[PATCH v1 1/1] serial: core: Restore sysfs fwnode information

[PATCH v1 1/1] serial: core: Restore sysfs fwnode information

[Andy Shevchenko]

The change that restores sysfs fwnode information does it only for OF cases.
Update the fix to cover all possible types of fwnodes.

Fixes: d36f0e9a0002 ("serial: core: restore of_node information in sysfs")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 drivers/tty/serial/serial_base_bus.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/drivers/tty/serial/serial_base_bus.c b/drivers/tty/serial/serial_base_bus.c
index 22749ab0428a..8e891984cdc0 100644
--- a/drivers/tty/serial/serial_base_bus.c
+++ b/drivers/tty/serial/serial_base_bus.c
@@ -13,7 +13,7 @@
 #include <linux/device.h>
 #include <linux/idr.h>
 #include <linux/module.h>
-#include <linux/of.h>
+#include <linux/property.h>
 #include <linux/serial_core.h>
 #include <linux/slab.h>
 #include <linux/spinlock.h>
@@ -60,6 +60,7 @@ void serial_base_driver_unregister(struct device_driver *driver)
 	driver_unregister(driver);
 }
 
+/* On failure the caller must put device @dev with put_device() */
 static int serial_base_device_init(struct uart_port *port,
 				   struct device *dev,
 				   struct device *parent_dev,
@@ -73,7 +74,8 @@ static int serial_base_device_init(struct uart_port *port,
 	dev->parent = parent_dev;
 	dev->bus = &serial_base_bus_type;
 	dev->release = release;
-	device_set_of_node_from_dev(dev, parent_dev);
+
+	device_set_node(dev, fwnode_handle_get(dev_fwnode(parent_dev)));
 
 	if (!serial_base_initialized) {
 		dev_dbg(port->dev, "uart_add_one_port() called before arch_initcall()?\n");
@@ -94,7 +96,7 @@ static void serial_base_ctrl_release(struct device *dev)
 {
 	struct serial_ctrl_device *ctrl_dev = to_serial_base_ctrl_device(dev);
 
-	of_node_put(dev->of_node);
+	fwnode_handle_put(dev_fwnode(dev));
 	kfree(ctrl_dev);
 }
 
@@ -142,7 +144,7 @@ static void serial_base_port_release(struct device *dev)
 {
 	struct serial_port_device *port_dev = to_serial_base_port_device(dev);
 
-	of_node_put(dev->of_node);
+	fwnode_handle_put(dev_fwnode(dev));
 	kfree(port_dev);
 }
 
-- 
2.50.1

Re: [PATCH v1 1/1] serial: core: Restore sysfs fwnode information

[Greg Kroah-Hartman]

On Thu, Nov 27, 2025 at 05:36:50PM +0100, Andy Shevchenko wrote:
> The change that restores sysfs fwnode information does it only for OF cases.
> Update the fix to cover all possible types of fwnodes.
> 
> Fixes: d36f0e9a0002 ("serial: core: restore of_node information in sysfs")
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---
>  drivers/tty/serial/serial_base_bus.c | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/tty/serial/serial_base_bus.c b/drivers/tty/serial/serial_base_bus.c
> index 22749ab0428a..8e891984cdc0 100644
> --- a/drivers/tty/serial/serial_base_bus.c
> +++ b/drivers/tty/serial/serial_base_bus.c
> @@ -13,7 +13,7 @@
>  #include <linux/device.h>
>  #include <linux/idr.h>
>  #include <linux/module.h>
> -#include <linux/of.h>
> +#include <linux/property.h>
>  #include <linux/serial_core.h>
>  #include <linux/slab.h>
>  #include <linux/spinlock.h>
> @@ -60,6 +60,7 @@ void serial_base_driver_unregister(struct device_driver *driver)
>  	driver_unregister(driver);
>  }
>  
> +/* On failure the caller must put device @dev with put_device() */
>  static int serial_base_device_init(struct uart_port *port,
>  				   struct device *dev,
>  				   struct device *parent_dev,
> @@ -73,7 +74,8 @@ static int serial_base_device_init(struct uart_port *port,
>  	dev->parent = parent_dev;
>  	dev->bus = &serial_base_bus_type;
>  	dev->release = release;
> -	device_set_of_node_from_dev(dev, parent_dev);
> +
> +	device_set_node(dev, fwnode_handle_get(dev_fwnode(parent_dev)));
>  
>  	if (!serial_base_initialized) {
>  		dev_dbg(port->dev, "uart_add_one_port() called before arch_initcall()?\n");
> @@ -94,7 +96,7 @@ static void serial_base_ctrl_release(struct device *dev)
>  {
>  	struct serial_ctrl_device *ctrl_dev = to_serial_base_ctrl_device(dev);
>  
> -	of_node_put(dev->of_node);
> +	fwnode_handle_put(dev_fwnode(dev));
>  	kfree(ctrl_dev);
>  }
>  
> @@ -142,7 +144,7 @@ static void serial_base_port_release(struct device *dev)
>  {
>  	struct serial_port_device *port_dev = to_serial_base_port_device(dev);
>  
> -	of_node_put(dev->of_node);
> +	fwnode_handle_put(dev_fwnode(dev));
>  	kfree(port_dev);
>  }
>  
> -- 
> 2.50.1
> 
> 

