On the i.MX8MN Beacon SOM, there is an RTC chip which is fed power
from the baseboard during power off. The SNVS RTC integrated into
the SoC is not fed power. Depending on the order the modules are
loaded, this can be a problem if the external RTC isn't rtc0.
Make the alias for rtc0 point to the external RTC all the time and
rtc1 point to the SVNS in order to correctly hold date/time over
a power-cycle.
Signed-off-by: Adam Ford <[email protected]>
diff --git a/arch/arm64/boot/dts/freescale/imx8mn-beacon-som.dtsi b/arch/arm64/boot/dts/freescale/imx8mn-beacon-som.dtsi
index 67e5e5b9ddea..2120e6485393 100644
--- a/arch/arm64/boot/dts/freescale/imx8mn-beacon-som.dtsi
+++ b/arch/arm64/boot/dts/freescale/imx8mn-beacon-som.dtsi
@@ -4,6 +4,11 @@
*/
/ {
+ aliases {
+ rtc0 = &rtc;
+ rtc1 = &snvs_rtc;
+ };
+
usdhc1_pwrseq: usdhc1_pwrseq {
compatible = "mmc-pwrseq-simple";
pinctrl-names = "default";
@@ -212,7 +217,7 @@ eeprom@50 {
reg = <0x50>;
};
- rtc@51 {
+ rtc: rtc@51 {
compatible = "nxp,pcf85263";
reg = <0x51>;
};
--
2.25.1
On Sun, Jan 10, 2021 at 05:53:53AM -0600, Adam Ford wrote:
> On the i.MX8MN Beacon SOM, there is an RTC chip which is fed power
> from the baseboard during power off. The SNVS RTC integrated into
> the SoC is not fed power. Depending on the order the modules are
> loaded, this can be a problem if the external RTC isn't rtc0.
>
> Make the alias for rtc0 point to the external RTC all the time and
> rtc1 point to the SVNS in order to correctly hold date/time over
> a power-cycle.
>
> Signed-off-by: Adam Ford <[email protected]>
Reviewed-by: Krzysztof Kozlowski <[email protected]>
Best regards,
Krzysztof
On Sun, Jan 10, 2021 at 05:53:53AM -0600, Adam Ford wrote:
> On the i.MX8MN Beacon SOM, there is an RTC chip which is fed power
> from the baseboard during power off. The SNVS RTC integrated into
> the SoC is not fed power. Depending on the order the modules are
> loaded, this can be a problem if the external RTC isn't rtc0.
>
> Make the alias for rtc0 point to the external RTC all the time and
> rtc1 point to the SVNS in order to correctly hold date/time over
> a power-cycle.
>
> Signed-off-by: Adam Ford <[email protected]>
Applied, thanks.