On Mon, 21 Jun 2021 at 22:20, Lorenzo Pieralisi
<lorenzo.pieralisi@xxxxxxx> wrote:
>
> On Fri, Jan 15, 2021 at 02:04:40PM +0800, Hongtao Wu wrote:
> > From: Hongtao Wu <billows.wu@xxxxxxxxxx>
> >
> > This series adds PCIe controller driver for Unisoc SoCs.
> > This controller is based on DesignWare PCIe IP.
> >
> > Signed-off-by: Hongtao Wu <billows.wu@xxxxxxxxxx>
> > ---
> > drivers/pci/controller/dwc/Kconfig | 12 ++
> > drivers/pci/controller/dwc/Makefile | 1 +
> > drivers/pci/controller/dwc/pcie-sprd.c | 293 +++++++++++++++++++++++++++++++++
> > 3 files changed, 306 insertions(+)
> > create mode 100644 drivers/pci/controller/dwc/pcie-sprd.c
> >
> > diff --git a/drivers/pci/controller/dwc/Kconfig b/drivers/pci/controller/dwc/Kconfig
> > index 22c5529..61f0b79 100644
> > --- a/drivers/pci/controller/dwc/Kconfig
> > +++ b/drivers/pci/controller/dwc/Kconfig
> > @@ -318,4 +318,16 @@ config PCIE_AL
> > required only for DT-based platforms. ACPI platforms with the
> > Annapurna Labs PCIe controller don't need to enable this.
> >
> > +config PCIE_SPRD
> > + tristate "Unisoc PCIe controller - Host Mode"
> > + depends on ARCH_SPRD || COMPILE_TEST
> > + depends on PCI_MSI_IRQ_DOMAIN
> > + select PCIE_DW_HOST
> > + help
> > + Unisoc PCIe controller uses the DesignWare core. It can be configured
> > + as an Endpoint (EP) or a Root complex (RC). In order to enable host
> > + mode (the controller works as RC), PCIE_SPRD must be selected.
> > + Say Y or M here if you want to PCIe RC controller support on Unisoc
> > + SoCs.
> > +
> > endmenu
> > diff --git a/drivers/pci/controller/dwc/Makefile b/drivers/pci/controller/dwc/Makefile
> > index a751553..eb546e9 100644
> > --- a/drivers/pci/controller/dwc/Makefile
> > +++ b/drivers/pci/controller/dwc/Makefile
> > @@ -20,6 +20,7 @@ obj-$(CONFIG_PCI_MESON) += pci-meson.o
> > obj-$(CONFIG_PCIE_TEGRA194) += pcie-tegra194.o
> > obj-$(CONFIG_PCIE_UNIPHIER) += pcie-uniphier.o
> > obj-$(CONFIG_PCIE_UNIPHIER_EP) += pcie-uniphier-ep.o
> > +obj-$(CONFIG_PCIE_SPRD) += pcie-sprd.o
> >
> > # The following drivers are for devices that use the generic ACPI
> > # pci_root.c driver but don't support standard ECAM config access.
> > diff --git a/drivers/pci/controller/dwc/pcie-sprd.c b/drivers/pci/controller/dwc/pcie-sprd.c
> > new file mode 100644
> > index 0000000..03e2064
> > --- /dev/null
> > +++ b/drivers/pci/controller/dwc/pcie-sprd.c
> > @@ -0,0 +1,293 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +/*
> > + * PCIe host controller driver for Unisoc SoCs
> > + *
> > + * Copyright (C) 2020-2021 Unisoc, Inc.