Hi,

This is the friendly patch-bot of Greg Kroah-Hartman.  You have sent him
a patch that has triggered this response.  He used to manually respond
to these common problems, but in order to save his sanity (he kept
writing the same thing over and over, yet to different people), I was
created.  Hopefully you will not take offence and will fix the problem
in your patch and resubmit it so that it can be accepted into the Linux
kernel tree.

You are receiving this message because of the following common error(s)
as indicated below:

- You have marked a patch with a "Fixes:" tag for a commit that is in an
  older released kernel, yet you do not have a cc: stable line in the
  signed-off-by area at all, which means that the patch will not be
  applied to any older kernel releases.  To properly fix this, please
  follow the documented rules in the
  Documentation/process/stable-kernel-rules.rst file for how to resolve
  this.

If you wish to discuss this problem further, or you have questions about
how to resolve this issue, please feel free to respond to this email and
Greg will reply once he has dug out from the pending patches received
from other developers.

thanks,

greg k-h's patch email bot

Re: [PATCH v1 1/1] serial: core: Restore sysfs fwnode information

[Cosmin Tanislav]

On 11/27/25 6:36 PM, Andy Shevchenko wrote:
> The change that restores sysfs fwnode information does it only for OF cases.
> Update the fix to cover all possible types of fwnodes.
> 
> Fixes: d36f0e9a0002 ("serial: core: restore of_node information in sysfs")
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---
>   drivers/tty/serial/serial_base_bus.c | 10 ++++++----
>   1 file changed, 6 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/tty/serial/serial_base_bus.c b/drivers/tty/serial/serial_base_bus.c
> index 22749ab0428a..8e891984cdc0 100644
> --- a/drivers/tty/serial/serial_base_bus.c
> +++ b/drivers/tty/serial/serial_base_bus.c
> @@ -13,7 +13,7 @@
>   #include <linux/device.h>
>   #include <linux/idr.h>
>   #include <linux/module.h>
> -#include <linux/of.h>
> +#include <linux/property.h>
>   #include <linux/serial_core.h>
>   #include <linux/slab.h>
>   #include <linux/spinlock.h>
> @@ -60,6 +60,7 @@ void serial_base_driver_unregister(struct device_driver *driver)
>   	driver_unregister(driver);
>   }
>   
> +/* On failure the caller must put device @dev with put_device() */
>   static int serial_base_device_init(struct uart_port *port,
>   				   struct device *dev,
>   				   struct device *parent_dev,
> @@ -73,7 +74,8 @@ static int serial_base_device_init(struct uart_port *port,
>   	dev->parent = parent_dev;
>   	dev->bus = &serial_base_bus_type;
>   	dev->release = release;
> -	device_set_of_node_from_dev(dev, parent_dev);
> +
> +	device_set_node(dev, fwnode_handle_get(dev_fwnode(parent_dev)));

device_set_of_node_from_dev() used to set dev->of_node_reused = true;
which made the pinctrl subsystem skip the logic inside
pinctrl_bind_pins().

Since you changed to device_set_node(), dev->of_node_reused is no longer
set to true and the following errors are issued because pinctrl will try
to request the same pins for all sub-devices.

