[PATCH v2 03/16] ACPI: property: Rework acpi_graph_get_next_endpoint()

Sakari Ailus posted 16 patches 1 week ago
[PATCH v2 03/16] ACPI: property: Rework acpi_graph_get_next_endpoint()
Posted by Sakari Ailus 1 week ago
Rework the code obtaining the next endpoint in
acpi_graph_get_next_endpoint(). The resulting code removes unnecessary
contitionals and should be easier to follow.

Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
---
 drivers/acpi/property.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/acpi/property.c b/drivers/acpi/property.c
index 3e85900080ac..5438592dc136 100644
--- a/drivers/acpi/property.c
+++ b/drivers/acpi/property.c
@@ -1399,14 +1399,15 @@ static struct fwnode_handle *acpi_graph_get_next_endpoint(
 	if (!port)
 		return NULL;
 
-	endpoint = acpi_get_next_subnode(port, prev);
-	while (!endpoint) {
-		port = acpi_get_next_subnode(fwnode, port);
-		if (!port)
+	do {
+		endpoint = acpi_get_next_subnode(port, prev);
+		if (endpoint)
 			break;
+
+		port = acpi_get_next_subnode(fwnode, port);
 		if (is_acpi_graph_node(port, "port"))
-			endpoint = acpi_get_next_subnode(port, NULL);
-	}
+			prev = NULL;
+	} while (port);
 
 	/*
 	 * The names of the endpoint nodes begin with "endpoint@" followed by
-- 
2.47.3
Re: [PATCH v2 03/16] ACPI: property: Rework acpi_graph_get_next_endpoint()
Posted by Laurent Pinchart 1 week ago
Hi Sakari,

On Wed, Sep 24, 2025 at 10:45:49AM +0300, Sakari Ailus wrote:
> Rework the code obtaining the next endpoint in
> acpi_graph_get_next_endpoint(). The resulting code removes unnecessary
> contitionals and should be easier to follow.
> 
> Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
> ---
>  drivers/acpi/property.c | 13 +++++++------
>  1 file changed, 7 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/acpi/property.c b/drivers/acpi/property.c
> index 3e85900080ac..5438592dc136 100644
> --- a/drivers/acpi/property.c
> +++ b/drivers/acpi/property.c
> @@ -1399,14 +1399,15 @@ static struct fwnode_handle *acpi_graph_get_next_endpoint(
>  	if (!port)
>  		return NULL;
>  
> -	endpoint = acpi_get_next_subnode(port, prev);
> -	while (!endpoint) {
> -		port = acpi_get_next_subnode(fwnode, port);
> -		if (!port)
> +	do {
> +		endpoint = acpi_get_next_subnode(port, prev);
> +		if (endpoint)
>  			break;
> +
> +		port = acpi_get_next_subnode(fwnode, port);
>  		if (is_acpi_graph_node(port, "port"))
> -			endpoint = acpi_get_next_subnode(port, NULL);
> -	}
> +			prev = NULL;

Isn't there an issue here ? If the next subnode of fwnode is not a port,
the next iteration of the do loop will attempt to get an endpoint from
that non-port node. Maybe the acpi_get_next_subnode() that will try to
get the endpoint from the non-port port will return NULL because prev
won't be a child of port, but that seems fragile.

I think the following would be easier to understand:

	do {
		endpoint = acpi_get_next_subnode(port, prev);
		if (endpoint)
			break;

		prev = NULL;

		do {
			port = acpi_get_next_subnode(fwnode, port);
		} while (port && !is_acpi_graph_node(port, "port"));
	} while (port);

> +	} while (port);
>  
>  	/*
>  	 * The names of the endpoint nodes begin with "endpoint@" followed by
> 

-- 
Regards,

Laurent Pinchart
Re: [PATCH v2 03/16] ACPI: property: Rework acpi_graph_get_next_endpoint()
Posted by Sakari Ailus 5 days, 14 hours ago
Hi Laurent,

Thank you for the review.

On Wed, Sep 24, 2025 at 12:39:34PM +0300, Laurent Pinchart wrote:
> Hi Sakari,
> 
> On Wed, Sep 24, 2025 at 10:45:49AM +0300, Sakari Ailus wrote:
> > Rework the code obtaining the next endpoint in
> > acpi_graph_get_next_endpoint(). The resulting code removes unnecessary
> > contitionals and should be easier to follow.
> > 
> > Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> > Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
> > ---
> >  drivers/acpi/property.c | 13 +++++++------
> >  1 file changed, 7 insertions(+), 6 deletions(-)
> > 
> > diff --git a/drivers/acpi/property.c b/drivers/acpi/property.c
> > index 3e85900080ac..5438592dc136 100644
> > --- a/drivers/acpi/property.c
> > +++ b/drivers/acpi/property.c
> > @@ -1399,14 +1399,15 @@ static struct fwnode_handle *acpi_graph_get_next_endpoint(
> >  	if (!port)
> >  		return NULL;
> >  
> > -	endpoint = acpi_get_next_subnode(port, prev);
> > -	while (!endpoint) {
> > -		port = acpi_get_next_subnode(fwnode, port);
> > -		if (!port)
> > +	do {
> > +		endpoint = acpi_get_next_subnode(port, prev);
> > +		if (endpoint)
> >  			break;
> > +
> > +		port = acpi_get_next_subnode(fwnode, port);
> >  		if (is_acpi_graph_node(port, "port"))
> > -			endpoint = acpi_get_next_subnode(port, NULL);
> > -	}
> > +			prev = NULL;
> 
> Isn't there an issue here ? If the next subnode of fwnode is not a port,
> the next iteration of the do loop will attempt to get an endpoint from
> that non-port node. Maybe the acpi_get_next_subnode() that will try to
> get the endpoint from the non-port port will return NULL because prev
> won't be a child of port, but that seems fragile.
> 
> I think the following would be easier to understand:
> 
> 	do {
> 		endpoint = acpi_get_next_subnode(port, prev);
> 		if (endpoint)
> 			break;
> 
> 		prev = NULL;
> 
> 		do {
> 			port = acpi_get_next_subnode(fwnode, port);
> 		} while (port && !is_acpi_graph_node(port, "port"));
> 	} while (port);

Yes, this indeed ensures port will be a port node. I'll use this in the
next version.

> 
> > +	} while (port);
> >  
> >  	/*
> >  	 * The names of the endpoint nodes begin with "endpoint@" followed by
> > 
> 

-- 
Regards,

Sakari Ailus