Move VM fault handler init to after all devices are registered.

During registration we must not call any function that depends on registered
data as the lists are not yet complete.  This includes signal allocation since
allocating shared GPU mapped memory depends on the list of GPUs.

Change-Id: I94d59e847802c546c2a5a0d9f55fe5ac3fd1d878

runtime/hsa-runtime/core/runtime/runtime.cpp

@@ -146,6 +146,8 @@ bool Runtime::IsOpen() {
          (Runtime::runtime_singleton_->ref_count_ != 0);
 }
 
+// Register agent information only.  Must not call anything that may use the registered information
+// since those tables are incomplete.
 void Runtime::RegisterAgent(Agent* agent) {
   // Record the agent in the node-to-agent reverse lookup table.
   agents_by_node_[agent->node_id()].push_back(agent);
@@ -225,10 +227,6 @@ void Runtime::RegisterAgent(Agent* agent) {
         start_svm_address_ =
             static_cast<uintptr_t>(svm_region->GetBaseAddress());
         end_svm_address_ = start_svm_address_ + svm_region->GetPhysicalSize();
-
-        // Bind VM fault handler when we detect the first GPU agent.
-        // TODO: validate if it works on APU.
-        BindVmFaultHandler();
       } else {
         start_svm_address_ = 0;
         end_svm_address_ = os::GetUserModeVirtualMemoryBase() +
@@ -1030,7 +1028,7 @@ void Runtime::AsyncEventsLoop(void*) {
 }
 
 void Runtime::BindVmFaultHandler() {
-  if (core::g_use_interrupt_wait) {
+  if (core::g_use_interrupt_wait && !gpu_agents_.empty()) {
     // Create memory event with manual reset to avoid racing condition
     // with driver in case of multiple concurrent VM faults.
     vm_fault_event_ =
@@ -1215,6 +1213,7 @@ hsa_status_t Runtime::Load() {
   if (!amd::Load()) {
     return HSA_STATUS_ERROR_OUT_OF_RESOURCES;
   }
+  BindVmFaultHandler();
 
   loader_ = amd::hsa::loader::Loader::Create(&loader_context_);