Skip to content

[OMPD] The LLVM OMPD ompd_get_task_in_parallel() function misspells "ompd_nvptx_parallel_info_t" #64

@jdelsign

Description

@jdelsign

The LLVM OMPD ompd_get_task_in_parallel() function has a typo in the CUDA device code. It misspells "ompd_nvptx_parallel_info_t" as "ompd_nvptx_paralel_info". Diff to fix it follows:

  -------------------------- libompd/src/omp-debug.cpp --------------------------
  index fae2b63..8e9b645 100644
  @@ -865,7 +865,7 @@ ompd_rc_t ompd_get_task_in_parallel(

     if (parallel_handle->ah->kind == OMPD_DEVICE_KIND_CUDA) {
       ret = TValue(context, parallel_handle->th)
  -              .cast("ompd_nvptx_paralel_info", 0,
  +              .cast("ompd_nvptx_parallel_info_t", 0,
		       OMPD_SEGMENT_CUDA_PTX_GLOBAL)
		 .access("parallel_tasks")
		 .cast("omptarget_nvptx_TaskDescr", 1,

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions