[PATCH 0/7] Add ti-sysc bus driver

[PATCH 0/7] Add ti-sysc bus driver

[Teresa Remmet]

Hello,

with the device tree update to v5.0-rc2 most of the barebox devices
could not be enabled or probed any more. The am33xx*.dtsi files
have been updated to the interconnect target module hierarchy.
To make barebox working again we need a ti-sysc bus driver to
register the child devices of the ti-sysc bus.

Teresa

Teresa Remmet (7):
  drivers: bus: Add ti-sysc bus driver
  arm: configs: omap_defconfig: Enable ti-sysc bus driver
  arm: configs: am33xx_mlo_defconfig: Enable ti-sysc bus driver
  drivers: of: Add function to enable and register a device by alias
  arm: mach-omap: am335x_generic: Enable nodes by alias where needed
  net: ethernet: cpsw: cpsw-phy-sel: Use phandle for phy sel
  arm: dts: am335x-phytec: Use phy-handle instead of phy_id

 arch/arm/configs/am335x_mlo_defconfig       |  2 +-
 arch/arm/configs/omap_defconfig             |  2 +-
 arch/arm/dts/am335x-phytec-phycard-som.dtsi |  7 ++++-
 arch/arm/dts/am335x-phytec-phycore-som.dtsi |  7 ++++-
 arch/arm/dts/am335x-phytec-phyflex-som.dtsi | 12 ++++++--
 arch/arm/mach-omap/am33xx_generic.c         |  8 ++---
 drivers/bus/Kconfig                         |  7 +++++
 drivers/bus/Makefile                        |  1 +
 drivers/bus/ti-sysc.c                       | 45 +++++++++++++++++++++++++++++
 drivers/net/cpsw.c                          | 17 +++++++----
 drivers/of/platform.c                       | 19 ++++++++++++
 include/of.h                                |  8 +++++
 12 files changed, 119 insertions(+), 16 deletions(-)
 create mode 100644 drivers/bus/ti-sysc.c

-- 
2.7.4


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Teresa Remmet]

Adds minimal support for the sysc interconnect target module found
on many TI SoCs. With this device tree includes have been rearagned.
We need the driver to probe the child devices of the bus.

Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
---
 drivers/bus/Kconfig   |  7 +++++++
 drivers/bus/Makefile  |  1 +
 drivers/bus/ti-sysc.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 53 insertions(+)
 create mode 100644 drivers/bus/ti-sysc.c

diff --git a/drivers/bus/Kconfig b/drivers/bus/Kconfig
index 202df59762e8..7d35a663fdb5 100644
--- a/drivers/bus/Kconfig
+++ b/drivers/bus/Kconfig
@@ -6,6 +6,13 @@ config BUS_OMAP_GPMC
 	depends on OMAP_GPMC
 	bool "TI OMAP/AM33xx GPMC support"
 
+config TI_SYSC
+	depends on ARCH_OMAP
+	bool "TI sysc interconnect target module driver"
+	help
+          Generic driver for Texas Instruments interconnect target module
+          found on many TI SoCs.
+
 config IMX_WEIM
 	depends on ARCH_IMX
 	bool "i.MX WEIM driver"
diff --git a/drivers/bus/Makefile b/drivers/bus/Makefile
index 4b7aa888aba8..ba5cee40636b 100644
--- a/drivers/bus/Makefile
+++ b/drivers/bus/Makefile
@@ -1,3 +1,4 @@
 obj-$(CONFIG_BUS_OMAP_GPMC)	+= omap-gpmc.o
 obj-$(CONFIG_IMX_WEIM)		+= imx-weim.o
 obj-$(CONFIG_MVEBU_MBUS)	+= mvebu-mbus.o
+obj-$(CONFIG_TI_SYSC)		+= ti-sysc.o
diff --git a/drivers/bus/ti-sysc.c b/drivers/bus/ti-sysc.c
new file mode 100644
index 000000000000..b42ee7c895c7
--- /dev/null
+++ b/drivers/bus/ti-sysc.c
@@ -0,0 +1,45 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2019 Phytec Messtechnik GmbH, Teresa Remmet <t.remmet@phytec.de>
+ */
+
+#include <common.h>
+#include <init.h>
+#include <of.h>
+#include <linux/err.h>
+
+static const struct of_device_id sysc_match_table[] = {
+	{ .compatible = "simple-bus", },
+	{ /* sentinel */ },
+};
+
+static int ti_sysc_probe(struct device_d *dev)
+{
+	int ret;
+
+	ret = of_platform_populate(dev->device_node, sysc_match_table, dev);
+	if (ret)
+		dev_err(dev, "%s fail to create devices.\n",
+					dev->device_node->full_name);
+	return ret;
+};
+
+static struct of_device_id ti_sysc_dt_ids[] = {
+	{ .compatible = "ti,sysc-omap4",},
+	{ .compatible = "ti,sysc-omap4-simple",},
+	{ .compatible = "ti,sysc-omap4-timer",},
+	{ .compatible = "ti,sysc-omap2",},
+	{ },
+};
+
+static struct driver_d ti_sysc_driver = {
+	.name = "ti-sysc",
+	.probe = ti_sysc_probe,
+	.of_compatible = DRV_OF_COMPAT(ti_sysc_dt_ids),
+};
+
+static int ti_sysc_init(void)
+{
+	return platform_driver_register(&ti_sysc_driver);
+}
+postcore_initcall(ti_sysc_init);
-- 
2.7.4


