rocr: revert back to old copy behaviour with no xgmi sdma engines

SDMA queue resources are limited when all SDMA copies are bottle necked
into 2 engines.  Callers will not be able to make the best decisions
to allocate queue resources fairly so have ROCr fallback to old round
robin behaviour dictated by KFD.

Change-Id: I93d52297976d74e20129c5eb1dcfbfa5aa5067a7

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

@@ -859,6 +859,11 @@ void GpuAgent::InitDma() {
           && isa_->GetStepping() < 10)
         rec_eng = -1;
 
+      // devices without dedicated xGMI SDMA engines should not target specific
+      // SDMA engines for queue creation as resources are limited
+      if (!properties_.NumSdmaXgmiEngines)
+        rec_eng = -1;
+
       auto ret = CreateBlitSdma(use_xgmi, rec_eng);
       if (ret != nullptr) return ret;
     }