Походу, это работает для V3 и V4. CH32V203F8P6, например. Но и у
неё, "из коробки" MRS, порядка 100нС. "The V4 series microprocessors support hardware single cycle automatic saving of 16 of the shaped Caller
Saved registers to an internal stack area that is not visible to the user. When an exception or interrupt returns,
the hardware single cycle automatically restores data from the internal stack area to the 16 shaped registers.
The hardware stack supports nesting with a maximum nesting depth of 3 levels. After a hardware stack
overflow, if a higher priority interrupt is still allowed to execute, the "field" is saved to the user stack area."