Implement the second set of SHA-1 functions in assembler; add some

macros in an attempt to keep the code reasonably clean.

2 files changed, 38 insertions, 36 deletions

diff --git a/modules/sha_x86/sha160.cpp b/modules/sha_x86/sha160.cpp
index fce5c9d04..d591eb6d8 100644
--- a/modules/sha_x86/sha160.cpp
+++ b/modules/sha_x86/sha160.cpp
@@ -77,20 +77,14 @@ void SHA_160::hash(const byte input[])
 #endif
 
    /*
-   F1(A,B,C,D,E,W[ 0]); F1(E,A,B,C,D,W[ 1]);   F1(D,E,A,B,C,W[ 2]);
-   F1(C,D,E,A,B,W[ 3]);   F1(B,C,D,E,A,W[ 4]);
-
-
-   F1(A,B,C,D,E,W[ 5]);
+   F1(A,B,C,D,E,W[ 0]);   F1(E,A,B,C,D,W[ 1]);   F1(D,E,A,B,C,W[ 2]);
+   F1(C,D,E,A,B,W[ 3]);   F1(B,C,D,E,A,W[ 4]);   F1(A,B,C,D,E,W[ 5]);
    F1(E,A,B,C,D,W[ 6]);   F1(D,E,A,B,C,W[ 7]);   F1(C,D,E,A,B,W[ 8]);
-   F1(B,C,D,E,A,W[ 9]);
-
-   F1(A,B,C,D,E,W[10]);   F1(E,A,B,C,D,W[11]);
+   F1(B,C,D,E,A,W[ 9]);   F1(A,B,C,D,E,W[10]);   F1(E,A,B,C,D,W[11]);
    F1(D,E,A,B,C,W[12]);   F1(C,D,E,A,B,W[13]);   F1(B,C,D,E,A,W[14]);
-
    F1(A,B,C,D,E,W[15]);   F1(E,A,B,C,D,W[16]);   F1(D,E,A,B,C,W[17]);
    F1(C,D,E,A,B,W[18]);   F1(B,C,D,E,A,W[19]);
-   */
+
 
    F2(A,B,C,D,E,W[20]);   F2(E,A,B,C,D,W[21]);   F2(D,E,A,B,C,W[22]);
    F2(C,D,E,A,B,W[23]);   F2(B,C,D,E,A,W[24]);   F2(A,B,C,D,E,W[25]);
@@ -99,6 +93,7 @@ void SHA_160::hash(const byte input[])
    F2(D,E,A,B,C,W[32]);   F2(C,D,E,A,B,W[33]);   F2(B,C,D,E,A,W[34]);
    F2(A,B,C,D,E,W[35]);   F2(E,A,B,C,D,W[36]);   F2(D,E,A,B,C,W[37]);
    F2(C,D,E,A,B,W[38]);   F2(B,C,D,E,A,W[39]);
+   */
 
    F3(A,B,C,D,E,W[40]);   F3(E,A,B,C,D,W[41]);   F3(D,E,A,B,C,W[42]);
    F3(C,D,E,A,B,W[43]);   F3(B,C,D,E,A,W[44]);   F3(A,B,C,D,E,W[45]);
diff --git a/modules/sha_x86/sha1core.S b/modules/sha_x86/sha1core.S
index 5a09cb00d..17352688d 100644
--- a/modules/sha_x86/sha1core.S
+++ b/modules/sha_x86/sha1core.S
@@ -53,6 +53,7 @@ sha160_core:
    movl 16(%ebp), %esi
 
 #define MAGIC1 $0x5A827999
+#define MAGIC2 $0x6ED9EBA1
 
 #define FUNC1(B, C, D, TEMP)   \
    movl C, TEMP                 ; \
@@ -65,9 +66,10 @@ sha160_core:
    xorl C, TEMP                 ; \
    xorl D, TEMP  
 
-#define F(A, B, C, D, E, TEMP, MSG, MAGIC, FUNC) \
+#define F(A, B, C, D, E, TEMP, MAGIC, FUNC) \
    FUNC(B, C, D, TEMP)          ; \
-   addl MSG, E                  ; \
+   addl 0(%edi), E            ; \
+   addl $4, %edi                ; \
    addl TEMP, E                 ; \
    addl MAGIC, E                ; \
    roll $5, A                   ; \
@@ -75,30 +77,35 @@ sha160_core:
    rorl $5, A                   ; \
    roll $30, B
 
