Sophgo CV1800 and SG2044 SoC platform glue
Small platform support code for Sophgo's RISC-V system-on-chip family, including the CV1800B used on low-cost Milk-V embedded boards (released 2022) and the SG2044 server-class chip launched in 2025. It provides the syscon and top-level register glue that downstream clock, RTC, and similar drivers hang off of.
recommendation
It should stay because this is brand-new code added in 2025 for current-generation Sophgo silicon that is still being sold and actively brought up upstream. Patches like the CV1800 rtcsys series were accepted into soc-for-next in May 2025, and SG2044 enablement is ongoing, so the directory reflects active vendor and community work rather than legacy cruft. Deployment is niche today (Milk-V boards and a handful of SG2044 servers) but the trajectory is growth, not decline.
repository signals
sources
- lists.openwall.net
CV1800 rtcsys support was posted upstream in March 2025 as a new driver, indicating this directory is newly added rather than legacy.
- spinics.net
The CV1800 rtcsys series was explicitly applied to soc-for-next on May 7, 2025, showing active upstream acceptance rather than removal discussion.
- lwn.net
The SG2044 support series includes the TOP syscon/clock work in April 2025, showing ongoing bring-up of new Sophgo server silicon.
- en.sophgo.com
Sophgo's own timeline says CV180 series released in 2022 and SG2044 released in 2025, so the hardware family is current in 2025.
- milkv.io
Milk-V lists CV1800B chips as sample-buy products, which is direct evidence that at least part of this family is still sold new.
- cateee.net
LKDDb shows CONFIG_SOPHGO_SG2044_TOPSYS present in Linux 6.16-6.19 and 7.0-rc+HEAD, confirming ongoing upstream presence.
codex reasoning notes (technical)
Local shell inspection showed two tiny platform/MFD stub drivers added in 2025 only, with no aging or bitrot pattern. `git -c safe.directory=... log --stat -- drivers/soc/sophgo` exposed exact lore-linked commit metadata; direct lore search tooling was unavailable, so I used web search to resolve those message IDs to public thread pages (openwall/spinics) and confirm one thread was applied to soc-for-next, with no removal discussion found. Vendor/current-market evidence came from web search to Sophgo's official timeline and Milk-V's chip sales page; LKDDb was obtained by web search for current kernel presence. This directory fronts current-generation Sophgo SoCs, has low but real niche deployment (embedded CV1800B boards and SG2044 servers), and has no natural replacement beyond the downstream child drivers it instantiates.