Add smoke test and benchmarks

smoke links crt0.o + libc.a with -nostdlib -static and proves the whole chain runs without the dynamic linker; bench_strlen and bench_syscall measure the string core and raw syscall round-trip cost via make bench.

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)

Co-authored-by: Sisyphus <[email protected]>

benchmarks/bench_strlen.c

@@ -0,0 +1,28 @@
+/*
+ * bench_strlen.c — how fast is our string core?
+ *
+ * Measures strlen() over a typical short string. 10M iterations keeps
+ * the loop overhead negligible; results are printed as ns per call.
+ */
+
+#include <string.h>
+
+#include "bench.h"
+
+#define ITERS 10000000UL
+
+int main(void)
+{
+    static const char s[] = "the quick brown fox jumps over the lazy dog";
+    volatile size_t sink = 0;
+    long long t0, t1;
+
+    t0 = bench_now_ns();
+    for (unsigned long i = 0; i < ITERS; i++)
+        sink += strlen(s);
+    t1 = bench_now_ns();
+
+    (void)sink; /* keep the loop observable */
+    bench_print_u64((unsigned long long)((t1 - t0) / ITERS));
+    return 0;
+}