From: Kory Maincent (Dent Project) <kory.maincent@bootlin.com>

Add support for a PSE device index to report the PSE controller index to
the user through ethtool. This will be useful for future support of power
domains and port priority management.

Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
---

changes in v2:
- new patch.
---
 drivers/net/pse-pd/pse_core.c | 23 ++++++++++++++++++-----
 include/linux/pse-pd/pse.h    |  4 ++++
 2 files changed, 22 insertions(+), 5 deletions(-)

diff --git a/drivers/net/pse-pd/pse_core.c b/drivers/net/pse-pd/pse_core.c
index 2906ce173f66..68297428f6b5 100644
--- a/drivers/net/pse-pd/pse_core.c
+++ b/drivers/net/pse-pd/pse_core.c
@@ -13,6 +13,7 @@
 
 static DEFINE_MUTEX(pse_list_mutex);
 static LIST_HEAD(pse_controller_list);
+static DEFINE_IDA(pse_ida);
 
 /**
  * struct pse_control - a PSE control
@@ -440,18 +441,22 @@ int pse_controller_register(struct pse_controller_dev *pcdev)
 
 	mutex_init(&pcdev->lock);
 	INIT_LIST_HEAD(&pcdev->pse_control_head);
+	ret = ida_alloc_max(&pse_ida, INT_MAX, GFP_KERNEL);
+	if (ret < 0)
+		return ret;
+	pcdev->id = ret;
 
 	if (!pcdev->nr_lines)
 		pcdev->nr_lines = 1;
 
 	ret = of_load_pse_pis(pcdev);
 	if (ret)
-		return ret;
+		goto free_pse_ida;
 
 	if (pcdev->ops->setup_pi_matrix) {
 		ret = pcdev->ops->setup_pi_matrix(pcdev);
 		if (ret)
-			return ret;
+			goto free_pse_ida;
 	}
 
 	/* Each regulator name len is pcdev dev name + 7 char +
@@ -468,15 +473,17 @@ int pse_controller_register(struct pse_controller_dev *pcdev)
 			continue;
 
 		reg_name = devm_kzalloc(pcdev->dev, reg_name_len, GFP_KERNEL);
-		if (!reg_name)
-			return -ENOMEM;
+		if (!reg_name) {
+			ret = -ENOMEM;
+			goto free_pse_ida;
+		}
 
 		snprintf(reg_name, reg_name_len, "pse-%s_pi%d",
 			 dev_name(pcdev->dev), i);
 
 		ret = devm_pse_pi_regulator_register(pcdev, reg_name, i);
 		if (ret)
-			return ret;
+			goto free_pse_ida;
 	}
 
 	mutex_lock(&pse_list_mutex);
@@ -484,6 +491,10 @@ int pse_controller_register(struct pse_controller_dev *pcdev)
 	mutex_unlock(&pse_list_mutex);
 
 	return 0;
+
+free_pse_ida:
+	ida_free(&pse_ida, pcdev->id);
+	return ret;
 }
 EXPORT_SYMBOL_GPL(pse_controller_register);
 
@@ -494,6 +505,7 @@ EXPORT_SYMBOL_GPL(pse_controller_register);
 void pse_controller_unregister(struct pse_controller_dev *pcdev)
 {
 	pse_release_pis(pcdev);
+	ida_free(&pse_ida, pcdev->id);
 	mutex_lock(&pse_list_mutex);
 	list_del(&pcdev->list);
 	mutex_unlock(&pse_list_mutex);
@@ -750,6 +762,7 @@ static int _pse_ethtool_get_status(struct pse_controller_dev *pcdev,
 		return -EOPNOTSUPP;
 	}
 
+	status->pse_id = pcdev->id;
 	return ops->ethtool_get_status(pcdev, id, extack, status);
 }
 
diff --git a/include/linux/pse-pd/pse.h b/include/linux/pse-pd/pse.h
index 85a08c349256..5312488cb3cf 100644
--- a/include/linux/pse-pd/pse.h
+++ b/include/linux/pse-pd/pse.h
@@ -31,6 +31,7 @@ struct pse_control_config {
 /**
  * struct pse_control_status - PSE control/channel status.
  *
+ * @pse_id: index number of the PSE. Set by PSE core.
  * @podl_admin_state: operational state of the PoDL PSE
  *	functions. IEEE 802.3-2018 30.15.1.1.2 aPoDLPSEAdminState
  * @podl_pw_status: power detection status of the PoDL PSE.
@@ -52,6 +53,7 @@ struct pse_control_config {
  *	ranges
  */
 struct pse_control_status {
+	u32 pse_id;
 	enum ethtool_podl_pse_admin_state podl_admin_state;
 	enum ethtool_podl_pse_pw_d_status podl_pw_status;
 	enum ethtool_c33_pse_admin_state c33_admin_state;
@@ -150,6 +152,7 @@ struct pse_pi {
  * @types: types of the PSE controller
  * @pi: table of PSE PIs described in this controller device
  * @no_of_pse_pi: flag set if the pse_pis devicetree node is not used
+ * @id: Index of the PSE
  */
 struct pse_controller_dev {
 	const struct pse_controller_ops *ops;
@@ -163,6 +166,7 @@ struct pse_controller_dev {
 	enum ethtool_pse_types types;
 	struct pse_pi *pi;
 	bool no_of_pse_pi;
+	int id;
 };
 
