Hi Rob,
Thanks for your review, some replies inline. All you pointed out will be
fixed in v3.
在 2021/1/20 22:55, Rob Herring 写道:
On Wed, Jan 20, 2021 at 06:16:58PM +0800, Simon Xue wrote:
pcie-dw-rockchip is based on DWC IP. But pcie-rockchip-host
is another IP which is only used for RK3399. So all the following
s/another/Cadence/
There's been a lot of reworking of the DWC drivers. You need to update
this based on that. Details inline.
DWC core really cover a lot, I think I need to refactor some codes.
It's more
reasonable to start link after all relative registers have set up. So,
dw_pcie_host_ops->host_init just do some set up works, link start should
postpone to dw_pcie_ops->start_link, and place rockchip_pcie_start_link in
the dw_pcie_ops->start_link call back.
non-RK3399 SoCs should use this driver.
Signed-off-by: Simon Xue <xxm@xxxxxxxxxxxxxx>
Signed-off-by: Shawn Lin <shawn.lin@xxxxxxxxxxxxxx>
---
drivers/pci/controller/Kconfig | 4 +-
drivers/pci/controller/dwc/Kconfig | 9 +
drivers/pci/controller/dwc/Makefile | 1 +
drivers/pci/controller/dwc/pcie-dw-rockchip.c | 370 ++++++++++++++++++
4 files changed, 382 insertions(+), 2 deletions(-)
create mode 100644 drivers/pci/controller/dwc/pcie-dw-rockchip.c
diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
index 64e2f5e379aa..48a7d62f97f3 100644
--- a/drivers/pci/controller/Kconfig
+++ b/drivers/pci/controller/Kconfig
@@ -219,7 +219,7 @@ config PCIE_ROCKCHIP_HOST
help
Say Y here if you want internal PCI support on Rockchip SoC.
There is 1 internal PCIe port available to support GEN2 with
- 4 slots.
+ 4 slots. Only for RK3399.
config PCIE_ROCKCHIP_EP
bool "Rockchip PCIe endpoint controller"
@@ -231,7 +231,7 @@ config PCIE_ROCKCHIP_EP
help
Say Y here if you want to support Rockchip PCIe controller in
endpoint mode on Rockchip SoC. There is 1 internal PCIe port
- available to support GEN2 with 4 slots.
+ available to support GEN2 with 4 slots. Only for RK3399.
I would make these changes a separate patch. I have a pending patch to
move this driver to the Cadence directory.
Ok, will remove this.
config PCIE_MEDIATEK
tristate "MediaTek PCIe controller"
diff --git a/drivers/pci/controller/dwc/Kconfig b/drivers/pci/controller/dwc/Kconfig
index 22c5529e9a65..110733d0c241 100644
--- a/drivers/pci/controller/dwc/Kconfig
+++ b/drivers/pci/controller/dwc/Kconfig
@@ -214,6 +214,15 @@ config PCIE_ARTPEC6_EP
Enables support for the PCIe controller in the ARTPEC-6 SoC to work in
endpoint mode. This uses the DesignWare core.
+config PCIE_DW_ROCKCHIP_HOST
PCIE_ROCKCHIP_DW_HOST to be more inline with other DWC config options.
+ bool "Rockchip DesignWare PCIe controller"
+ select PCIE_DW
+ select PCIE_DW_HOST
+ depends on ARCH_ROCKCHIP
|| COMPILE_TEST
+ depends on OF
+ help
+ Enables support for the DW PCIe controller in the Rockchip SoC.
