Am 2022-01-30 um 02:38 schrieb Tao Zhou:
Replace reset queue for specific PASID with unmap all queues, reset
queue could break CP scheduler.
Signed-off-by: Tao Zhou <[email protected]>
The change looks reasonable, based on what kfd_process_vm_fault does.
But the function name is now a bit misleading. Maybe rename it to
something more general, e.g. kfd_process_mem_fault or kfd_dqm_evict_pasid.
Regards,
Felix
---
drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
index e8bc28009c22..dca0b5fac1db 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
@@ -109,8 +109,7 @@ static void event_interrupt_poison_consumption(struct
kfd_dev *dev,
switch (source_id) {
case SOC15_INTSRC_SQ_INTERRUPT_MSG:
- if (dev->dqm->ops.reset_queues)
- ret = dev->dqm->ops.reset_queues(dev->dqm, pasid);
+ ret = kfd_process_vm_fault(dev->dqm, pasid);
break;
case SOC15_INTSRC_SDMA_ECC:
default:
