Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1693348 > unrolled thread

[PATCH 0/5] PCI: MediaTek: Add support for new generation host controller

Started by<honghui.zhang@mediatek.com>
First post2017-07-21 04:40 +0200
Last post2017-07-21 04:40 +0200
Articles 4 — 1 participant

Back to article view | Back to linux.kernel


Contents

  [PATCH 0/5] PCI: MediaTek: Add support for new generation host controller <honghui.zhang@mediatek.com> - 2017-07-21 04:40 +0200
    [PATCH 3/5] dt-bindings: PCI: rename and cleanup MediaTek binding text <honghui.zhang@mediatek.com> - 2017-07-21 04:40 +0200
    [PATCH 1/5] PCI: mediatek: Add a structure to abstract the controller generations <honghui.zhang@mediatek.com> - 2017-07-21 04:40 +0200
    [PATCH 2/5] PCI: mediatek: switch to use platform_get_resource_byname() <honghui.zhang@mediatek.com> - 2017-07-21 04:40 +0200

#1693348 — [PATCH 0/5] PCI: MediaTek: Add support for new generation host controller

From<honghui.zhang@mediatek.com>
Date2017-07-21 04:40 +0200
Subject[PATCH 0/5] PCI: MediaTek: Add support for new generation host controller
Message-ID<u5w1b-7nv-3@gated-at.bofh.it>
From: Honghui Zhang <honghui.zhang@mediatek.com>

MediaTek's PCIe host controller has two generation HWs, the new
generation HW has two root ports, it shares most probing flow with the
legacy controller. But the read/write config space logical is different
from the lagacy controller.
This patchset abstract the common probing flow, and add support for the
new generation controller.

Ryder Lee (5):
  PCI: mediatek: Add a structure to abstract the controller generations
  PCI: mediatek: switch to use platform_get_resource_byname()
  dt-bindings: PCI: rename and cleanup MediaTek binding text
  PCI: mediatek: Add new generation controller support
  dt-bindings: PCI: add support for new generation controller

 ...{mediatek,mt7623-pcie.txt => mediatek-pcie.txt} | 113 +++-
 drivers/pci/host/Kconfig                           |   5 +-
 drivers/pci/host/pcie-mediatek.c                   | 578 +++++++++++++++++++--
 3 files changed, 629 insertions(+), 67 deletions(-)
 rename Documentation/devicetree/bindings/pci/{mediatek,mt7623-pcie.txt => mediatek-pcie.txt} (59%)

-- 
2.6.4

[toc] | [next] | [standalone]


#1693349 — [PATCH 3/5] dt-bindings: PCI: rename and cleanup MediaTek binding text

From<honghui.zhang@mediatek.com>
Date2017-07-21 04:40 +0200
Subject[PATCH 3/5] dt-bindings: PCI: rename and cleanup MediaTek binding text
Message-ID<u5w1b-7nv-7@gated-at.bofh.it>
In reply to#1693348
From: Ryder Lee <ryder.lee@mediatek.com>

In order to accommodate other SoC generations, this patch updates filename
to make it more generic, regroups specific properties by SoCs, and removes
redundant descriptions.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
---
 ...{mediatek,mt7623-pcie.txt => mediatek-pcie.txt} | 29 +++++++++++-----------
 1 file changed, 14 insertions(+), 15 deletions(-)
 rename Documentation/devicetree/bindings/pci/{mediatek,mt7623-pcie.txt => mediatek-pcie.txt} (91%)

diff --git a/Documentation/devicetree/bindings/pci/mediatek,mt7623-pcie.txt b/Documentation/devicetree/bindings/pci/mediatek-pcie.txt
similarity index 91%
rename from Documentation/devicetree/bindings/pci/mediatek,mt7623-pcie.txt
rename to Documentation/devicetree/bindings/pci/mediatek-pcie.txt
index fe80dda..294f4a3 100644
--- a/Documentation/devicetree/bindings/pci/mediatek,mt7623-pcie.txt
+++ b/Documentation/devicetree/bindings/pci/mediatek-pcie.txt
@@ -1,18 +1,13 @@
-MediaTek Gen2 PCIe controller which is available on MT7623 series SoCs
-
-PCIe subsys supports single root complex (RC) with 3 Root Ports. Each root
-ports supports a Gen2 1-lane Link and has PIPE interface to PHY.
+MediaTek Gen2 PCIe controller
 