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 2/7] arm: configs: omap_defconfig: Enable ti-sysc bus driver

[Teresa Remmet]

Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
---
 arch/arm/configs/omap_defconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/arm/configs/omap_defconfig b/arch/arm/configs/omap_defconfig
index 128027a6406b..8615283453d4 100644
--- a/arch/arm/configs/omap_defconfig
+++ b/arch/arm/configs/omap_defconfig
@@ -97,7 +97,6 @@ CONFIG_CMD_BOOTCHOOSER=y
 CONFIG_NET=y
 CONFIG_NET_NFS=y
 CONFIG_NET_NETCONSOLE=y
-CONFIG_NET_RESOLV=y
 CONFIG_OF_BAREBOX_DRIVERS=y
 CONFIG_DRIVER_SERIAL_NS16550=y
 CONFIG_DRIVER_SERIAL_NS16550_OMAP_EXTENSIONS=y
@@ -141,6 +140,7 @@ CONFIG_WATCHDOG_OMAP=y
 CONFIG_GPIO_GENERIC_PLATFORM=y
 CONFIG_PINCTRL_SINGLE=y
 CONFIG_BUS_OMAP_GPMC=y
+CONFIG_TI_SYSC=y
 CONFIG_FS_EXT4=y
 CONFIG_FS_TFTP=y
 CONFIG_FS_NFS=y
-- 
2.7.4


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 3/7] arm: configs: am33xx_mlo_defconfig: Enable ti-sysc bus driver

[Teresa Remmet]

Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
---
 arch/arm/configs/am335x_mlo_defconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/arm/configs/am335x_mlo_defconfig b/arch/arm/configs/am335x_mlo_defconfig
index 58034b6cea15..b58b71a85937 100644
--- a/arch/arm/configs/am335x_mlo_defconfig
+++ b/arch/arm/configs/am335x_mlo_defconfig
@@ -8,7 +8,6 @@ CONFIG_MACH_PHYTEC_SOM_AM335X=y
 CONFIG_THUMB2_BAREBOX=y
 # CONFIG_MEMINFO is not set
 CONFIG_MMU=y
-CONFIG_TEXT_BASE=0x0
 CONFIG_BAREBOX_MAX_PBLX_SIZE=0x1b400
 CONFIG_MALLOC_SIZE=0x0
 CONFIG_MALLOC_TLSF=y
@@ -38,6 +37,7 @@ CONFIG_MCI=y
 CONFIG_MCI_OMAP_HSMMC=y
 CONFIG_PINCTRL_SINGLE=y
 CONFIG_BUS_OMAP_GPMC=y
+CONFIG_TI_SYSC=y
 # CONFIG_FS_DEVFS is not set
 CONFIG_FS_FAT=y
 CONFIG_FS_FAT_LFN=y
-- 
2.7.4


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 4/7] drivers: of: Add function to enable and register a device by alias

[Teresa Remmet]

In some cases node names are not unique and passing the full path is really long.
So make add a new device by passing the alias string possible.

Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
---
 drivers/of/platform.c | 19 +++++++++++++++++++
 include/of.h          |  8 ++++++++
 2 files changed, 27 insertions(+)

diff --git a/drivers/of/platform.c b/drivers/of/platform.c
index ef8969ca8b19..d3795d799a13 100644
--- a/drivers/of/platform.c
+++ b/drivers/of/platform.c
@@ -207,6 +207,25 @@ struct device_d *of_device_enable_and_register_by_name(const char *name)
 }
 EXPORT_SYMBOL(of_device_enable_and_register_by_name);
 
+/**
+ * of_device_enable_and_register_by_alias - Enable and register device by alias
+ * @name: alias of the device node
+ *
+ * Returns pointer to created platform device, or NULL if a device was not
+ * registered. Unavailable devices will not get registered.
+ */
+struct device_d *of_device_enable_and_register_by_alias(const char *alias)
+{
+	struct device_node *node;
+
+	node = of_find_node_by_alias(NULL, alias);
+	if (!node)
+		return NULL;
+
+	return of_device_enable_and_register(node);
+}
+EXPORT_SYMBOL(of_device_enable_and_register_by_alias);
+
 #ifdef CONFIG_ARM_AMBA
 static struct device_d *of_amba_device_create(struct device_node *np)
 {
diff --git a/include/of.h b/include/of.h
index 184acb474136..b5f54dd4e5b5 100644
--- a/include/of.h
+++ b/include/of.h
@@ -255,6 +255,8 @@ extern int of_platform_populate(struct device_node *root,
 extern struct device_d *of_find_device_by_node(struct device_node *np);
 extern struct device_d *of_device_enable_and_register(struct device_node *np);
 extern struct device_d *of_device_enable_and_register_by_name(const char *name);
+extern struct device_d *of_device_enable_and_register_by_alias(
+							const char *alias);
 
 struct cdev *of_parse_partition(struct cdev *cdev, struct device_node *node);
 int of_parse_partitions(struct cdev *cdev, struct device_node *node);
@@ -670,6 +672,12 @@ static inline struct device_d *of_device_enable_and_register_by_name(
 	return NULL;
 }
 
+static inline struct device_d *of_device_enable_and_register_by_alias(
+				const char *alias)
+{
+	return NULL;
+}
+
 static inline int of_register_fixup(int (*fixup)(struct device_node *, void *),
 				void *context)
 {
-- 
2.7.4


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 5/7] arm: mach-omap: am335x_generic: Enable nodes by alias where needed

[Teresa Remmet]

With the interconnet target module hierarchy the node names used to enable boot
devices are not unique any more. Find those nodes by alias now.

Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
---
 arch/arm/mach-omap/am33xx_generic.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/arch/arm/mach-omap/am33xx_generic.c b/arch/arm/mach-omap/am33xx_generic.c
index fe3c4a8b175a..7577df761cb7 100644
--- a/arch/arm/mach-omap/am33xx_generic.c
+++ b/arch/arm/mach-omap/am33xx_generic.c
@@ -474,18 +474,18 @@ int am33xx_of_register_bootdevice(void)
 	switch (bootsource_get()) {
 	case BOOTSOURCE_MMC:
 		if (bootsource_get_instance() == 0)
-			dev = of_device_enable_and_register_by_name("mmc@48060000");
+			dev = of_device_enable_and_register_by_alias("mmc0");
 		else
-			dev = of_device_enable_and_register_by_name("mmc@481d8000");
+			dev = of_device_enable_and_register_by_alias("mmc1");
 		break;
 	case BOOTSOURCE_NAND:
 		dev = of_device_enable_and_register_by_name("gpmc@50000000");
 		break;
 	case BOOTSOURCE_SPI:
-		dev = of_device_enable_and_register_by_name("spi@48030000");
+		dev = of_device_enable_and_register_by_alias("spi0");
 		break;
 	case BOOTSOURCE_NET:
-		dev = of_device_enable_and_register_by_name("ethernet@4a100000");
+		dev = of_device_enable_and_register_by_alias("ethernet0");
 		break;
 	default:
 		/* Use nand fallback */
-- 
2.7.4


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 6/7] net: ethernet: cpsw: cpsw-phy-sel: Use phandle for phy sel

[Teresa Remmet]

With the usage of the interconnect target module hierarchy in device tree the
cpsw-phy-sel node moved to the system control module. An phandle has
been added instead to the device tree include.

Try to use the cpsw-phy-sel phandle first then fall back looking for the
cpsw-phy-sel child.

