[PATCH 4/5] PCI: stm32: Add PCIe endpoint support for STM32MP25

[Bjorn Helgaas]

On Tue, Nov 12, 2024 at 05:19:24PM +0100, Christian Bruel wrote:
> Add driver to configure the STM32MP25 SoC PCIe Gen2 controller based on the
> DesignWare PCIe core in endpoint mode.
> Uses the common reference clock provided by the host.

> +++ b/drivers/pci/controller/dwc/Kconfig

> +config PCIE_STM32_EP
> +	tristate "STMicroelectronics STM32MP25 PCIe Controller (endpoint mode)"
> +	depends on ARCH_STM32 || COMPILE_TEST
> +	depends on PCI_ENDPOINT
> +	select PCIE_DW_EP
> +	help
> +	  Enables endpoint support for DesignWare core based PCIe controller in found
> +	  in STM32MP25 SoC.
> +
> +	  This driver can also be built as a module. If so, the module
> +	  will be called pcie-stm32-ep.

Move as for the host mode entry.

> +++ b/drivers/pci/controller/dwc/pcie-stm32-ep.c

> +static const struct of_device_id stm32_pcie_ep_of_match[] = {
> +	{ .compatible = "st,stm32mp25-pcie-ep" },
> +	{},
> +};

Move next to stm32_pcie_ep_driver.

> +static void stm32_pcie_ep_init(struct dw_pcie_ep *ep)
> +{
> +	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> +	struct stm32_pcie *stm32_pcie = to_stm32_pcie(pci);
> +	enum pci_barno bar;
> +
> +	for (bar = BAR_0; bar <= PCI_STD_NUM_BARS; bar++)

Most users just use "bar = 0".  BAR_0 is 0, but there's no real
connection with PCI_STD_NUM_BARS, so I think 0 is probably better.

Looks like this should be "bar < PCI_STD_NUM_BARS"?

> +		dw_pcie_ep_reset_bar(pci, bar);
> +
> +	/* Defer Completion Requests until link started */

Not sure what a Completion Request is.  Is this some internal STM or
DWC thing?  Or is this related to Request Retry Status completions for
config requests?

> +	regmap_update_bits(stm32_pcie->regmap, SYSCFG_PCIECR,
> +			   STM32MP25_PCIECR_REQ_RETRY_EN,
> +			   STM32MP25_PCIECR_REQ_RETRY_EN);
> +}

> +static int stm32_pcie_raise_irq(struct dw_pcie_ep *ep, u8 func_no,
> +				unsigned int type, u16 interrupt_num)
> +{
> +	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> +
> +	switch (type) {
> +	case PCI_IRQ_INTX:
> +		return dw_pcie_ep_raise_intx_irq(ep, func_no);
> +	case PCI_IRQ_MSI:
> +		return dw_pcie_ep_raise_msi_irq(ep, func_no, interrupt_num);
> +	default:
> +		dev_err(pci->dev, "UNKNOWN IRQ type\n");
> +		return -EINVAL;
> +	}
> +
> +	return 0;

Is the compiler not smart enough to notice that this is unreachable?

> +static void stm32_pcie_perst_deassert(struct dw_pcie *pci)
> +{
> +	struct stm32_pcie *stm32_pcie = to_stm32_pcie(pci);
> +	struct device *dev = pci->dev;
> +	struct dw_pcie_ep *ep = &pci->ep;
> +	int ret;
> +
> +	if (stm32_pcie->link_status == STM32_PCIE_EP_LINK_ENABLED) {
> +		dev_dbg(pci->dev, "Link is already enabled\n");
> +		return;
> +	}
> +
> +	dev_dbg(dev, "PERST de-asserted by host. Starting link training\n");
> +
> +	ret = pm_runtime_resume_and_get(dev);
> +	if (ret < 0) {
> +		dev_err(dev, "pm runtime resume failed: %d\n", ret);
> +		return;
> +	}
> +
> +	ret = stm32_pcie_enable_resources(stm32_pcie);
> +	if (ret) {
> +		dev_err(dev, "Failed to enable resources: %d\n", ret);
> +		pm_runtime_put_sync(dev);
> +		return;
> +	}
> +
> +	ret = dw_pcie_ep_init_registers(ep);
> +	if (ret) {
> +		dev_err(dev, "Failed to complete initialization: %d\n", ret);
> +		stm32_pcie_disable_resources(stm32_pcie);
> +		pm_runtime_put_sync(dev);
> +		return;
> +	}
> +
> +	pci_epc_init_notify(ep->epc);
> +
> +	ret = stm32_pcie_enable_link(pci);
> +	if (ret) {
> +		dev_err(dev, "PCIe Cannot establish link: %d\n", ret);
> +		stm32_pcie_disable_resources(stm32_pcie);
> +		pm_runtime_put_sync(dev);
> +		return;
> +	}
> +
> +	stm32_pcie->link_status = STM32_PCIE_EP_LINK_ENABLED;
> +}

