[PATCH] drivers/rtc/interface.c: check the validation of rtc_time in __rtc_read_time

[PATCH] drivers/rtc/interface.c: check the validation of rtc_time in __rtc_read_time

[Hyogi Gim]

Some of rtc devices always return '0' when rtc_class_ops.read_time is called.
So if rtc_time isn't verified in callback, rtc interface cannot know whether
rtc_time is valid.

Check rtc_time by using 'rtc_valid_tm' in '__rtc_read_time'. And add the
message for debugging.

Signed-off-by: Hyogi Gim <hyogi.gim@lge.com>
---
 drivers/rtc/interface.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/rtc/interface.c b/drivers/rtc/interface.c
index 5b2717f..818ea97 100644
--- a/drivers/rtc/interface.c
+++ b/drivers/rtc/interface.c
@@ -30,6 +30,14 @@ static int __rtc_read_time(struct rtc_device *rtc, struct rtc_time *tm)
 	else {
 		memset(tm, 0, sizeof(struct rtc_time));
 		err = rtc->ops->read_time(rtc->dev.parent, tm);
+		if (err < 0) {
+			dev_err(&rtc->dev, "read_time: fail to read\n");
+			return err;
+		}
+
+		err = rtc_valid_tm(tm);
+		if (err < 0)
+			dev_err(&rtc->dev, "read_time: rtc_time isn't valid\n");
 	}
 	return err;
 }
-- 
1.8.3.2

Re: [PATCH] drivers/rtc/interface.c: check the validation of rtc_time in __rtc_read_time

[Andrew Morton]

On Wed, 8 Oct 2014 11:05:29 +0900 Hyogi Gim <ciogenis@gmail.com> wrote:

> Some of rtc devices always return '0' when rtc_class_ops.read_time is called.
> So if rtc_time isn't verified in callback, rtc interface cannot know whether
> rtc_time is valid.

We should fix the buggy .read_time() implementations...

> Check rtc_time by using 'rtc_valid_tm' in '__rtc_read_time'. And add the
> message for debugging.
> 

Re: [PATCH] drivers/rtc/interface.c: check the validation of rtc_time in __rtc_read_time

[Hyogi Gim]

On Tue, Oct 14, 2014 at 03:21:31PM -0700, Andrew Morton wrote:
> On Wed, 8 Oct 2014 11:05:29 +0900 Hyogi Gim <ciogenis@gmail.com> wrote:
> 
> > Some of rtc devices always return '0' when rtc_class_ops.read_time is called.
> > So if rtc_time isn't verified in callback, rtc interface cannot know whether
> > rtc_time is valid.
> 
> We should fix the buggy .read_time() implementations...
> 

Thanks for comments.
I'm studying rtc framework and trying to find a problem.
This is a minor bug. But, as you know, it can be a big problem for
newbies like me.

Even though it was late, I send you reply because a question.
If this patch is merged, rtc_valid_tm() in rtc_class_ops.read_time
callback is not necessary anymore.
So, I think each rtc device driver should change the code.
I can change the code. But, test is a little different.

In this case, how can I progress this modification?
Can I just send the notification for rtc_valid_tm() to the driver
manufacturer? or, request test for the code that I've modified?

-- 
Hyogi