> > + *
> > + * Author: Hongtao Wu <Billows.Wu@xxxxxxxxxx>
> > + */
> > +
> > +#include <linux/delay.h>
> > +#include <linux/gpio/consumer.h>
> > +#include <linux/interrupt.h>
> > +#include <linux/mfd/syscon.h>
> > +#include <linux/module.h>
> > +#include <linux/of_device.h>
> > +#include <linux/of_irq.h>
> > +#include <linux/platform_device.h>
> > +#include <linux/property.h>
> > +#include <linux/regmap.h>
> > +
> > +#include "pcie-designware.h"
> > +
> > +/* aon apb syscon */
> > +#define IPA_ACCESS_CFG 0xcd8
> > +#define AON_ACCESS_PCIE_EN BIT(1)
> > +
> > +/* pmu apb syscon */
> > +#define SNPS_PCIE3_SLP_CTRL 0xac
> > +#define PERST_N_ASSERT BIT(1)
> > +#define PERST_N_AUTO_EN BIT(0)
> > +#define PD_PCIE_CFG_0 0x3e8
> > +#define PCIE_FORCE_SHUTDOWN BIT(25)
> > +
> > +#define PCIE_SS_REG_BASE 0xE00
> > +#define APB_CLKFREQ_TIMEOUT 0x4
> > +#define BUSERR_EN BIT(12)
> > +#define APB_TIMER_DIS BIT(10)
> > +#define APB_TIMER_LIMIT GENMASK(31, 16)
> > +
> > +#define PE0_GEN_CTRL_3 0x58
> > +#define LTSSM_EN BIT(0)
> > +
> > +struct sprd_pcie_soc_data {
> > + u32 syscon_offset;
> > +};
> > +
> > +static const struct sprd_pcie_soc_data ums9520_syscon_data = {
> > + .syscon_offset = 0x1000, /* The offset of set/clear register */
> > +};
> > +
> > +struct sprd_pcie {
> > + u32 syscon_offset;
> > + struct device *dev;
> > + struct dw_pcie *pci;
> > + struct regmap *aon_map;
> > + struct regmap *pmu_map;
> > + const struct sprd_pcie_soc_data *socdata;
> > +};
> > +
> > +enum sprd_pcie_syscon_type {
> > + normal_syscon, /* it's not a set/clear register */
> > + set_syscon, /* set a set/clear register */
> > + clr_syscon, /* clear a set/clear register */
> > +};
> > +
> > +static void sprd_pcie_buserr_enable(struct dw_pcie *pci)
>
> I think you should add a comment to explain what this function
> enables - it is unclear.
Yes, I agree, maybe also need to change to a more reasonable name :)
I will think about it.
>
> > +{
> > + u32 val;
> > +
> > + val = dw_pcie_readl_dbi(pci, PCIE_SS_REG_BASE + APB_CLKFREQ_TIMEOUT);
> > + val &= ~APB_TIMER_DIS;
> > + val |= BUSERR_EN;
> > + val |= APB_TIMER_LIMIT & (0x1f4 << 16);
> > + dw_pcie_writel_dbi(pci, PCIE_SS_REG_BASE + APB_CLKFREQ_TIMEOUT, val);
> > +}
> > +
> > +static void sprd_pcie_ltssm_enable(struct dw_pcie *pci, bool enable)
> > +{
> > + u32 val;
> > +
> > + val = dw_pcie_readl_dbi(pci, PCIE_SS_REG_BASE + PE0_GEN_CTRL_3);
> > + if (enable)
> > + dw_pcie_writel_dbi(pci, PCIE_SS_REG_BASE + PE0_GEN_CTRL_3,
> > + val | LTSSM_EN);
> > + else
> > + dw_pcie_writel_dbi(pci, PCIE_SS_REG_BASE + PE0_GEN_CTRL_3,
> > + val & ~LTSSM_EN);
> > +}
> > +
> > +static int sprd_pcie_syscon_set(struct sprd_pcie *ctrl, struct regmap *map,
> > + u32 reg, u32 mask, u32 val,
> > + enum sprd_pcie_syscon_type type)
> > +{
> > + int ret = 0;
> > + u32 read_val;
> > + u32 offset = ctrl->syscon_offset;
> > + struct device *dev = ctrl->pci->dev;
> > +
> > + /*
> > + * Each set/clear register has three registers:
