lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250813230733.GA299608@bhelgaas>
Date: Wed, 13 Aug 2025 18:07:33 -0500
From: Bjorn Helgaas <helgaas@...nel.org>
To: Shradha Todi <shradha.t@...sung.com>
Cc: linux-pci@...r.kernel.org, devicetree@...r.kernel.org,
	linux-arm-kernel@...ts.infradead.org,
	linux-samsung-soc@...r.kernel.org, linux-kernel@...r.kernel.org,
	linux-phy@...ts.infradead.org, mani@...nel.org,
	lpieralisi@...nel.org, kwilczynski@...nel.org, robh@...nel.org,
	bhelgaas@...gle.com, jingoohan1@...il.com, krzk+dt@...nel.org,
	conor+dt@...nel.org, alim.akhtar@...sung.com, vkoul@...nel.org,
	kishon@...nel.org, arnd@...db.de, m.szyprowski@...sung.com,
	jh80.chung@...sung.com, pankaj.dubey@...sung.com
Subject: Re: [PATCH v3 05/12] PCI: exynos: Add resource ops, soc variant and
 device mode

On Mon, Aug 11, 2025 at 09:16:31PM +0530, Shradha Todi wrote:
> Some resources might differ based on platforms and we need platform
> specific functions to initialize or alter them. For better code
> re-usability, making a separate res_ops which will hold all such
> function pointers or other resource specific data. Include ops like
>  - init_regulator (initialize the regulator data)
>  - pcie_irq_handler (interrupt handler for PCIe)
>  - set_device_mode (set device mode to EP or RC)
> 
> Some operations maybe specific to certain SoCs and not applicable
> to others. For such use cases, adding an SoC variant data field
> which can be used to distinguish between the variants.
> 
> Some SoCs may have dual-role PCIe controller which can work as
> RC or EP. Add device_mode to store the role and take decisions
> accordingly.
> 
> Make enable/disable of regulator and initialization of IRQ as
> common functions to be used by all Samsung SoCs.

As hinted above, this patch ends up being a mixture of several things
that makes this kind of hard to review.  Separating these into their
own patches would make it easier.

> Suggested-by: Pankaj Dubey <pankaj.dubey@...sung.com>
> Signed-off-by: Shradha Todi <shradha.t@...sung.com>
> ---
>  drivers/pci/controller/dwc/pci-exynos.c | 143 +++++++++++++++++++-----
>  1 file changed, 116 insertions(+), 27 deletions(-)
> 
> diff --git a/drivers/pci/controller/dwc/pci-exynos.c b/drivers/pci/controller/dwc/pci-exynos.c
> index c830b20d54f0..ef1f42236575 100644
> --- a/drivers/pci/controller/dwc/pci-exynos.c
> +++ b/drivers/pci/controller/dwc/pci-exynos.c
> @@ -49,10 +49,18 @@
>  #define EXYNOS_PCIE_ELBI_SLV_ARMISC		0x120
>  #define EXYNOS_PCIE_ELBI_SLV_DBI_ENABLE		BIT(21)
>  
> +/* to store different SoC variants of Samsung */
> +enum samsung_pcie_variants {
> +	EXYNOS_5433,
> +};
> +
>  struct samsung_pcie_pdata {
>  	struct pci_ops				*pci_ops;
>  	const struct dw_pcie_ops		*dwc_ops;
>  	const struct dw_pcie_host_ops		*host_ops;
> +	const struct samsung_res_ops		*res_ops;
> +	unsigned int				soc_variant;
> +	enum dw_pcie_device_mode		device_mode;
>  };
>  
>  struct exynos_pcie {
> @@ -61,7 +69,14 @@ struct exynos_pcie {
>  	const struct samsung_pcie_pdata	*pdata;
>  	struct clk_bulk_data		*clks;
>  	struct phy			*phy;
> -	struct regulator_bulk_data	supplies[2];
> +	struct regulator_bulk_data	*supplies;
> +	int				supplies_cnt;
> +};
> +
> +struct samsung_res_ops {
> +	int (*init_regulator)(struct exynos_pcie *ep);
> +	irqreturn_t (*pcie_irq_handler)(int irq, void *arg);
> +	void (*set_device_mode)(struct exynos_pcie *ep);
>  };
>  
>  static void exynos_pcie_writel(void __iomem *base, u32 val, u32 reg)
> @@ -74,6 +89,31 @@ static u32 exynos_pcie_readl(void __iomem *base, u32 reg)
>  	return readl(base + reg);
>  }
>  
> +static int samsung_regulator_enable(struct exynos_pcie *ep)
> +{
> +	int ret;
> +
> +	if (ep->supplies_cnt == 0)
> +		return 0;
> +
> +	ret = regulator_bulk_enable(ep->supplies_cnt, ep->supplies);
> +
> +	return ret;