> [    0.472463] pinctrl-rzt2h 802c0000.pinctrl: pin P27_4 already requested by 80005000.serial; cannot claim for 80005000.serial:0
> [    0.472485] pinctrl-rzt2h 802c0000.pinctrl: error -EINVAL: pin-220 (80005000.serial:0)
> [    0.472501] pinctrl-rzt2h 802c0000.pinctrl: error -EINVAL: could not request pin 220 (P27_4) from group sci0-pins on device pinctrl-rzt2h
> [    0.472515] ctrl 80005000.serial:0: Error applying setting, reverse things back
> [    0.472711] pinctrl-rzt2h 802c0000.pinctrl: pin P27_4 already requested by 80005000.serial; cannot claim for 80005000.serial:0.0
> [    0.472725] pinctrl-rzt2h 802c0000.pinctrl: error -EINVAL: pin-220 (80005000.serial:0.0)
> [    0.472738] pinctrl-rzt2h 802c0000.pinctrl: error -EINVAL: could not request pin 220 (P27_4) from group sci0-pins on device pinctrl-rzt2h
> [    0.472752] port 80005000.serial:0.0: Error applying setting, reverse things back

Maybe we should set

dev->of_node_reused = true;

manually right above the call to device_set_node(), grouped with the
rest of dev initialization?

>   
>   	if (!serial_base_initialized) {
>   		dev_dbg(port->dev, "uart_add_one_port() called before arch_initcall()?\n");
> @@ -94,7 +96,7 @@ static void serial_base_ctrl_release(struct device *dev)
>   {
>   	struct serial_ctrl_device *ctrl_dev = to_serial_base_ctrl_device(dev);
>   
> -	of_node_put(dev->of_node);
> +	fwnode_handle_put(dev_fwnode(dev));
>   	kfree(ctrl_dev);
>   }
>   
> @@ -142,7 +144,7 @@ static void serial_base_port_release(struct device *dev)
>   {
>   	struct serial_port_device *port_dev = to_serial_base_port_device(dev);
>   
> -	of_node_put(dev->of_node);
> +	fwnode_handle_put(dev_fwnode(dev));
>   	kfree(port_dev);
>   }
>   

Re: [PATCH v1 1/1] serial: core: Restore sysfs fwnode information

[Marek Szyprowski]

On 27.11.2025 17:36, Andy Shevchenko wrote:
> The change that restores sysfs fwnode information does it only for OF cases.
> Update the fix to cover all possible types of fwnodes.
>
> Fixes: d36f0e9a0002 ("serial: core: restore of_node information in sysfs")
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---

This patch landed in today's linux-next as commit 24ec03cc5512 ("serial: 
core: Restore sysfs fwnode information"). In my tests I found that it 
breaks booting of most of my test boards (ARM 32 and 64 bit). 
Unfortunately I cannot provide anything useful besides the information 
that booting stops and system doesn't reach shell. There is nothing 
suspicious in the kernel logs. I suspect a memory trashing. Reverting 
$subject on top of linux-next fixes booting.

>   drivers/tty/serial/serial_base_bus.c | 10 ++++++----
>   1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/tty/serial/serial_base_bus.c b/drivers/tty/serial/serial_base_bus.c
> index 22749ab0428a..8e891984cdc0 100644
> --- a/drivers/tty/serial/serial_base_bus.c
> +++ b/drivers/tty/serial/serial_base_bus.c
> @@ -13,7 +13,7 @@
>   #include <linux/device.h>
>   #include <linux/idr.h>
>   #include <linux/module.h>
> -#include <linux/of.h>
> +#include <linux/property.h>
>   #include <linux/serial_core.h>
>   #include <linux/slab.h>
>   #include <linux/spinlock.h>
> @@ -60,6 +60,7 @@ void serial_base_driver_unregister(struct device_driver *driver)
>   	driver_unregister(driver);
>   }
>   
> +/* On failure the caller must put device @dev with put_device() */
>   static int serial_base_device_init(struct uart_port *port,
>   				   struct device *dev,
>   				   struct device *parent_dev,
> @@ -73,7 +74,8 @@ static int serial_base_device_init(struct uart_port *port,
>   	dev->parent = parent_dev;
>   	dev->bus = &serial_base_bus_type;
>   	dev->release = release;
> -	device_set_of_node_from_dev(dev, parent_dev);
> +
> +	device_set_node(dev, fwnode_handle_get(dev_fwnode(parent_dev)));
>   
>   	if (!serial_base_initialized) {
>   		dev_dbg(port->dev, "uart_add_one_port() called before arch_initcall()?\n");
> @@ -94,7 +96,7 @@ static void serial_base_ctrl_release(struct device *dev)
>   {
>   	struct serial_ctrl_device *ctrl_dev = to_serial_base_ctrl_device(dev);
>   
> -	of_node_put(dev->of_node);
> +	fwnode_handle_put(dev_fwnode(dev));
>   	kfree(ctrl_dev);
>   }
>   
> @@ -142,7 +144,7 @@ static void serial_base_port_release(struct device *dev)
>   {
>   	struct serial_port_device *port_dev = to_serial_base_port_device(dev);
>   
> -	of_node_put(dev->of_node);
> +	fwnode_handle_put(dev_fwnode(dev));
>   	kfree(port_dev);
>   }
>   

