[PATCH v2 0/3] introduce board-driver support

[PATCH v2 0/3] introduce board-driver support

[Oleksij Rempel]

changes v3:
- move devinfo patch to the first place
- devinfo: provide better commit message 

changes v2:
- move root device registration to the of_probe()
- port one board as example
- fix issue with devinfo

Oleksij Rempel (3):
  devinfo: do not dump the device node for the root node
  of: base: register DT root as device
  ARM: embest-riotboard: port board file to the driver model

 arch/arm/boards/embest-riotboard/board.c | 18 +++++++++++++-----
 commands/devinfo.c                       |  2 +-
 drivers/of/base.c                        |  5 +++++
 3 files changed, 19 insertions(+), 6 deletions(-)

-- 
2.27.0


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

[PATCH v3 1/3] devinfo: do not dump the device node for the root node

[Oleksij Rempel]

Calling the devinfo against a device which is linked to some devicetree
node weill result a devictree dump of this node. For example:

barebox@Protonic PRTI6Q board:/ devinfo sound-spdif.of
Bus: platform
Device node: /sound-spdif
sound-spdif {
        compatible = "fsl,imx-audio-spdif";
        model = "imx-spdif";
        spdif-controller = <0x7a>;
        spdif-in;
        spdif-out;
};

Calling same command on a device wich is linked to the root node of
devicetree, for example "machine.of", will trigger a dump fo complete
devicetree. Since the same functionality is provided by the "of_dump"
command, it is better to limit devinfo on the root device. After this
patch we would get following output:

barebox@Protonic PRTI6Q board:/ devinfo machine.of
Driver: board-protonic-imx6
Bus: platform
Parameters:
  boardid: 0 (type: uint32)
  boardrev: 1 (type: uint32)

In this example, "boardid" and "boardrev" are variabled provided by this
this specific board driver.

Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
---
 commands/devinfo.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/commands/devinfo.c b/commands/devinfo.c
index 81956b1cc0..018a7c4167 100644
--- a/commands/devinfo.c
+++ b/commands/devinfo.c
@@ -99,7 +99,7 @@ static int do_devinfo(int argc, char *argv[])
 			printf("\n");
 		}
 #ifdef CONFIG_OFDEVICE
-		if (dev->device_node) {
+		if (dev->device_node && dev->device_node != of_get_root_node()) {
 			printf("Device node: %s\n", dev->device_node->full_name);
 			of_print_nodes(dev->device_node, 0);
 		}
-- 
2.27.0


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

Re: [PATCH v3 1/3] devinfo: do not dump the device node for the root node

[Lucas Stach]

Am Donnerstag, den 16.07.2020, 08:46 +0200 schrieb Oleksij Rempel:
> Calling the devinfo against a device which is linked to some devicetree
> node weill result a devictree dump of this node. For example:
> 
> barebox@Protonic PRTI6Q board:/ devinfo sound-spdif.of
> Bus: platform
> Device node: /sound-spdif
> sound-spdif {
>         compatible = "fsl,imx-audio-spdif";
>         model = "imx-spdif";
>         spdif-controller = <0x7a>;
>         spdif-in;
>         spdif-out;
> };
> 
> Calling same command on a device wich is linked to the root node of
> devicetree, for example "machine.of", will trigger a dump fo complete
> devicetree.

Taking a step back from the immediate issue with the root node: is it
even a good idea to dump the whole subtree of nodes in the devinfo?
Most likely someone using devinfo is only interested in the of_node of
the device and maybe some subnodes that don't have a device on their
own.

So wouldn't it make more sense to only dump the node of the device in
question and subnodes without a compatible? This way we could remove
the special case for the root node and still get some useful
information in devinfo, like the board compatible property.

Regards,
Lucas

>  Since the same functionality is provided by the "of_dump"
> command, it is better to limit devinfo on the root device. After this
> patch we would get following output:
> 
> barebox@Protonic PRTI6Q board:/ devinfo machine.of
> Driver: board-protonic-imx6
> Bus: platform
> Parameters:
>   boardid: 0 (type: uint32)
>   boardrev: 1 (type: uint32)
> 
> In this example, "boardid" and "boardrev" are variabled provided by this
> this specific board driver.
> 
> Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
> ---
>  commands/devinfo.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/commands/devinfo.c b/commands/devinfo.c
> index 81956b1cc0..018a7c4167 100644
> --- a/commands/devinfo.c
> +++ b/commands/devinfo.c
> @@ -99,7 +99,7 @@ static int do_devinfo(int argc, char *argv[])
>  			printf("\n");
>  		}
>  #ifdef CONFIG_OFDEVICE
> -		if (dev->device_node) {
> +		if (dev->device_node && dev->device_node != of_get_root_node()) {
>  			printf("Device node: %s\n", dev->device_node->full_name);
>  			of_print_nodes(dev->device_node, 0);
>  		}


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

Re: [PATCH v3 1/3] devinfo: do not dump the device node for the root node

[Oleksij Rempel]

[-- Attachment #1.1: Type: text/plain, Size: 1763 bytes --]

