STMicroelectronics STM32MP Clock and Reset Controllers
Provides the on-chip clock tree and peripheral reset control (the RCC block) for STMicroelectronics' STM32MP family of Cortex-A Linux-capable microprocessors, including the STM32MP1, MP13, MP21, and MP25 lines used in industrial gateways, HMIs, and embedded systems from roughly 2019 onward through brand-new parts launched in 2026.
recommendation
It should stay in the kernel because it serves STMicroelectronics' current STM32MP1, MP13, MP21, and MP25 application processors, all of which ST is actively selling for industrial and embedded Linux use in 2025. Upstream commit activity is healthy, with new MP21 SoC support landing in 2025 and further fixes queued into 2026, so this is a maintained driver tracking newly launched silicon rather than a legacy hold-over.
repository signals
sources
- git.kernel.org
Upstream activity is recent and substantive, including new STM32MP21 support in 2025 and core/rate-conversion fixes in 2026; this does not look abandoned.
- git.kernel.org
The directory is the in-tree RCC clock/reset driver family for multiple current STM32MP SoC generations, not a legacy one-off.
- st.com
STM32MP13 is an active Linux-capable MPU family marketed for industrial and embedded applications, indicating ongoing deployment demand.
- st.com
STM32MP25 is a current STM32MP2 product line with evaluation hardware and active ecosystem support, showing the driver covers newly sold hardware.
- blog.st.com
ST launched STM32MP21x on January 5, 2026, confirming this driver family maps to newly introduced hardware rather than obsolete devices.
codex reasoning notes (technical)
Local shell inspection (`rg`, `sed`) showed real platform-driver code and Kconfig entries for stm32mp13x/mp15x/mp21x/mp25x. Local shell `git log` showed ongoing work through 2026-01-08 plus 2025 introduction of STM32MP21 support. Direct lore evidence could not be retrieved in this environment because the advertised `lore-http` MCP server was unavailable and `lei` was not installed, so no removal-thread hit was found; absent any removal evidence and given active commits/new SoCs, `keep` is the defensible outcome. Source acquisition: kernel.org URLs by canonical recall; ST URLs from web search results `turn0search0`, `turn0search1`, `turn0search6`, and `turn0search3`.