[MSan] factor userspace-specific declarations into createUserspaceApi(). NFC

[glider]

Summary

This patch introduces createUserspaceApi() that creates function/global declarations for symbols used by MSan in the userspace.
This is a step towards the upcoming KMSAN implementation patch.

[eugenis]

lib/Transforms/Instrumentation/MemorySanitizer.cpp

596

This should probably be a property of a module, not a compiler invocation.

[glider]

lib/Transforms/Instrumentation/MemorySanitizer.cpp

596

We used to check for WarningFn value here, and are just using a special bool instead now.
It might be cleaner to make CallbacksInitialized a property of class MemorySanitizer, but how can it be a property of a module?

[eugenis]

lib/Transforms/Instrumentation/MemorySanitizer.cpp

596

Just consider that this bool can become true only once in the compiler process. No matter how many modules are created.

Moving it to class MemorySanitizer sounds good.

In fact, a FunctionPass is not supposed to create new functions (which AFAIK includes function declarations) outside of doInitialization(). MSan breaks this assumption in a bunch of places...

[glider]

Moved CallbackInitialized into class MemorySanitizer

[glider]

lib/Transforms/Instrumentation/MemorySanitizer.cpp

596

Just consider that this bool can become true only once in the compiler process. No matter how many modules are created.

IIRC we don't do that in other tools either, do we?

Moving it to class MemorySanitizer sounds good.

Done

In fact, a FunctionPass is not supposed to create new functions (which AFAIK includes function declarations) outside of doInitialization(). MSan breaks this assumption in a bunch of places...

We only create getAndInsertFunction() from initializeCallbacks().
Wonder if it's possible to just call it from doInitialization()

596

We only create getAndInsertFunction() from initializeCallbacks().

s/create/call

[eugenis]

lib/Transforms/Instrumentation/MemorySanitizer.cpp

596

IIRC we don't do that in other tools either, do we?

Do what?

We only create getAndInsertFunction() from initializeCallbacks().
Wonder if it's possible to just call it from doInitialization()

Yes, but only if at least on function in the module has sanitize_memory attribute.
We also mess with other functions attributes in visitCallSite(), and that can not be done ahead of time because some other function pass may revert the change.

[glider]

Landed r337155.

lib/Transforms/Instrumentation/MemorySanitizer.cpp

 private:
   friend struct MemorySanitizerVisitor;
   friend struct VarArgAMD64Helper;
   friend struct VarArgMIPS64Helper;
   friend struct VarArgAArch64Helper;
   friend struct VarArgPowerPC64Helper;
 
   void initializeCallbacks(Module &M);
+  void createUserspaceApi(Module &M);
 
   /// Track origins (allocation points) of uninitialized values.
   int TrackOrigins;
   bool Recover;
 
   LLVMContext *C;
   Type *IntptrTy;
   Type *OriginTy;
private:
   /// Thread-local shadow storage for va_arg overflow area
   /// (x86_64-specific).
   GlobalVariable *VAArgOverflowSizeTLS;
 
   /// Thread-local space used to pass origin value to the UMR reporting
   /// function.
   GlobalVariable *OriginTLS;
 
+  /// Are the instrumentation callbacks set up?
+  bool CallbacksInitialized = false;
+
   /// The run-time callback to print a warning.
-  Value *WarningFn = nullptr;
+  Value *WarningFn;
 
   // These arrays are indexed by log2(AccessSize).
   Value *MaybeWarningFn[kNumberOfAccessSizes];
   Value *MaybeStoreOriginFn[kNumberOfAccessSizes];
 
   /// Run-time helper that generates a new origin value for a stack
   /// allocation.
   Value *MsanSetAllocaOrigin4Fn;

 /// frame ID, so the string needs to be mutable.
 static GlobalVariable *createPrivateNonConstGlobalForString(Module &M,
                                                             StringRef Str) {
   Constant *StrConst = ConstantDataArray::getString(M.getContext(), Str);
   return new GlobalVariable(M, StrConst->getType(), /*isConstant=*/false,
                             GlobalValue::PrivateLinkage, StrConst, "");
 }
 
