[AsyncAlloc][CUDA] Don't set pool maxSize if CUDA version < 12.2

unified-runtime/source/adapters/cuda/usm.cpp

@@ -431,13 +431,24 @@ ur_usm_pool_handle_t_::ur_usm_pool_handle_t_(ur_context_handle_t Context,
     case UR_STRUCTURE_TYPE_USM_POOL_LIMITS_DESC: {
       const ur_usm_pool_limits_desc_t *Limits =
           reinterpret_cast<const ur_usm_pool_limits_desc_t *>(BaseDesc);
+#if CUDA_VERSION >= 12020
+      // maxSize as a member of CUmemPoolProps was introduced in CUDA 12.2.
       MemPoolProps.maxSize =
           Limits->maxPoolableSize; // CUDA lazily reserves memory for pools in
                                    // 32MB chunks. maxSize is elevated to the
                                    // next 32MB multiple. Each 32MB chunk is
                                    // only reserved when it's needed for the
                                    // first time (cuMemAllocFromPoolAsync).
-
+#else
+      // Only error if the user set a value >0 for the maximum size.
+      // Otherwise, do nothing.
+      if (Limits->maxPoolableSize > 0) {
+        setErrorMessage(
+            "The memory pool maximum size feature requires CUDA 12.2 or later.",
+            UR_RESULT_ERROR_ADAPTER_SPECIFIC);
+        throw UsmAllocationException(UR_RESULT_ERROR_ADAPTER_SPECIFIC);
+      }
+#endif
       maxSize = Limits->maxPoolableSize;
       size_t chunkSize = 33554432; // 32MB
       size_t remainder = Limits->maxPoolableSize % chunkSize;
@@ -485,8 +496,8 @@ urUSMPoolCreateExp(ur_context_handle_t Context, ur_device_handle_t Device,
   try {
     *pPool = reinterpret_cast<ur_usm_pool_handle_t>(
         new ur_usm_pool_handle_t_(Context, Device, pPoolDesc));
-  } catch (ur_result_t err) {
-    return err;
+  } catch (const UsmAllocationException &Ex) {
+    return Ex.getError();
   } catch (...) {
     return UR_RESULT_ERROR_UNKNOWN;
   }