On Tue, Sep 22, 2026 at 11:46:24PM +0800, Zhuxu Ran wrote: > The upstream devicetrees describe the CV1800 RTC subsystem with the > "sophgo,cv1800b-rtc" compatible string. The RTC seconds counter and the > system-reset controls share the same register range. > > Add an RTC driver which directly matches this node, enables its clock, > and implements reading and setting the 32-bit seconds counter. Bind the > CV1800B sysreset driver as a child from the RTC driver's bind callback so > both functions can be instantiated from the same devicetree node. > > Initialize the RTC and sysreset register maps from their shared devicetree > node instead of using hard-coded MMIO addresses. Remove the board-level > sysreset binding from Milk-V Duo. Boards using the upstream RTC node can > now instantiate both functions without board-specific binding code. > > Signed-off-by: Zhuxu Ran <[email protected]>
This patch looks quite large, and comparing to other parts of the series, it adds new drivers and modifies existing driver massively, so I think it makes sense to separate it out of the series, i.e., send a separate series converting sysreset-related ABI to match upstream first. > --- > board/sophgo/milkv_duo/board.c | 5 -- > configs/milkv_duo_256m_defconfig | 2 + > configs/milkv_duo_defconfig | 2 + > configs/sipeed_licheerv_nano_defconfig | 2 + > drivers/rtc/Kconfig | 9 +++ > drivers/rtc/Makefile | 1 + > drivers/rtc/rtc-cv1800.c | 122 > +++++++++++++++++++++++++++++++++ > drivers/sysreset/Kconfig | 6 +- > drivers/sysreset/sysreset_cv1800b.c | 86 ++++++++++++++--------- This involves multiple subsystems, please separate rtc/sysreset/ board-level changes into different patches. Bisectability might be broken for one or two commits, but I think it would be much clear to handle. Regards, Yao Zi
