This removes all but the generic C RC4 implementation. At this point we
want to optimize for size/simplicity rather than speed.

See also upstream's 3e9e810f2e which
removed the RC4_CHUNK code and standardized on RC4_INDEX. A
since-removed comment says that it was implemented for "pre-21164a Alpha
CPUs don't have byte load/store instructions" and helps with SPARC and
MIPS.

This also removes all the manual loop unrolling.

Change-Id: I91135568483260b2e1e675f190fb00ce8f9eff3d
Reviewed-on: https://boringssl-review.googlesource.com/10720
Commit-Queue: David Benjamin <davidben@google.com>
Commit-Queue: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>