On Monday 22 January 2018 04:23 PM, Ladislav Michl wrote:

Dear Claudiu,

On Mon, Jan 22, 2018 at 11:17:08AM +0200, Claudiu Beznea wrote:

quoted

On 17.01.2018 23:47, Ladislav Michl wrote:

quoted

@@ -334,18 +348,18 @@ static int pwm_omap_dmtimer_probe(struct platform_device *pdev)
 
 	mutex_init(&omap->mutex);
 
-	status = pwmchip_add(&omap->chip);
-	if (status < 0) {
+	ret = pwmchip_add(&omap->chip);
+	if (ret < 0) {
 		dev_err(&pdev->dev, "failed to register PWM\n");
-		omap->pdata->free(omap->dm_timer);
-		ret = status;
-		goto put;
+		goto free;
 	}
 
 	platform_set_drvdata(pdev, omap);

From documentation: "of_parse_phandle(): Returns the device_node pointer with refcount

incremented. Use of_node_put() on it when done."
In case of success the of_node_put() should also be called as I see.

Based on you previous suggestions found here:
https://patchwork.kernel.org/patch/10140209/
I'd say this fix belongs to patch which introduces of_node_put() in the error
path. I'll then rebase this patches on top of the fix.

Agreed. I missed the success path as i assumed only error paths needed
te of_node_put(). I will post v8 of this patch alone as other patches
can be left untouched. Hope that is okay.

Regards,
Keerthy

Thank you,
	ladis