Cvs missed adapted module itself, here it goes...
authorAndy Polyakov <appro@openssl.org>
Tue, 3 May 2005 23:03:31 +0000 (23:03 +0000)
committerAndy Polyakov <appro@openssl.org>
Tue, 3 May 2005 23:03:31 +0000 (23:03 +0000)
crypto/md5/asm/md5-x86_64.pl

index f570f05..c36a7fe 100755 (executable)
@@ -105,19 +105,17 @@ EOF
 }
 
 my $output = shift;
-open STDOUT,">$output" or die "can't open $output: $!";
+open STDOUT,"| $^X ../perlasm/x86_64-xlate.pl $output";
 
 $code .= <<EOF;
 .text
 .align 16
 
 .globl md5_block_asm_host_order
-.type md5_block_asm_host_order,\@function
+.type md5_block_asm_host_order,\@function,3
 md5_block_asm_host_order:
        push    %rbp
        push    %rbx
-       push    %r12
-       push    %r13
        push    %r14
        push    %r15
 
@@ -131,7 +129,6 @@ md5_block_asm_host_order:
        mov     1*4(%rbp),      %ebx    # ebx = ctx->B
        mov     2*4(%rbp),      %ecx    # ecx = ctx->C
        mov     3*4(%rbp),      %edx    # edx = ctx->D
-       push    %rbp                    # save ctx
        # end is 'rdi'
        # ptr is 'rsi'
        # A is 'eax'
@@ -140,10 +137,10 @@ md5_block_asm_host_order:
        # D is 'edx'
 
        cmp     %rdi,           %rsi            # cmp end with ptr
-       je      1f                              # jmp if ptr == end
+       je      .Lend                           # jmp if ptr == end
 
        # BEGIN of loop over 16-word blocks
-2:     # save old values of A, B, C, D
+.Lloop:        # save old values of A, B, C, D
        mov     %eax,           %r8d
        mov     %ebx,           %r9d
        mov     %ecx,           %r14d
@@ -226,10 +223,10 @@ $code .= <<EOF;
        # loop control
        add     \$64,           %rsi            # ptr += 64
        cmp     %rdi,           %rsi            # cmp end with ptr
-       jb      2b                              # jmp if ptr < end
+       jb      .Lloop                          # jmp if ptr < end
        # END of loop over 16-word blocks
 
-1:     pop     %rbp                            # restore ctx
+.Lend:
        mov     %eax,           0*4(%rbp)       # ctx->A = A
        mov     %ebx,           1*4(%rbp)       # ctx->B = B
        mov     %ecx,           2*4(%rbp)       # ctx->C = C
@@ -237,13 +234,12 @@ $code .= <<EOF;
 
        pop     %r15
        pop     %r14
-       pop     %r13
-       pop     %r12
        pop     %rbx
        pop     %rbp
        ret
-.L_md5_block_asm_host_order_end:
-.size md5_block_asm_host_order,.L_md5_block_asm_host_order_end-md5_block_asm_host_order
+.size md5_block_asm_host_order,.-md5_block_asm_host_order
 EOF
 
 print $code;
+
+close STDOUT;