Add benchmarks for u32/u64 functions.

This allows us to see the effect of any specicalized implementations for
`getrandom::u32` or `getrandom::u64`. As expected, on Linux (which just
uses the default implementation in `utils.rs`) there is no change:
```
test bench_u32                 ... bench:         196.50 ns/iter (+/- 4.85) = 20 MB/s
test bench_u32_via_fill        ... bench:         198.25 ns/iter (+/- 1.78) = 20 MB/s
test bench_u64                 ... bench:         196.95 ns/iter (+/- 2.99) = 40 MB/s
test bench_u64_via_fill        ... bench:         197.62 ns/iter (+/- 2.24) = 40 MB/s
```
but when using the `rdrand` backend (which is specialized), there is a
mesurable difference.
```
test bench_u32                 ... bench:          16.84 ns/iter (+/- 0.09) = 250 MB/s
test bench_u32_via_fill        ... bench:          18.40 ns/iter (+/- 0.28) = 222 MB/s
test bench_u64                 ... bench:          16.62 ns/iter (+/- 0.06) = 500 MB/s
test bench_u64_via_fill        ... bench:          17.70 ns/iter (+/- 0.08) = 470 MB/s
```

Signed-off-by: Joe Richey <[email protected]>

benches/buffer.rs

@@ -1,7 +1,10 @@
 #![feature(test, maybe_uninit_uninit_array_transpose)]
 extern crate test;
 
-use std::mem::MaybeUninit;
+use std::{
+    mem::{size_of, MaybeUninit},
+    slice,
+};
 
 // Call getrandom on a zero-initialized stack buffer
 #[inline(always)]
@@ -19,6 +22,53 @@ fn bench_fill_uninit<const N: usize>() {
     test::black_box(buf);
 }
 
+#[bench]
+pub fn bench_u32(b: &mut test::Bencher) {
+    #[inline(never)]
+    fn inner() -> u32 {
+        getrandom::u32().unwrap()
+    }
+    b.bytes = 4;
+    b.iter(inner);
+}
+#[bench]
+pub fn bench_u32_via_fill(b: &mut test::Bencher) {
+    #[inline(never)]
+    fn inner() -> u32 {
+        let mut res = MaybeUninit::<u32>::uninit();
+        let dst: &mut [MaybeUninit<u8>] =
+            unsafe { slice::from_raw_parts_mut(res.as_mut_ptr().cast(), size_of::<u32>()) };
+        getrandom::fill_uninit(dst).unwrap();
+        unsafe { res.assume_init() }
+    }
+    b.bytes = 4;
+    b.iter(inner);
+}
+
+#[bench]
+pub fn bench_u64(b: &mut test::Bencher) {
+    #[inline(never)]
+    fn inner() -> u64 {
+        getrandom::u64().unwrap()
+    }
+    b.bytes = 8;
+    b.iter(inner);
+}
+
+#[bench]
+pub fn bench_u64_via_fill(b: &mut test::Bencher) {
+    #[inline(never)]
+    fn inner() -> u64 {
+        let mut res = MaybeUninit::<u64>::uninit();
+        let dst: &mut [MaybeUninit<u8>] =
+            unsafe { slice::from_raw_parts_mut(res.as_mut_ptr().cast(), size_of::<u64>()) };
+        getrandom::fill_uninit(dst).unwrap();
+        unsafe { res.assume_init() }
+    }
+    b.bytes = 8;
+    b.iter(inner);
+}
+
 // We benchmark using #[inline(never)] "inner" functions for two reasons:
 //  - Avoiding inlining reduces a source of variance when running benchmarks.
 //  - It is _much_ easier to get the assembly or IR for the inner loop.