Optimization for Poly1305

Rearranging this code seems to let both GCC and Clang do a little
better on the core loop, 4-7% depending on buffer size on my i7-6700k

1 files changed, 1 insertions, 1 deletions

diff --git a/news.rst b/news.rst
index e4c7b0b7a..1fe7f4468 100644
--- a/news.rst
+++ b/news.rst
@@ -77,7 +77,7 @@ Version 2.8.0, Not Yet Released
 * Correct bugs which would cause CFB, OCB, and GCM modes to crash when they
   were used in an unkeyed state. (GH #1639)
 
-* Optimizations for SM4
+* Optimizations for SM4 and Poly1305
 
 * Avoid a cache side channel in the AES key schedule