-/// Insert extern declaration of runtime-provided functions and globals.
-void MemorySanitizer::initializeCallbacks(Module &M) {
-  // Only do this once.
-  if (WarningFn)
-    return;
-
+/// Insert declarations for userspace-specific functions and globals.
+void MemorySanitizer::createUserspaceApi(Module &M) {
   IRBuilder<> IRB(*C);
   // Create the callback.
   // FIXME: this function should have "Cold" calling conv,
   // which is not yet implemented.
   StringRef WarningFnName = Recover ? "__msan_warning"
                                     : "__msan_warning_noreturn";
   WarningFn = M.getOrInsertFunction(WarningFnName, IRB.getVoidTy());
 
+  // Create the global TLS variables.
+  RetvalTLS = new GlobalVariable(
+      M, ArrayType::get(IRB.getInt64Ty(), kRetvalTLSSize / 8), false,
+      GlobalVariable::ExternalLinkage, nullptr, "__msan_retval_tls", nullptr,
+      GlobalVariable::InitialExecTLSModel);
+
+  RetvalOriginTLS = new GlobalVariable(
+      M, OriginTy, false, GlobalVariable::ExternalLinkage, nullptr,
+      "__msan_retval_origin_tls", nullptr, GlobalVariable::InitialExecTLSModel);
+
+  ParamTLS = new GlobalVariable(
+      M, ArrayType::get(IRB.getInt64Ty(), kParamTLSSize / 8), false,
+      GlobalVariable::ExternalLinkage, nullptr, "__msan_param_tls", nullptr,
+      GlobalVariable::InitialExecTLSModel);
+
+  ParamOriginTLS = new GlobalVariable(
+      M, ArrayType::get(OriginTy, kParamTLSSize / 4), false,
+      GlobalVariable::ExternalLinkage, nullptr, "__msan_param_origin_tls",
+      nullptr, GlobalVariable::InitialExecTLSModel);
+
+  VAArgTLS = new GlobalVariable(
+      M, ArrayType::get(IRB.getInt64Ty(), kParamTLSSize / 8), false,
+      GlobalVariable::ExternalLinkage, nullptr, "__msan_va_arg_tls", nullptr,
+      GlobalVariable::InitialExecTLSModel);
+  VAArgOverflowSizeTLS = new GlobalVariable(
+      M, IRB.getInt64Ty(), false, GlobalVariable::ExternalLinkage, nullptr,
+      "__msan_va_arg_overflow_size_tls", nullptr,
+      GlobalVariable::InitialExecTLSModel);
+  OriginTLS = new GlobalVariable(
+      M, IRB.getInt32Ty(), false, GlobalVariable::ExternalLinkage, nullptr,
+      "__msan_origin_tls", nullptr, GlobalVariable::InitialExecTLSModel);
+
   for (size_t AccessSizeIndex = 0; AccessSizeIndex < kNumberOfAccessSizes;
        AccessSizeIndex++) {
     unsigned AccessSize = 1 << AccessSizeIndex;
     std::string FunctionName = "__msan_maybe_warning_" + itostr(AccessSize);
     MaybeWarningFn[AccessSizeIndex] = M.getOrInsertFunction(
         FunctionName, IRB.getVoidTy(), IRB.getIntNTy(AccessSize * 8),
         IRB.getInt32Ty());
 
     FunctionName = "__msan_maybe_store_origin_" + itostr(AccessSize);
     MaybeStoreOriginFn[AccessSizeIndex] = M.getOrInsertFunction(
         FunctionName, IRB.getVoidTy(), IRB.getIntNTy(AccessSize * 8),
         IRB.getInt8PtrTy(), IRB.getInt32Ty());
   }
 
