Hi,
The RTC used in the A31 and later SoC has an accuracy issue, which is
already significant even after a couple of hours.
This is due to the fact that the oscillator used by default is an internal
and very inaccurate one.
A first attempt at fixing that by switching to the external oscillator was
done in the patch "rtc: sun6i: Switch to the external oscillator". However,
it turned out to be problematic since it was tracked properly in the clock
framework, which might lead to some clocks being disabled, even though
their devices were not notified.
This is a second attempt, this time by making it part of the CCF. It
turned out to be a bit more complicated than one would expect since the mux
found inside the RTC also controls one of the input of the main clock unit.
Therefore, it needs to be probed before the main clock unit driver.
Let me know what you think,
Maxime
Changes from v1:
- Force the muxing to the external oscillator even for old DT
- Make the patches a bit more stable friendly
- Change the bindings to have a cell of 1, to account for the external
output of the oscillator
- Split the driver remove() removal into a separate patch and switched to
devm
- Reordered the patches
Maxime Ripard (7):
rtc: sun6i: Disable the build as a module
rtc: sun6i: Add some locking
rtc: sun6i: Switch to the external oscillator
rtc: sun6i: Expose the 32kHz oscillator
rtc: sun6i: Switch to devm_rtc_device_register
ARM: sun8i: a23/a33: Enable the real LOSC and use it
ARM: sun8i: a23/a33: Add the oscillators accuracy
Documentation/devicetree/bindings/rtc/sun6i-rtc.txt | 10 +-
arch/arm/boot/dts/sun8i-a23-a33.dtsi | 15 +-
drivers/rtc/Kconfig | 2 +-
drivers/rtc/rtc-sun6i.c | 182 ++++++++++---
4 files changed, 174 insertions(+), 35 deletions(-)
base-commit: 99cef370ac9939df2aeb16c96d07e842b2fa8201
--
git-series 0.8.11
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
