remove fldcw/fstcw from Win64 builds

_MCW_PC (Precision control) is not supported on x64:
https://docs.microsoft.com/en-us/cpp/c-runtime-library/reference/control87-controlfp-control87-2

The x87 FPU is not used on Win64 or ARM so setting the x87 control word
is not necessary. The SSE/SSE2 and ARM FPUs don't have a precision
control - the precision is embedded in each instruction - so the need to
set the control word is also gone.

BUG=webm:1500

Change-Id: I014513282a7dc320d1cdeaec48249d98a66bf09f

vpx_ports/vpx_ports.mk

@@ -21,10 +21,6 @@ ifeq ($(ARCH_X86),yes)
 PORTS_SRCS-$(HAVE_MMX) += emms_mmx.c
 endif
 
-ifeq ($(ARCH_X86_64),yes)
-PORTS_SRCS-$(CONFIG_MSVS) += float_control_word.asm
-endif
-
 ifeq ($(ARCH_X86)$(ARCH_X86_64),yes)
 PORTS_SRCS-yes += x86.h
 PORTS_SRCS-yes += x86_abi_support.asm

vpx_ports/x86.h

@@ -295,11 +295,11 @@ static unsigned short x87_get_control_word(void) {
   return mode;
 }
 #elif ARCH_X86_64
-/* No fldcw intrinsics on Windows x64, punt to external asm */
-extern void vpx_winx64_fldcw(unsigned short mode);
-extern unsigned short vpx_winx64_fstcw(void);
-#define x87_set_control_word vpx_winx64_fldcw
-#define x87_get_control_word vpx_winx64_fstcw
+// Unsupported on Win64:
+// https://docs.microsoft.com/en-us/cpp/c-runtime-library/reference/control87-controlfp-control87-2
+// _MCW_PC (Precision control) (Not supported on ARM or x64 platforms.)
+static void x87_set_control_word(unsigned int mode) { (void)mode; }
+static unsigned int x87_get_control_word(void) { return 0; }
 #else
 static void x87_set_control_word(unsigned short mode) {
   __asm { fldcw mode }
@@ -313,6 +313,17 @@ static unsigned short x87_get_control_word(void) {
 
 static INLINE unsigned int x87_set_double_precision(void) {
   unsigned int mode = x87_get_control_word();
+  // Intel 64 and IA-32 Architectures Developer's Manual: Vol. 1
+  // https://www.intel.com/content/dam/www/public/us/en/documents/manuals/64-ia-32-architectures-software-developer-vol-1-manual.pdf
+  // 8.1.5.2 Precision Control Field
+  // Bits 8 and 9 (0x300) of the x87 FPU Control Word ("Precision Control")
+  // determine the number of bits used in floating point calculations. To match
+  // later SSE instructions restrict x87 operations to Double Precision (0x200).
+  // Precision                     PC Field
+  // Single Precision (24-Bits)    00B
+  // Reserved                      01B
+  // Double Precision (53-Bits)    10B
+  // Extended Precision (64-Bits)  11B
   x87_set_control_word((mode & ~0x300) | 0x200);
   return mode;
 }