Index: openssl/crypto/sha/asm/sha1-x86_64.pl |
=================================================================== |
--- openssl/crypto/sha/asm/sha1-x86_64.pl (revision 105093) |
+++ openssl/crypto/sha/asm/sha1-x86_64.pl (working copy) |
@@ -29,14 +29,18 @@ |
# Xeon P4 +65% +0% 9.9 |
# Core2 +60% +10% 7.0 |
-$output=shift; |
+$flavour = shift; |
+$output = shift; |
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; } |
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/); |
+ |
$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1; |
( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or |
( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or |
die "can't locate x86_64-xlate.pl"; |
-open STDOUT,"| $^X $xlate $output"; |
+open STDOUT,"| $^X $xlate $flavour $output"; |
$ctx="%rdi"; # 1st arg |
$inp="%rsi"; # 2nd arg |
@@ -69,13 +73,14 @@ |
push %rbx |
push %rbp |
push %r12 |
- mov %rsp,%rax |
+ mov %rsp,%r11 |
mov %rdi,$ctx # reassigned argument |
sub \$`8+16*4`,%rsp |
mov %rsi,$inp # reassigned argument |
and \$-64,%rsp |
mov %rdx,$num # reassigned argument |
- mov %rax,`16*4`(%rsp) |
+ mov %r11,`16*4`(%rsp) |
+.Lprologue: |
mov 0($ctx),$A |
mov 4($ctx),$B |
@@ -88,10 +93,12 @@ |
sub EPILOGUE { |
my $func=shift; |
$code.=<<___; |
- mov `16*4`(%rsp),%rsp |
- pop %r12 |
- pop %rbp |
- pop %rbx |
+ mov `16*4`(%rsp),%rsi |
+ mov (%rsi),%r12 |
+ mov 8(%rsi),%rbp |
+ mov 16(%rsi),%rbx |
+ lea 24(%rsi),%rsp |
+.Lepilogue: |
ret |
.size $func,.-$func |
___ |
@@ -233,8 +240,110 @@ |
&EPILOGUE("sha1_block_data_order"); |
$code.=<<___; |
.asciz "SHA1 block transform for x86_64, CRYPTOGAMS by <appro\@openssl.org>" |
+.align 16 |
___ |
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame, |
+# CONTEXT *context,DISPATCHER_CONTEXT *disp) |
+if ($win64) { |
+$rec="%rcx"; |
+$frame="%rdx"; |
+$context="%r8"; |
+$disp="%r9"; |
+ |
+$code.=<<___; |
+.extern __imp_RtlVirtualUnwind |
+.type se_handler,\@abi-omnipotent |
+.align 16 |
+se_handler: |
+ push %rsi |
+ push %rdi |
+ push %rbx |
+ push %rbp |
+ push %r12 |
+ push %r13 |
+ push %r14 |
+ push %r15 |
+ pushfq |
+ sub \$64,%rsp |
+ |
+ mov 120($context),%rax # pull context->Rax |
+ mov 248($context),%rbx # pull context->Rip |
+ |
+ lea .Lprologue(%rip),%r10 |
+ cmp %r10,%rbx # context->Rip<.Lprologue |
+ jb .Lin_prologue |
+ |
+ mov 152($context),%rax # pull context->Rsp |
+ |
+ lea .Lepilogue(%rip),%r10 |
+ cmp %r10,%rbx # context->Rip>=.Lepilogue |
+ jae .Lin_prologue |
+ |
+ mov `16*4`(%rax),%rax # pull saved stack pointer |
+ lea 24(%rax),%rax |
+ |
+ mov -8(%rax),%rbx |
+ mov -16(%rax),%rbp |
+ mov -24(%rax),%r12 |
+ mov %rbx,144($context) # restore context->Rbx |
+ mov %rbp,160($context) # restore context->Rbp |
+ mov %r12,216($context) # restore context->R12 |
+ |
+.Lin_prologue: |
+ mov 8(%rax),%rdi |
+ mov 16(%rax),%rsi |
+ mov %rax,152($context) # restore context->Rsp |
+ mov %rsi,168($context) # restore context->Rsi |
+ mov %rdi,176($context) # restore context->Rdi |
+ |
+ mov 40($disp),%rdi # disp->ContextRecord |
+ mov $context,%rsi # context |
+ mov \$154,%ecx # sizeof(CONTEXT) |
+ .long 0xa548f3fc # cld; rep movsq |
+ |
+ mov $disp,%rsi |
+ xor %rcx,%rcx # arg1, UNW_FLAG_NHANDLER |
+ mov 8(%rsi),%rdx # arg2, disp->ImageBase |
+ mov 0(%rsi),%r8 # arg3, disp->ControlPc |
+ mov 16(%rsi),%r9 # arg4, disp->FunctionEntry |
+ mov 40(%rsi),%r10 # disp->ContextRecord |
+ lea 56(%rsi),%r11 # &disp->HandlerData |
+ lea 24(%rsi),%r12 # &disp->EstablisherFrame |
+ mov %r10,32(%rsp) # arg5 |
+ mov %r11,40(%rsp) # arg6 |
+ mov %r12,48(%rsp) # arg7 |
+ mov %rcx,56(%rsp) # arg8, (NULL) |
+ call *__imp_RtlVirtualUnwind(%rip) |
+ |
+ mov \$1,%eax # ExceptionContinueSearch |
+ add \$64,%rsp |
+ popfq |
+ pop %r15 |
+ pop %r14 |
+ pop %r13 |
+ pop %r12 |
+ pop %rbp |
+ pop %rbx |
+ pop %rdi |
+ pop %rsi |
+ ret |
+.size se_handler,.-se_handler |
+ |
+.section .pdata |
+.align 4 |
+ .rva .LSEH_begin_sha1_block_data_order |
+ .rva .LSEH_end_sha1_block_data_order |
+ .rva .LSEH_info_sha1_block_data_order |
+ |
+.section .xdata |
+.align 8 |
+.LSEH_info_sha1_block_data_order: |
+ .byte 9,0,0,0 |
+ .rva se_handler |
+___ |
+} |
+ |
#################################################################### |
$code =~ s/\`([^\`]*)\`/eval $1/gem; |