ART: Target-dependent stack overflow, less check elision

Refactor the separate stack overflow reserved sizes from thread.h
into instruction_set.h and make sure they're used in the compiler.

Refactor the decision on when to elide stack overflow checks:
especially with large interpreter stack frames, it is not a good
idea to elide checks when the frame size is even close to the
reserved size. Currently enforce checks when the frame size is
>= 2KB, but make sure that frame sizes 1KB and below will elide
the checks (number from experience).

Bug: 15728765
Change-Id: I016bfd3d8218170cbccbd123ed5e2203db167c06
diff --git a/compiler/dex/quick/x86/call_x86.cc b/compiler/dex/quick/x86/call_x86.cc
index dd5dab2..28195ab 100644
--- a/compiler/dex/quick/x86/call_x86.cc
+++ b/compiler/dex/quick/x86/call_x86.cc
@@ -235,8 +235,8 @@
    * We can safely skip the stack overflow check if we're
    * a leaf *and* our frame size < fudge factor.
    */
-  const bool skip_overflow_check = (mir_graph_->MethodIsLeaf() &&
-      (static_cast<size_t>(frame_size_) < Thread::kStackOverflowReservedBytes));
+  const bool skip_overflow_check = mir_graph_->MethodIsLeaf() &&
+      !IsLargeFrame(frame_size_, Gen64Bit() ? kX86_64 : kX86);
   NewLIR0(kPseudoMethodEntry);
   /* Spill core callee saves */
   SpillCoreRegs();