This patch is based on the upstream kernel patch:
18eb8aea7fb2 ("net: ethernet: cpsw-phy-sel: prefer phandle for phy sel")

Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
---
 drivers/net/cpsw.c | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/drivers/net/cpsw.c b/drivers/net/cpsw.c
index 432004ec9ca8..65f71c6fce5a 100644
--- a/drivers/net/cpsw.c
+++ b/drivers/net/cpsw.c
@@ -1053,6 +1053,7 @@ static int cpsw_probe_dt(struct cpsw_priv *priv)
 {
 	struct device_d *dev = priv->dev;
 	struct device_node *np = dev->device_node, *child;
+	struct device_node *physel;
 	int ret, i = 0;
 
 	ret = of_property_read_u32(np, "slaves", &priv->num_slaves);
@@ -1061,13 +1062,17 @@ static int cpsw_probe_dt(struct cpsw_priv *priv)
 
 	priv->slaves = xzalloc(sizeof(struct cpsw_slave) * priv->num_slaves);
 
-	for_each_child_of_node(np, child) {
-		if (of_device_is_compatible(child, "ti,am3352-cpsw-phy-sel")) {
-			ret = cpsw_phy_sel_init(priv, child);
-			if (ret)
-				return ret;
-		}
+	physel = of_parse_phandle(dev->device_node, "cpsw-phy-sel", 0);
+	if (!physel) {
+		physel = of_get_child_by_name(dev->device_node, "cpsw-phy-sel");
+		if (!physel)
+			dev_err(dev, "Phy mode node not found\n");
+	}
+	ret = cpsw_phy_sel_init(priv, physel);
+	if (ret)
+		return ret;
 
+	for_each_child_of_node(np, child) {
 		if (of_device_is_compatible(child, "ti,davinci_mdio")) {
 			ret = of_pinctrl_select_state_default(child);
 			if (ret)
-- 
2.7.4


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 7/7] arm: dts: am335x-phytec: Use phy-handle instead of phy_id

[Teresa Remmet]

Update am335x-phytec som device trees and use phy-handle instead of phy_id.

Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
---
 arch/arm/dts/am335x-phytec-phycard-som.dtsi |  7 ++++++-
 arch/arm/dts/am335x-phytec-phycore-som.dtsi |  7 ++++++-
 arch/arm/dts/am335x-phytec-phyflex-som.dtsi | 12 ++++++++++--
 3 files changed, 22 insertions(+), 4 deletions(-)

diff --git a/arch/arm/dts/am335x-phytec-phycard-som.dtsi b/arch/arm/dts/am335x-phytec-phycard-som.dtsi
index 6a0442916bcb..2320ca18070f 100644
--- a/arch/arm/dts/am335x-phytec-phycard-som.dtsi
+++ b/arch/arm/dts/am335x-phytec-phycard-som.dtsi
@@ -122,6 +122,10 @@
 	pinctrl-names = "default";
 	pinctrl-0 = <&davinci_mdio_default>;
 	status = "okay";
+
+	phy0: ethernet-phy@0 {
+		reg = <0>;
+	};
 };
 
 &phy_sel {
@@ -129,8 +133,9 @@
 };
 
 &cpsw_emac0 {
-	phy_id = <&davinci_mdio>, <0>;
+	phy-handle = <&phy0>;
 	phy-mode = "rmii";
+	dual_emac_res_vlan = <1>;
 };
 
 &mac {
diff --git a/arch/arm/dts/am335x-phytec-phycore-som.dtsi b/arch/arm/dts/am335x-phytec-phycore-som.dtsi
index cb2d30a15ccd..0601f5ab7ba5 100644
--- a/arch/arm/dts/am335x-phytec-phycore-som.dtsi
+++ b/arch/arm/dts/am335x-phytec-phycore-som.dtsi
@@ -244,6 +244,10 @@
 	pinctrl-names = "default";
 	pinctrl-0 = <&davinci_mdio_default>;
 	status = "okay";
+
+	phy0: ethernet-phy@0 {
+		reg = <0>;
+	};
 };
 
 &phy_sel {
@@ -251,8 +255,9 @@
 };
 
 &cpsw_emac0 {
-	phy_id = <&davinci_mdio>, <0>;
+	phy-handle = <&phy0>;
 	phy-mode = "rmii";
+	dual_emac_res_vlan = <1>;
 };
 
