[PATCH 5/5] drm/amdgpu: add graceful VM fault handling

[Christian König]

Next step towards HMM support. For now just silence the retry fault and
optionally redirect the request to the dummy page.

Signed-off-by: Christian König <christian.koenig at amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 59 ++++++++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h |  2 +
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c  |  4 ++
 3 files changed, 65 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index a2d18fe46691..6f8c7cc07e8b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -3122,3 +3122,62 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm)
 		}
 	}
 }
+
+/**
+ * amdgpu_vm_handle_fault - graceful handling of VM faults.
+ * @adev: amdgpu device pointer
+ * @pasid: PASID of the VM
+ * @addr: Address of the fault
+ *
+ * Try to gracefully handle a VM fault. Return true if the fault was handled and
+ * shouldn't be reported any more.
+ */
+bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, unsigned int pasid,
+			    uint64_t addr)
+{
+	struct amdgpu_ring *ring = &adev->sdma.instance[0].page;
+	uint64_t value, flags;
+	struct amdgpu_vm *vm;
+	long r;
+
+	if (!ring->sched.ready)
+		return false;
+
+	spin_lock(&adev->vm_manager.pasid_lock);
+	vm = idr_find(&adev->vm_manager.pasid_idr, pasid);
+	spin_unlock(&adev->vm_manager.pasid_lock);
+
+	if (!vm)
+		return false;
+
+	r = amdgpu_bo_reserve(vm->root.base.bo, true);
+	if (r)
+		return false;
+
+	addr /= AMDGPU_GPU_PAGE_SIZE;
+	flags = AMDGPU_PTE_VALID | AMDGPU_PTE_SNOOPED |
+		AMDGPU_PTE_SYSTEM;
+
+	if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_NEVER) {
+		/* Redirect the access to the dummy page */
+		value = adev->dummy_page_addr;
+		flags |= AMDGPU_PTE_EXECUTABLE | AMDGPU_PTE_READABLE |
+			AMDGPU_PTE_WRITEABLE;
+	} else {
+		value = 0;
+	}
+
+	r = amdgpu_vm_bo_update_mapping(adev, vm, true, NULL, addr, addr + 1,
+					flags, value, NULL, NULL);
+	if (r)
+		goto error;
+
+	r = amdgpu_vm_update_pdes(adev, vm, true);
+
+error:
+	amdgpu_bo_unreserve(vm->root.base.bo);
+	if (r < 0)
+		DRM_ERROR("Can't handle page fault (%ld)\n", r);
+
+	return false;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 7227c8da1a2a..d0a30614f70a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -408,6 +408,8 @@ void amdgpu_vm_check_compute_bug(struct amdgpu_device *adev);
 
 void amdgpu_vm_get_task_info(struct amdgpu_device *adev, unsigned int pasid,
 			     struct amdgpu_task_info *task_info);
+bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, unsigned int pasid,
+			    uint64_t addr);
 
 void amdgpu_vm_set_task_info(struct amdgpu_vm *vm);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index bd5d36944481..faf6d64cd4d5 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -324,6 +324,10 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 		return 1; /* This also prevents sending it to KFD */
 
 	/* If it's the first fault for this address, process it normally */
+	if (retry_fault && !in_interrupt() &&
+	    amdgpu_vm_handle_fault(adev, entry->pasid, addr))
+		return 1; /* This also prevents sending it to KFD */
+
 	if (!amdgpu_sriov_vf(adev)) {
 		status = RREG32(hub->vm_l2_pro_fault_status);
 		WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1);
-- 
2.17.1