crypto/rc4: remove assembler implementations

This CL removes the RC4 assembler implementations.
RC4 is broken and should not be used for encryption
anymore. Therefore it's not worth maintaining
platform-specific assembler implementations.

The native Go implementation may be slower
or faster depending on the CPU:

name       old time/op   new time/op   delta
RC4_128-4    256ns ± 0%    196ns ± 0%  -23.78%  (p=0.029 n=4+4)
RC4_1K-4    2.38µs ± 0%   1.54µs ± 0%  -35.22%  (p=0.029 n=4+4)
RC4_8K-4    19.4µs ± 1%   12.0µs ± 0%  -38.35%  (p=0.029 n=4+4)

name       old speed     new speed     delta
RC4_128-4  498MB/s ± 0%  654MB/s ± 0%  +31.12%  (p=0.029 n=4+4)
RC4_1K-4   431MB/s ± 0%  665MB/s ± 0%  +54.34%  (p=0.029 n=4+4)
RC4_8K-4   418MB/s ± 1%  677MB/s ± 0%  +62.18%  (p=0.029 n=4+4)

vendor_id	: GenuineIntel
cpu family	: 6
model		: 142
model name	: Intel(R) Core(TM) i5-7Y54 CPU @ 1.20GHz
stepping	: 9
microcode	: 0x84
cpu MHz		: 800.036
cache size	: 4096 KB

name       old time/op   new time/op   delta
RC4_128-4    235ns ± 1%    431ns ± 0%  +83.00%  (p=0.000 n=10+10)
RC4_1K-4    1.74µs ± 0%   3.41µs ± 0%  +96.74%  (p=0.000 n=10+10)
RC4_8K-4    13.6µs ± 1%   26.8µs ± 0%  +97.58%   (p=0.000 n=10+9)

name       old speed     new speed     delta
RC4_128-4  543MB/s ± 0%  297MB/s ± 1%  -45.29%  (p=0.000 n=10+10)
RC4_1K-4   590MB/s ± 0%  300MB/s ± 0%  -49.16%  (p=0.000 n=10+10)
RC4_8K-4   596MB/s ± 1%  302MB/s ± 0%  -49.39%   (p=0.000 n=10+9)

vendor_id       : GenuineIntel
cpu family      : 6
model           : 63
model name      : Intel(R) Xeon(R) CPU @ 2.30GHz
stepping        : 0
microcode       : 0x1
cpu MHz         : 2300.000
cache size      : 46080 KB

Fixes #25417

Change-Id: I4124037154aaaa8e48d300c23974f125b6055a1c
Reviewed-on: https://go-review.googlesource.com/130397
Run-TryBot: Filippo Valsorda <[email protected]>
Reviewed-by: Brad Fitzpatrick <[email protected]>
Reviewed-by: Filippo Valsorda <[email protected]>
TryBot-Result: Gobot Gobot <[email protected]>

Author: Andreas Auernhammer

src/crypto/rc4/rc4.go

@@ -54,12 +54,9 @@ func (c *Cipher) Reset() {
 	c.i, c.j = 0, 0
 }
 
-// xorKeyStreamGeneric sets dst to the result of XORing src with the
-// key stream. Dst and src must overlap entirely or not at all.
-//
-// This is the pure Go version. rc4_{amd64,386,arm}* contain assembly
-// implementations. This is here for tests and to prevent bitrot.
-func (c *Cipher) xorKeyStreamGeneric(dst, src []byte) {
+// XORKeyStream sets dst to the result of XORing src with the key stream.
+// Dst and src must overlap entirely or not at all.
+func (c *Cipher) XORKeyStream(dst, src []byte) {
 	if len(src) == 0 {
 		return
 	}