From: Pádraig Brady <P@draigBrady.com>
Date: Tue, 25 Feb 2020 13:30:39 +0000 (+0000)
Subject: b2sum: sync better with upstream
X-Git-Tag: v8.32~16
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=ead8464e3acffdd6eedf44ad623bb00c943302dc;p=thirdparty%2Fcoreutils.git

b2sum: sync better with upstream

* src/blake2/blake2-impl.h: Sync load16() implementation,
which doesn't change code generation.
Also leverage (builtin) memcpy to more efficiently
move data on little endian systems,
giving a 2% win with GCC 9.2.1 on an i3-2310M.
---

diff --git a/src/blake2/blake2-impl.h b/src/blake2/blake2-impl.h
index 32ff316334..7ecc8a615a 100644
--- a/src/blake2/blake2-impl.h
+++ b/src/blake2/blake2-impl.h
@@ -15,6 +15,13 @@
 #ifndef BLAKE2_IMPL_H
 #define BLAKE2_IMPL_H
 
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+#ifndef WORDS_BIGENDIAN
+# define NATIVE_LITTLE_ENDIAN 1
+#endif
+
 #include <stdint.h>
 #include <string.h>
 
@@ -72,8 +79,8 @@ static BLAKE2_INLINE uint16_t load16( const void *src )
   return w;
 #else
   const uint8_t *p = ( const uint8_t * )src;
-  return (( uint16_t )( p[0] ) <<  0) |
-         (( uint16_t )( p[1] ) <<  8) ;
+  return ( uint16_t )((( uint32_t )( p[0] ) <<  0) |
+                      (( uint32_t )( p[1] ) <<  8));
 #endif
 }