Switch benchmarks to match unit test namespacing. (#3268)

Specifically, rather than nesting them in `Carbon::Testing`, nest them
in `Carbon::Foo` for whatever component they're benchmarking. All our
current benchmarks are lexer benchmarks so its `Carbon::Lex`.

This makes even more sense for benchmarks than unittests I think.

toolchain/lex/tokenized_buffer_benchmark.cpp

@@ -15,12 +15,9 @@
 #include "toolchain/lex/token_kind.h"
 #include "toolchain/lex/tokenized_buffer.h"
 
-namespace Carbon::Testing {
+namespace Carbon::Lex {
 namespace {
 
-using Lex::TokenizedBuffer;
-using Lex::TokenKind;
-
 // A large value for measurement stability without making benchmarking too slow.
 // Needs to be a multiple of 100 so we can easily divide it up into percentages,
 // and 1% itself needs to not be too tiny. This makes 100,000 a great balance.
@@ -383,4 +380,4 @@ BENCHMARK(BM_ValidMix)
     ->Args({75, 10});
 
 }  // namespace
-}  // namespace Carbon::Testing
+}  // namespace Carbon::Lex