On Thu, Jul 16, 2020 at 11:40:46AM +0200, Lucas Stach wrote:
> Am Donnerstag, den 16.07.2020, 08:46 +0200 schrieb Oleksij Rempel:
> > Calling the devinfo against a device which is linked to some devicetree
> > node weill result a devictree dump of this node. For example:
> > 
> > barebox@Protonic PRTI6Q board:/ devinfo sound-spdif.of
> > Bus: platform
> > Device node: /sound-spdif
> > sound-spdif {
> >         compatible = "fsl,imx-audio-spdif";
> >         model = "imx-spdif";
> >         spdif-controller = <0x7a>;
> >         spdif-in;
> >         spdif-out;
> > };
> > 
> > Calling same command on a device wich is linked to the root node of
> > devicetree, for example "machine.of", will trigger a dump fo complete
> > devicetree.
> 
> Taking a step back from the immediate issue with the root node: is it
> even a good idea to dump the whole subtree of nodes in the devinfo?
> Most likely someone using devinfo is only interested in the of_node of
> the device and maybe some subnodes that don't have a device on their
> own.
> 
> So wouldn't it make more sense to only dump the node of the device in
> question and subnodes without a compatible? This way we could remove
> the special case for the root node and still get some useful
> information in devinfo, like the board compatible property.

Hm... i have nothing against it. Then probably it is better to exclude
this patch for now.

Regards,
Oleksij
-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

[-- Attachment #2: Type: text/plain, Size: 149 bytes --]

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

[PATCH v3 2/3] of: base: register DT root as device

[Oleksij Rempel]

A usual board file contains at least one of_machine_is_compatible().
Some of the have a rather long list with complicated version logic.

To avoid own implementation for driver management, register the root node
of devicetree as platform device. So, the main platform bus can attach
proper board driver. After this patch a typical board.c file can reuse
existing driver infrastructure.

After this patch, you will be able to see all registered board drivers
with drvinfo as fallow:
...
board-embest-riot
board-protonic-imx6
    machine.of
...

With devinfo, you'll be able to get some board specific information,
if this is implemented:
barebox@Protonic PRTI6Q board:/ devinfo machine.of
Driver: board-protonic-imx6
Bus: platform
Parameters:
  boardid: 0 (type: uint32)
  boardrev: 1 (type: uint32)

Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
---
 drivers/of/base.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/of/base.c b/drivers/of/base.c
index 4c633bcd49..c48bf7a7fb 100644
--- a/drivers/of/base.c
+++ b/drivers/of/base.c
@@ -2133,6 +2133,7 @@ static void of_probe_memory(void)
 int of_probe(void)
 {
 	struct device_node *firmware;
+	struct device_d *dev;
 
 	if(!root_node)
 		return -ENODEV;
@@ -2149,6 +2150,10 @@ int of_probe(void)
 	if (firmware)
 		of_platform_populate(firmware, NULL, NULL);
 
+	dev = of_platform_device_create(root_node, NULL);
+	if (dev)
+		dev_set_name(dev, "%s.of", "machine");
+
 	of_clk_init(root_node, NULL);
 	of_platform_populate(root_node, of_default_bus_match_table, NULL);
 
-- 
2.27.0


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

[PATCH v3 3/3] ARM: embest-riotboard: port board file to the driver model

[Oleksij Rempel]

This patch can be used as example for the new board-driver
functionality.

Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
---
 arch/arm/boards/embest-riotboard/board.c | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)

diff --git a/arch/arm/boards/embest-riotboard/board.c b/arch/arm/boards/embest-riotboard/board.c
index 2e0cc9f0ab..746ecf0562 100644
--- a/arch/arm/boards/embest-riotboard/board.c
+++ b/arch/arm/boards/embest-riotboard/board.c
@@ -51,11 +51,8 @@ static int ar8035_phy_fixup(struct phy_device *dev)
 	return 0;
 }
 
-static int riotboard_device_init(void)
+static int riotboard_probe(struct device_d *dev)
 {
-	if (!of_machine_is_compatible("riot,imx6s-riotboard"))
-		return 0;
-
 	phy_register_fixup_for_uid(0x004dd072, 0xffffffef, ar8035_phy_fixup);
 
 	imx6_bbu_internal_mmc_register_handler("emmc", "/dev/mmc3.barebox",
@@ -65,4 +62,15 @@ static int riotboard_device_init(void)
 
 	return 0;
 }
-device_initcall(riotboard_device_init);
+
+static const struct of_device_id riotboard_of_match[] = {
+	{ .compatible = "riot,imx6s-riotboard" },
+	{},
+};
+
+static struct driver_d riotboard_driver = {
+	.name = "board-embest-riot",
+	.probe = riotboard_probe,
+	.of_compatible = DRV_OF_COMPAT(riotboard_of_match),
+};
+device_platform_driver(riotboard_driver);
-- 
2.27.0


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

[PATCH v2 0/3] introduce board-driver support

[Oleksij Rempel]

changes v2:
- move root device registration to the of_probe()
- port one board as example
- fix issue with devinfo

Oleksij Rempel (3):
  of: base: register DT root as device
  ARM: embest-riotboard: port board file to the driver model
  devinfo: do not dump the device node for the root node

 arch/arm/boards/embest-riotboard/board.c | 18 +++++++++++++-----
 commands/devinfo.c                       |  2 +-
 drivers/of/base.c                        |  5 +++++
 3 files changed, 19 insertions(+), 6 deletions(-)

-- 
2.27.0


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