Unless you want a warning here, there's no need for "ret":

  return regulator_bulk_enable(ep->supplies_cnt, ep->supplies);

> +}
> +
> +static void samsung_regulator_disable(struct exynos_pcie *ep)
> +{
> +	struct device *dev = ep->pci.dev;
> +	int ret;
> +
> +	if (ep->supplies_cnt == 0)
> +		return;
> +
> +	ret = regulator_bulk_disable(ep->supplies_cnt, ep->supplies);
> +	if (ret)
> +		dev_warn(dev, "failed to disable regulators: %d\n", ret);
> +}
> +
>  static void exynos_pcie_sideband_dbi_w_mode(struct exynos_pcie *ep, bool on)
>  {
>  	u32 val;
> @@ -244,7 +284,26 @@ static const struct dw_pcie_host_ops exynos_pcie_host_ops = {
>  	.init = exynos_pcie_host_init,
>  };
>  
> -static int exynos_add_pcie_port(struct exynos_pcie *ep,
> +static int exynos_init_regulator(struct exynos_pcie *ep)
> +{
> +	struct device *dev = ep->pci.dev;
> +	int ret = 0;
> +
> +	ep->supplies_cnt = 2;
> +
> +	ep->supplies = devm_kcalloc(dev, ep->supplies_cnt, sizeof(*ep->supplies), GFP_KERNEL);
> +	if (!ep->supplies)
> +		return -ENOMEM;
> +
> +	ep->supplies[0].supply = "vdd18";
> +	ep->supplies[1].supply = "vdd10";
> +
> +	ret = devm_regulator_bulk_get(dev, ep->supplies_cnt, ep->supplies);
> +
> +	return ret;

No need for ret.

> +}
> +
> +static int samsung_irq_init(struct exynos_pcie *ep,
>  				       struct platform_device *pdev)
>  {
>  	struct dw_pcie *pci = &ep->pci;
> @@ -256,22 +315,15 @@ static int exynos_add_pcie_port(struct exynos_pcie *ep,
>  	if (pp->irq < 0)
>  		return pp->irq;
>  
> -	ret = devm_request_irq(dev, pp->irq, exynos_pcie_irq_handler,
> +	ret = devm_request_irq(dev, pp->irq, ep->pdata->res_ops->pcie_irq_handler,
>  			       IRQF_SHARED, "exynos-pcie", ep);
>  	if (ret) {
>  		dev_err(dev, "failed to request irq\n");
>  		return ret;
>  	}
>  
> -	pp->ops = &exynos_pcie_host_ops;
>  	pp->msi_irq[0] = -ENODEV;
>  
> -	ret = dw_pcie_host_init(pp);
> -	if (ret) {
> -		dev_err(dev, "failed to initialize host\n");
> -		return ret;
> -	}
> -
>  	return 0;
>  }
>  
> @@ -282,6 +334,11 @@ static const struct dw_pcie_ops exynos_dw_pcie_ops = {
>  	.start_link = exynos_pcie_start_link,
>  };
>  
> +static const struct samsung_res_ops exynos_res_ops_data = {
> +	.init_regulator		= exynos_init_regulator,
> +	.pcie_irq_handler	= exynos_pcie_irq_handler,
> +};
> +
>  static int exynos_pcie_probe(struct platform_device *pdev)
>  {
>  	struct device *dev = &pdev->dev;
> @@ -313,28 +370,46 @@ static int exynos_pcie_probe(struct platform_device *pdev)
>  	if (ret < 0)
>  		return ret;
>  
> -	ep->supplies[0].supply = "vdd18";
> -	ep->supplies[1].supply = "vdd10";
> -	ret = devm_regulator_bulk_get(dev, ARRAY_SIZE(ep->supplies),
> -				      ep->supplies);
> -	if (ret)
> -		return ret;
> +	if (pdata->res_ops && pdata->res_ops->init_regulator) {
> +		ret = ep->pdata->res_ops->init_regulator(ep);
> +		if (ret)
> +			return ret;
> +	}
>  
> -	ret = regulator_bulk_enable(ARRAY_SIZE(ep->supplies), ep->supplies);
> +	ret = samsung_regulator_enable(ep);
>  	if (ret)
>  		return ret;
>  
>  	platform_set_drvdata(pdev, ep);
>  
> -	ret = exynos_add_pcie_port(ep, pdev);
> -	if (ret < 0)
> -		goto fail_probe;
> +	if (pdata->res_ops && pdata->res_ops->set_device_mode)
> +		pdata->res_ops->set_device_mode(ep);
> +
> +	switch (ep->pdata->device_mode) {
> +	case DW_PCIE_RC_TYPE:
> +		ret = samsung_irq_init(ep, pdev);
> +		if (ret)
> +			goto fail_regulator;
> +
> +		ep->pci.pp.ops = pdata->host_ops;
> +
> +		ret = dw_pcie_host_init(&ep->pci.pp);
> +		if (ret < 0)
> +			goto fail_phy_init;
> +
> +		break;
> +	default:
> +		dev_err(dev, "invalid device type\n");
> +		ret = -EINVAL;
> +		goto fail_regulator;
> +	}
>  
>  	return 0;
>  
> -fail_probe:
> +fail_phy_init:
>  	phy_exit(ep->phy);
> -	regulator_bulk_disable(ARRAY_SIZE(ep->supplies), ep->supplies);
> +fail_regulator:
> +	samsung_regulator_disable(ep);
>  
>  	return ret;
>  }
> @@ -343,21 +418,29 @@ static void exynos_pcie_remove(struct platform_device *pdev)
>  {
>  	struct exynos_pcie *ep = platform_get_drvdata(pdev);
>  
> +	if (ep->pdata->device_mode == DW_PCIE_EP_TYPE)
> +		return;
>  	dw_pcie_host_deinit(&ep->pci.pp);
> -	exynos_pcie_assert_core_reset(ep);
> +	if (ep->pdata->soc_variant == EXYNOS_5433)
> +		exynos_pcie_assert_core_reset(ep);

