Re: [PATCH 2/2] regulator: userspace-consumer: Multiple regulators

From: Zev Weiss
Date: Thu Apr 20 2023 - 07:00:44 EST


On Thu, Apr 20, 2023 at 01:46:14AM PDT, Naresh Solanki wrote:
Hi Zev,

On 20-04-2023 05:32 am, Zev Weiss wrote:
On Tue, Apr 18, 2023 at 07:50:51AM PDT, Naresh Solanki wrote:
Use property regulator-supplies to determine all regulator
supplies.
This is useful in case of a connector having 2 or more supplies.
Example: PCIe connector on mainboard can be powered by 12V & 3.3V
suplies.

Signed-off-by: Naresh Solanki <Naresh.Solanki@xxxxxxxxxxxxx>
---
drivers/regulator/userspace-consumer.c | 19 +++++++++++++++----
1 file changed, 15 insertions(+), 4 deletions(-)

diff --git a/drivers/regulator/userspace-consumer.c b/drivers/regulator/userspace-consumer.c
index 97f075ed68c9..0bb49547b926 100644
--- a/drivers/regulator/userspace-consumer.c
+++ b/drivers/regulator/userspace-consumer.c
@@ -120,7 +120,10 @@ static int regulator_userspace_consumer_probe(struct platform_device *pdev)
    struct regulator_userspace_consumer_data tmpdata;
    struct regulator_userspace_consumer_data *pdata;
    struct userspace_consumer_data *drvdata;
-    int ret;
+    struct device_node *np = pdev->dev.of_node;
+    struct property *prop;
+    const char *supply;
+    int ret, count = 0;

    pdata = dev_get_platdata(&pdev->dev);
    if (!pdata) {
@@ -131,11 +134,19 @@ static int regulator_userspace_consumer_probe(struct platform_device *pdev)
        memset(pdata, 0, sizeof(*pdata));

        pdata->no_autoswitch = true;
-        pdata->num_supplies = 1;
-        pdata->supplies = devm_kzalloc(&pdev->dev, sizeof(*pdata->supplies), GFP_KERNEL);
+        pdata->num_supplies = of_property_count_strings(np, "regulator-supplies");
+        if (pdata->num_supplies < 0) {
+            dev_err(&pdev->dev, "could not parse property regulator-supplies");
+            return -EINVAL;
+        }
+        pdata->supplies = devm_kzalloc(&pdev->dev,
+                           sizeof(*pdata->supplies) * pdata->num_supplies,
+                           GFP_KERNEL);

AFAICT this doesn't appear to implement the "vout" default specified in the dt-binding patch?
The "regulator-supplies" property will hold the default value of "vout" unless specified otherwise. As a result, the string enumeration retrieves the value of "vout" by default, and the "vout-supply" property is utilized for the regulator.


With the disclaimer that I'm not a DT expert, that's not my understanding of how DT works. I don't think the 'default' value specified in the binding forces the fdt to always include that value if it's not present in the dts (since I'm pretty sure dtc doesn't even look at the binding to know that a default exists when compiling the dts); rather, it's information meant to be used by the software implementing support for that device (e.g. a driver for it) about what value to assume if the property isn't present in the fdt.


Also, since the core of the userspace-consumer driver itself already supports multiple regulators, it might be nice for the subject line to mention DT supplies or something a bit more specifically.
Sure. How about 'Support multiple supplies' ?

I meant that it should explicitly mention "DT" (or perhaps "OF"). The driver's structure has supported multiple supplies since it was first introduced in 2009, so "Support multiple supplies" sounds like this commit is adding functionality that was already there. What this patch is doing is connecting that existing support to the OF support logic so that it can be used in a device-tree context.


        if (!pdata->supplies)
            return -ENOMEM;
-        pdata->supplies[0].supply = "vout";
+
+        of_property_for_each_string(np, "regulator-supplies", prop, supply)
+            pdata->supplies[count++].supply = supply;
    }

    if (pdata->num_supplies < 1) {
--
2.39.1


Regards,
Naresh