[llvm] [BOLT] Simplify RAState helpers (NFCI) (PR #162820)

[Peter Waller via llvm-commits]

================
@@ -81,10 +86,15 @@ void InsertNegateRAState::coverFunctionFragmentStart(BinaryFunction &BF,
       });
   // If a function is already split in the input, the first FF can also start
   // with Signed state. This covers that scenario as well.
-  if (BC.MIB->isRASigned(*((*FirstNonEmpty)->begin()))) {
+  auto RAState = BC.MIB->getRAState(*(*FirstNonEmpty)->begin());
+  if (!RAState) {
+    BC.errs() << "BOLT-ERROR: unknown RAState after inferUnknownStates "
+              << " in function " << BF.getPrintName() << "\n";
+    return;
----------------
peterwaller-arm wrote:

>From https://llvm.org/docs/CodingStandards.html:

> The ``llvm_unreachable`` function can be used to document areas of control flow
> that should never be entered if the program invariants hold:
> 
> .. code-block:: c++
> 
>   enum { Foo, Bar, Baz } X = foo();
> 
>   switch (X) {
>     case Foo: /* Handle Foo */; break;
>     case Bar: /* Handle Bar */; break;
>     default:
>       llvm_unreachable("X should be Foo or Bar here");
>   }
> 
> Additionally, ``reportFatalInternalError`` can be used to report invariant
> violations even in builds that do not enable assertions:
> 
> .. code-block:: c++
> 
>   if (VerifyFooAnalysis && !Foo.verify()) {
>     reportFatalInternalError("Analysis 'foo' not preserved");
>   }

The coding standards suggest llvm_unreachable may only be visible in assertions builds. If there is a risk this could happen in the real world, a fatal internal error would be preferable. Bolt doesn't use this function, the standard appears to be to simply exit if there is nothing more to be done.

https://github.com/llvm/llvm-project/pull/162820