bench: add criterion benchmarks for protocol, bandwidth, TCP RX scan, and EC-SRP5

Adds four Criterion.rs benchmark suites to measure hot-path performance
and demonstrate the impact of Sprints 1–3 optimizations:

- benches/protocol.rs    — Command & StatusMessage serialize/deserialize
- benches/bandwidth.rs   — BandwidthState atomics, budget, interval math
- benches/tcp_rx_scan.rs — memchr SIMD scan vs naive O(n) loop (55× faster
                           on 256KB buffers with status at end)
- benches/ecsrp5.rs      — WCurve::new() heavy math vs cached LazyLock
                           (~123,000× faster access)

Also adds BENCHMARKS.md with usage instructions and example results.

Visibility changes (bench-only):
- scan_status_message is now pub (was #[cfg(test)] only)
- WCurve and WCURVE are now pub in ecsrp5.rs

dev-dependencies: criterion + pprof (optional flamegraph support)

benches/tcp_rx_scan.rs

@@ -0,0 +1,100 @@
+use criterion::{black_box, criterion_group, criterion_main, BenchmarkId, Criterion, Throughput};
+use btest_rs::client::scan_status_message;
+use btest_rs::protocol::STATUS_MSG_TYPE;
+
+/// Naive O(n) byte-by-byte scan — the old implementation.
+fn naive_scan(buf: &[u8]) -> Option<u8> {
+    const STATUS_MSG_SIZE: usize = 12;
+    if buf.len() < STATUS_MSG_SIZE {
+        return None;
+    }
+    for i in 0..=(buf.len() - STATUS_MSG_SIZE) {
+        if buf[i] == STATUS_MSG_TYPE && buf[i + 1] >= 0x80 {
+            return Some((buf[i + 1] & 0x7F).min(100));
+        }
+    }
+    None
+}
+
+fn make_buffer(size: usize, status_at: Option<usize>) -> Vec<u8> {
+    let mut buf = vec![0u8; size];
+    if let Some(pos) = status_at {
+        buf[pos] = STATUS_MSG_TYPE;
+        buf[pos + 1] = 0x80 | 50; // CPU = 50%
+    }
+    buf
+}
+
+fn bench_scan_all_zeros(c: &mut Criterion) {
+    let mut group = c.benchmark_group("tcp_rx_scan_all_zeros");
+    for size in [4096, 65536, 262144] {
+        let buf = make_buffer(size, None);
+        group.throughput(Throughput::Bytes(size as u64));
+        group.bench_with_input(BenchmarkId::new("naive", size), &buf, |b, buf| {
+            b.iter(|| black_box(naive_scan(black_box(buf))))
+        });
+        group.bench_with_input(BenchmarkId::new("memchr", size), &buf, |b, buf| {
+            b.iter(|| black_box(scan_status_message(black_box(&[]), black_box(buf))))
+        });
+    }
+    group.finish();
+}
+
+fn bench_scan_status_at_start(c: &mut Criterion) {
+    let mut group = c.benchmark_group("tcp_rx_scan_status_at_start");
+    for size in [4096, 65536, 262144] {
+        let buf = make_buffer(size, Some(0));
+        group.throughput(Throughput::Bytes(size as u64));
+        group.bench_with_input(BenchmarkId::new("naive", size), &buf, |b, buf| {
+            b.iter(|| black_box(naive_scan(black_box(buf))))
+        });
+        group.bench_with_input(BenchmarkId::new("memchr", size), &buf, |b, buf| {
+            b.iter(|| black_box(scan_status_message(black_box(&[]), black_box(buf))))
+        });
+    }
+    group.finish();
+}
+
+fn bench_scan_status_at_end(c: &mut Criterion) {
+    let mut group = c.benchmark_group("tcp_rx_scan_status_at_end");
+    for size in [4096, 65536, 262144] {
+        let buf = make_buffer(size, Some(size - 12));
+        group.throughput(Throughput::Bytes(size as u64));
+        group.bench_with_input(BenchmarkId::new("naive", size), &buf, |b, buf| {
+            b.iter(|| black_box(naive_scan(black_box(buf))))
+        });
+        group.bench_with_input(BenchmarkId::new("memchr", size), &buf, |b, buf| {
+            b.iter(|| black_box(scan_status_message(black_box(&[]), black_box(buf))))
+        });
+    }
+    group.finish();
+}
+
+fn bench_scan_split_message(c: &mut Criterion) {
+    // Simulate a status message split across two reads:
+    // carry has first 5 bytes, buf has remaining 7 bytes
+    let mut carry = vec![0u8; 5];
+    carry[0] = STATUS_MSG_TYPE;
+    carry[1] = 0x80 | 75;
+    let buf = vec![0u8; 7];
+
+    c.bench_function("scan_split_5_7", |b| {
+        b.iter(|| black_box(scan_status_message(black_box(&carry), black_box(&buf))))
+    });
+
+    // Split with 2 bytes in carry (status type + cpu byte), 10 in buf
+    let carry_2 = vec![STATUS_MSG_TYPE, 0x80 | 33];
+    let buf_10 = vec![0u8; 10];
+    c.bench_function("scan_split_2_10", |b| {
+        b.iter(|| black_box(scan_status_message(black_box(&carry_2), black_box(&buf_10))))
+    });
+}
+
+criterion_group!(
+    tcp_rx_scan_benches,
+    bench_scan_all_zeros,
+    bench_scan_status_at_start,
+    bench_scan_status_at_end,
+    bench_scan_split_message
+);
+criterion_main!(tcp_rx_scan_benches);