 &mac {
diff --git a/arch/arm/dts/am335x-phytec-phyflex-som.dtsi b/arch/arm/dts/am335x-phytec-phyflex-som.dtsi
index 465d3fa16cd7..4d0a91398843 100644
--- a/arch/arm/dts/am335x-phytec-phyflex-som.dtsi
+++ b/arch/arm/dts/am335x-phytec-phyflex-som.dtsi
@@ -197,6 +197,14 @@
 	pinctrl-names = "default", "sleep";
 	pinctrl-0 = <&davinci_mdio_default>;
 	status = "okay";
+
+	phy0: ethernet-phy@0 {
+		reg = <1>;
+	};
+
+	phy1: ethernet-phy@1 {
+		reg = <2>;
+	};
 };
 
 &phy_sel {
@@ -204,7 +212,7 @@
 };
 
 &cpsw_emac0 {
-	phy_id = <&davinci_mdio>, <1>;
+	phy-handle = <&phy0>;
 	phy-mode = "rgmii";
 	dual_emac_res_vlan = <1>;
 
@@ -221,7 +229,7 @@
 };
 
 &cpsw_emac1 {
-	phy_id = <&davinci_mdio>, <2>;
+	phy-handle = <&phy1>;
 	phy-mode = "rmii";
 	dual_emac_res_vlan = <2>;
 	status = "disabled";
-- 
2.7.4


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Sam Ravnborg]

Hi Teresa.

On Mon, Feb 11, 2019 at 04:16:34PM +0100, Teresa Remmet wrote:
> Adds minimal support for the sysc interconnect target module found
> on many TI SoCs. With this device tree includes have been rearagned.
s/rearagned/rearranged"

> We need the driver to probe the child devices of the bus.
> 
> Signed-off-by: Teresa Remmet <t.remmet@phytec.de>

The implementation is simple, which is always a good thing.
There is so far nothing TI specific in the implementation except
for the of_device_id entries.
Would it be an idea to rename this to a simple-bus driver.
And then we can add relevant ID's as other may need the same functionality.
?

	Sam

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Andrey Smirnov]

On Mon, Feb 11, 2019 at 7:16 AM Teresa Remmet <t.remmet@phytec.de> wrote:
>
> Adds minimal support for the sysc interconnect target module found
> on many TI SoCs. With this device tree includes have been rearagned.
> We need the driver to probe the child devices of the bus.
>
> Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
> ---
>  drivers/bus/Kconfig   |  7 +++++++
>  drivers/bus/Makefile  |  1 +
>  drivers/bus/ti-sysc.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 53 insertions(+)
>  create mode 100644 drivers/bus/ti-sysc.c
>
> diff --git a/drivers/bus/Kconfig b/drivers/bus/Kconfig
> index 202df59762e8..7d35a663fdb5 100644
> --- a/drivers/bus/Kconfig
> +++ b/drivers/bus/Kconfig
> @@ -6,6 +6,13 @@ config BUS_OMAP_GPMC
>         depends on OMAP_GPMC
>         bool "TI OMAP/AM33xx GPMC support"
>
> +config TI_SYSC
> +       depends on ARCH_OMAP
> +       bool "TI sysc interconnect target module driver"
> +       help
> +          Generic driver for Texas Instruments interconnect target module
> +          found on many TI SoCs.
> +
>  config IMX_WEIM
>         depends on ARCH_IMX
>         bool "i.MX WEIM driver"
> diff --git a/drivers/bus/Makefile b/drivers/bus/Makefile
> index 4b7aa888aba8..ba5cee40636b 100644
> --- a/drivers/bus/Makefile
> +++ b/drivers/bus/Makefile
> @@ -1,3 +1,4 @@
>  obj-$(CONFIG_BUS_OMAP_GPMC)    += omap-gpmc.o
>  obj-$(CONFIG_IMX_WEIM)         += imx-weim.o
>  obj-$(CONFIG_MVEBU_MBUS)       += mvebu-mbus.o
> +obj-$(CONFIG_TI_SYSC)          += ti-sysc.o
> diff --git a/drivers/bus/ti-sysc.c b/drivers/bus/ti-sysc.c
> new file mode 100644
> index 000000000000..b42ee7c895c7
> --- /dev/null
> +++ b/drivers/bus/ti-sysc.c
> @@ -0,0 +1,45 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/*
> + * Copyright (C) 2019 Phytec Messtechnik GmbH, Teresa Remmet <t.remmet@phytec.de>
> + */
> +
> +#include <common.h>
> +#include <init.h>
> +#include <of.h>
> +#include <linux/err.h>
> +
> +static const struct of_device_id sysc_match_table[] = {
> +       { .compatible = "simple-bus", },
> +       { /* sentinel */ },
> +};
> +
> +static int ti_sysc_probe(struct device_d *dev)
> +{
> +       int ret;
> +
> +       ret = of_platform_populate(dev->device_node, sysc_match_table, dev);

Is sysc_match_table really necessary? AFAIK that argument can be NULL
unless on of the child nodes needs to be treated like a platform bus
as well.

