crypto/sha1: speed up sha1 for very small blocks

For very small blocks significant time is spent in checkSum function,
adding necessary padding. Instead of writing it byte by byte, copy
encoding/binary PutUint functions, which are compiled into single mov.

name            old time/op    new time/op    delta
Hash8Bytes-6       344ns ± 0%     310ns ± 0%   -9.78%  (p=0.000 n=10+9)
Hash320Bytes-6    1.28µs ± 0%    1.25µs ± 0%   -2.58%  (p=0.000 n=10+10)
Hash1K-6          2.51µs ± 0%    2.47µs ± 0%   -1.67%  (p=0.000 n=10+10)
Hash8K-6          15.8µs ± 0%    15.7µs ± 1%   -0.21%  (p=0.023 n=10+10)

name            old speed      new speed      delta
Hash8Bytes-6    23.2MB/s ± 0%  25.7MB/s ± 0%  +10.77%  (p=0.000 n=10+9)
Hash320Bytes-6   249MB/s ± 0%   256MB/s ± 0%   +2.65%  (p=0.000 n=10+10)
Hash1K-6         408MB/s ± 0%   414MB/s ± 0%   +1.70%  (p=0.000 n=10+10)

Change-Id: I3975ee929465c7dd137d0ca757ad3792a004e1a3
Reviewed-on: https://go-review.googlesource.com/54391
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Giovanni Bajo <rasky@develer.com>
Reviewed-by: Filippo Valsorda <hi@filippo.io>

src/crypto/sha1/sha1.go

@@ -104,9 +104,7 @@ func (d *digest) checkSum() [Size]byte {
 
 	// Length in bits.
 	len <<= 3
-	for i := uint(0); i < 8; i++ {
-		tmp[i] = byte(len >> (56 - 8*i))
-	}
+	putUint64(tmp[:], len)
 	d.Write(tmp[0:8])
 
 	if d.nx != 0 {
@@ -114,12 +112,12 @@ func (d *digest) checkSum() [Size]byte {
 	}
 
 	var digest [Size]byte
-	for i, s := range d.h {
-		digest[i*4] = byte(s >> 24)
-		digest[i*4+1] = byte(s >> 16)
-		digest[i*4+2] = byte(s >> 8)
-		digest[i*4+3] = byte(s)
-	}
+
+	putUint32(digest[0:], d.h[0])
+	putUint32(digest[4:], d.h[1])
+	putUint32(digest[8:], d.h[2])
+	putUint32(digest[12:], d.h[3])
+	putUint32(digest[16:], d.h[4])
 
 	return digest
 }
@@ -199,3 +197,23 @@ func Sum(data []byte) [Size]byte {
 	d.Write(data)
 	return d.checkSum()
 }
+
+func putUint64(x []byte, s uint64) {
+	_ = x[7]
+	x[0] = byte(s >> 56)
+	x[1] = byte(s >> 48)
+	x[2] = byte(s >> 40)
+	x[3] = byte(s >> 32)
+	x[4] = byte(s >> 24)
+	x[5] = byte(s >> 16)
+	x[6] = byte(s >> 8)
+	x[7] = byte(s)
+}
+
+func putUint32(x []byte, s uint32) {
+	_ = x[3]
+	x[0] = byte(s >> 24)
+	x[1] = byte(s >> 16)
+	x[2] = byte(s >> 8)
+	x[3] = byte(s)
+}