+
config PCIE_INTEL_GW
bool "Intel Gateway PCIe host controller support"
depends on OF && (X86 || COMPILE_TEST)
diff --git a/drivers/pci/controller/dwc/Makefile b/drivers/pci/controller/dwc/Makefile
index a751553fa0db..9c7048d2be17 100644
--- a/drivers/pci/controller/dwc/Makefile
+++ b/drivers/pci/controller/dwc/Makefile
@@ -13,6 +13,7 @@ obj-$(CONFIG_PCI_LAYERSCAPE_EP) += pci-layerscape-ep.o
obj-$(CONFIG_PCIE_QCOM) += pcie-qcom.o
obj-$(CONFIG_PCIE_ARMADA_8K) += pcie-armada8k.o
obj-$(CONFIG_PCIE_ARTPEC6) += pcie-artpec6.o
+obj-$(CONFIG_PCIE_DW_ROCKCHIP_HOST) += pcie-dw-rockchip.o
obj-$(CONFIG_PCIE_INTEL_GW) += pcie-intel-gw.o
obj-$(CONFIG_PCIE_KIRIN) += pcie-kirin.o
obj-$(CONFIG_PCIE_HISI_STB) += pcie-histb.o
diff --git a/drivers/pci/controller/dwc/pcie-dw-rockchip.c b/drivers/pci/controller/dwc/pcie-dw-rockchip.c
new file mode 100644
index 000000000000..e41cd6041c3f
--- /dev/null
+++ b/drivers/pci/controller/dwc/pcie-dw-rockchip.c
@@ -0,0 +1,370 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * PCIe host controller driver for Rockchip SoCs
+ *
+ * Copyright (C) 2021 Rockchip Electronics Co., Ltd.
+ * http://www.rock-chips.com
+ *
+ * Author: Simon Xue <xxm@xxxxxxxxxxxxxx>
+ */
+
+#include <linux/clk.h>
+#include <linux/gpio/consumer.h>
+#include <linux/mfd/syscon.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/phy/phy.h>
+#include <linux/platform_device.h>
+#include <linux/regmap.h>
+#include <linux/reset.h>
+
+#include "pcie-designware.h"
+
+/*
+ * The upper 16 bits of PCIE_CLIENT_CONFIG are a write
+ * mask for the lower 16 bits. This allows atomic updates
+ * of the register without locking.
+ */
+#define HIWORD_UPDATE(mask, val) (((mask) << 16) | (val))
+#define HIWORD_UPDATE_BIT(val) HIWORD_UPDATE(val, val)
+
+#define to_rockchip_pcie(x) dev_get_drvdata((x)->dev)
+
+#define PCIE_CLIENT_RC_MODE HIWORD_UPDATE_BIT(0x40)
+#define PCIE_CLIENT_ENABLE_LTSSM HIWORD_UPDATE_BIT(0xc)
+#define PCIE_SMLH_LINKUP BIT(16)
+#define PCIE_RDLH_LINKUP BIT(17)
+#define PCIE_L0S_ENTRY 0x11
+#define PCIE_CLIENT_GENERAL_CONTROL 0x0
+#define PCIE_CLIENT_GENERAL_DEBUG 0x104
+#define PCIE_CLIENT_HOT_RESET_CTRL 0x180
+#define PCIE_CLIENT_LTSSM_STATUS 0x300
+#define PCIE_CLIENT_DBG_FIFO_MODE_CON 0x310
+#define PCIE_CLIENT_DBG_FIFO_PTN_HIT_D0 0x320
+#define PCIE_CLIENT_DBG_FIFO_PTN_HIT_D1 0x324
+#define PCIE_CLIENT_DBG_FIFO_TRN_HIT_D0 0x328
+#define PCIE_CLIENT_DBG_FIFO_TRN_HIT_D1 0x32c
+#define PCIE_CLIENT_DBG_FIFO_STATUS 0x350
+#define PCIE_CLIENT_DBG_TRANSITION_DATA 0xffff0000
+#define PCIE_CLIENT_DBF_EN 0xffff0003
+#define PCIE_LTSSM_ENABLE_ENHANCE BIT(4)
+
+struct rockchip_pcie {
+ struct dw_pcie *pci;
Make this a struct, not a pointer. Then 1 alloc instead of 2.
+ void __iomem *apb_base;
+ struct phy *phy;
+ struct clk_bulk_data *clks;
+ unsigned int clk_cnt;
+ struct reset_control *rst;
+ struct gpio_desc *rst_gpio;
+ struct pcie_port pp;
Move this up to 2nd element.
No need to define pp, struct dw_pcie already has one, will remove this.
+ struct regulator *vpcie3v3;
+};
+
+static inline int rockchip_pcie_readl_apb(struct rockchip_pcie *rockchip,
+ u32 reg)
You can drop inline on all of these. The compiler does that anyways with
static functions.
