SWDEV-465041 - Add support for user events with DD (#321)

* SWDEV-465041 - Add support for user events with DD

User events can be replaced with HSA signals. Add the interface
to allocate HSA signal for user events and update the status on
CL_COMPLETE.
Force pinned path with DD to avoid blocking calls. Pinned memory
can be released only when the command is complete.
Simplify device enqueue path to use generic kernel arg buffer and
signals

* Fix notifyCmdQueue() logic for OCL

* Avoid blocking calls in OCL with DD

* Add event  destruciton in a case of the failure.

rocclr/device/rocm/rocdevice.cpp

@@ -3464,6 +3464,25 @@ void Device::ReleaseGlobalSignal(void* signal) const {
   }
 }
 
+// ================================================================================================
+bool Device::CreateUserEvent(amd::UserEvent* event) const {
+  std::unique_ptr<ProfilingSignal> signal(new ProfilingSignal());
+  if ((signal == nullptr) ||
+      (HSA_STATUS_SUCCESS != hsa_signal_create(0, 0, nullptr, &signal->signal_))) {
+    return false;
+  }
+  hsa_signal_silent_store_relaxed(signal->signal_, kInitSignalValueOne);
+  event->SetHwEvent(signal.release());
+  return true;
+}
+
+// ================================================================================================
+void Device::SetUserEvent(amd::UserEvent* event) const {
+  auto signal = reinterpret_cast<ProfilingSignal*>(event->HwEvent());
+  assert(signal != nullptr && "Can't have user event without hw event!");
+  hsa_signal_silent_store_relaxed(signal->signal_, 0);
+}
+
 // ================================================================================================
 bool Device::IsValidAllocation(const void* dev_ptr, size_t size, hsa_amd_pointer_info_t* ptr_info) {
   // Query ptr type to see if it's a HMM allocation