crypto/sha1, crypto/sha256, crypto/sha512: use copy for partial block

Use copy rather than a hand rolled loop when moving a partial input
block to the scratch area. This results in a reasonable performance
gain when partial blocks are written.

Benchmarks on Intel(R) Xeon(R) CPU X5650  @ 2.67GHz with Go amd64:

       benchmark               old MB/s     new MB/s  speedup
SHA1   BenchmarkHash8Bytes        18.37        22.80    1.24x
SHA256 BenchmarkHash8Bytes        11.86        13.78    1.16x
SHA512 BenchmarkHash8Bytes         4.51         5.24    1.16x

       benchmark              old ns/op    new ns/op    delta
SHA1   BenchmarkHash8Bytes          435          350  -19.54%
SHA256 BenchmarkHash8Bytes          674          580  -13.95%
SHA512 BenchmarkHash8Bytes         1772         1526  -13.88%

R=agl, dave, bradfitz
CC=golang-codereviews
https://golang.org/cl/35840044

src/pkg/crypto/sha1/sha1.go

@@ -62,16 +62,10 @@ func (d *digest) Write(p []byte) (nn int, err error) {
 	nn = len(p)
 	d.len += uint64(nn)
 	if d.nx > 0 {
-		n := len(p)
-		if n > chunk-d.nx {
-			n = chunk - d.nx
-		}
-		for i := 0; i < n; i++ {
-			d.x[d.nx+i] = p[i]
-		}
+		n := copy(d.x[d.nx:], p)
 		d.nx += n
 		if d.nx == chunk {
-			block(d, d.x[0:])
+			block(d, d.x[:])
 			d.nx = 0
 		}
 		p = p[n:]

src/pkg/crypto/sha256/sha256.go

@@ -106,16 +106,10 @@ func (d *digest) Write(p []byte) (nn int, err error) {
 	nn = len(p)
 	d.len += uint64(nn)
 	if d.nx > 0 {
-		n := len(p)
-		if n > chunk-d.nx {
-			n = chunk - d.nx
-		}
-		for i := 0; i < n; i++ {
-			d.x[d.nx+i] = p[i]
-		}
+		n := copy(d.x[d.nx:], p)
 		d.nx += n
 		if d.nx == chunk {
-			block(d, d.x[0:])
+			block(d, d.x[:])
 			d.nx = 0
 		}
 		p = p[n:]

src/pkg/crypto/sha512/sha512.go

@@ -106,16 +106,10 @@ func (d *digest) Write(p []byte) (nn int, err error) {
 	nn = len(p)
 	d.len += uint64(nn)
 	if d.nx > 0 {
-		n := len(p)
-		if n > chunk-d.nx {
-			n = chunk - d.nx
-		}
-		for i := 0; i < n; i++ {
-			d.x[d.nx+i] = p[i]
-		}
+		n := copy(d.x[d.nx:], p)
 		d.nx += n
 		if d.nx == chunk {
-			block(d, d.x[0:])
+			block(d, d.x[:])
 			d.nx = 0
 		}
 		p = p[n:]