+{
+ return readl(rockchip->apb_base + reg);
+}
+
+static inline void rockchip_pcie_writel_apb(struct rockchip_pcie *rockchip,
+ u32 reg, u32 val)
Use the same order as writel. val then reg.
+{
+ writel(val, rockchip->apb_base + reg);
+}
+
+static inline void rockchip_pcie_set_mode(struct rockchip_pcie *rockchip)
Set what mode? Just drop this wrapper. The register access tells me more
than the function name does.
+{
+ rockchip_pcie_writel_apb(rockchip, PCIE_CLIENT_GENERAL_CONTROL,
+ PCIE_CLIENT_RC_MODE);
+}
+
+static inline void rockchip_pcie_enable_ltssm(struct rockchip_pcie *rockchip)
+{
+ rockchip_pcie_writel_apb(rockchip, PCIE_CLIENT_GENERAL_CONTROL,
+ PCIE_CLIENT_ENABLE_LTSSM);
+}
+
+static int rockchip_pcie_link_up(struct dw_pcie *pci)
+{
+ struct rockchip_pcie *rockchip = to_rockchip_pcie(pci);
+ u32 val = rockchip_pcie_readl_apb(rockchip, PCIE_CLIENT_LTSSM_STATUS);
+
+ if ((val & (PCIE_RDLH_LINKUP | PCIE_SMLH_LINKUP)) == 0x30000 &&
+ (val & GENMASK(5, 0)) == PCIE_L0S_ENTRY)
+ return 1;
Just to confirm, the PORT_LOGIC registers for this don't work?
Yes, we need to check the link status from our own SoC's register.
+
+ return 0;
+}
+
+static void rockchip_pcie_establish_link(struct dw_pcie *pci)
s/establish/start/
+{
+ struct rockchip_pcie *rockchip = to_rockchip_pcie(pci);
+
+ if (dw_pcie_link_up(pci)) {
The DWC core does this for you now.
+ dev_err(pci->dev, "link already up\n");
+ return;
+ }
+
+ /* Reset device */
+ gpiod_set_value_cansleep(rockchip->rst_gpio, 0);
+ msleep(100);
+ gpiod_set_value_cansleep(rockchip->rst_gpio, 1);
+
+ rockchip_pcie_enable_ltssm(rockchip);
+}
+
+static void rockchip_pcie_enable_debug(struct rockchip_pcie *rockchip)
+{
+ rockchip_pcie_writel_apb(rockchip, PCIE_CLIENT_DBG_FIFO_PTN_HIT_D0,
+ PCIE_CLIENT_DBG_TRANSITION_DATA);
+ rockchip_pcie_writel_apb(rockchip, PCIE_CLIENT_DBG_FIFO_PTN_HIT_D1,
+ PCIE_CLIENT_DBG_TRANSITION_DATA);
+ rockchip_pcie_writel_apb(rockchip, PCIE_CLIENT_DBG_FIFO_TRN_HIT_D0,
+ PCIE_CLIENT_DBG_TRANSITION_DATA);
+ rockchip_pcie_writel_apb(rockchip, PCIE_CLIENT_DBG_FIFO_TRN_HIT_D1,
+ PCIE_CLIENT_DBG_TRANSITION_DATA);
+ rockchip_pcie_writel_apb(rockchip, PCIE_CLIENT_DBG_FIFO_MODE_CON,
+ PCIE_CLIENT_DBF_EN);
+}
+
+static void rockchip_pcie_debug_dump(struct rockchip_pcie *rockchip)
+{
+ u32 loop;
+ struct dw_pcie *pci = rockchip->pci;
+
+ dev_dbg(pci->dev, "ltssm = 0x%x\n",
+ rockchip_pcie_readl_apb(rockchip, PCIE_CLIENT_LTSSM_STATUS));
+ for (loop = 0; loop < 64; loop++)
+ dev_dbg(pci->dev, "fifo_status = 0x%x\n",
+ rockchip_pcie_readl_apb(rockchip,
+ PCIE_CLIENT_DBG_FIFO_STATUS));
+}
Please document what this debug stuff does. I tend to think it should be
removed.
Ok, will remove this.