> +       if (ret)
> +               dev_err(dev, "%s fail to create devices.\n",
> +                                       dev->device_node->full_name);
> +       return ret;
> +};
> +
> +static struct of_device_id ti_sysc_dt_ids[] = {
> +       { .compatible = "ti,sysc-omap4",},
> +       { .compatible = "ti,sysc-omap4-simple",},
> +       { .compatible = "ti,sysc-omap4-timer",},
> +       { .compatible = "ti,sysc-omap2",},
> +       { },
> +};
> +
> +static struct driver_d ti_sysc_driver = {
> +       .name = "ti-sysc",
> +       .probe = ti_sysc_probe,
> +       .of_compatible = DRV_OF_COMPAT(ti_sysc_dt_ids),
> +};
> +
> +static int ti_sysc_init(void)
> +{
> +       return platform_driver_register(&ti_sysc_driver);
> +}
> +postcore_initcall(ti_sysc_init);

postcore_platform_driver()?

Thanks,
Andrey Smirnov

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Sascha Hauer]

Hi Sam,

On Mon, Feb 11, 2019 at 04:41:38PM +0100, Sam Ravnborg wrote:
> Hi Teresa.
> 
> On Mon, Feb 11, 2019 at 04:16:34PM +0100, Teresa Remmet wrote:
> > Adds minimal support for the sysc interconnect target module found
> > on many TI SoCs. With this device tree includes have been rearagned.
> s/rearagned/rearranged"
> 
> > We need the driver to probe the child devices of the bus.
> > 
> > Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
> 
> The implementation is simple, which is always a good thing.
> There is so far nothing TI specific in the implementation except
> for the of_device_id entries.
> Would it be an idea to rename this to a simple-bus driver.
> And then we can add relevant ID's as other may need the same functionality.
> ?

I'm not sure if it's worth it. It could also happen that we'll need to do
some TI specific initialization in the currently not-so-generic driver
in which case we would have to fork from the generic driver later.

I'm fine with either way.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Sascha Hauer]

On Mon, Feb 11, 2019 at 04:16:34PM +0100, Teresa Remmet wrote:
> Adds minimal support for the sysc interconnect target module found
> on many TI SoCs. With this device tree includes have been rearagned.
> We need the driver to probe the child devices of the bus.
> 
> Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
> ---
>  drivers/bus/Kconfig   |  7 +++++++
>  drivers/bus/Makefile  |  1 +
>  drivers/bus/ti-sysc.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 53 insertions(+)
>  create mode 100644 drivers/bus/ti-sysc.c
> 
> diff --git a/drivers/bus/Kconfig b/drivers/bus/Kconfig
> index 202df59762e8..7d35a663fdb5 100644
> --- a/drivers/bus/Kconfig
> +++ b/drivers/bus/Kconfig
> @@ -6,6 +6,13 @@ config BUS_OMAP_GPMC
>  	depends on OMAP_GPMC
>  	bool "TI OMAP/AM33xx GPMC support"
>  
> +config TI_SYSC
> +	depends on ARCH_OMAP
> +	bool "TI sysc interconnect target module driver"
> +	help
> +          Generic driver for Texas Instruments interconnect target module
> +          found on many TI SoCs.

Can we default 'y' this option or select it from the SoCs that need it?

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Teresa Remmet]

Hello Andrey,

Am Montag, den 11.02.2019, 18:57 -0800 schrieb Andrey Smirnov:
> On Mon, Feb 11, 2019 at 7:16 AM Teresa Remmet <t.remmet@phytec.de>
> wrote:
> > 
> > Adds minimal support for the sysc interconnect target module found
> > on many TI SoCs. With this device tree includes have been
> > rearagned.
> > We need the driver to probe the child devices of the bus.
> > 
> > Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
> > ---
> > 
> > +
> > +static const struct of_device_id sysc_match_table[] = {
> > +       { .compatible = "simple-bus", },
> > +       { /* sentinel */ },
> > +};
> > +
> > +static int ti_sysc_probe(struct device_d *dev)
> > +{
> > +       int ret;
> > +
> > +       ret = of_platform_populate(dev->device_node,
> > sysc_match_table, dev);
> 
> Is sysc_match_table really necessary? AFAIK that argument can be NULL
> unless on of the child nodes needs to be treated like a platform bus
> as well.

