Change labs to std::abs for Windows ARM64

Because Windows ARM64 is in LLP64 data model, the long type as input parameter
for labs is still 32-bit. std::abs has overloads for both long and long long type, which
works for both LP64 and LLP64 data model (see below links).

https://docs.microsoft.com/en-us/cpp/c-runtime-library/reference/abs-labs-llabs-abs64
https://en.cppreference.com/w/cpp/numeric/math/abs

Bug: chromium:893460
Change-Id: I477d22d550bd92bf9dcc974ac84b4fe5222c1e13
Reviewed-on: https://chromium-review.googlesource.com/c/1448715Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Jaroslav Sevcik <jarin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#59358}

test/cctest/test-assembler-arm64.cc

@@ -14531,7 +14531,7 @@ static void AbsHelperX(int64_t value) {
   __ Mov(x1, value);
 
   if (value != kXMinInt) {
-    expected = labs(value);
+    expected = std::abs(value);
 
     Label next;
     // The result is representable.
@@ -14541,7 +14541,7 @@ static void AbsHelperX(int64_t value) {
     __ Bind(&next);
     __ Abs(x13, x1, nullptr, &done);
   } else {
-    // labs is undefined for kXMinInt but our implementation in the
+    // std::abs is undefined for kXMinInt but our implementation in the
     // MacroAssembler will return kXMinInt in such a case.
     expected = kXMinInt;