+
+static int rockchip_pcie_host_init(struct pcie_port *pp)
+{
+ struct dw_pcie *pci = to_dw_pcie_from_pp(pp);
+ struct rockchip_pcie *rockchip = to_rockchip_pcie(pci);
+
+ dw_pcie_setup_rc(pp);
The DWC core does this now after host_init.
+
+ rockchip_pcie_enable_debug(rockchip);
+
+ rockchip_pcie_establish_link(pci);
+ dw_pcie_wait_for_link(pci);
The DWC core does link handling now.
+
+ rockchip_pcie_debug_dump(rockchip);
+
+ return 0;
+}
+
+static const struct dw_pcie_host_ops rockchip_pcie_host_ops = {
+ .host_init = rockchip_pcie_host_init,
+};
+
+static int rockchip_add_pcie_port(struct rockchip_pcie *rockchip)
Remove this wrapper and move to probe.
+{
+ int ret;
+ struct dw_pcie *pci = rockchip->pci;
+ struct pcie_port *pp = &pci->pp;
+
+ pp->ops = &rockchip_pcie_host_ops;
+
+ rockchip_pcie_set_mode(rockchip);
+
+ ret = dw_pcie_host_init(pp);
+ if (ret)
+ return ret;
+
+ return 0;
+}
+
+static void rockchip_pcie_clk_deinit(struct rockchip_pcie *rockchip)
+{
Another pointless wrapper.
+ clk_bulk_disable_unprepare(rockchip->clk_cnt, rockchip->clks);
+}
+
+static int rockchip_pcie_clk_init(struct rockchip_pcie *rockchip)
+{
+ struct device *dev = rockchip->pci->dev;
+ int ret;
+
+ ret = devm_clk_bulk_get_all(dev, &rockchip->clks);
+ if (ret < 0)
+ return ret;
+
+ rockchip->clk_cnt = ret;
+
+ ret = clk_bulk_prepare_enable(rockchip->clk_cnt, rockchip->clks);
+ if (ret)
+ return ret;
+
+ return 0;
+}
+
+static int rockchip_pcie_resource_get(struct platform_device *pdev,
+ struct rockchip_pcie *rockchip)
+{
+ struct dw_pcie *pci = rockchip->pci;
+
+ pci->dbi_base = devm_platform_ioremap_resource_byname(pdev, "pcie-dbi");
The established name is 'dbi' and the DWC core handles it for you.
+ if (IS_ERR(pci->dbi_base))
+ return PTR_ERR(pci->dbi_base);
+
+ rockchip->apb_base = devm_platform_ioremap_resource_byname(pdev,
+ "pcie-apb");
Just 'apb' for the name.
+ if (IS_ERR(rockchip->apb_base))
+ return PTR_ERR(rockchip->apb_base);
+
+ rockchip->rst_gpio = devm_gpiod_get_optional(&pdev->dev, "reset",
+ GPIOD_OUT_HIGH);
+ if (IS_ERR(rockchip->rst_gpio))
+ return PTR_ERR(rockchip->rst_gpio);
+
+ return 0;
+}
+
+static int rockchip_pcie_phy_init(struct rockchip_pcie *rockchip)
+{
+ int ret;
+ struct device *dev = rockchip->pci->dev;
+
+ rockchip->phy = devm_phy_get(dev, "pcie-phy");
+ if (IS_ERR(rockchip->phy))
+ return dev_err_probe(dev, PTR_ERR(rockchip->phy),
+ "missing phy\n");
+
+ ret = phy_init(rockchip->phy);
+ if (ret < 0)
+ return ret;
+
+ phy_power_on(rockchip->phy);
+
+ return 0;
+}
+
+static void rockchip_pcie_phy_deinit(struct rockchip_pcie *rockchip)
+{
+ phy_exit(rockchip->phy);
+ phy_power_off(rockchip->phy);
+}
+
+static int rockchip_pcie_reset_control_release(struct rockchip_pcie *rockchip)
+{
+ struct device *dev = rockchip->pci->dev;
+ int ret;
+
+ rockchip->rst = devm_reset_control_array_get_exclusive(dev);
+ if (IS_ERR(rockchip->rst))
+ return dev_err_probe(dev, PTR_ERR(rockchip->rst),
+ "failed to get reset lines\n");
+
+ ret = reset_control_deassert(rockchip->rst);
+
+ return ret;
+}
+
+static void rockchip_pcie_fast_link_setup(struct rockchip_pcie *rockchip)
+{
+ u32 val;
+
+ /* LTSSM EN ctrl mode */
+ val = rockchip_pcie_readl_apb(rockchip, PCIE_CLIENT_HOT_RESET_CTRL);
+ val |= PCIE_LTSSM_ENABLE_ENHANCE | (PCIE_LTSSM_ENABLE_ENHANCE << 16);
+ rockchip_pcie_writel_apb(rockchip, PCIE_CLIENT_HOT_RESET_CTRL, val);
+}
+
+static const struct of_device_id rockchip_pcie_of_match[] = {
+ { .compatible = "rockchip,rk3568-pcie", },
+ { /* sentinel */ },
+};
Move this next to MODULE_DEVICE_TABLE.