My first attempted was using NULL but this does not work. As for
example the pinmux node (pinmux@800) is child of a simple-bus. And the
simple-bus itself is a child of a ti-sysc node. See dts/src/arm/am33xx-
l4.dtsi.
The kernel does it the same way. 

> 
> > +       if (ret)
> > +               dev_err(dev, "%s fail to create devices.\n",
> > +                                       dev->device_node-
> > >full_name);
> > +       return ret;
> > +};
> > +
> > +static struct of_device_id ti_sysc_dt_ids[] = {
> > +       { .compatible = "ti,sysc-omap4",},
> > +       { .compatible = "ti,sysc-omap4-simple",},
> > +       { .compatible = "ti,sysc-omap4-timer",},
> > +       { .compatible = "ti,sysc-omap2",},
> > +       { },
> > +};
> > +
> > +static struct driver_d ti_sysc_driver = {
> > +       .name = "ti-sysc",
> > +       .probe = ti_sysc_probe,
> > +       .of_compatible = DRV_OF_COMPAT(ti_sysc_dt_ids),
> > +};
> > +
> > +static int ti_sysc_init(void)
> > +{
> > +       return platform_driver_register(&ti_sysc_driver);
> > +}
> > +postcore_initcall(ti_sysc_init);
> 
> postcore_platform_driver()?

Thanks! I will update this.

Teresa

> 
> Thanks,
> Andrey Smirnov


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Teresa Remmet]

Hello Sascha and Sam, 
 
Am Dienstag, den 12.02.2019, 09:09 +0100 schrieb Sascha Hauer:
> Hi Sam,
> 
> On Mon, Feb 11, 2019 at 04:41:38PM +0100, Sam Ravnborg wrote:
> > Hi Teresa.
> > 
> > On Mon, Feb 11, 2019 at 04:16:34PM +0100, Teresa Remmet wrote:
> > > Adds minimal support for the sysc interconnect target module
> > > found
> > > on many TI SoCs. With this device tree includes have been
> > > rearagned.
> > 
> > s/rearagned/rearranged"
> > 
> > > We need the driver to probe the child devices of the bus.
> > > 
> > > Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
> > 
> > The implementation is simple, which is always a good thing.
> > There is so far nothing TI specific in the implementation except
> > for the of_device_id entries.
> > Would it be an idea to rename this to a simple-bus driver.
> > And then we can add relevant ID's as other may need the same
> > functionality.
> > ?
> 
> I'm not sure if it's worth it. It could also happen that we'll need
> to do
> some TI specific initialization in the currently not-so-generic
> driver
> in which case we would have to fork from the generic driver later.

Yes there might be some special initialization added later. That's what
I thought, too. At the moment the child devices of a simple bus are
populated in of_probe(). 
If there is any need for a such a generic driver right now. I will
rename it.

Teresa

> 
> I'm fine with either way.
> 
> Sascha
> 


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Teresa Remmet]

Hello Sascha,

Am Dienstag, den 12.02.2019, 09:10 +0100 schrieb Sascha Hauer:
> On Mon, Feb 11, 2019 at 04:16:34PM +0100, Teresa Remmet wrote:
> > Adds minimal support for the sysc interconnect target module found
> > on many TI SoCs. With this device tree includes have been
> > rearagned.
> > We need the driver to probe the child devices of the bus.
> > 
> > Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
> > ---
> >  drivers/bus/Kconfig   |  7 +++++++
> >  drivers/bus/Makefile  |  1 +
> >  drivers/bus/ti-sysc.c | 45
> > +++++++++++++++++++++++++++++++++++++++++++++
> >  3 files changed, 53 insertions(+)
> >  create mode 100644 drivers/bus/ti-sysc.c
> > 
> > diff --git a/drivers/bus/Kconfig b/drivers/bus/Kconfig
> > index 202df59762e8..7d35a663fdb5 100644
> > --- a/drivers/bus/Kconfig
> > +++ b/drivers/bus/Kconfig
> > @@ -6,6 +6,13 @@ config BUS_OMAP_GPMC
> >  	depends on OMAP_GPMC
> >  	bool "TI OMAP/AM33xx GPMC support"
> >  
> > +config TI_SYSC
> > +	depends on ARCH_OMAP
> > +	bool "TI sysc interconnect target module driver"
> > +	help
> > +          Generic driver for Texas Instruments interconnect target
> > module
> > +          found on many TI SoCs.
> 
> Can we default 'y' this option or select it from the SoCs that need
> it?

yes, this would be better. As barebox will not start without it. I will
update it.

