Am 03.03.20 um 13:50 schrieb Nirmoy Das:
[SNIP]
struct amdgpu_mec {
struct amdgpu_bo *hpd_eop_obj;
u64 hpd_eop_gpu_addr;
@@ -280,8 +290,9 @@ struct amdgpu_gfx {
uint32_t num_gfx_sched;
unsigned num_gfx_rings;
struct amdgpu_ring compute_ring[AMDGPU_MAX_COMPUTE_RINGS];
+ struct drm_gpu_scheduler **compute_prio_sched[AMDGPU_GFX_PIPE_PRIO_MAX];
struct drm_gpu_scheduler *compute_sched[AMDGPU_MAX_COMPUTE_RINGS];
- uint32_t num_compute_sched;
+ uint32_t num_compute_sched[AMDGPU_GFX_PIPE_PRIO_MAX];
unsigned num_compute_rings;
struct amdgpu_irq_src eop_irq;
struct amdgpu_irq_src priv_reg_irq;
Well my question is why we we need compute_prio_sched here?
Can't we just design that as:
struct drm_gpu_scheduler
*compute_sched[AMDGPU_GFX_PIPE_PRIO_MAX][AMDGPU_MAX_HI_COMPUTE_RINGS];
uint32_t num_compute_sched[AMDGPU_GFX_PIPE_PRIO_MAX];
I mean the drm_gpu_scheduler * array doesn't needs to be constructed by
the context code in the first place.
Or am I missing something?
Regards,
Christian.
_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
