etm-trace: sync docs with the final firmware behavioretmtrace-rp2350

Copilot and the claude review on #3851 both caught boards.md describing
two intermediate states of this branch rather than its result: the
DBGPAUSE clear is present (restored on measurement - J-Link does not
clear it and the frozen us-timer hangs sleep_ms after any debug
session), the TX-only console fallback is gone (the carrier routes the
console to GP12/13), and TRACE_ETM additionally carries the compile-time
trace-pin conflict checks. The ozone reference header repeated the
TX-only claim; both now state what family.c actually does.

0 files changed, 0 insertions, 0 deletions