> +		return PTR_ERR(cpucp->rx_base);
> +
> +	cpucp->irq = platform_get_irq(pdev, 0);
> +	if (cpucp->irq < 0)
> +		return cpucp->irq;
> +
> +	mbox = &cpucp->mbox;
> +	mbox->dev = cpucp->dev;
> +	mbox->num_chans = 1;
> +	mbox->chans = &cpucp->chan;
> +	mbox->ops = &cpucp_mbox_chan_ops;
> +	mbox->of_xlate = qcom_cpucp_mbox_xlate;
> +	mbox->txdone_irq = false;
> +	mbox->txdone_poll = false;
> +
> +	ret = devm_mbox_controller_register(&pdev->dev, mbox);
> +	if (ret)
> +		return ret;
> +
> +	ret = devm_request_irq(&pdev->dev, cpucp->irq, qcom_cpucp_mbox_irq_fn,
> +			       IRQF_TRIGGER_HIGH | IRQF_NO_SUSPEND, "qcom_cpucp_mbox", cpucp);

Please rely on trigger from Devicetree.

> +	if (ret < 0) {
> +		dev_err(&pdev->dev, "Failed to register the irq: %d\n", ret);
> +		return ret;
> +	}
> +
> +	platform_set_drvdata(pdev, cpucp);

drvdata seems to be unused.

Regards,
Bjorn
> +
> +	return 0;
> +}
> +
> +static const struct of_device_id qcom_cpucp_mbox_of_match[] = {
> +	{ .compatible = "qcom,cpucp-mbox"},
> +	{}
> +};
> +MODULE_DEVICE_TABLE(of, qcom_cpucp_mbox_of_match);
> +
> +static struct platform_driver qcom_cpucp_mbox_driver = {
> +	.probe = qcom_cpucp_mbox_probe,
> +	.driver = {
> +		.name = "qcom_cpucp_mbox",
> +		.of_match_table = qcom_cpucp_mbox_of_match,
> +	},
> +};
> +module_platform_driver(qcom_cpucp_mbox_driver);
> +
> +MODULE_AUTHOR("Gaurav Kohli <gkohli@codeaurora.org>");
> +MODULE_AUTHOR("Sibi Sankar <quic_sibis@qti.qualcomm.com>");
> +MODULE_DESCRIPTION("Qualcomm Technologies, Inc. CPUSS Control Processor Mailbox driver");
> +MODULE_LICENSE("GPL");
> -- 
> 2.7.4
>