 Required properties:
-- compatible: Should contain "mediatek,mt7623-pcie".
+- compatible: Should contain one of the following string:
+	"mediatek,mt7623-pcie"
+	"mediatek,mt2701-pcie"
 - device_type: Must be "pci"
 - reg: Base addresses and lengths of the PCIe controller.
 - #address-cells: Address representation for root ports (must be 3)
 - #size-cells: Size representation for root ports (must be 2)
-- #interrupt-cells: Size representation for interrupts (must be 1)
-- interrupt-map-mask and interrupt-map: Standard PCI IRQ mapping properties
-  Please refer to the standard PCI bus binding document for a more detailed
-  explanation.
 - clocks: Must contain an entry for each entry in clock-names.
   See ../clocks/clock-bindings.txt for details.
 - clock-names: Must include the following entries:
@@ -20,12 +15,6 @@ Required properties:
   - sys_ck0 :for clock of Port0
   - sys_ck1 :for clock of Port1
   - sys_ck2 :for clock of Port2
-- resets: Must contain an entry for each entry in reset-names.
-  See ../reset/reset.txt for details.
-- reset-names: Must include the following entries:
-  - pcie-rst0 :port0 reset
-  - pcie-rst1 :port1 reset
-  - pcie-rst2 :port2 reset
 - phys: List of PHY specifiers (used by generic PHY framework).
 - phy-names : Must be "pcie-phy0", "pcie-phy1", "pcie-phyN".. based on the
   number of PHYs as specified in *phys* property.
@@ -34,6 +23,16 @@ Required properties:
 - bus-range: Range of bus numbers associated with this controller.
 - ranges: Ranges for the PCI memory and I/O regions.
 
+Required properties for MT7623/MT2701:
+- #interrupt-cells: Size representation for interrupts (must be 1)
+- interrupt-map-mask and interrupt-map: Standard PCI IRQ mapping properties
+  Please refer to the standard PCI bus binding document for a more detailed
+  explanation.
+- resets: Must contain an entry for each entry in reset-names.
+  See ../reset/reset.txt for details.
+- reset-names: Must be "pcie-rst0", "pcie-rst1", "pcie-rstN".. based on the
+  number of root ports.
+
 In addition, the device tree node must have sub-nodes describing each
 PCIe port interface, having the following mandatory properties:
 
-- 
2.6.4

[toc] | [prev] | [next] | [standalone]


#1693350 — [PATCH 1/5] PCI: mediatek: Add a structure to abstract the controller generations

From<honghui.zhang@mediatek.com>
Date2017-07-21 04:40 +0200
Subject[PATCH 1/5] PCI: mediatek: Add a structure to abstract the controller generations
Message-ID<u5w1c-7nv-11@gated-at.bofh.it>
In reply to#1693348
From: Ryder Lee <ryder.lee@mediatek.com>

Introduce a structure "mtk_pcie_soc" to abstract the differences between
controller generations, and the .startup() hook is used to encapsulate
some SoC-dependent related setting. In doing so, the common code which
will be reused by future chips.

In addition, we change the approaches to waiting Gen2 training by using
readl_poll_timeout() calls.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Honghui Zhang <honghui.zhang@mediatek.com>
---
 drivers/pci/host/pcie-mediatek.c | 81 +++++++++++++++++++++++-----------------
 1 file changed, 47 insertions(+), 34 deletions(-)

diff --git a/drivers/pci/host/pcie-mediatek.c b/drivers/pci/host/pcie-mediatek.c
index 5a9d858..d6ac342 100644
--- a/drivers/pci/host/pcie-mediatek.c
+++ b/drivers/pci/host/pcie-mediatek.c
@@ -16,6 +16,7 @@
 
 #include <linux/clk.h>
 #include <linux/delay.h>
+#include <linux/iopoll.h>
 #include <linux/kernel.h>
 #include <linux/of_address.h>
 #include <linux/of_pci.h>
@@ -63,6 +64,18 @@
 #define PCIE_FC_CREDIT_MASK	(GENMASK(31, 31) | GENMASK(28, 16))
 #define PCIE_FC_CREDIT_VAL(x)	((x) << 16)
 
