derp: use AvailableBuffer for WriteFrameHeader, consolidate tests (#19101)

Use bufio.Writer.AvailableBuffer to write the frame header directly
into bufio's internal buffer as a single append+Write, avoiding 5
separate WriteByte calls. Fall back to the existing writeUint32
byte-at-a-time path when the buffer has insufficient space.

```
name                  old ns/op  new ns/op  speedup
WriteFrameHeader-8    18.8       7.8        ~2.4x
(0 allocs/op in both)
```

Add TestWriteFrameHeader with correctness
checks, allocation assertions, and coverage of both fast and slow
write paths. Move BenchmarkReadFrameHeader from client_test.go to
derp_test.go alongside BenchmarkWriteFrameHeader, co-located with
the functions under test.

Updates tailscale/corp#38509

Signed-off-by: Mike O'Driscoll <mikeo@tailscale.com>

derp/client_test.go

@@ -6,7 +6,6 @@ package derp
 import (
 	"bufio"
 	"bytes"
-	"io"
 	"net"
 	"reflect"
 	"sync"
@@ -126,33 +125,6 @@ func TestClientSendPong(t *testing.T) {
 	}
 }
 
-func BenchmarkWriteUint32(b *testing.B) {
-	w := bufio.NewWriter(io.Discard)
-	b.ReportAllocs()
-	b.ResetTimer()
-	for range b.N {
-		writeUint32(w, 0x0ba3a)
-	}
-}
-
-type nopRead struct{}
-
-func (r nopRead) Read(p []byte) (int, error) {
-	return len(p), nil
-}
-
-func BenchmarkReadFrameHeader(b *testing.B) {
-	r := bufio.NewReader(nopRead{})
-	b.ReportAllocs()
-	b.ResetTimer()
-	for range b.N {
-		_, _, err := ReadFrameHeader(r)
-		if err != nil {
-			b.Fatal(err)
-		}
-	}
-}
-
 type countWriter struct {
 	mu     sync.Mutex
 	writes int