What does a "don't-touch subtree" constraint actually protect during CTS, and when would you use it?

Short Answer

The don't-touch subtree constraint selectively preserves a portion of the clock tree at a particular clock pin -- useful, for example, at a mux between two clocks feeding different downstream paths, where you want CTS free to optimize everywhere except that specific already-correct structure. It's a scoped protection, not a whole-tree freeze.

Technical Explanation

• The don't-touch subtree constraint selectively preserves a portion of the clock tree at a particular clock pin, not the whole tree.
• A concrete use case: at a mux between two clocks (e.g. CLK1/CLK2) feeding different downstream paths, you may want CTS free to optimize everywhere else while leaving that specific mux structure exactly as-is.
• This is scoped protection, distinct from a global dont_touch on the whole clock network -- you're protecting one already-correct structure, not opting the entire tree out of optimization.
• No verified dedicated command syntax was found for a "don't-touch subtree" scoped to one clock pin -- the likely mechanism is a set_dont_touch or mark_clock_trees -dont_touch variant, but the exact flag should be verified against the ICC2 Implementation User Guide rather than assumed.

Follow-up Question & Model Response

"Why would a clock mux specifically be a common place to need this kind of scoped protection?"

Candidate Model Response: A clock mux is exactly where two independently-timed clock paths converge into one structure feeding different downstream domains -- getting that convergence point right by hand and then protecting it from CTS's automated optimization is more reliable than trusting the algorithm to preserve a correctness property it doesn't know is special.