Optimize and cleanup 32-bit and 64-bit endianess transforms

* cipher/bithelp.h (bswap32, bswap64, le_bswap32, be_bswap32)
(le_bswap64, be_bswap64): New.
* cipher/bufhelp.h (buf_get_be32, buf_get_le32, buf_put_le32)
(buf_put_be32, buf_get_be64, buf_get_le64, buf_put_be64)
(buf_put_le64): New.
* cipher/blowfish.c (do_encrypt_block, do_decrypt_block): Use new
endian conversion helpers.
(do_bf_setkey): Turn endian specific code to generic.
* cipher/camellia.c (GETU32, PUTU32): Use new endian conversion
helpers.
* cipher/cast5.c (rol): Remove, use rol from bithelp.
(F1, F2, F3): Fix to use rol from bithelp.
(do_encrypt_block, do_decrypt_block, do_cast_setkey): Use new endian
conversion helpers.
* cipher/des.c (READ_64BIT_DATA, WRITE_64BIT_DATA): Ditto.
* cipher/md4.c (transform, md4_final): Ditto.
* cipher/md5.c (transform, md5_final): Ditto.
* cipher/rmd160.c (transform, rmd160_final): Ditto.
* cipher/salsa20.c (LE_SWAP32, LE_READ_UINT32): Ditto.
* cipher/scrypt.c (READ_UINT64, LE_READ_UINT64, LE_SWAP32): Ditto.
* cipher/seed.c (GETU32, PUTU32): Ditto.
* cipher/serpent.c (byte_swap_32): Remove.
(serpent_key_prepare, serpent_encrypt_internal)
(serpent_decrypt_internal): Use new endian conversion helpers.
* cipher/sha1.c (transform, sha1_final): Ditto.
* cipher/sha256.c (transform, sha256_final): Ditto.
* cipher/sha512.c (__transform, sha512_final): Ditto.
* cipher/stribog.c (transform, stribog_final): Ditto.
* cipher/tiger.c (transform, tiger_final): Ditto.
* cipher/twofish.c (INPACK, OUTUNPACK): Ditto.
* cipher/whirlpool.c (buffer_to_block, block_to_buffer): Ditto.
* configure.ac (gcry_cv_have_builtin_bswap32): Check for compiler
provided __builtin_bswap32.
(gcry_cv_have_builtin_bswap64): Check for compiler provided
__builtin_bswap64.
--

Patch add helper functions that provide conversions to/from integers and
buffers of different endianess. Benefits are code cleanup and optimization
for architectures that have byte-swaping instructions and/or can do fast
unaligned memory accesses.

Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>

1 files changed, 10 insertions, 38 deletions

diff --git a/cipher/tiger.c b/cipher/tiger.c
index 8f5959b8..df16098e 100644
--- a/cipher/tiger.c
+++ b/cipher/tiger.c
@@ -28,6 +28,8 @@
 #include "g10lib.h"
 #include "cipher.h"
 #include "hash-common.h"
+#include "bithelp.h"
+#include "bufhelp.h"
 
 /* We really need a 64 bit type for this code.  */
 #ifdef HAVE_U64_TYPEDEF
@@ -697,24 +699,10 @@ transform ( void *ctx, const unsigned char *data )
   TIGER_CONTEXT *hd = ctx;
   u64 a,b,c,aa,bb,cc;
   u64 x[8];
-#ifdef WORDS_BIGENDIAN
-#define MKWORD(d,n) \
-		(  ((u64)(d)[8*(n)+7]) << 56 | ((u64)(d)[8*(n)+6]) << 48  \
-		 | ((u64)(d)[8*(n)+5]) << 40 | ((u64)(d)[8*(n)+4]) << 32  \
-		 | ((u64)(d)[8*(n)+3]) << 24 | ((u64)(d)[8*(n)+2]) << 16  \
-		 | ((u64)(d)[8*(n)+1]) << 8  | ((u64)(d)[8*(n)	])	 )
-  x[0] = MKWORD(data, 0);
-  x[1] = MKWORD(data, 1);
-  x[2] = MKWORD(data, 2);
-  x[3] = MKWORD(data, 3);
-  x[4] = MKWORD(data, 4);
-  x[5] = MKWORD(data, 5);
-  x[6] = MKWORD(data, 6);
-  x[7] = MKWORD(data, 7);
-#undef MKWORD
-#else
-  memcpy( &x[0], data, 64 );
-#endif
+  int i;
+
+  for ( i = 0; i < 8; i++ )
+    x[i] = buf_get_le64(data + i * 8);
 
   /* save */
   a = aa = hd->a;
@@ -783,30 +771,14 @@ tiger_final( void *context )
       memset(hd->bctx.buf, 0, 56 ); /* fill next block with zeroes */
     }
   /* append the 64 bit count */
-  hd->bctx.buf[56] = lsb	   ;
-  hd->bctx.buf[57] = lsb >>  8;
-  hd->bctx.buf[58] = lsb >> 16;
-  hd->bctx.buf[59] = lsb >> 24;
-  hd->bctx.buf[60] = msb	   ;
-  hd->bctx.buf[61] = msb >>  8;
-  hd->bctx.buf[62] = msb >> 16;
-  hd->bctx.buf[63] = msb >> 24;
+  buf_put_le32(hd->bctx.buf + 56, lsb);
+  buf_put_le32(hd->bctx.buf + 60, msb);
   burn = transform( hd, hd->bctx.buf );
   _gcry_burn_stack (burn);
 
   p = hd->bctx.buf;
-#ifdef WORDS_BIGENDIAN
-#define X(a) do { *(u64*)p = hd->a ; p += 8; } while(0)
-#else /* little endian */
-#define X(a) do { *p++ = hd->a >> 56; *p++ = hd->a >> 48; \
-	          *p++ = hd->a >> 40; *p++ = hd->a >> 32; \
-	          *p++ = hd->a >> 24; *p++ = hd->a >> 16; \
-	          *p++ = hd->a >>  8; *p++ = hd->a;       } while(0)
-#endif
-#define Y(a) do { *p++ = hd->a      ; *p++ = hd->a >> 8;  \
-	          *p++ = hd->a >> 16; *p++ = hd->a >> 24; \
-	          *p++ = hd->a >> 32; *p++ = hd->a >> 40; \
-	          *p++ = hd->a >> 48; *p++ = hd->a >> 56; } while(0)
+#define X(a) do { *(u64*)p = be_bswap64(hd->a); p += 8; } while(0)
+#define Y(a) do { *(u64*)p = le_bswap64(hd->a); p += 8; } while(0)
   if (hd->variant == 0)
     {
       X(a);