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland

Re: [PATCH v1 1/1] serial: core: Restore sysfs fwnode information

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 858 bytes --]

On Thu, Dec 18, 2025 at 04:29:55PM +0100, Marek Szyprowski wrote:
> On 27.11.2025 17:36, Andy Shevchenko wrote:
> > The change that restores sysfs fwnode information does it only for OF cases.
> > Update the fix to cover all possible types of fwnodes.

> This patch landed in today's linux-next as commit 24ec03cc5512 ("serial: 
> core: Restore sysfs fwnode information"). In my tests I found that it 
> breaks booting of most of my test boards (ARM 32 and 64 bit). 
> Unfortunately I cannot provide anything useful besides the information 
> that booting stops and system doesn't reach shell. There is nothing 
> suspicious in the kernel logs. I suspect a memory trashing. Reverting 
> $subject on top of linux-next fixes booting.

I'm also seeing this in my lab and Arm's lab, there are a few systems
that survive but it's a small minority.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH v1 1/1] serial: core: Restore sysfs fwnode information

[Marek Szyprowski]

On 18.12.2025 17:54, Mark Brown wrote:
> On Thu, Dec 18, 2025 at 04:29:55PM +0100, Marek Szyprowski wrote:
>> On 27.11.2025 17:36, Andy Shevchenko wrote:
>>> The change that restores sysfs fwnode information does it only for OF cases.
>>> Update the fix to cover all possible types of fwnodes.
>> This patch landed in today's linux-next as commit 24ec03cc5512 ("serial:
>> core: Restore sysfs fwnode information"). In my tests I found that it
>> breaks booting of most of my test boards (ARM 32 and 64 bit).
>> Unfortunately I cannot provide anything useful besides the information
>> that booting stops and system doesn't reach shell. There is nothing
>> suspicious in the kernel logs. I suspect a memory trashing. Reverting
>> $subject on top of linux-next fixes booting.
> I'm also seeing this in my lab and Arm's lab, there are a few systems
> that survive but it's a small minority.

I have a few spare minutes and spent them analyzing this issue.

This is somehow related to dev->of_node_reused device property and its 
check in pinctrl_bind_pins() in drivers/base/pinctrl.c.

The following hack/workaround fixes the observed boot issues:

diff --git a/drivers/tty/serial/serial_base_bus.c 
b/drivers/tty/serial/serial_base_bus.c
index 8e891984cdc0..f3332a5e134c 100644
--- a/drivers/tty/serial/serial_base_bus.c
+++ b/drivers/tty/serial/serial_base_bus.c
@@ -76,6 +76,7 @@ static int serial_base_device_init(struct uart_port *port,
         dev->release = release;

         device_set_node(dev, fwnode_handle_get(dev_fwnode(parent_dev)));
+       dev->of_node_reused = true;

         if (!serial_base_initialized) {
                 dev_dbg(port->dev, "uart_add_one_port() called before 
arch_initcall()?\n");


If I then remove the dev->of_node_reused check in pinctrl_bind_pins(), 
the affected boards don't boot again the same way. I hope this helps 
fixing this issue.