-    F(%eax, %ebx, %ecx, %edx, %esi, %ebp,  0(%edi), MAGIC1, FUNC1)
-    F(%esi, %eax, %ebx, %ecx, %edx, %ebp,  4(%edi), MAGIC1, FUNC1)
-    F(%edx, %esi, %eax, %ebx, %ecx, %ebp,  8(%edi), MAGIC1, FUNC1)
-    F(%ecx, %edx, %esi, %eax, %ebx, %ebp, 12(%edi), MAGIC1, FUNC1)
-    F(%ebx, %ecx, %edx, %esi, %eax, %ebp, 16(%edi), MAGIC1, FUNC1)
-
-    F(%eax, %ebx, %ecx, %edx, %esi, %ebp, 20(%edi), MAGIC1, FUNC1)
-    F(%esi, %eax, %ebx, %ecx, %edx, %ebp, 24(%edi), MAGIC1, FUNC1)
-    F(%edx, %esi, %eax, %ebx, %ecx, %ebp, 28(%edi), MAGIC1, FUNC1)
-    F(%ecx, %edx, %esi, %eax, %ebx, %ebp, 32(%edi), MAGIC1, FUNC1)
-    F(%ebx, %ecx, %edx, %esi, %eax, %ebp, 36(%edi), MAGIC1, FUNC1)
-
-    F(%eax, %ebx, %ecx, %edx, %esi, %ebp, 40(%edi), MAGIC1, FUNC1)
-    F(%esi, %eax, %ebx, %ecx, %edx, %ebp, 44(%edi), MAGIC1, FUNC1)
-    F(%edx, %esi, %eax, %ebx, %ecx, %ebp, 48(%edi), MAGIC1, FUNC1)
-    F(%ecx, %edx, %esi, %eax, %ebx, %ebp, 52(%edi), MAGIC1, FUNC1)
-    F(%ebx, %ecx, %edx, %esi, %eax, %ebp, 56(%edi), MAGIC1, FUNC1)
-
-    F(%eax, %ebx, %ecx, %edx, %esi, %ebp, 60(%edi), MAGIC1, FUNC1)
-    F(%esi, %eax, %ebx, %ecx, %edx, %ebp, 64(%edi), MAGIC1, FUNC1)
-    F(%edx, %esi, %eax, %ebx, %ecx, %ebp, 68(%edi), MAGIC1, FUNC1)
-    F(%ecx, %edx, %esi, %eax, %ebx, %ebp, 72(%edi), MAGIC1, FUNC1)
-    F(%ebx, %ecx, %edx, %esi, %eax, %ebp, 76(%edi), MAGIC1, FUNC1)
- 
+#define F1(A, B, C, D, E, TEMP) \
+   F(A, B, C, D, E, TEMP, MAGIC1, FUNC1)
+
+#define F2(A, B, C, D, E, TEMP) \
+   F(A, B, C, D, E, TEMP, MAGIC2, FUNC2)
+
+#define F_BLOCK(F) \
+    F(%eax, %ebx, %ecx, %edx, %esi, %ebp) ; \
+    F(%esi, %eax, %ebx, %ecx, %edx, %ebp)  ; \
+    F(%edx, %esi, %eax, %ebx, %ecx, %ebp)  ; \
+    F(%ecx, %edx, %esi, %eax, %ebx, %ebp)  ; \
+    F(%ebx, %ecx, %edx, %esi, %eax, %ebp)
+/*
+    F1(%eax, %ebx, %ecx, %edx, %esi, %ebp)
+    F1(%esi, %eax, %ebx, %ecx, %edx, %ebp)
+    F1(%edx, %esi, %eax, %ebx, %ecx, %ebp)
+    F1(%ecx, %edx, %esi, %eax, %ebx, %ebp)
+    F1(%ebx, %ecx, %edx, %esi, %eax, %ebp)
+*/
+   F_BLOCK(F1)
+   F_BLOCK(F1)
+   F_BLOCK(F1)
+   F_BLOCK(F1)
+
+   F_BLOCK(F2)
+   F_BLOCK(F2)
+   F_BLOCK(F2)
+   F_BLOCK(F2)
+
    movl 20(%esp), %ebp
    movl %eax, 0(%ebp)
    movl %ebx, 4(%ebp)