+struct mtk_pcie_port;
+
+/**
+ * struct mtk_pcie_soc - differentiate between host generations
+ * @ops: pointer to configuration access functions
+ * @startup: pointer to controller setting functions
+ */
+struct mtk_pcie_soc {
+	struct pci_ops *ops;
+	int (*startup)(struct mtk_pcie_port *port);
+};
+
 /**
  * struct mtk_pcie_port - PCIe port information
  * @base: IO mapped register base
@@ -96,6 +109,7 @@ struct mtk_pcie_port {
  * @busn: bus range
  * @offset: IO / Memory offset
  * @ports: pointer to PCIe port information
+ * @soc: pointer to SoC-dependent operations
  */
 struct mtk_pcie {
 	struct device *dev;
@@ -111,13 +125,9 @@ struct mtk_pcie {
 		resource_size_t io;
 	} offset;
 	struct list_head ports;
+	const struct mtk_pcie_soc *soc;
 };
 
-static inline bool mtk_pcie_link_up(struct mtk_pcie_port *port)
-{
-	return !!(readl(port->base + PCIE_LINK_STATUS) & PCIE_PORT_LINKUP);
-}
-
 static void mtk_pcie_subsys_powerdown(struct mtk_pcie *pcie)
 {
 	struct device *dev = pcie->dev;
@@ -171,12 +181,30 @@ static struct pci_ops mtk_pcie_ops = {
 	.write = pci_generic_config_write,
 };
 
-static void mtk_pcie_configure_rc(struct mtk_pcie_port *port)
+static int mtk_pcie_startup_ports(struct mtk_pcie_port *port)
 {
 	struct mtk_pcie *pcie = port->pcie;
 	u32 func = PCI_FUNC(port->index << 3);
 	u32 slot = PCI_SLOT(port->index << 3);
 	u32 val;
+	int err;
+
+	/* assert port PERST_N */
+	val = readl(pcie->base + PCIE_SYS_CFG);
+	val |= PCIE_PORT_PERST(port->index);
+	writel(val, pcie->base + PCIE_SYS_CFG);
+
+	/* de-assert port PERST_N */
+	val = readl(pcie->base + PCIE_SYS_CFG);
+	val &= ~PCIE_PORT_PERST(port->index);
+	writel(val, pcie->base + PCIE_SYS_CFG);
+
+	/* 100ms timeout value should be enough for Gen1/2 training */
+	err = readl_poll_timeout(port->base + PCIE_LINK_STATUS, val,
+				 !!(val & PCIE_PORT_LINKUP), 20,
+				 100 * USEC_PER_MSEC);
+	if (err)
+		return -ETIMEDOUT;
 
 	/* enable interrupt */
 	val = readl(pcie->base + PCIE_INT_ENABLE);
@@ -209,30 +237,14 @@ static void mtk_pcie_configure_rc(struct mtk_pcie_port *port)
 	writel(PCIE_CONF_ADDR(PCIE_FTS_NUM, func, slot, 0),
 	       pcie->base + PCIE_CFG_ADDR);
 	writel(val, pcie->base + PCIE_CFG_DATA);
-}
 
-static void mtk_pcie_assert_ports(struct mtk_pcie_port *port)
-{
-	struct mtk_pcie *pcie = port->pcie;
-	u32 val;
-
-	/* assert port PERST_N */
-	val = readl(pcie->base + PCIE_SYS_CFG);
-	val |= PCIE_PORT_PERST(port->index);
-	writel(val, pcie->base + PCIE_SYS_CFG);
-
-	/* de-assert port PERST_N */
-	val = readl(pcie->base + PCIE_SYS_CFG);
-	val &= ~PCIE_PORT_PERST(port->index);
-	writel(val, pcie->base + PCIE_SYS_CFG);
-
-	/* PCIe v2.0 need at least 100ms delay to train from Gen1 to Gen2 */
-	msleep(100);
+	return 0;
 }
 
 static void mtk_pcie_enable_ports(struct mtk_pcie_port *port)
 {
-	struct device *dev = port->pcie->dev;
+	struct mtk_pcie *pcie = port->pcie;
+	struct device *dev = pcie->dev;
 	int err;
 
 	err = clk_prepare_enable(port->sys_ck);
@@ -250,13 +262,8 @@ static void mtk_pcie_enable_ports(struct mtk_pcie_port *port)
 		goto err_phy_on;
 	}
 
-	mtk_pcie_assert_ports(port);
-
-	/* if link up, then setup root port configuration space */
-	if (mtk_pcie_link_up(port)) {
-		mtk_pcie_configure_rc(port);
+	if (!pcie->soc->startup(port))
 		return;
-	}
 
 	dev_info(dev, "Port%d link down\n", port->index);
 
