mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] regulator: renesas-usb-vbus-regulator: Fix module autoloading
@ 2026-11-16  8:44 Nhat Tran
  2026-11-16  9:00 ` Biju Das
  2026-11-16  9:06 ` Biju Das
  0 siblings, 2 replies; 4+ messages in thread
From: Nhat Tran @ 2026-11-16  8:44 UTC (permalink / raw)
  To: Liam Girdwood, Mark Brown
  Cc: Biju Das, linux-renesas-soc, linux-kernel, Nhat Tran

The driver can be built as a module, and Kconfig makes it a module by
default when RESET_RZG2L_USBPHY_CTRL is a module. However the driver has
neither a platform_device_id table nor a MODULE_ALIAS(), so no module
alias is generated and udev never loads it.

The device is created by the rzg2l-usbphy-ctrl reset driver with
platform_device_set_of_node_from_dev(), which marks the OF node as
reused, so the uevent carries the platform modalias
"platform:rzg2l-usb-vbus-regulator" rather than an OF one.

Without the regulator, the USB0 PHY waits forever for its vbus supply,
and the whole USB0 port stays unusable. On an RZ/G2L SMARC EVK:

  # cat /sys/kernel/debug/devices_deferred
  11c50100.usb    platform: supplier 11c50200.usb-phy not ready
  11c50000.usb    platform: supplier 11c50200.usb-phy not ready
  11c60000.usb    platform: supplier 11c50200.usb-phy not ready
  11c50200.usb-phy

Loading the module by hand makes all of them probe.

Add a platform_device_id table and export it with MODULE_DEVICE_TABLE()
so that the module is loaded automatically.

Fixes: 84fbd6198766 ("regulator: Add Renesas RZ/G2L USB VBUS regulator driver")
Signed-off-by: Nhat Tran <nhat.tranminhk2@gmail.com>
---
 drivers/regulator/renesas-usb-vbus-regulator.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/regulator/renesas-usb-vbus-regulator.c b/drivers/regulator/renesas-usb-vbus-regulator.c
index 9ba791bd72ec..65a58bf5ffa4 100644
--- a/drivers/regulator/renesas-usb-vbus-regulator.c
+++ b/drivers/regulator/renesas-usb-vbus-regulator.c
@@ -55,12 +55,19 @@ static int rzg2l_usb_vbus_regulator_probe(struct platform_device *pdev)
 	return 0;
 }
 
+static const struct platform_device_id rzg2l_usb_vbus_regulator_ids[] = {
+	{ .name = "rzg2l-usb-vbus-regulator" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(platform, rzg2l_usb_vbus_regulator_ids);
+
 static struct platform_driver rzg2l_usb_vbus_regulator_driver = {
 	.probe = rzg2l_usb_vbus_regulator_probe,
 	.driver	= {
 		.name = "rzg2l-usb-vbus-regulator",
 		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
 	},
+	.id_table = rzg2l_usb_vbus_regulator_ids,
 };
 module_platform_driver(rzg2l_usb_vbus_regulator_driver);
 
-- 
2.25.1


^ permalink raw reply	[flat|nested] 4+ messages in thread

* RE: [PATCH] regulator: renesas-usb-vbus-regulator: Fix module autoloading
  2026-11-16  8:44 [PATCH] regulator: renesas-usb-vbus-regulator: Fix module autoloading Nhat Tran
@ 2026-11-16  9:00 ` Biju Das
  2026-11-16  9:06 ` Biju Das
  1 sibling, 0 replies; 4+ messages in thread
From: Biju Das @ 2026-11-16  9:00 UTC (permalink / raw)
  To: Nhat Tran, Liam Girdwood, Mark Brown; +Cc: linux-renesas-soc, linux-kernel

Hi Nhat Tran,

Thanks for the patch.

