1 files changed, 20 insertions, 70 deletions

diff --git a/src/lib/libcrypto/x86_64cpuid.pl b/src/lib/libcrypto/x86_64cpuid.pl
index 7b7b93b223..b771a8539d 100644
--- a/src/lib/libcrypto/x86_64cpuid.pl
+++ b/src/lib/libcrypto/x86_64cpuid.pl
@@ -7,23 +7,15 @@ 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 $flavour $output";
-
-($arg1,$arg2,$arg3,$arg4)=$win64?("%rcx","%rdx","%r8", "%r9") : # Win64 order
-                                 ("%rdi","%rsi","%rdx","%rcx"); # Unix order
+open STDOUT,"| $^X ${dir}perlasm/x86_64-xlate.pl $flavour $output";
 
+if ($win64)     { $arg1="%rcx"; $arg2="%rdx"; }
+else            { $arg1="%rdi"; $arg2="%rsi"; }
 print<<___;
+#include <machine/asm.h>
 .extern         OPENSSL_cpuid_setup
-.hidden         OPENSSL_cpuid_setup
 .section        .init
-        call    OPENSSL_cpuid_setup
-
-.hidden OPENSSL_ia32cap_P
-.comm   OPENSSL_ia32cap_P,8,4
+        call    PIC_PLT(OPENSSL_cpuid_setup)
 
 .text
 
@@ -55,7 +47,7 @@ OPENSSL_rdtsc:
 .type   OPENSSL_ia32_cpuid,\@abi-omnipotent
 .align  16
 OPENSSL_ia32_cpuid:
-        mov     %rbx,%r8                # save %rbx
+        mov     %rbx,%r8
 
         xor     %eax,%eax
         cpuid
@@ -87,15 +79,7 @@ OPENSSL_ia32_cpuid:
         # AMD specific
         mov     \$0x80000000,%eax
         cpuid
-        cmp     \$0x80000001,%eax
-        jb      .Lintel
-        mov     %eax,%r10d
-        mov     \$0x80000001,%eax
-        cpuid
-        or      %ecx,%r9d
-        and     \$0x00000801,%r9d       # isolate AMD XOP bit, 1<<11
-
-        cmp     \$0x80000008,%r10d
+        cmp     \$0x80000008,%eax
         jb      .Lintel
 
         mov     \$0x80000008,%eax
@@ -106,12 +90,12 @@ OPENSSL_ia32_cpuid:
         mov     \$1,%eax
         cpuid
         bt      \$28,%edx               # test hyper-threading bit
-        jnc     .Lgeneric
+        jnc     .Ldone
         shr     \$16,%ebx               # number of logical processors
         cmp     %r10b,%bl
-        ja      .Lgeneric
+        ja      .Ldone
         and     \$0xefffffff,%edx       # ~(1<<28)
-        jmp     .Lgeneric
+        jmp     .Ldone
 
 .Lintel:
         cmp     \$4,%r11d
@@ -128,47 +112,30 @@ OPENSSL_ia32_cpuid:
 .Lnocacheinfo:
         mov     \$1,%eax
         cpuid
-        and     \$0xbfefffff,%edx       # force reserved bits to 0
         cmp     \$0,%r9d
         jne     .Lnotintel
-        or      \$0x40000000,%edx       # set reserved bit#30 on Intel CPUs
+        or      \$0x00100000,%edx       # use reserved 20th bit to engage RC4_CHAR
         and     \$15,%ah
         cmp     \$15,%ah                # examine Family ID
-        jne     .Lnotintel
-        or      \$0x00100000,%edx       # set reserved bit#20 to engage RC4_CHAR
+        je      .Lnotintel
+        or      \$0x40000000,%edx       # use reserved bit to skip unrolled loop
 .Lnotintel:
         bt      \$28,%edx               # test hyper-threading bit
-        jnc     .Lgeneric
+        jnc     .Ldone
         and     \$0xefffffff,%edx       # ~(1<<28)
         cmp     \$0,%r10d
-        je      .Lgeneric
+        je      .Ldone
 
         or      \$0x10000000,%edx       # 1<<28
         shr     \$16,%ebx
         cmp     \$1,%bl                 # see if cache is shared
-        ja      .Lgeneric
+        ja      .Ldone
         and     \$0xefffffff,%edx       # ~(1<<28)
-.Lgeneric:
-        and     \$0x00000800,%r9d       # isolate AMD XOP flag
-        and     \$0xfffff7ff,%ecx
-        or      %ecx,%r9d               # merge AMD XOP flag
-
-        mov     %edx,%r10d              # %r9d:%r10d is copy of %ecx:%edx
-        bt      \$27,%r9d               # check OSXSAVE bit
-        jnc     .Lclear_avx
-        xor     %ecx,%ecx               # XCR0
-        .byte   0x0f,0x01,0xd0          # xgetbv
-        and     \$6,%eax                # isolate XMM and YMM state support
-        cmp     \$6,%eax
-        je      .Ldone
-.Lclear_avx:
-        mov     \$0xefffe7ff,%eax       # ~(1<<28|1<<12|1<<11)
-        and     %eax,%r9d               # clear AVX, FMA and AMD XOP bits
 .Ldone:
-        shl     \$32,%r9
-        mov     %r10d,%eax
-        mov     %r8,%rbx                # restore %rbx
-        or      %r9,%rax
+        shl     \$32,%rcx
+        mov     %edx,%eax
+        mov     %r8,%rbx
+        or      %rcx,%rax
         ret
 .size   OPENSSL_ia32_cpuid,.-OPENSSL_ia32_cpuid
 
@@ -263,21 +230,4 @@ OPENSSL_wipe_cpu:
 .size   OPENSSL_wipe_cpu,.-OPENSSL_wipe_cpu
 ___
 
-print<<___;
-.globl  OPENSSL_ia32_rdrand
-.type   OPENSSL_ia32_rdrand,\@abi-omnipotent
-.align  16
-OPENSSL_ia32_rdrand:
-        mov     \$8,%ecx
-.Loop_rdrand:
-        rdrand  %rax
-        jc      .Lbreak_rdrand
-        loop    .Loop_rdrand
-.Lbreak_rdrand:
-        cmp     \$0,%rax
-        cmove   %rcx,%rax
-        ret
-.size   OPENSSL_ia32_rdrand,.-OPENSSL_ia32_rdrand
-___
-
 close STDOUT;   # flush