   MsanSetAllocaOrigin4Fn = M.getOrInsertFunction(
     "__msan_set_alloca_origin4", IRB.getVoidTy(), IRB.getInt8PtrTy(), IntptrTy,
     IRB.getInt8PtrTy(), IntptrTy);
   MsanPoisonStackFn =
       M.getOrInsertFunction("__msan_poison_stack", IRB.getVoidTy(),
                             IRB.getInt8PtrTy(), IntptrTy);
+}
+
+/// Insert extern declaration of runtime-provided functions and globals.
+void MemorySanitizer::initializeCallbacks(Module &M) {
+  // Only do this once.
+  if (CallbacksInitialized)
+    return;
+
+  IRBuilder<> IRB(*C);
+  // Initialize callbacks that are common for kernel and userspace
+  // instrumentation.
   MsanChainOriginFn = M.getOrInsertFunction(
     "__msan_chain_origin", IRB.getInt32Ty(), IRB.getInt32Ty());
   MemmoveFn = M.getOrInsertFunction(
     "__msan_memmove", IRB.getInt8PtrTy(), IRB.getInt8PtrTy(),
     IRB.getInt8PtrTy(), IntptrTy);
   MemcpyFn = M.getOrInsertFunction(
     "__msan_memcpy", IRB.getInt8PtrTy(), IRB.getInt8PtrTy(), IRB.getInt8PtrTy(),
     IntptrTy);
   MemsetFn = M.getOrInsertFunction(
     "__msan_memset", IRB.getInt8PtrTy(), IRB.getInt8PtrTy(), IRB.getInt32Ty(),
     IntptrTy);
-
-  // Create globals.
-  RetvalTLS = new GlobalVariable(
-    M, ArrayType::get(IRB.getInt64Ty(), kRetvalTLSSize / 8), false,
-    GlobalVariable::ExternalLinkage, nullptr, "__msan_retval_tls", nullptr,
-    GlobalVariable::InitialExecTLSModel);
-  RetvalOriginTLS = new GlobalVariable(
-    M, OriginTy, false, GlobalVariable::ExternalLinkage, nullptr,
-    "__msan_retval_origin_tls", nullptr, GlobalVariable::InitialExecTLSModel);
-
-  ParamTLS = new GlobalVariable(
-    M, ArrayType::get(IRB.getInt64Ty(), kParamTLSSize / 8), false,
-    GlobalVariable::ExternalLinkage, nullptr, "__msan_param_tls", nullptr,
-    GlobalVariable::InitialExecTLSModel);
-  ParamOriginTLS = new GlobalVariable(
-    M, ArrayType::get(OriginTy, kParamTLSSize / 4), false,
-    GlobalVariable::ExternalLinkage, nullptr, "__msan_param_origin_tls",
-    nullptr, GlobalVariable::InitialExecTLSModel);
-
-  VAArgTLS = new GlobalVariable(
-    M, ArrayType::get(IRB.getInt64Ty(), kParamTLSSize / 8), false,
-    GlobalVariable::ExternalLinkage, nullptr, "__msan_va_arg_tls", nullptr,
-    GlobalVariable::InitialExecTLSModel);
-  VAArgOverflowSizeTLS = new GlobalVariable(
-    M, IRB.getInt64Ty(), false, GlobalVariable::ExternalLinkage, nullptr,
-    "__msan_va_arg_overflow_size_tls", nullptr,
-    GlobalVariable::InitialExecTLSModel);
-  OriginTLS = new GlobalVariable(
-    M, IRB.getInt32Ty(), false, GlobalVariable::ExternalLinkage, nullptr,
-    "__msan_origin_tls", nullptr, GlobalVariable::InitialExecTLSModel);
-
   // We insert an empty inline asm after __msan_report* to avoid callback merge.
   EmptyAsm = InlineAsm::get(FunctionType::get(IRB.getVoidTy(), false),
                             StringRef(""), StringRef(""),
                             /*hasSideEffects=*/true);
+
+  createUserspaceApi(M);
+  CallbacksInitialized = true;
 }
 
 /// Module-level initialization.
 ///
 /// inserts a call to __msan_init to the module's constructor list.
 bool MemorySanitizer::doInitialization(Module &M) {
   auto &DL = M.getDataLayout();