> -----Original Message-----
> From: Nhat Tran <nhat.tranminhk2@gmail.com>
> Sent: 16 November 2026 08:44
> Subject: [PATCH] regulator: renesas-usb-vbus-regulator: Fix module autoloading
> 
> 
> The driver can be built as a module, and Kconfig makes it a module by default when
> RESET_RZG2L_USBPHY_CTRL is a module. However the driver has neither a platform_device_id table nor a
> MODULE_ALIAS(), so no module alias is generated and udev never loads it.
> 
> The device is created by the rzg2l-usbphy-ctrl reset driver with
> platform_device_set_of_node_from_dev(), which marks the OF node as reused, so the uevent carries the
> platform modalias "platform:rzg2l-usb-vbus-regulator" rather than an OF one.
> 
> Without the regulator, the USB0 PHY waits forever for its vbus supply, and the whole USB0 port stays
> unusable. On an RZ/G2L SMARC EVK:
> 
>   # cat /sys/kernel/debug/devices_deferred
>   11c50100.usb    platform: supplier 11c50200.usb-phy not ready
>   11c50000.usb    platform: supplier 11c50200.usb-phy not ready
>   11c60000.usb    platform: supplier 11c50200.usb-phy not ready
>   11c50200.usb-phy
> 
> Loading the module by hand makes all of them probe.
> 
> Add a platform_device_id table and export it with MODULE_DEVICE_TABLE() so that the module is loaded
> automatically.

A patch is already mainlined [1] with these changes as part of RZ/G3L regulator support.

[1]
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=e0d3ff3fe71b2773913c43e5f759473cd003c9af

Cheers,
Biju

> 
> Fixes: 84fbd6198766 ("regulator: Add Renesas RZ/G2L USB VBUS regulator driver")
> Signed-off-by: Nhat Tran <nhat.tranminhk2@gmail.com>
> ---
>  drivers/regulator/renesas-usb-vbus-regulator.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/drivers/regulator/renesas-usb-vbus-regulator.c b/drivers/regulator/renesas-usb-vbus-
> regulator.c
> index 9ba791bd72ec..65a58bf5ffa4 100644
> --- a/drivers/regulator/renesas-usb-vbus-regulator.c
> +++ b/drivers/regulator/renesas-usb-vbus-regulator.c
> @@ -55,12 +55,19 @@ static int rzg2l_usb_vbus_regulator_probe(struct platform_device *pdev)
>         return 0;
>  }
> 
> +static const struct platform_device_id rzg2l_usb_vbus_regulator_ids[] = {
> +       { .name = "rzg2l-usb-vbus-regulator" },

Have you tested this patch properly? 

I had an issue in probing the device related to id_table match due to string length
restriction in [2]

[2]https://elixir.bootlin.com/linux/v7.1.2/source/include/linux/mod_devicetable.h#L604

Cheers,
Biju

> +       { /* sentinel */ }
> +};
> +MODULE_DEVICE_TABLE(platform, rzg2l_usb_vbus_regulator_ids);
> +
>  static struct platform_driver rzg2l_usb_vbus_regulator_driver = {
>         .probe = rzg2l_usb_vbus_regulator_probe,
>         .driver = {
>                 .name = "rzg2l-usb-vbus-regulator",
>                 .probe_type = PROBE_PREFER_ASYNCHRONOUS,
>         },
> +       .id_table = rzg2l_usb_vbus_regulator_ids,
>  };
>  module_platform_driver(rzg2l_usb_vbus_regulator_driver);
> 
> --
> 2.25.1


^ permalink raw reply	[flat|nested] 4+ messages in thread

* RE: [PATCH] regulator: renesas-usb-vbus-regulator: Fix module autoloading
  2026-11-16  8:44 [PATCH] regulator: renesas-usb-vbus-regulator: Fix module autoloading Nhat Tran
  2026-11-16  9:00 ` Biju Das
@ 2026-11-16  9:06 ` Biju Das
  2026-11-16  9:17   ` Nhat Tran
  1 sibling, 1 reply; 4+ messages in thread
From: Biju Das @ 2026-11-16  9:06 UTC (permalink / raw)
  To: Nhat Tran, Liam Girdwood, Mark Brown; +Cc: linux-renesas-soc, linux-kernel

Hi Nhat Tran,

Thanks for the patch.