> > + * reg: base register
> > + * reg + offset: set register
> > + * reg + offset * 2: clear register
> > + */
> > + switch (type) {
> > + case normal_syscon:
> > + ret = regmap_read(map, reg, &read_val);
> > + if (ret) {
> > + dev_err(dev, "failed to read register 0x%x\n", reg);
> > + return ret;
> > + }
> > + read_val &= ~mask;
> > + read_val |= (val & mask);
> > + ret = regmap_write(map, reg, read_val);
> > + break;
> > + case set_syscon:
> > + reg = reg + offset;
> > + ret = regmap_write(map, reg, val);
> > + break;
> > + case clr_syscon:
> > + reg = reg + offset * 2;
> > + ret = regmap_write(map, reg, val);
> > + break;
> > + default:
> > + break;
> > + }
> > +
> > + if (ret)
> > + dev_err(dev, "failed to write register 0x%x\n", reg);
> > +
> > + return ret;
> > +}
> > +
> > +static int sprd_pcie_perst_assert(struct sprd_pcie *ctrl)
> > +{
> > + return sprd_pcie_syscon_set(ctrl, ctrl->pmu_map, SNPS_PCIE3_SLP_CTRL,
> > + PERST_N_ASSERT, PERST_N_ASSERT, set_syscon);
> > +}
> > +
> > +static int sprd_pcie_perst_deassert(struct sprd_pcie *ctrl)
> > +{
> > + int ret;
> > +
> > + ret = sprd_pcie_syscon_set(ctrl, ctrl->pmu_map, SNPS_PCIE3_SLP_CTRL,
> > + PERST_N_ASSERT, 0, clr_syscon);
> > + usleep_range(2000, 3000);
>
> This delay seems arbitrary.
Yes, I guess it came out from some tests result, I need to double
check with my colleagues.
> The PCI specifications require a 100ms
> after issuing PERST before link training is enabled.
>
> Have a look at e.g. drivers/pci/controller/pci-aardvark.c
>
> > +
> > + return ret;
> > +}
> > +
> > +static int sprd_pcie_power_on(struct platform_device *pdev)
> > +{
> > + int ret;
> > + struct sprd_pcie *ctrl = platform_get_drvdata(pdev);
> > + struct dw_pcie *pci = ctrl->pci;
> > +
> > + ret = sprd_pcie_syscon_set(ctrl, ctrl->aon_map, PD_PCIE_CFG_0,
> > + PCIE_FORCE_SHUTDOWN, 0, clr_syscon);
> > + if (ret)
> > + return ret;
> > +
> > + ret = sprd_pcie_syscon_set(ctrl, ctrl->aon_map, IPA_ACCESS_CFG,
> > + AON_ACCESS_PCIE_EN, AON_ACCESS_PCIE_EN,
> > + set_syscon);
> > + if (ret)
> > + return ret;
> > +
> > + ret = sprd_pcie_perst_deassert(ctrl);
> > + if (ret)
> > + return ret;
> > +
> > + sprd_pcie_buserr_enable(pci);
> > + sprd_pcie_ltssm_enable(pci, true);
> > +
> > + return ret;
> > +}
> > +
> > +static int sprd_pcie_power_off(struct platform_device *pdev)
> > +{
> > + struct sprd_pcie *ctrl = platform_get_drvdata(pdev);
> > + struct dw_pcie *pci = ctrl->pci;
> > +
> > + sprd_pcie_ltssm_enable(pci, false);
> > +
> > + sprd_pcie_perst_assert(ctrl);
> > + sprd_pcie_syscon_set(ctrl, ctrl->aon_map, PD_PCIE_CFG_0,
> > + PCIE_FORCE_SHUTDOWN, PCIE_FORCE_SHUTDOWN,
> > + set_syscon);
> > + sprd_pcie_syscon_set(ctrl, ctrl->aon_map, IPA_ACCESS_CFG,
> > + AON_ACCESS_PCIE_EN, 0, clr_syscon);
> > +
> > + return 0;
> > +}
> > +
> > +static int sprd_add_pcie_port(struct platform_device *pdev)
> > +{
> > + struct sprd_pcie *ctrl = platform_get_drvdata(pdev);
> > + struct dw_pcie *pci = ctrl->pci;
> > + struct pcie_port *pp = &pci->pp;
