cryptocore: prefetch nonces in 512-byte blocks

On my machine, reading 512-byte blocks from /dev/urandom
(same via getentropy syscall) is a lot faster in terms of
throughput:

Blocksize    Throughput
 16          28.18 MB/s
512          83.75 MB/s

For a single-threaded streaming write, this drops the CPU usage of
nonceGenerator.Get to almost 1/3:

        flat  flat%   sum%        cum   cum%
Before     0     0% 95.08%      0.35s  2.92%  github.com/rfjakob/gocryptfs/internal/cryptocore.(*nonceGenerator).Get
After  0.01s 0.092% 92.34%      0.13s  1.20%  github.com/rfjakob/gocryptfs/internal/cryptocore.(*nonceGenerator).Get

This change makes the nonce reading single-threaded, which may
hurt massively-parallel writes.

1 files changed, 40 insertions, 0 deletions

diff --git a/internal/cryptocore/randprefetch_test.go b/internal/cryptocore/randprefetch_test.go
new file mode 100644
index 0000000..2a568f3
--- /dev/null
+++ b/internal/cryptocore/randprefetch_test.go
@@ -0,0 +1,40 @@
+package cryptocore
+
+import (
+	"bytes"
+	"compress/flate"
+	"runtime"
+	"sync"
+	"testing"
+)
+
+// TestRandPrefetch hammers the randPrefetcher with 100 goroutines and verifies
+// that the result is incompressible
+func TestRandPrefetch(t *testing.T) {
+	runtime.GOMAXPROCS(10)
+	p := 100
+	l := 200
+	vec := make([][]byte, p)
+	var wg sync.WaitGroup
+	for i := 0; i < p; i++ {
+		wg.Add(1)
+		go func(i int) {
+			var tmp []byte
+			for x := 0; x < l; x++ {
+				tmp = append(tmp, randPrefetcher.read(l)...)
+			}
+			vec[i] = tmp
+			wg.Done()
+		}(i)
+	}
+	wg.Wait()
+	var b bytes.Buffer
+	fw, _ := flate.NewWriter(&b, flate.BestCompression)
+	for _, v := range vec {
+		fw.Write(v)
+	}
+	fw.Close()
+	if b.Len() < p*l*l {
+		t.Errorf("random data should be incompressible, but: in=%d compressed=%d\n", p*l*l, b.Len())
+	}
+}