> +static int stm32_add_pcie_ep(struct stm32_pcie *stm32_pcie,
> +			     struct platform_device *pdev)
> +{
> +	struct dw_pcie *pci = stm32_pcie->pci;
> +	struct dw_pcie_ep *ep = &pci->ep;
> +	struct device *dev = &pdev->dev;
> +	int ret;
> +
> +	ret = regmap_update_bits(stm32_pcie->regmap, SYSCFG_PCIECR,
> +				 STM32MP25_PCIECR_TYPE_MASK,
> +				 STM32MP25_PCIECR_EP);
> +	if (ret)
> +		return ret;
> +
> +	ret = pm_runtime_resume_and_get(dev);
> +	if (ret < 0) {
> +		dev_err(dev, "pm runtime resume failed: %d\n", ret);
> +		return ret;
> +	}
> +
> +	reset_control_assert(stm32_pcie->rst);
> +	reset_control_deassert(stm32_pcie->rst);
> +
> +	ep->ops = &stm32_pcie_ep_ops;
> +
> +	ret = dw_pcie_ep_init(ep);
> +	if (ret) {
> +		dev_err(dev, "failed to initialize ep: %d\n", ret);
> +		pm_runtime_put_sync(dev);
> +		return ret;
> +	}
> +
> +	ret = stm32_pcie_enable_resources(stm32_pcie);
> +	if (ret) {
> +		dev_err(dev, "failed to enable resources: %d\n", ret);
> +		dw_pcie_ep_deinit(ep);
> +		pm_runtime_put_sync(dev);
> +		return ret;
> +	}
> +
> +	ret = dw_pcie_ep_init_registers(ep);
> +	if (ret) {
> +		dev_err(dev, "Failed to initialize DWC endpoint registers\n");
> +		stm32_pcie_disable_resources(stm32_pcie);
> +		dw_pcie_ep_deinit(ep);
> +		pm_runtime_put_sync(dev);
> +		return ret;
> +	}

Consider gotos for the error cases with a cleanup block at the end.
There's a fair bit of repetition there as more things get initialized,
and it's error-prone to extend this in the future.

Same applies in stm32_pcie_perst_deassert().

> +	pci_epc_init_notify(ep->epc);
> +
> +	return 0;
> +}
> +
> +static int stm32_pcie_probe(struct platform_device *pdev)
> +{
> +	struct stm32_pcie *stm32_pcie;
> +	struct dw_pcie *dw;
> +	struct device *dev = &pdev->dev;
> +	int ret;
> +
> +	stm32_pcie = devm_kzalloc(dev, sizeof(*stm32_pcie), GFP_KERNEL);
> +	if (!stm32_pcie)
> +		return -ENOMEM;
> +
> +	dw = devm_kzalloc(dev, sizeof(*dw), GFP_KERNEL);
> +	if (!dw)
> +		return -ENOMEM;

Add blank line here.

> +	stm32_pcie->pci = dw;

> +static struct platform_driver stm32_pcie_ep_driver = {
> +	.probe = stm32_pcie_probe,
> +	.remove_new = stm32_pcie_remove,

.remove().

> +	.driver = {
> +		.name = "stm32-ep-pcie",
> +		.of_match_table = stm32_pcie_ep_of_match,
> +	},
> +};