regulator: return set_mode is same mode is requested

On Mon, 2010-05-17 at 17:34 +0200, Mark Brown wrote:
> This doesn't seem like the right error handling - if the driver has a
> set_mode() you'd *expect* it to have a get_mode() but there's no need
> for it to be a strict requirement.
True. In such a case, even a valid request would be lost! So now
in the updated patch:
 - check if get_mode is present to avoid oops;
 - if get_mode is not present, proceed anyways for the request.

Here is the updated patch:

>From bad0d5eb51ef84be5b100e3dd0f5a590ea0529b6 Mon Sep 17 00:00:00 2001
From: Sundar R Iyer <[email protected]>
Date: Fri, 14 May 2010 15:14:17 +0530
Subject: [PATCH 1/1] regulator: return set_mode when same mode is requested

save I/O costs by returning when the same mode is
requested for the regulator

Cc: Liam Girdwood <[email protected]>
Cc: Mark Brown <[email protected]>
Acked-by: Linus Walleij <[email protected]>
Signed-off-by: Sundar R Iyer <[email protected]>
Acked-by: Mark Brown <[email protected]>
Signed-off-by: Liam Girdwood <[email protected]>

Author: Sundar R Iyer

drivers/regulator/core.c

@@ -1745,6 +1745,7 @@ int regulator_set_mode(struct regulator *regulator, unsigned int mode)
 {
 	struct regulator_dev *rdev = regulator->rdev;
 	int ret;
+	int regulator_curr_mode;
 
 	mutex_lock(&rdev->mutex);
 
@@ -1754,6 +1755,15 @@ int regulator_set_mode(struct regulator *regulator, unsigned int mode)
 		goto out;
 	}
 
+	/* return if the same mode is requested */
+	if (rdev->desc->ops->get_mode) {
+		regulator_curr_mode = rdev->desc->ops->get_mode(rdev);
+		if (regulator_curr_mode == mode) {
+			ret = 0;
+			goto out;
+		}
+	}
+
 	/* constraints check */
 	ret = regulator_check_mode(rdev, mode);
 	if (ret < 0)