I think it's a mistake to test for specific soc_variants.  Better to
test a *property* (or use a function pointer than can be SoC-specific)
because it's likely that several SoCs will share the same property.

>  	phy_power_off(ep->phy);
>  	phy_exit(ep->phy);
> -	regulator_bulk_disable(ARRAY_SIZE(ep->supplies), ep->supplies);
> +	samsung_regulator_disable(ep);
>  }
>  
>  static int exynos_pcie_suspend_noirq(struct device *dev)
>  {
>  	struct exynos_pcie *ep = dev_get_drvdata(dev);
> +	struct dw_pcie *pci = &ep->pci;
>  
> -	exynos_pcie_assert_core_reset(ep);
> +	if (ep->pdata->device_mode == DW_PCIE_EP_TYPE)
> +		return 0;
> +
> +	if (ep->pdata->soc_variant == EXYNOS_5433)
> +		exynos_pcie_assert_core_reset(ep);
>  	phy_power_off(ep->phy);
>  	phy_exit(ep->phy);
> -	regulator_bulk_disable(ARRAY_SIZE(ep->supplies), ep->supplies);
> +	samsung_regulator_disable(ep);
>  
>  	return 0;
>  }
> @@ -369,7 +452,10 @@ static int exynos_pcie_resume_noirq(struct device *dev)
>  	struct dw_pcie_rp *pp = &pci->pp;
>  	int ret;
>  
> -	ret = regulator_bulk_enable(ARRAY_SIZE(ep->supplies), ep->supplies);
> +	if (ep->pdata->device_mode == DW_PCIE_EP_TYPE)
> +		return 0;
> +
> +	ret = samsung_regulator_enable(ep);
>  	if (ret)
>  		return ret;
>  
> @@ -389,6 +475,9 @@ static const struct samsung_pcie_pdata exynos_5433_pcie_rc_pdata = {
>  	.dwc_ops		= &exynos_dw_pcie_ops,
>  	.pci_ops		= &exynos_pci_ops,
>  	.host_ops		= &exynos_pcie_host_ops,
> +	.res_ops		= &exynos_res_ops_data,
> +	.soc_variant		= EXYNOS_5433,
> +	.device_mode		= DW_PCIE_RC_TYPE,
>  };
>  
>  static const struct of_device_id exynos_pcie_of_match[] = {
> -- 
> 2.49.0
> 

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