@@ -480,7 +487,7 @@ static int mtk_pcie_register_host(struct pci_host_bridge *host)
 
 	host->busnr = pcie->busn.start;
 	host->dev.parent = pcie->dev;
-	host->ops = &mtk_pcie_ops;
+	host->ops = pcie->soc->ops;
 	host->map_irq = of_irq_parse_and_map_pci;
 	host->swizzle_irq = pci_common_swizzle;
 
@@ -513,6 +520,7 @@ static int mtk_pcie_probe(struct platform_device *pdev)
 	pcie = pci_host_bridge_priv(host);
 
 	pcie->dev = dev;
+	pcie->soc = of_device_get_match_data(dev);
 	platform_set_drvdata(pdev, pcie);
 	INIT_LIST_HEAD(&pcie->ports);
 
@@ -537,9 +545,14 @@ static int mtk_pcie_probe(struct platform_device *pdev)
 	return err;
 }
 
+static struct mtk_pcie_soc mtk_pcie_soc_v1 = {
+	.ops = &mtk_pcie_ops,
+	.startup = mtk_pcie_startup_ports,
+};
+
 static const struct of_device_id mtk_pcie_ids[] = {
-	{ .compatible = "mediatek,mt7623-pcie"},
-	{ .compatible = "mediatek,mt2701-pcie"},
+	{ .compatible = "mediatek,mt7623-pcie", .data = &mtk_pcie_soc_v1 },
+	{ .compatible = "mediatek,mt2701-pcie", .data = &mtk_pcie_soc_v1 },
 	{},
 };
 
-- 
2.6.4

[toc] | [prev] | [next] | [standalone]


#1693352 — [PATCH 2/5] PCI: mediatek: switch to use platform_get_resource_byname()

From<honghui.zhang@mediatek.com>
Date2017-07-21 04:40 +0200
Subject[PATCH 2/5] PCI: mediatek: switch to use platform_get_resource_byname()
Message-ID<u5w1d-7nv-19@gated-at.bofh.it>
In reply to#1693348
From: Ryder Lee <ryder.lee@mediatek.com>

This is a transitional patch. We currently use platfarm_get_resource() for
retrieving the IOMEM resources, but there might be some chips don't have
subsys/shared registers part, which depends on platform design, and these
will be introduced in further patches.

Switch this function to use the platform_get_resource_byname() so that the
binding can be agnostic of the resource order.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
---
 drivers/pci/host/pcie-mediatek.c | 17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

diff --git a/drivers/pci/host/pcie-mediatek.c b/drivers/pci/host/pcie-mediatek.c
index d6ac342..5e0a2ee2 100644
--- a/drivers/pci/host/pcie-mediatek.c
+++ b/drivers/pci/host/pcie-mediatek.c
@@ -295,7 +295,8 @@ static int mtk_pcie_parse_ports(struct mtk_pcie *pcie,
 		return err;
 	}
 
-	regs = platform_get_resource(pdev, IORESOURCE_MEM, index + 1);
+	snprintf(name, sizeof(name), "port%d", index);
+	regs = platform_get_resource_byname(pdev, IORESOURCE_MEM, name);
 	port->base = devm_ioremap_resource(dev, regs);
 	if (IS_ERR(port->base)) {
 		dev_err(dev, "failed to map port%d base\n", index);
@@ -336,12 +337,14 @@ static int mtk_pcie_subsys_powerup(struct mtk_pcie *pcie)
 	struct resource *regs;
 	int err;
 
-	/* get shared registers */
-	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	pcie->base = devm_ioremap_resource(dev, regs);
-	if (IS_ERR(pcie->base)) {
-		dev_err(dev, "failed to map shared register\n");
-		return PTR_ERR(pcie->base);
+	/* get shared registers, which are optional */
+	regs = platform_get_resource_byname(pdev, IORESOURCE_MEM, "subsys");
+	if (regs) {
+		pcie->base = devm_ioremap_resource(dev, regs);
+		if (IS_ERR(pcie->base)) {
+			dev_err(dev, "failed to map shared register\n");
+			return PTR_ERR(pcie->base);
+		}
 	}
 
 	pcie->free_ck = devm_clk_get(dev, "free_ck");
-- 
2.6.4

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web