> > +
> > + return dw_pcie_host_init(pp);
> > +}
> > +
> > +static int sprd_pcie_probe(struct platform_device *pdev)
> > +{
> > + struct device *dev = &pdev->dev;
> > + struct sprd_pcie *ctrl;
> > + struct dw_pcie *pci;
> > + int ret;
> > +
> > + ctrl = devm_kzalloc(dev, sizeof(*ctrl), GFP_KERNEL);
> > + if (!ctrl)
> > + return -ENOMEM;
> > +
> > + ctrl->socdata =
> > + (struct sprd_pcie_soc_data *)of_device_get_match_data(dev);
> > + if (!ctrl->socdata) {
> > + dev_warn(dev,
> > + "using the default set/clear register offset address");
> > + ctrl->syscon_offset = 0x1000;
> > + }
> > + ctrl->syscon_offset = ctrl->socdata->syscon_offset;
> > +
> > + ctrl->aon_map = syscon_regmap_lookup_by_phandle(dev->of_node,
> > + "sprd, regmap-aon");
> > + if (IS_ERR(ctrl->aon_map)) {
> > + dev_err(dev, "failed to get syscon regmap aon\n");
> > + ret = PTR_ERR(ctrl->aon_map);
> > + goto err;
> > + }
> > +
> > + ctrl->pmu_map = syscon_regmap_lookup_by_phandle(dev->of_node,
> > + "sprd, regmap-pmu");
> > + if (IS_ERR(ctrl->pmu_map)) {
> > + dev_err(dev, "failed to get syscon regmap pmu\n");
> > + ret = PTR_ERR(ctrl->pmu_map);
> > + goto err;
> > + }
> > +
> > + pci = ctrl->pci;
> > + pci->dev = dev;
> > +
> > + platform_set_drvdata(pdev, ctrl);
> > +
> > + ret = sprd_pcie_power_on(pdev);
> > + if (ret < 0) {
> > + dev_err(dev, "failed to power on, return %d\n",
> > + ret);
> > + goto err_power_off;
>
> This looks a bit weird but I believe that's just a naming convention.
>
> It looks like if we *fail* to power the controller on we power it
> off, which does not make much sense.
>
> I guess the goal in err_power_off is to clean-up what was done in
> sprd_pcie_power_on() - more below.
Yes, maybe we need to give it a more suitable name as well.
>
> > + }
> > +
> > + ret = sprd_add_pcie_port(pdev);
> > + if (ret) {
> > + dev_warn(dev, "failed to initialize RC controller\n");
> > + return ret;
>
> If we go by the same logic, we shouldn't return but rather
>
> goto err_power_off;
Right, I will address this.
Many thanks for your review,
Chunyan
>
> on error here. Am I missing something ?
>
> Lorenzo
>
> > + }
> > +
> > + return 0;
> > +
> > +err_power_off:
> > + sprd_pcie_power_off(pdev);
> > +err:
> > + return ret;
> > +}
> > +
> > +static int sprd_pcie_remove(struct platform_device *pdev)
> > +{
> > + sprd_pcie_power_off(pdev);
> > +
> > + return 0;
> > +}
> > +
> > +static const struct of_device_id sprd_pcie_of_match[] = {
> > + {
> > + .compatible = "sprd,ums9520-pcie",
> > + .data = &ums9520_syscon_data,
> > + },
> > + {},
> > +};
> > +
> > +static struct platform_driver sprd_pcie_driver = {
> > + .probe = sprd_pcie_probe,
> > + .remove = sprd_pcie_remove,
> > + .driver = {
> > + .name = "sprd-pcie",
> > + .of_match_table = sprd_pcie_of_match,
> > + },
> > +};
> > +
> > +module_platform_driver(sprd_pcie_driver);
> > +
> > +MODULE_DESCRIPTION("Unisoc PCIe host controller driver");
> > +MODULE_LICENSE("GPL");
> > --
> > 2.7.4
> >