+
+static const struct dw_pcie_ops dw_pcie_ops = {
+ .link_up = rockchip_pcie_link_up,
+};
+
+static int rockchip_pcie_probe(struct platform_device *pdev)
+{
+ struct device *dev = &pdev->dev;
+ struct rockchip_pcie *rockchip;
+ struct dw_pcie *pci;
+ int ret;
+
+ rockchip = devm_kzalloc(dev, sizeof(*rockchip), GFP_KERNEL);
+ if (!rockchip)
+ return -ENOMEM;
+
+ pci = devm_kzalloc(dev, sizeof(*pci), GFP_KERNEL);
+ if (!pci)
+ return -ENOMEM;
+
+ pci->dev = dev;
+ pci->ops = &dw_pcie_ops;
+
+ rockchip->pci = pci;
+
+ ret = rockchip_pcie_resource_get(pdev, rockchip);
+ if (ret)
+ return ret;
+
+ /* DON'T MOVE ME: must be enable before phy init */
+ rockchip->vpcie3v3 = devm_regulator_get_optional(dev, "vpcie3v3");
+ if (IS_ERR(rockchip->vpcie3v3)) {
+ if (PTR_ERR(rockchip->vpcie3v3) != -ENODEV)
I think that cannot happen since it is optional.
+ return PTR_ERR(rockchip->vpcie3v3);
Use dev_err_probe()
+ dev_info(dev, "no vpcie3v3 regulator found\n");
So? It's optional.
Yes, vpcie3v3 is enabled by software or by hardware directly. If get
NULL, means
enabled by hardware directly.
+ }
+
+ if (!IS_ERR(rockchip->vpcie3v3)) {
You shouldn't need this check.
+ ret = regulator_enable(rockchip->vpcie3v3);
+ if (ret) {
+ dev_err(pci->dev, "fail to enable vpcie3v3 regulator\n");
+ return ret;
+ }
+ }
+
+ ret = rockchip_pcie_phy_init(rockchip);
+ if (ret)
+ goto disable_regulator;
+
+ ret = rockchip_pcie_reset_control_release(rockchip);
+ if (ret)
+ goto deinit_phy;
+
+ ret = rockchip_pcie_clk_init(rockchip);
+ if (ret)
+ goto deinit_phy;
+
+ rockchip_pcie_fast_link_setup(rockchip);
+
+ platform_set_drvdata(pdev, rockchip);
Move this up next to 'rockchip' alloc.
+
+ ret = rockchip_add_pcie_port(rockchip);
+ if (ret)
+ goto deinit_clk;
+
+ return 0;
+
+deinit_clk:
+ rockchip_pcie_clk_deinit(rockchip);
+deinit_phy:
+ rockchip_pcie_phy_deinit(rockchip);
+disable_regulator:
+ if (!IS_ERR(rockchip->vpcie3v3))
Shouldn't need this check.
+ regulator_disable(rockchip->vpcie3v3);
+
+ return ret;
+}
+
+MODULE_DEVICE_TABLE(of, rockchip_pcie_of_match);
+
+static struct platform_driver rockchip_pcie_driver = {
+ .driver = {
+ .name = "rk-pcie",
+ .of_match_table = rockchip_pcie_of_match,
+ .suppress_bind_attrs = true,
+ },
+ .probe = rockchip_pcie_probe,
+};
+
+builtin_platform_driver(rockchip_pcie_driver);
--
2.25.1