 #if IS_ENABLED(CONFIG_PSE_CONTROLLER)

-- 
2.34.1

On Wed, Oct 30, 2024 at 05:53:07PM +0100, Kory Maincent wrote:

...
>  /**
>   * struct pse_control - a PSE control
> @@ -440,18 +441,22 @@ int pse_controller_register(struct pse_controller_dev *pcdev)
>  
>  	mutex_init(&pcdev->lock);
>  	INIT_LIST_HEAD(&pcdev->pse_control_head);
> +	ret = ida_alloc_max(&pse_ida, INT_MAX, GFP_KERNEL);

s/INT_MAX/U32_MAX

...
>  struct pse_control_status {
> +	u32 pse_id;
>  	enum ethtool_podl_pse_admin_state podl_admin_state;
...
>  struct pse_controller_dev {
>  	const struct pse_controller_ops *ops;
> @@ -163,6 +166,7 @@ struct pse_controller_dev {
>  	enum ethtool_pse_types types;
>  	struct pse_pi *pi;
>  	bool no_of_pse_pi;
> +	int id;

        u32 id

It would be better to have one type for all variables.

-- 
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 |

On Thu, Oct 31, 2024 at 07:27:59AM +0100, Oleksij Rempel wrote:
> On Wed, Oct 30, 2024 at 05:53:07PM +0100, Kory Maincent wrote:
> 
> ...
> >  /**
> >   * struct pse_control - a PSE control
> > @@ -440,18 +441,22 @@ int pse_controller_register(struct pse_controller_dev *pcdev)
> >  
> >  	mutex_init(&pcdev->lock);
> >  	INIT_LIST_HEAD(&pcdev->pse_control_head);
> > +	ret = ida_alloc_max(&pse_ida, INT_MAX, GFP_KERNEL);
> 
> s/INT_MAX/U32_MAX

 * Return: The allocated ID, or %-ENOMEM if memory could not be allocated,
 * or %-ENOSPC if there are no free IDs.

static inline int ida_alloc_max(struct ida *ida, unsigned int max, gfp_t gfp)

We need to be careful here, at least theoretically. Assuming a 32 bit
system, and you pass it U32_MAX, how does it return values in the
range S32_MAX..U32_MAX when it also needs to be able to return
negative numbers as errors?

I think the correct value to pass is S32_MAX, because it will always
fit in a u32, and there is space left for negative values for errors.

But this is probably theoretical, no real system should have that many
controllers.

	Andrew

On Thu, 31 Oct 2024 22:28:29 +0100
Andrew Lunn <andrew@lunn.ch> wrote:

> On Thu, Oct 31, 2024 at 07:27:59AM +0100, Oleksij Rempel wrote:
> > On Wed, Oct 30, 2024 at 05:53:07PM +0100, Kory Maincent wrote:
> > 
> > ...  
> > >  /**
> > >   * struct pse_control - a PSE control
> > > @@ -440,18 +441,22 @@ int pse_controller_register(struct
> > > pse_controller_dev *pcdev) 
> > >  	mutex_init(&pcdev->lock);
> > >  	INIT_LIST_HEAD(&pcdev->pse_control_head);
> > > +	ret = ida_alloc_max(&pse_ida, INT_MAX, GFP_KERNEL);  
> > 
> > s/INT_MAX/U32_MAX  
> 
>  * Return: The allocated ID, or %-ENOMEM if memory could not be allocated,
>  * or %-ENOSPC if there are no free IDs.
> 
> static inline int ida_alloc_max(struct ida *ida, unsigned int max, gfp_t gfp)
> 
> We need to be careful here, at least theoretically. Assuming a 32 bit
> system, and you pass it U32_MAX, how does it return values in the
> range S32_MAX..U32_MAX when it also needs to be able to return
> negative numbers as errors?
> 
> I think the correct value to pass is S32_MAX, because it will always
> fit in a u32, and there is space left for negative values for errors.
> 
> But this is probably theoretical, no real system should have that many
> controllers.

Indeed you are right we might have issue between S32_MAX and U32_MAX if we want
to return errors.
Small question, is S32_MAX better than INT_MAX? Is there a point to limit it to
32 bits?

Regards,
-- 
Köry Maincent, Bootlin
Embedded Linux and kernel engineering
https://bootlin.com