drm/amdgpu: Put drm_dev_enter/exit outside hot codepath

We hit soft hang while doing memory pressure test on one numa system.
After a qucik look, this is because kfd invalid/valid userptr memory
frequently with process_info lock hold.
Looks like update page table mapping use too much cpu time.

perf top says below,
75.81%  [kernel]       [k] __srcu_read_unlock
 6.19%  [amdgpu]       [k] amdgpu_gmc_set_pte_pde
 3.56%  [kernel]       [k] __srcu_read_lock
 2.20%  [amdgpu]       [k] amdgpu_vm_cpu_update
 2.20%  [kernel]       [k] __sg_page_iter_dma_next
 2.15%  [drm]          [k] drm_dev_enter
 1.70%  [drm]          [k] drm_prime_sg_to_dma_addr_array
 1.18%  [kernel]       [k] __sg_alloc_table_from_pages
 1.09%  [drm]          [k] drm_dev_exit

So move drm_dev_enter/exit outside gmc code, instead let caller do it.
They are gart_unbind, gart_map, vm_clear_bo, vm_update_pdes and
gmc_init_pdb0. vm_bo_update_mapping already calls it.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-and-tested-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

1 files changed, 20 insertions, 8 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 6b15cad78de9..a1ddf74bbdba 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -800,7 +800,7 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
 	struct amdgpu_bo *bo = &vmbo->bo;
 	unsigned entries, ats_entries;
 	uint64_t addr;
-	int r;
+	int r, idx;
 
 	/* Figure out our place in the hierarchy */
 	if (ancestor->parent) {
@@ -845,9 +845,12 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
 			return r;
 	}
 
+	if (!drm_dev_enter(&adev->ddev, &idx))
+		return -ENODEV;
+
 	r = vm->update_funcs->map_table(vmbo);
 	if (r)
-		return r;
+		goto exit;
 
 	memset(&params, 0, sizeof(params));
 	params.adev = adev;
@@ -856,7 +859,7 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
 
 	r = vm->update_funcs->prepare(&params, NULL, AMDGPU_SYNC_EXPLICIT);
 	if (r)
-		return r;
+		goto exit;
 
 	addr = 0;
 	if (ats_entries) {
@@ -872,7 +875,7 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
 		r = vm->update_funcs->update(&params, vmbo, addr, 0, ats_entries,
 					     value, flags);
 		if (r)
-			return r;
+			goto exit;
 
 		addr += ats_entries * 8;
 	}
@@ -895,10 +898,13 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
 		r = vm->update_funcs->update(&params, vmbo, addr, 0, entries,
 					     value, flags);
 		if (r)
-			return r;
+			goto exit;
 	}
 
-	return vm->update_funcs->commit(&params, NULL);
+	r = vm->update_funcs->commit(&params, NULL);
+exit:
+	drm_dev_exit(idx);
+	return r;
 }
 
 /**
@@ -1384,11 +1390,14 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
 			  struct amdgpu_vm *vm, bool immediate)
 {
 	struct amdgpu_vm_update_params params;
-	int r;
+	int r, idx;
 
 	if (list_empty(&vm->relocated))
 		return 0;
 
+	if (!drm_dev_enter(&adev->ddev, &idx))
+		return -ENODEV;
+
 	memset(&params, 0, sizeof(params));
 	params.adev = adev;
 	params.vm = vm;
@@ -1396,7 +1405,7 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
 
 	r = vm->update_funcs->prepare(&params, NULL, AMDGPU_SYNC_EXPLICIT);
 	if (r)
-		return r;
+		goto exit;
 
 	while (!list_empty(&vm->relocated)) {
 		struct amdgpu_vm_bo_base *entry;
@@ -1414,10 +1423,13 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
 	r = vm->update_funcs->commit(&params, &vm->last_update);
 	if (r)
 		goto error;
+	drm_dev_exit(idx);
 	return 0;
 
 error:
 	amdgpu_vm_invalidate_pds(adev, vm);
+exit:
+	drm_dev_exit(idx);
 	return r;
 }