[BOLT] Check if symbol is in data area of function (#160143)

There are cases in which `getEntryIDForSymbol` is called, where the
given Symbol is in a constant island, and so BOLT can not find its
function. This causes BOLT to reach `llvm_unreachable("symbol not
found")` and crash. This patch adds a check that avoids this crash.

bolt/include/bolt/Core/BinaryFunction.h

@@ -683,7 +683,8 @@ private:
   ///
   /// Prefer to use BinaryContext::getFunctionForSymbol(EntrySymbol, &ID)
   /// instead of calling this function directly.
-  uint64_t getEntryIDForSymbol(const MCSymbol *EntrySymbol) const;
+  std::optional<uint64_t>
+  getEntryIDForSymbol(const MCSymbol *EntrySymbol) const;
 
   /// If the function represents a secondary split function fragment, set its
   /// parent fragment to \p BF.

bolt/lib/Core/BinaryContext.cpp

@@ -2523,8 +2523,10 @@ BinaryFunction *BinaryContext::getFunctionForSymbol(const MCSymbol *Symbol,
     return nullptr;
 
   BinaryFunction *BF = BFI->second;
-  if (EntryDesc)
-    *EntryDesc = BF->getEntryIDForSymbol(Symbol);
+  if (EntryDesc) {
+    std::optional<uint64_t> EntryID = BF->getEntryIDForSymbol(Symbol);
+    *EntryDesc = EntryID.value_or(0);
+  }
 
   return BF;
 }

bolt/lib/Core/BinaryFunction.cpp

@@ -1908,7 +1908,7 @@ bool BinaryFunction::scanExternalRefs() {
 }
 
 bool BinaryFunction::validateInternalBranches() {
-  if (!isSimple() || TrapsOnEntry)
+  if (!hasInstructions() || !isSimple() || TrapsOnEntry)
     return true;
 
   for (const auto &KV : Labels) {
@@ -3885,8 +3885,9 @@ MCSymbol *BinaryFunction::getSymbolForEntryID(uint64_t EntryID) {
   return nullptr;
 }
 
-uint64_t BinaryFunction::getEntryIDForSymbol(const MCSymbol *Symbol) const {
-  if (!isMultiEntry())
+std::optional<uint64_t>
+BinaryFunction::getEntryIDForSymbol(const MCSymbol *Symbol) const {
+  if (!isMultiEntry() || !Symbol)
     return 0;
 
   for (const MCSymbol *FunctionSymbol : getSymbols())
@@ -3912,8 +3913,7 @@ uint64_t BinaryFunction::getEntryIDForSymbol(const MCSymbol *Symbol) const {
       return NumEntries;
     ++NumEntries;
   }
-
-  llvm_unreachable("symbol not found");
+  return std::nullopt;
 }
 
 bool BinaryFunction::forEachEntryPoint(EntryPointCallbackTy Callback) const {

bolt/test/AArch64/check-symbol-area.s

@@ -0,0 +1,40 @@
+// This test checks that when looking for a function corresponding to a
+// symbol, BOLT is not looking through a data area (constant island).
+
+# RUN: llvm-mc -filetype=obj -triple aarch64-unknown-unknown %s -o %t.o
+# RUN: %clang %cflags %t.o -o %t.exe -Wl,-q
+# RUN: llvm-bolt %t.exe -o %t.bolt 2>&1 | FileCheck %s
+
+// Before adding a check for constant islands, BOLT would exit with an error
+// of the form: "symbol not found" and throw an LLVM UNREACHABLE error.
+# CHECK-NOT: symbol not found
+# CHECK-NOT: UNREACHABLE
+
+// Now BOLT throws a warning and does not crash.
+# CHECK: BOLT-WARNING: corrupted control flow detected in function main{{.*}}:
+# CHECK-SAME: an external branch/call targets an invalid instruction in
+# CHECK-SAME: function foo{{.*}} at address {{.*}}; ignoring both functions
+
+# CHECK: BOLT-WARNING: ignoring entry point at address 0x{{[0-9a-f]+}} in constant island of function foo{{.*}}
+
+.text
+.global main
+.type main, %function
+main:
+        add     x0, x1, x1
+        bl      first_block
+        ret
+
+.type foo, %function
+foo:
+   nop
+$d:
+first_block:
+        add     x0, x1, x1
+        bl      second_block
+        ret
+
+$x:
+second_block:
+        add     x0, x1, x1
+        ret