[benchmark] Use -Os to match firmware optimization level

CodSpeed benchmarks were building with -O2, while all firmware
targets (ESP8266, ESP32, LibreTiny) use -Os. This mismatch means
the benchmarks cannot detect inlining regressions that affect real
devices — GCC under -O2 inlines functions that -Os outlines due to
its size-conscious cost model.

Switch to -Os with -ffunction-sections/-fdata-sections for proper
dead-code stripping (needed because -Os preserves references that
-O2 optimizes away at compile time).

script/cpp_benchmark.py

@@ -27,7 +27,10 @@ STUBS_DIR: Path = Path(root_path) / "tests" / "benchmarks" / "stubs"
 
 PLATFORMIO_OPTIONS = {
     "build_flags": [
+        "-Os",  # match firmware optimization level (detects inlining regressions)
         "-g",  # debug symbols for profiling
+        "-ffunction-sections",  # required for dead-code stripping with -Os
+        "-fdata-sections",  # required for dead-code stripping with -Os
         "-DUSE_BENCHMARK",  # disable WarnIfComponentBlockingGuard in finish()
         f"-I{STUBS_DIR}",  # stub headers for ESP32-only components
     ],