| Message ID | 1365135708-23886-2-git-send-email-nm@ti.com (mailing list archive) |
|---|---|
| State | Changes Requested, archived |
| Headers | show |
On Thu, Apr 04, 2013 at 11:21:47PM -0500, Nishanth Menon wrote: > commit 6d191a5fc7a969d972f1681e1c23781aecb06a61 > (regulator: core: Don't defer probe if there's no DT binding for a supply) > > Attempted to differentiate between regulator_get() with an actual > DT binding for the supply and when there is none to avoid unnecessary > deferal. So, this is an extremely long and hence difficult to understand and follow commit message which manages to miss out mentioning the core issue which is that we're ignoring the return value from lookup_dev(). I had to actually look at the code to understand. What should be being said here is that the ret value supplied by regulator_dev_lookup() is being ignored by _regulator_get(). > mutex_unlock(®ulator_list_mutex); > - return regulator; > + return ret ? ERR_PTR(ret) : regulator; Please implement this so it looks like the rest of the function - everywhere else in the function we just make regulator an ERR_PTR() and goto out, we should do the same.
diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c index e3661c2..139d86a 100644 --- a/drivers/regulator/core.c +++ b/drivers/regulator/core.c @@ -1229,7 +1229,7 @@ static struct regulator *_regulator_get(struct device *dev, const char *id, struct regulator_dev *rdev; struct regulator *regulator = ERR_PTR(-EPROBE_DEFER); const char *devname = NULL; - int ret; + int ret = 0; if (id == NULL) { pr_err("get() with no identifier\n"); @@ -1266,7 +1266,7 @@ static struct regulator *_regulator_get(struct device *dev, const char *id, #endif mutex_unlock(®ulator_list_mutex); - return regulator; + return ret ? ERR_PTR(ret) : regulator; found: if (rdev->exclusive) {
commit 6d191a5fc7a969d972f1681e1c23781aecb06a61 (regulator: core: Don't defer probe if there's no DT binding for a supply) Attempted to differentiate between regulator_get() with an actual DT binding for the supply and when there is none to avoid unnecessary deferal. In cases where a driver, such as cpufreq driver, need to handle (based on platform) DT nodes which may or may not have regulator supply binding, it is necessary to differentiate the return value. When there is regulator supply node, defering the probe is beneficial as sequence of dependent regulator probe may not be predictable. However, where is no regulator supply binding, it can operate without the same as needed. So, to provide an appropriate return result, when we cannot find an regulator in regulator_dev_lookup and we run out of options, use the regulator_dev_lookup result. This helps the caller to help make informed decision. Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Mark Brown <broonie@opensource.wolfsonmicro.com> Signed-off-by: Nishanth Menon <nm@ti.com> --- drivers/regulator/core.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)