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland

Re: [PATCH v1 1/1] serial: core: Restore sysfs fwnode information

[Dmitry Baryshkov]

On Mon, Dec 22, 2025 at 11:55:26AM +0100, Marek Szyprowski wrote:
> On 18.12.2025 17:54, Mark Brown wrote:
> > On Thu, Dec 18, 2025 at 04:29:55PM +0100, Marek Szyprowski wrote:
> >> On 27.11.2025 17:36, Andy Shevchenko wrote:
> >>> The change that restores sysfs fwnode information does it only for OF cases.
> >>> Update the fix to cover all possible types of fwnodes.
> >> This patch landed in today's linux-next as commit 24ec03cc5512 ("serial:
> >> core: Restore sysfs fwnode information"). In my tests I found that it
> >> breaks booting of most of my test boards (ARM 32 and 64 bit).
> >> Unfortunately I cannot provide anything useful besides the information
> >> that booting stops and system doesn't reach shell. There is nothing
> >> suspicious in the kernel logs. I suspect a memory trashing. Reverting
> >> $subject on top of linux-next fixes booting.
> > I'm also seeing this in my lab and Arm's lab, there are a few systems
> > that survive but it's a small minority.
> 
> I have a few spare minutes and spent them analyzing this issue.
> 
> This is somehow related to dev->of_node_reused device property and its 
> check in pinctrl_bind_pins() in drivers/base/pinctrl.c.
> 
> The following hack/workaround fixes the observed boot issues:
> 
> diff --git a/drivers/tty/serial/serial_base_bus.c 
> b/drivers/tty/serial/serial_base_bus.c
> index 8e891984cdc0..f3332a5e134c 100644
> --- a/drivers/tty/serial/serial_base_bus.c
> +++ b/drivers/tty/serial/serial_base_bus.c
> @@ -76,6 +76,7 @@ static int serial_base_device_init(struct uart_port *port,
>          dev->release = release;
> 
>          device_set_node(dev, fwnode_handle_get(dev_fwnode(parent_dev)));
> +       dev->of_node_reused = true;
> 
>          if (!serial_base_initialized) {
>                  dev_dbg(port->dev, "uart_add_one_port() called before 
> arch_initcall()?\n");
> 
> 
> If I then remove the dev->of_node_reused check in pinctrl_bind_pins(), 
> the affected boards don't boot again the same way. I hope this helps 
> fixing this issue.

I can confirm that this patch fixes the breakage on Qualcomm devices.
Please send it as a proper patch (though I'd say, the flag should only
be set only if dev->of_node != NULL).

An alternative is to introduce dev_set_node_from_dev() which mimics both
dev_set_node() and device_set_of_node_from_dev().

Greg, what would be your preference?

> 
> Best regards
> -- 
> Marek Szyprowski, PhD
> Samsung R&D Institute Poland
> 

-- 
With best wishes
Dmitry

Re: [PATCH v1 1/1] serial: core: Restore sysfs fwnode information

[Andy Shevchenko]

On Mon, Dec 29, 2025 at 05:18:14AM +0200, Dmitry Baryshkov wrote:
> On Mon, Dec 22, 2025 at 11:55:26AM +0100, Marek Szyprowski wrote:

...

> I can confirm that this patch fixes the breakage on Qualcomm devices.
> Please send it as a proper patch (though I'd say, the flag should only
> be set only if dev->of_node != NULL).
> 
> An alternative is to introduce dev_set_node_from_dev() which mimics both
> dev_set_node() and device_set_of_node_from_dev().

While I am on vacation, I have a brief look at the report and fix and the
discussion. There are several ongoing tasks that may collide with this:

- the desire to split device.h, so we have something like device/fwnode.h with
all related APIs listed there

- taking these, it might be good to have device_set_used_node() as a combination
of the above and also device_is_node_used() to be used in its solely user of pin
control core

- the proper understanding of this flag to begin with and how it may be related
to ACPI-based platforms (or non-DT in general)

- with that done, it might be a room to see if the flag actually should be
somewhere in fwnode structure or elsewhere

- or we can revisit the whole design and see if we can get rid of the flag

+Cc: Linus for searching for his comments, thoughts, etc.

-- 
With Best Regards,
Andy Shevchenko