[llvm] [Offload] Implement the remaining initial Offload API (PR #122106)

Joseph Huber via llvm-commits llvm-commits at lists.llvm.org
Fri Feb 21 05:01:24 PST 2025


================
@@ -1088,26 +1089,33 @@ olEnqueueKernelLaunch_val(ol_queue_handle_t Queue, ol_kernel_handle_t Kernel,
       return OL_ERRC_INVALID_NULL_HANDLE;
     }
 
+    if (NULL == ArgumentsData) {
+      return OL_ERRC_INVALID_NULL_POINTER;
+    }
+
     if (NULL == LaunchSizeArgs) {
       return OL_ERRC_INVALID_NULL_POINTER;
     }
   }
 
-  return olEnqueueKernelLaunch_impl(Queue, Kernel, LaunchSizeArgs, EventOut);
+  return olEnqueueKernelLaunch_impl(Queue, Kernel, ArgumentsData, ArgumentsSize,
+                                    LaunchSizeArgs, EventOut);
 }
 OL_APIEXPORT ol_result_t OL_APICALL
 olEnqueueKernelLaunch(ol_queue_handle_t Queue, ol_kernel_handle_t Kernel,
+                      const void *ArgumentsData, size_t ArgumentsSize,
                       const ol_kernel_launch_size_args_t *LaunchSizeArgs,
                       ol_event_handle_t *EventOut) {
   if (offloadConfig().TracingEnabled) {
     std::cerr << "---> olEnqueueKernelLaunch";
----------------
jhuber6 wrote:

```suggestion
    std::cerr << "---> olEnqueueKernelLaunch";
```
No standard streams, use `llvm::errs()` or potentially even a custom logger class.

https://github.com/llvm/llvm-project/pull/122106


More information about the llvm-commits mailing list