> -----Original Message-----
> From: Nhat Tran <nhat.tranminhk2@gmail.com>
> Sent: 16 November 2026 08:44
> Subject: [PATCH] regulator: renesas-usb-vbus-regulator: Fix module autoloading
> 
> 
> The driver can be built as a module, and Kconfig makes it a module by default when
> RESET_RZG2L_USBPHY_CTRL is a module. However the driver has neither a platform_device_id table nor a
> MODULE_ALIAS(), so no module alias is generated and udev never loads it.
> 
> The device is created by the rzg2l-usbphy-ctrl reset driver with
> platform_device_set_of_node_from_dev(), which marks the OF node as reused, so the uevent carries the
> platform modalias "platform:rzg2l-usb-vbus-regulator" rather than an OF one.
> 
> Without the regulator, the USB0 PHY waits forever for its vbus supply, and the whole USB0 port stays
> unusable. On an RZ/G2L SMARC EVK:
> 
>   # cat /sys/kernel/debug/devices_deferred
>   11c50100.usb    platform: supplier 11c50200.usb-phy not ready
>   11c50000.usb    platform: supplier 11c50200.usb-phy not ready
>   11c60000.usb    platform: supplier 11c50200.usb-phy not ready
>   11c50200.usb-phy
> 
> Loading the module by hand makes all of them probe.
> 
> Add a platform_device_id table and export it with MODULE_DEVICE_TABLE() so that the module is loaded
> automatically.
> 
> Fixes: 84fbd6198766 ("regulator: Add Renesas RZ/G2L USB VBUS regulator driver")
> Signed-off-by: Nhat Tran <nhat.tranminhk2@gmail.com>
> ---
>  drivers/regulator/renesas-usb-vbus-regulator.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/drivers/regulator/renesas-usb-vbus-regulator.c b/drivers/regulator/renesas-usb-vbus-
> regulator.c
> index 9ba791bd72ec..65a58bf5ffa4 100644
> --- a/drivers/regulator/renesas-usb-vbus-regulator.c
> +++ b/drivers/regulator/renesas-usb-vbus-regulator.c
> @@ -55,12 +55,19 @@ static int rzg2l_usb_vbus_regulator_probe(struct platform_device *pdev)
>         return 0;
>  }
> 
> +static const struct platform_device_id rzg2l_usb_vbus_regulator_ids[] = {
> +       { .name = "rzg2l-usb-vbus-regulator" },
> +       { /* sentinel */ }

Have you tested this patch properly? 

I had an issue in probing the device related to id_table match due to string length restriction in [1].

So, to avoid ABI breakage a patch is introduced in [2] and
Similar changes done in [3].

We cannot apply fixes due to ABI brakage.

So do you need this patch as [3] is already mainlined?

[1]https://elixir.bootlin.com/linux/v7.1.2/source/include/linux/mod_devicetable.h#L604
[2]https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=26599cbf46fb1a72d4c7789d08a127a3d6bc59d0
[3]https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=e0d3ff3fe71b2773913c43e5f759473cd003c9af


Cheers,
Biju

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] regulator: renesas-usb-vbus-regulator: Fix module autoloading
  2026-11-16  9:06 ` Biju Das
@ 2026-11-16  9:17   ` Nhat Tran
  0 siblings, 0 replies; 4+ messages in thread
From: Nhat Tran @ 2026-11-16  9:17 UTC (permalink / raw)
  To: Biju Das; +Cc: Liam Girdwood, Mark Brown, linux-renesas-soc, linux-kernel

Hi Biju,

On Mon, Nov 16, 2026 at 09:00 UTC, Biju Das wrote:
> A patch is already mainlined [1] with these changes as part of RZ/G3L
> regulator support.

Sorry, I missed that series. Please drop this patch.

On Mon, Nov 16, 2026 at 09:06 UTC, Biju Das wrote:
>> +static const struct platform_device_id rzg2l_usb_vbus_regulator_ids[] = {
>> +	{ .name = "rzg2l-usb-vbus-regulator" },
>
> Have you tested this patch properly?
>
> I had an issue in probing the device related to id_table match due to
> string length restriction in [2]

You are right. "rzg2l-usb-vbus-regulator" is 24 characters long, so it
does not fit in name[PLATFORM_NAME_SIZE] together with its terminating
NUL.

I tested it on an RZ/G2L SMARC EVK with the driver built as a module:
the module was loaded automatically and USB0 probed. But that only
worked by luck: driver_data is 0 and directly follows name[], so
strcmp() stops there. With a non-zero driver_data, as in your RZ/G3L
patch, the match fails. Renaming the device, as you did, is the right
fix.

Thanks for the review,
Nhat

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2026-11-16  9:18 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-16  8:44 [PATCH] regulator: renesas-usb-vbus-regulator: Fix module autoloading Nhat Tran
2026-11-16  9:00 ` Biju Das
2026-11-16  9:06 ` Biju Das
2026-11-16  9:17   ` Nhat Tran

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®