|
@@ -193,13 +193,13 @@ static bool amdgpu_fence_activity(struct amdgpu_ring *ring)
|
|
|
* xchg of the current process.
|
|
|
*
|
|
|
* More over for this to go in infinite loop there need to be
|
|
|
- * continuously new fence signaled ie radeon_fence_read needs
|
|
|
+ * continuously new fence signaled ie amdgpu_fence_read needs
|
|
|
* to return a different value each time for both the currently
|
|
|
* polling process and the other process that xchg the last_seq
|
|
|
* btw atomic read and xchg of the current process. And the
|
|
|
* value the other process set as last seq must be higher than
|
|
|
* the seq value we just read. Which means that current process
|
|
|
- * need to be interrupted after radeon_fence_read and before
|
|
|
+ * need to be interrupted after amdgpu_fence_read and before
|
|
|
* atomic xchg.
|
|
|
*
|
|
|
* To be even more safe we count the number of time we loop and
|