On 07/04/2018 11:53 AM, Alexander Graf wrote:
On 07/04/2018 09:36 AM, AKASHI Takahiro wrote:
This patch is missing a patch description. I'm not the maintainer of the rtc code base so it's not my call, but I personally just reject all patches with empty patch descriptions ;).
And thanks a lot for doing the conversion! I think it's a very good step forward.
Signed-off-by: AKASHI Takahiro takahiro.akashi@linaro.org
...snip...
/*
- Reset the RTC. We set the date back to 1970-01-01.
- Get the current time from the RTC
*/ -void rtc_reset(void) +static int pl031_rtc_get(struct udevice *dev, struct rtc_time *tm) { - RTC_WRITE_REG(RTC_LR, 0x00); - if(!pl031_initted) - rtc_init(); + struct pl031_rtc_platdata *pdata = dev_get_platdata(dev); + ulong tim;
+ if (!tm) { + puts("Error getting the date/time\n"); + return -1; + }
+ if (!pl031_initted)
In theory with dm you can now have multiple instances of the device, right? So we can no longer have a global variable that indicates if a device is initialized. Instead, this needs to move into device private data.
I think the initialization code in rtc_init() should be move to the .probe callback instead, so there's no need to keep the bool aroun.