Make snmalloc build on Windows with Clang

Fixes a few places where Clang complains about Windows specific code,
and also uses macros supported by Clang on Windows.  A few places
separating platform and compiler specific code, as MSVC and WIN32 were
used interchangably previously.

src/ds/aba.h

@@ -97,7 +97,7 @@ namespace snmalloc
 #  if defined(_MSC_VER) && defined(SNMALLOC_VA_BITS_64)
       return _InterlockedCompareExchange128(
         (volatile __int64*)&linked,
-        expect.aba + 1,
+        (__int64)(expect.aba + (uintptr_t)1),
         (__int64)value,
         (__int64*)&expect);
 #  else

src/ds/bits.h

@@ -138,13 +138,16 @@ namespace snmalloc
 
     inline size_t ctz(size_t x)
     {
-#if defined(_MSC_VER)
+#if __has_builtin(__builtin_ctzl)
+      return static_cast<size_t>(__builtin_ctzl(x));
+#elif defined(_MSC_VER)
 #  ifdef SNMALLOC_VA_BITS_64
       return _tzcnt_u64(x);
 #  else
       return _tzcnt_u32((uint32_t)x);
 #  endif
 #else
+      // Probably GCC at this point.
       return static_cast<size_t>(__builtin_ctzl(x));
 #endif
     }

src/ds/defines.h

@@ -1,6 +1,6 @@
 #pragma once
 
-#ifdef _MSC_VER
+#if defined(_MSC_VER) && !defined(__clang__)
 #  define ALWAYSINLINE __forceinline
 #  define NOINLINE __declspec(noinline)
 #  define likely(x) !!(x)