Teresa

> 
> Sascha
> 


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Sam Ravnborg]

Hi Teresa.

> If there is any need for a such a generic driver right now. I will
> rename it.

I see no need for one for now.
So follow Sasha's recommendation and keep it TI specific
is the best choice.

	Sam

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Andrey Smirnov]

On Tue, Feb 12, 2019 at 5:13 AM Teresa Remmet <t.remmet@phytec.de> wrote:
>
> Hello Andrey,
>
> Am Montag, den 11.02.2019, 18:57 -0800 schrieb Andrey Smirnov:
> > On Mon, Feb 11, 2019 at 7:16 AM Teresa Remmet <t.remmet@phytec.de>
> > wrote:
> > >
> > > Adds minimal support for the sysc interconnect target module found
> > > on many TI SoCs. With this device tree includes have been
> > > rearagned.
> > > We need the driver to probe the child devices of the bus.
> > >
> > > Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
> > > ---
> > >
> > > +
> > > +static const struct of_device_id sysc_match_table[] = {
> > > +       { .compatible = "simple-bus", },
> > > +       { /* sentinel */ },
> > > +};
> > > +
> > > +static int ti_sysc_probe(struct device_d *dev)
> > > +{
> > > +       int ret;
> > > +
> > > +       ret = of_platform_populate(dev->device_node,
> > > sysc_match_table, dev);
> >
> > Is sysc_match_table really necessary? AFAIK that argument can be NULL
> > unless on of the child nodes needs to be treated like a platform bus
> > as well.
>
> My first attempted was using NULL but this does not work. As for
> example the pinmux node (pinmux@800) is child of a simple-bus. And the
> simple-bus itself is a child of a ti-sysc node. See dts/src/arm/am33xx-
> l4.dtsi.
> The kernel does it the same way.
>

Ah, I see, your use-case does require child nodes to be treated as a
platform bus. FWIW, there's also globally visible
of_default_bus_match_table that can probably replace sysc_match_table.

Thanks,
Andrey Smirnov

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 1/7] drivers: bus: Add ti-sysc bus driver

[Teresa Remmet]

Am Dienstag, den 12.02.2019, 16:56 -0800 schrieb Andrey Smirnov:
> On Tue, Feb 12, 2019 at 5:13 AM Teresa Remmet <t.remmet@phytec.de>
> wrote:
> > 
> > Hello Andrey,
> > 
> > Am Montag, den 11.02.2019, 18:57 -0800 schrieb Andrey Smirnov:
> > > On Mon, Feb 11, 2019 at 7:16 AM Teresa Remmet <t.remmet@phytec.de
> > > >
> > > wrote:
> > > > 
> > > > Adds minimal support for the sysc interconnect target module
> > > > found
> > > > on many TI SoCs. With this device tree includes have been
> > > > rearagned.
> > > > We need the driver to probe the child devices of the bus.
> > > > 
> > > > Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
> > > > ---
> > > > 
> > > > +
> > > > +static const struct of_device_id sysc_match_table[] = {
> > > > +       { .compatible = "simple-bus", },
> > > > +       { /* sentinel */ },
> > > > +};
> > > > +
> > > > +static int ti_sysc_probe(struct device_d *dev)
> > > > +{
> > > > +       int ret;
> > > > +
> > > > +       ret = of_platform_populate(dev->device_node,
> > > > sysc_match_table, dev);
> > > 
> > > Is sysc_match_table really necessary? AFAIK that argument can be
> > > NULL
> > > unless on of the child nodes needs to be treated like a platform
> > > bus
> > > as well.
> > 
> > My first attempted was using NULL but this does not work. As for
> > example the pinmux node (pinmux@800) is child of a simple-bus. And
> > the
> > simple-bus itself is a child of a ti-sysc node. See
> > dts/src/arm/am33xx-
> > l4.dtsi.
> > The kernel does it the same way.
> > 
> 
> Ah, I see, your use-case does require child nodes to be treated as a
> platform bus. FWIW, there's also globally visible
> of_default_bus_match_table that can probably replace
> sysc_match_table.

Yes, will add this in v2. 

Thanks,
Teresa


> 
> Thanks,
> Andrey Smirnov


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox