From 4671078eb8e390bd44c458e2f482fbb61f5bc612 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Mon, 21 Sep 2020 15:55:39 +0200
Subject: drm/amdgpu: switch over to the new pin interface
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Stop using TTM_PL_FLAG_NO_EVICT.

Signed-off-by: Christian König <christian.koenig@amd.com>
Tested-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Link: https://patchwork.freedesktop.org/patch/391617/?series=81973&rev=1
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 41 ++++++++----------------------
 1 file changed, 11 insertions(+), 30 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_object.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index ac043baac05d..63e9c5793c30 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -78,7 +78,7 @@ static void amdgpu_bo_destroy(struct ttm_buffer_object *tbo)
 	struct amdgpu_device *adev = amdgpu_ttm_adev(tbo->bdev);
 	struct amdgpu_bo *bo = ttm_to_amdgpu_bo(tbo);
 
-	if (bo->pin_count > 0)
+	if (bo->tbo.pin_count > 0)
 		amdgpu_bo_subtract_pin_size(bo);
 
 	amdgpu_bo_kunmap(bo);
@@ -721,7 +721,7 @@ int amdgpu_bo_validate(struct amdgpu_bo *bo)
 	uint32_t domain;
 	int r;
 
-	if (bo->pin_count)
+	if (bo->tbo.pin_count)
 		return 0;
 
 	domain = bo->preferred_domains;
@@ -918,13 +918,13 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 	 */
 	domain = amdgpu_bo_get_preferred_pin_domain(adev, domain);
 
-	if (bo->pin_count) {
+	if (bo->tbo.pin_count) {
 		uint32_t mem_type = bo->tbo.mem.mem_type;
 
 		if (!(domain & amdgpu_mem_type_to_domain(mem_type)))
 			return -EINVAL;
 
-		bo->pin_count++;
+		ttm_bo_pin(&bo->tbo);
 
 		if (max_offset != 0) {
 			u64 domain_start = amdgpu_ttm_domain_start(adev,
@@ -955,7 +955,6 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 		if (!bo->placements[i].lpfn ||
 		    (lpfn && lpfn < bo->placements[i].lpfn))
 			bo->placements[i].lpfn = lpfn;
-		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	}
 
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
@@ -964,7 +963,7 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 		goto error;
 	}
 
-	bo->pin_count = 1;
+	ttm_bo_pin(&bo->tbo);
 
 	domain = amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type);
 	if (domain == AMDGPU_GEM_DOMAIN_VRAM) {
@@ -1006,34 +1005,16 @@ int amdgpu_bo_pin(struct amdgpu_bo *bo, u32 domain)
  * Returns:
  * 0 for success or a negative error code on failure.
  */
-int amdgpu_bo_unpin(struct amdgpu_bo *bo)
+void amdgpu_bo_unpin(struct amdgpu_bo *bo)
 {
-	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	struct ttm_operation_ctx ctx = { false, false };
-	int r, i;
-
-	if (WARN_ON_ONCE(!bo->pin_count)) {
-		dev_warn(adev->dev, "%p unpin not necessary\n", bo);
-		return 0;
-	}
-	bo->pin_count--;
-	if (bo->pin_count)
-		return 0;
+	ttm_bo_unpin(&bo->tbo);
+	if (bo->tbo.pin_count)
+		return;
 
 	amdgpu_bo_subtract_pin_size(bo);
 
 	if (bo->tbo.base.import_attach)
 		dma_buf_unpin(bo->tbo.base.import_attach);
-
-	for (i = 0; i < bo->placement.num_placement; i++) {
-		bo->placements[i].lpfn = 0;
-		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
-	}
-	r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
-	if (unlikely(r))
-		dev_err(adev->dev, "%p validate failed for unpin\n", bo);
-
-	return r;
 }
 
 /**
@@ -1385,7 +1366,7 @@ int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 		return 0;
 
 	/* Can't move a pinned BO to visible VRAM */
-	if (abo->pin_count > 0)
+	if (abo->tbo.pin_count > 0)
 		return -EINVAL;
 
 	/* hurrah the memory is not visible ! */
@@ -1489,7 +1470,7 @@ u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo)
 {
 	WARN_ON_ONCE(bo->tbo.mem.mem_type == TTM_PL_SYSTEM);
 	WARN_ON_ONCE(!dma_resv_is_locked(bo->tbo.base.resv) &&
-		     !bo->pin_count && bo->tbo.type != ttm_bo_type_kernel);
+		     !bo->tbo.pin_count && bo->tbo.type != ttm_bo_type_kernel);
 	WARN_ON_ONCE(bo->tbo.mem.start == AMDGPU_BO_INVALID_OFFSET);
 	WARN_ON_ONCE(bo->tbo.mem.mem_type == TTM_PL_VRAM &&
 		     !(bo->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS));
-- 
cgit v1.2.3


From d3ef581afa5e6a65cc3a40de2e62901f17b8cebc Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Fri, 25 Sep 2020 15:21:49 +0200
Subject: drm/amdgpu: stop using TTMs fault callback
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Implement the fault handler ourself using the provided TTM functions.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/392324/
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 20 +++++++--------
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.h |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    | 40 +++++++++++++++++++++++++++---
 3 files changed, 46 insertions(+), 16 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_object.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 63e9c5793c30..80bc7177cd45 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -1341,19 +1341,14 @@ void amdgpu_bo_release_notify(struct ttm_buffer_object *bo)
  * Returns:
  * 0 for success or a negative error code on failure.
  */
-int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
+vm_fault_t amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
 	struct ttm_operation_ctx ctx = { false, false };
-	struct amdgpu_bo *abo;
+	struct amdgpu_bo *abo = ttm_to_amdgpu_bo(bo);
 	unsigned long offset, size;
 	int r;
 
-	if (!amdgpu_bo_is_amdgpu_bo(bo))
-		return 0;
-
-	abo = ttm_to_amdgpu_bo(bo);
-
 	/* Remember that this BO was accessed by the CPU */
 	abo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 
@@ -1367,7 +1362,7 @@ int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 
 	/* Can't move a pinned BO to visible VRAM */
 	if (abo->tbo.pin_count > 0)
-		return -EINVAL;
+		return VM_FAULT_SIGBUS;
 
 	/* hurrah the memory is not visible ! */
 	atomic64_inc(&adev->num_vram_cpu_page_faults);
@@ -1379,15 +1374,18 @@ int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 	abo->placement.busy_placement = &abo->placements[1];
 
 	r = ttm_bo_validate(bo, &abo->placement, &ctx);
-	if (unlikely(r != 0))
-		return r;
+	if (unlikely(r == -EBUSY || r == -ERESTARTSYS))
+		return VM_FAULT_NOPAGE;
+	else if (unlikely(r))
+		return VM_FAULT_SIGBUS;
 
 	offset = bo->mem.start << PAGE_SHIFT;
 	/* this should never happen */
 	if (bo->mem.mem_type == TTM_PL_VRAM &&
 	    (offset + size) > adev->gmc.visible_vram_size)
-		return -EINVAL;
+		return VM_FAULT_SIGBUS;
 
+	ttm_bo_move_to_lru_tail_unlocked(bo);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
index e91750e43448..132e5f955180 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
@@ -284,7 +284,7 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
 			   bool evict,
 			   struct ttm_resource *new_mem);
 void amdgpu_bo_release_notify(struct ttm_buffer_object *bo);
-int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo);
+vm_fault_t amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo);
 void amdgpu_bo_fence(struct amdgpu_bo *bo, struct dma_fence *fence,
 		     bool shared);
 int amdgpu_bo_sync_wait_resv(struct amdgpu_device *adev, struct dma_resv *resv,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index d3bd2fd448be..399961035ae6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1708,7 +1708,6 @@ static struct ttm_bo_driver amdgpu_bo_driver = {
 	.verify_access = &amdgpu_verify_access,
 	.move_notify = &amdgpu_bo_move_notify,
 	.release_notify = &amdgpu_bo_release_notify,
-	.fault_reserve_notify = &amdgpu_bo_fault_reserve_notify,
 	.io_mem_reserve = &amdgpu_ttm_io_mem_reserve,
 	.io_mem_pfn = amdgpu_ttm_io_mem_pfn,
 	.access_memory = &amdgpu_ttm_access_memory,
@@ -2088,15 +2087,48 @@ void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev, bool enable)
 	adev->mman.buffer_funcs_enabled = enable;
 }
 
+static vm_fault_t amdgpu_ttm_fault(struct vm_fault *vmf)
+{
+	struct ttm_buffer_object *bo = vmf->vma->vm_private_data;
+	vm_fault_t ret;
+
+	ret = ttm_bo_vm_reserve(bo, vmf);
+	if (ret)
+		return ret;
+
+	ret = amdgpu_bo_fault_reserve_notify(bo);
+	if (ret)
+		goto unlock;
+
+	ret = ttm_bo_vm_fault_reserved(vmf, vmf->vma->vm_page_prot,
+				       TTM_BO_VM_NUM_PREFAULT, 1);
+	if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
+		return ret;
+
+unlock:
+	dma_resv_unlock(bo->base.resv);
+	return ret;
+}
+
+static struct vm_operations_struct amdgpu_ttm_vm_ops = {
+	.fault = amdgpu_ttm_fault,
+	.open = ttm_bo_vm_open,
+	.close = ttm_bo_vm_close,
+	.access = ttm_bo_vm_access
+};
+
 int amdgpu_mmap(struct file *filp, struct vm_area_struct *vma)
 {
 	struct drm_file *file_priv = filp->private_data;
 	struct amdgpu_device *adev = drm_to_adev(file_priv->minor->dev);
+	int r;
 
-	if (adev == NULL)
-		return -EINVAL;
+	r = ttm_bo_mmap(filp, vma, &adev->mman.bdev);
+	if (unlikely(r != 0))
+		return r;
 
-	return ttm_bo_mmap(filp, vma, &adev->mman.bdev);
+	vma->vm_ops = &amdgpu_ttm_vm_ops;
+	return 0;
 }
 
 int amdgpu_copy_buffer(struct amdgpu_ring *ring, uint64_t src_offset,
-- 
cgit v1.2.3


From 15c90a1fbcb1621be0ba492bc5d7b88edf3d2e22 Mon Sep 17 00:00:00 2001
From: Huang Rui <ray.huang@amd.com>
Date: Fri, 28 Aug 2020 17:54:31 +0800
Subject: drm/amdgpu: get the correct vram type for van gogh

This patch is to get the correct vram type from atombios for van gogh.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c | 4 ++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c       | 1 +
 2 files changed, 5 insertions(+)

(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_object.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index b4df6460e45a..5f5ce705d20e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -149,6 +149,10 @@ static int convert_atom_mem_type_to_vram_type(struct amdgpu_device *adev,
 		case LpDdr4MemType:
 			vram_type = AMDGPU_VRAM_TYPE_DDR4;
 			break;
+		case Ddr5MemType:
+		case LpDdr5MemType:
+			vram_type = AMDGPU_VRAM_TYPE_DDR5;
+			break;
 		default:
 			vram_type = AMDGPU_VRAM_TYPE_UNKNOWN;
 			break;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 887a4da5eefc..07e81aca9c34 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -1063,6 +1063,7 @@ static const char *amdgpu_vram_names[] = {
 	"DDR3",
 	"DDR4",
 	"GDDR6",
+	"DDR5"
 };
 
 /**
-- 
cgit v1.2.3


From 4ce032d64c2a30cf5b23c57b3328d5d2dab99a1f Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Thu, 1 Oct 2020 15:21:00 +0200
Subject: drm/ttm: nuke ttm_bo_evict_mm and rename mgr function v3
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Make it more clear what the resource manager function
does and nuke the wrapper function.

v2: nuke the wrapper
v3: fix typo in radeon, rebased

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> (v2)
Link: https://patchwork.freedesktop.org/patch/393914/
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c   |  5 ++++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c   |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c    |  6 +++++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c  |  2 +-
 drivers/gpu/drm/nouveau/nouveau_drm.c         |  5 ++++-
 drivers/gpu/drm/nouveau/nouveau_ttm.c         |  4 ++--
 drivers/gpu/drm/qxl/qxl_object.c              | 10 ++++++++--
 drivers/gpu/drm/radeon/radeon_object.c        |  6 +++++-
 drivers/gpu/drm/ttm/ttm_bo.c                  | 18 ------------------
 drivers/gpu/drm/ttm/ttm_range_manager.c       |  2 +-
 drivers/gpu/drm/ttm/ttm_resource.c            | 10 +++++-----
 drivers/gpu/drm/vmwgfx/vmwgfx_drv.c           | 12 +++++++++---
 drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c |  2 +-
 drivers/gpu/drm/vmwgfx/vmwgfx_thp.c           |  2 +-
 include/drm/ttm/ttm_bo_api.h                  | 20 --------------------
 include/drm/ttm/ttm_resource.h                |  4 ++--
 16 files changed, 49 insertions(+), 61 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_object.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index abe0c2729e1c..a5e08dc54e4a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -1319,6 +1319,7 @@ static int amdgpu_debugfs_evict_gtt(struct seq_file *m, void *data)
 	struct drm_info_node *node = (struct drm_info_node *)m->private;
 	struct drm_device *dev = node->minor->dev;
 	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct ttm_resource_manager *man;
 	int r;
 
 	r = pm_runtime_get_sync(dev->dev);
@@ -1327,7 +1328,9 @@ static int amdgpu_debugfs_evict_gtt(struct seq_file *m, void *data)
 		return r;
 	}
 
-	seq_printf(m, "(%d)\n", ttm_bo_evict_mm(&adev->mman.bdev, TTM_PL_TT));
+	man = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT);
+	r = ttm_resource_manager_evict_all(&adev->mman.bdev, man);
+	seq_printf(m, "(%d)\n", r);
 
 	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c
index f203e4a6a3f2..1721739def84 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c
@@ -136,7 +136,7 @@ void amdgpu_gtt_mgr_fini(struct amdgpu_device *adev)
 
 	ttm_resource_manager_set_used(man, false);
 
-	ret = ttm_resource_manager_force_list_clean(&adev->mman.bdev, man);
+	ret = ttm_resource_manager_evict_all(&adev->mman.bdev, man);
 	if (ret)
 		return;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 80bc7177cd45..8b96e7aaeff1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -1029,6 +1029,8 @@ void amdgpu_bo_unpin(struct amdgpu_bo *bo)
  */
 int amdgpu_bo_evict_vram(struct amdgpu_device *adev)
 {
+	struct ttm_resource_manager *man;
+
 	/* late 2.6.33 fix IGP hibernate - we need pm ops to do this correct */
 #ifndef CONFIG_HIBERNATION
 	if (adev->flags & AMD_IS_APU) {
@@ -1036,7 +1038,9 @@ int amdgpu_bo_evict_vram(struct amdgpu_device *adev)
 		return 0;
 	}
 #endif
-	return ttm_bo_evict_mm(&adev->mman.bdev, TTM_PL_VRAM);
+
+	man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
+	return ttm_resource_manager_evict_all(&adev->mman.bdev, man);
 }
 
 static const char *amdgpu_vram_names[] = {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
index 01c1171afbe0..7747be644dd0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
@@ -212,7 +212,7 @@ void amdgpu_vram_mgr_fini(struct amdgpu_device *adev)
 
 	ttm_resource_manager_set_used(man, false);
 
-	ret = ttm_resource_manager_force_list_clean(&adev->mman.bdev, man);
+	ret = ttm_resource_manager_evict_all(&adev->mman.bdev, man);
 	if (ret)
 		return;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 72640bca1617..d141a5f004af 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -820,6 +820,7 @@ static int
 nouveau_do_suspend(struct drm_device *dev, bool runtime)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct ttm_resource_manager *man;
 	int ret;
 
 	nouveau_svm_suspend(drm);
@@ -836,7 +837,9 @@ nouveau_do_suspend(struct drm_device *dev, bool runtime)
 	}
 
 	NV_DEBUG(drm, "evicting buffers...\n");
-	ttm_bo_evict_mm(&drm->ttm.bdev, TTM_PL_VRAM);
+
+	man = ttm_manager_type(&drm->ttm.bdev, TTM_PL_VRAM);
+	ttm_resource_manager_evict_all(&drm->ttm.bdev, man);
 
 	NV_DEBUG(drm, "waiting for kernel channels to go idle...\n");
 	if (drm->cechan) {
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index edf3bb89a47f..04b95277c73a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -222,7 +222,7 @@ nouveau_ttm_fini_vram(struct nouveau_drm *drm)
 
 	if (drm->client.device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
 		ttm_resource_manager_set_used(man, false);
-		ttm_resource_manager_force_list_clean(&drm->ttm.bdev, man);
+		ttm_resource_manager_evict_all(&drm->ttm.bdev, man);
 		ttm_resource_manager_cleanup(man);
 		ttm_set_driver_manager(&drm->ttm.bdev, TTM_PL_VRAM, NULL);
 		kfree(man);
@@ -267,7 +267,7 @@ nouveau_ttm_fini_gtt(struct nouveau_drm *drm)
 		ttm_range_man_fini(&drm->ttm.bdev, TTM_PL_TT);
 	else {
 		ttm_resource_manager_set_used(man, false);
-		ttm_resource_manager_force_list_clean(&drm->ttm.bdev, man);
+		ttm_resource_manager_evict_all(&drm->ttm.bdev, man);
 		ttm_resource_manager_cleanup(man);
 		ttm_set_driver_manager(&drm->ttm.bdev, TTM_PL_TT, NULL);
 		kfree(man);
diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c
index c8b67e7a3f02..940e99354f49 100644
--- a/drivers/gpu/drm/qxl/qxl_object.c
+++ b/drivers/gpu/drm/qxl/qxl_object.c
@@ -351,10 +351,16 @@ int qxl_bo_check_id(struct qxl_device *qdev, struct qxl_bo *bo)
 
 int qxl_surf_evict(struct qxl_device *qdev)
 {
-	return ttm_bo_evict_mm(&qdev->mman.bdev, TTM_PL_PRIV);
+	struct ttm_resource_manager *man;
+
+	man = ttm_manager_type(&qdev->mman.bdev, TTM_PL_PRIV);
+	return ttm_resource_manager_evict_all(&qdev->mman.bdev, man);
 }
 
 int qxl_vram_evict(struct qxl_device *qdev)
 {
-	return ttm_bo_evict_mm(&qdev->mman.bdev, TTM_PL_VRAM);
+	struct ttm_resource_manager *man;
+
+	man = ttm_manager_type(&qdev->mman.bdev, TTM_PL_VRAM);
+	return ttm_resource_manager_evict_all(&qdev->mman.bdev, man);
 }
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index 8c285eb118f9..ad0e6e9ef922 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -402,6 +402,9 @@ void radeon_bo_unpin(struct radeon_bo *bo)
 
 int radeon_bo_evict_vram(struct radeon_device *rdev)
 {
+	struct ttm_bo_device *bdev = &rdev->mman.bdev;
+	struct ttm_resource_manager *man;
+
 	/* late 2.6.33 fix IGP hibernate - we need pm ops to do this correct */
 #ifndef CONFIG_HIBERNATION
 	if (rdev->flags & RADEON_IS_IGP) {
@@ -410,7 +413,8 @@ int radeon_bo_evict_vram(struct radeon_device *rdev)
 			return 0;
 	}
 #endif
-	return ttm_bo_evict_mm(&rdev->mman.bdev, TTM_PL_VRAM);
+	man = ttm_manager_type(bdev, TTM_PL_VRAM);
+	return ttm_resource_manager_evict_all(bdev, man);
 }
 
 void radeon_bo_force_delete(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 88d215de9ae1..6eb151dd840d 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1249,24 +1249,6 @@ size_t ttm_bo_dma_acc_size(struct ttm_bo_device *bdev,
 }
 EXPORT_SYMBOL(ttm_bo_dma_acc_size);
 
-int ttm_bo_evict_mm(struct ttm_bo_device *bdev, unsigned mem_type)
-{
-	struct ttm_resource_manager *man = ttm_manager_type(bdev, mem_type);
-
-	if (mem_type == 0 || mem_type >= TTM_NUM_MEM_TYPES) {
-		pr_err("Illegal memory manager memory type %u\n", mem_type);
-		return -EINVAL;
-	}
-
-	if (!man) {
-		pr_err("Memory type %u has not been initialized\n", mem_type);
-		return 0;
-	}
-
-	return ttm_resource_manager_force_list_clean(bdev, man);
-}
-EXPORT_SYMBOL(ttm_bo_evict_mm);
-
 static void ttm_bo_global_kobj_release(struct kobject *kobj)
 {
 	struct ttm_bo_global *glob =
diff --git a/drivers/gpu/drm/ttm/ttm_range_manager.c b/drivers/gpu/drm/ttm/ttm_range_manager.c
index 1da0e277c511..ea77919569a2 100644
--- a/drivers/gpu/drm/ttm/ttm_range_manager.c
+++ b/drivers/gpu/drm/ttm/ttm_range_manager.c
@@ -149,7 +149,7 @@ int ttm_range_man_fini(struct ttm_bo_device *bdev,
 
 	ttm_resource_manager_set_used(man, false);
 
-	ret = ttm_resource_manager_force_list_clean(bdev, man);
+	ret = ttm_resource_manager_evict_all(bdev, man);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/ttm/ttm_resource.c b/drivers/gpu/drm/ttm/ttm_resource.c
index b325b9264203..4ebc043e2867 100644
--- a/drivers/gpu/drm/ttm/ttm_resource.c
+++ b/drivers/gpu/drm/ttm/ttm_resource.c
@@ -75,16 +75,16 @@ void ttm_resource_manager_init(struct ttm_resource_manager *man,
 EXPORT_SYMBOL(ttm_resource_manager_init);
 
 /*
- * ttm_resource_manager_force_list_clean
+ * ttm_resource_manager_evict_all
  *
  * @bdev - device to use
  * @man - manager to use
  *
- * Force all the objects out of a memory manager until clean.
+ * Evict all the objects out of a memory manager until it is empty.
  * Part of memory manager cleanup sequence.
  */
-int ttm_resource_manager_force_list_clean(struct ttm_bo_device *bdev,
-					  struct ttm_resource_manager *man)
+int ttm_resource_manager_evict_all(struct ttm_bo_device *bdev,
+				   struct ttm_resource_manager *man)
 {
 	struct ttm_operation_ctx ctx = {
 		.interruptible = false,
@@ -126,7 +126,7 @@ int ttm_resource_manager_force_list_clean(struct ttm_bo_device *bdev,
 
 	return 0;
 }
-EXPORT_SYMBOL(ttm_resource_manager_force_list_clean);
+EXPORT_SYMBOL(ttm_resource_manager_evict_all);
 
 /**
  * ttm_resource_manager_debug
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index bdb7a5e96560..10a054d25485 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -468,7 +468,10 @@ out_no_query_bo:
 	if (dev_priv->cman)
 		vmw_cmdbuf_remove_pool(dev_priv->cman);
 	if (dev_priv->has_mob) {
-		(void) ttm_bo_evict_mm(&dev_priv->bdev, VMW_PL_MOB);
+		struct ttm_resource_manager *man;
+
+		man = ttm_manager_type(&dev_priv->bdev, VMW_PL_MOB);
+		ttm_resource_manager_evict_all(&dev_priv->bdev, man);
 		vmw_otables_takedown(dev_priv);
 	}
 	if (dev_priv->cman)
@@ -501,7 +504,10 @@ static void vmw_release_device_early(struct vmw_private *dev_priv)
 		vmw_cmdbuf_remove_pool(dev_priv->cman);
 
 	if (dev_priv->has_mob) {
-		ttm_bo_evict_mm(&dev_priv->bdev, VMW_PL_MOB);
+		struct ttm_resource_manager *man;
+
+		man = ttm_manager_type(&dev_priv->bdev, VMW_PL_MOB);
+		ttm_resource_manager_evict_all(&dev_priv->bdev, man);
 		vmw_otables_takedown(dev_priv);
 	}
 }
@@ -1257,7 +1263,7 @@ void vmw_svga_disable(struct vmw_private *dev_priv)
 	if (ttm_resource_manager_used(man)) {
 		ttm_resource_manager_set_used(man, false);
 		spin_unlock(&dev_priv->svga_lock);
-		if (ttm_bo_evict_mm(&dev_priv->bdev, TTM_PL_VRAM))
+		if (ttm_resource_manager_evict_all(&dev_priv->bdev, man))
 			DRM_ERROR("Failed evicting VRAM buffers.\n");
 		vmw_write(dev_priv, SVGA_REG_ENABLE,
 			  SVGA_REG_ENABLE_HIDE |
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c b/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
index db64c3a90285..9e8a3e337b96 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
@@ -143,7 +143,7 @@ void vmw_gmrid_man_fini(struct vmw_private *dev_priv, int type)
 
 	ttm_resource_manager_set_used(man, false);
 
-	ttm_resource_manager_force_list_clean(&dev_priv->bdev, man);
+	ttm_resource_manager_evict_all(&dev_priv->bdev, man);
 
 	ttm_resource_manager_cleanup(man);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c b/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c
index 63fe7da4cbf4..328d332a55e2 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c
@@ -149,7 +149,7 @@ void vmw_thp_fini(struct vmw_private *dev_priv)
 
 	ttm_resource_manager_set_used(man, false);
 
-	ret = ttm_resource_manager_force_list_clean(&dev_priv->bdev, man);
+	ret = ttm_resource_manager_evict_all(&dev_priv->bdev, man);
 	if (ret)
 		return;
 	spin_lock(&rman->lock);
diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h
index c96a25d571c8..6f5ced7d83fe 100644
--- a/include/drm/ttm/ttm_bo_api.h
+++ b/include/drm/ttm/ttm_bo_api.h
@@ -448,26 +448,6 @@ int ttm_bo_init(struct ttm_bo_device *bdev, struct ttm_buffer_object *bo,
 		struct sg_table *sg, struct dma_resv *resv,
 		void (*destroy) (struct ttm_buffer_object *));
 
-/**
- * ttm_bo_evict_mm
- *
- * @bdev: Pointer to a ttm_bo_device struct.
- * @mem_type: The memory type.
- *
- * Evicts all buffers on the lru list of the memory type.
- * This is normally part of a VT switch or an
- * out-of-memory-space-due-to-fragmentation handler.
- * The caller must make sure that there are no other processes
- * currently validating buffers, and can do that by taking the
- * struct ttm_bo_device::ttm_lock in write mode.
- *
- * Returns:
- * -EINVAL: Invalid or uninitialized memory type.
- * -ERESTARTSYS: The call was interrupted by a signal while waiting to
- * evict a buffer.
- */
-int ttm_bo_evict_mm(struct ttm_bo_device *bdev, unsigned mem_type);
-
 /**
  * ttm_kmap_obj_virtual
  *
diff --git a/include/drm/ttm/ttm_resource.h b/include/drm/ttm/ttm_resource.h
index 0e172d94a0c1..1b2f56163c6c 100644
--- a/include/drm/ttm/ttm_resource.h
+++ b/include/drm/ttm/ttm_resource.h
@@ -228,8 +228,8 @@ void ttm_resource_free(struct ttm_buffer_object *bo, struct ttm_resource *res);
 void ttm_resource_manager_init(struct ttm_resource_manager *man,
 			       unsigned long p_size);
 
-int ttm_resource_manager_force_list_clean(struct ttm_bo_device *bdev,
-					  struct ttm_resource_manager *man);
+int ttm_resource_manager_evict_all(struct ttm_bo_device *bdev,
+				   struct ttm_resource_manager *man);
 
 void ttm_resource_manager_debug(struct ttm_resource_manager *man,
 				struct drm_printer *p);
-- 
cgit v1.2.3


From ce65b874001d756294e0b7cf06c51137af964f38 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Wed, 30 Sep 2020 16:44:16 +0200
Subject: drm/ttm: nuke caching placement flags
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Changing the caching on the fly never really worked
flawlessly.

So stop this completely and just let drivers specific the
desired caching in the tt or bus object.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Link: https://patchwork.freedesktop.org/patch/394256/
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 20 ++++---------
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    | 12 ++------
 drivers/gpu/drm/drm_gem_vram_helper.c      |  7 ++---
 drivers/gpu/drm/nouveau/nouveau_bo.c       | 36 ++++++-----------------
 drivers/gpu/drm/qxl/qxl_object.c           | 10 +++----
 drivers/gpu/drm/qxl/qxl_ttm.c              |  2 +-
 drivers/gpu/drm/radeon/radeon_object.c     | 46 ++++++------------------------
 drivers/gpu/drm/radeon/radeon_ttm.c        | 18 +++---------
 drivers/gpu/drm/ttm/ttm_agp_backend.c      |  2 +-
 drivers/gpu/drm/ttm/ttm_bo.c               | 44 +++-------------------------
 drivers/gpu/drm/ttm/ttm_bo_util.c          | 10 ++-----
 drivers/gpu/drm/ttm/ttm_tt.c               | 29 -------------------
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c | 30 +++++++++----------
 include/drm/ttm/ttm_placement.h            | 14 ---------
 include/drm/ttm/ttm_tt.h                   | 15 ----------
 15 files changed, 61 insertions(+), 234 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_object.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 8b96e7aaeff1..1aa516429c80 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -137,7 +137,7 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
 		places[c].mem_type = TTM_PL_VRAM;
-		places[c].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED;
+		places[c].flags = 0;
 
 		if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED)
 			places[c].lpfn = visible_pfn;
@@ -154,11 +154,6 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 		places[c].lpfn = 0;
 		places[c].mem_type = TTM_PL_TT;
 		places[c].flags = 0;
-		if (flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
-			places[c].flags |= TTM_PL_FLAG_WC |
-				TTM_PL_FLAG_UNCACHED;
-		else
-			places[c].flags |= TTM_PL_FLAG_CACHED;
 		c++;
 	}
 
@@ -167,11 +162,6 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 		places[c].lpfn = 0;
 		places[c].mem_type = TTM_PL_SYSTEM;
 		places[c].flags = 0;
-		if (flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
-			places[c].flags |= TTM_PL_FLAG_WC |
-				TTM_PL_FLAG_UNCACHED;
-		else
-			places[c].flags |= TTM_PL_FLAG_CACHED;
 		c++;
 	}
 
@@ -179,7 +169,7 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
 		places[c].mem_type = AMDGPU_PL_GDS;
-		places[c].flags = TTM_PL_FLAG_UNCACHED;
+		places[c].flags = 0;
 		c++;
 	}
 
@@ -187,7 +177,7 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
 		places[c].mem_type = AMDGPU_PL_GWS;
-		places[c].flags = TTM_PL_FLAG_UNCACHED;
+		places[c].flags = 0;
 		c++;
 	}
 
@@ -195,7 +185,7 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
 		places[c].mem_type = AMDGPU_PL_OA;
-		places[c].flags = TTM_PL_FLAG_UNCACHED;
+		places[c].flags = 0;
 		c++;
 	}
 
@@ -203,7 +193,7 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
 		places[c].mem_type = TTM_PL_SYSTEM;
-		places[c].flags = TTM_PL_MASK_CACHING;
+		places[c].flags = 0;
 		c++;
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 5b56a66063fd..8cdec58b9106 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -92,7 +92,7 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo,
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = TTM_PL_SYSTEM,
-		.flags = TTM_PL_MASK_CACHING
+		.flags = 0
 	};
 
 	/* Don't handle scatter gather BOs */
@@ -538,19 +538,13 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, bool evict,
 	placements.fpfn = 0;
 	placements.lpfn = 0;
 	placements.mem_type = TTM_PL_TT;
-	placements.flags = TTM_PL_MASK_CACHING;
+	placements.flags = 0;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx);
 	if (unlikely(r)) {
 		pr_err("Failed to find GTT space for blit from VRAM\n");
 		return r;
 	}
 
-	/* set caching flags */
-	r = ttm_tt_set_placement_caching(bo->ttm, tmp_mem.placement);
-	if (unlikely(r)) {
-		goto out_cleanup;
-	}
-
 	r = ttm_tt_populate(bo->bdev, bo->ttm, ctx);
 	if (unlikely(r))
 		goto out_cleanup;
@@ -599,7 +593,7 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo, bool evict,
 	placements.fpfn = 0;
 	placements.lpfn = 0;
 	placements.mem_type = TTM_PL_TT;
-	placements.flags = TTM_PL_MASK_CACHING;
+	placements.flags = 0;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx);
 	if (unlikely(r)) {
 		pr_err("Failed to find GTT space for blit to VRAM\n");
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index b9e7ce1adf25..7aeb5daf2805 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -147,15 +147,12 @@ static void drm_gem_vram_placement(struct drm_gem_vram_object *gbo,
 
 	if (pl_flag & DRM_GEM_VRAM_PL_FLAG_VRAM) {
 		gbo->placements[c].mem_type = TTM_PL_VRAM;
-		gbo->placements[c++].flags = TTM_PL_FLAG_WC |
-					     TTM_PL_FLAG_UNCACHED |
-					     invariant_flags;
+		gbo->placements[c++].flags = invariant_flags;
 	}
 
 	if (pl_flag & DRM_GEM_VRAM_PL_FLAG_SYSTEM || !c) {
 		gbo->placements[c].mem_type = TTM_PL_SYSTEM;
-		gbo->placements[c++].flags = TTM_PL_MASK_CACHING |
-					     invariant_flags;
+		gbo->placements[c++].flags = invariant_flags;
 	}
 
 	gbo->placement.num_placement = c;
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index cb878c0e8276..4ccb3329014b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -343,37 +343,23 @@ nouveau_bo_new(struct nouveau_cli *cli, u64 size, int align,
 }
 
 static void
-set_placement_list(struct nouveau_drm *drm, struct ttm_place *pl, unsigned *n,
-		   uint32_t domain, uint32_t flags)
+set_placement_list(struct ttm_place *pl, unsigned *n, uint32_t domain)
 {
 	*n = 0;
 
 	if (domain & NOUVEAU_GEM_DOMAIN_VRAM) {
-		struct nvif_mmu *mmu = &drm->client.mmu;
-		const u8 type = mmu->type[drm->ttm.type_vram].type;
-
 		pl[*n].mem_type = TTM_PL_VRAM;
-		pl[*n].flags = flags & ~TTM_PL_FLAG_CACHED;
-
-		/* Some BARs do not support being ioremapped WC */
-		if (drm->client.device.info.family >= NV_DEVICE_INFO_V0_TESLA &&
-		    type & NVIF_MEM_UNCACHED)
-			pl[*n].flags &= ~TTM_PL_FLAG_WC;
-
+		pl[*n].flags = 0;
 		(*n)++;
 	}
 	if (domain & NOUVEAU_GEM_DOMAIN_GART) {
 		pl[*n].mem_type = TTM_PL_TT;
-		pl[*n].flags = flags;
-
-		if (drm->agp.bridge)
-			pl[*n].flags &= ~TTM_PL_FLAG_CACHED;
-
+		pl[*n].flags = 0;
 		(*n)++;
 	}
 	if (domain & NOUVEAU_GEM_DOMAIN_CPU) {
 		pl[*n].mem_type = TTM_PL_SYSTEM;
-		pl[(*n)++].flags = flags;
+		pl[(*n)++].flags = 0;
 	}
 }
 
@@ -415,18 +401,14 @@ void
 nouveau_bo_placement_set(struct nouveau_bo *nvbo, uint32_t domain,
 			 uint32_t busy)
 {
-	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct ttm_placement *pl = &nvbo->placement;
-	uint32_t flags = nvbo->force_coherent ? TTM_PL_FLAG_UNCACHED :
-						TTM_PL_MASK_CACHING;
 
 	pl->placement = nvbo->placements;
-	set_placement_list(drm, nvbo->placements, &pl->num_placement,
-			   domain, flags);
+	set_placement_list(nvbo->placements, &pl->num_placement, domain);
 
 	pl->busy_placement = nvbo->busy_placements;
-	set_placement_list(drm, nvbo->busy_placements, &pl->num_busy_placement,
-			   domain | busy, flags);
+	set_placement_list(nvbo->busy_placements, &pl->num_busy_placement,
+			   domain | busy);
 
 	set_placement_range(nvbo, domain);
 }
@@ -888,7 +870,7 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict,
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = TTM_PL_TT,
-		.flags = TTM_PL_MASK_CACHING
+		.flags = 0
 	};
 	struct ttm_placement placement;
 	struct ttm_resource tmp_reg;
@@ -930,7 +912,7 @@ nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict,
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = TTM_PL_TT,
-		.flags = TTM_PL_MASK_CACHING
+		.flags = 0
 	};
 	struct ttm_placement placement;
 	struct ttm_resource tmp_reg;
diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c
index 940e99354f49..547d46c14d56 100644
--- a/drivers/gpu/drm/qxl/qxl_object.c
+++ b/drivers/gpu/drm/qxl/qxl_object.c
@@ -64,21 +64,21 @@ void qxl_ttm_placement_from_domain(struct qxl_bo *qbo, u32 domain)
 	qbo->placement.busy_placement = qbo->placements;
 	if (domain == QXL_GEM_DOMAIN_VRAM) {
 		qbo->placements[c].mem_type = TTM_PL_VRAM;
-		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | pflag;
+		qbo->placements[c++].flags = pflag;
 	}
 	if (domain == QXL_GEM_DOMAIN_SURFACE) {
 		qbo->placements[c].mem_type = TTM_PL_PRIV;
-		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | pflag;
+		qbo->placements[c++].flags = pflag;
 		qbo->placements[c].mem_type = TTM_PL_VRAM;
-		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | pflag;
+		qbo->placements[c++].flags = pflag;
 	}
 	if (domain == QXL_GEM_DOMAIN_CPU) {
 		qbo->placements[c].mem_type = TTM_PL_SYSTEM;
-		qbo->placements[c++].flags = TTM_PL_MASK_CACHING | pflag;
+		qbo->placements[c++].flags = pflag;
 	}
 	if (!c) {
 		qbo->placements[c].mem_type = TTM_PL_SYSTEM;
-		qbo->placements[c++].flags = TTM_PL_MASK_CACHING;
+		qbo->placements[c++].flags = 0;
 	}
 	qbo->placement.num_placement = c;
 	qbo->placement.num_busy_placement = c;
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index 61eb06dbbce8..e3ed20215f18 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -56,7 +56,7 @@ static void qxl_evict_flags(struct ttm_buffer_object *bo,
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = TTM_PL_SYSTEM,
-		.flags = TTM_PL_MASK_CACHING
+		.flags = 0
 	};
 
 	if (!qxl_ttm_bo_is_qxl_bo(bo)) {
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index ad0e6e9ef922..ab81e35cb060 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -113,57 +113,29 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain)
 			rbo->placements[c].fpfn =
 				rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
 			rbo->placements[c].mem_type = TTM_PL_VRAM;
-			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
-						     TTM_PL_FLAG_UNCACHED;
+			rbo->placements[c++].flags = 0;
 		}
 
 		rbo->placements[c].fpfn = 0;
 		rbo->placements[c].mem_type = TTM_PL_VRAM;
-		rbo->placements[c++].flags = TTM_PL_FLAG_WC |
-					     TTM_PL_FLAG_UNCACHED;
+		rbo->placements[c++].flags = 0;
 	}
 
 	if (domain & RADEON_GEM_DOMAIN_GTT) {
-		if (rbo->flags & RADEON_GEM_GTT_UC) {
-			rbo->placements[c].fpfn = 0;
-			rbo->placements[c].mem_type = TTM_PL_TT;
-			rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED;
-
-		} else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
-			   (rbo->rdev->flags & RADEON_IS_AGP)) {
-			rbo->placements[c].fpfn = 0;
-			rbo->placements[c].mem_type = TTM_PL_TT;
-			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
-				TTM_PL_FLAG_UNCACHED;
-		} else {
-			rbo->placements[c].fpfn = 0;
-			rbo->placements[c].mem_type = TTM_PL_TT;
-			rbo->placements[c++].flags = TTM_PL_FLAG_CACHED;
-		}
+		rbo->placements[c].fpfn = 0;
+		rbo->placements[c].mem_type = TTM_PL_TT;
+		rbo->placements[c++].flags = 0;
 	}
 
 	if (domain & RADEON_GEM_DOMAIN_CPU) {
-		if (rbo->flags & RADEON_GEM_GTT_UC) {
-			rbo->placements[c].fpfn = 0;
-			rbo->placements[c].mem_type = TTM_PL_SYSTEM;
-			rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED;
-
-		} else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
-		    rbo->rdev->flags & RADEON_IS_AGP) {
-			rbo->placements[c].fpfn = 0;
-			rbo->placements[c].mem_type = TTM_PL_SYSTEM;
-			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
-				TTM_PL_FLAG_UNCACHED;
-		} else {
-			rbo->placements[c].fpfn = 0;
-			rbo->placements[c].mem_type = TTM_PL_SYSTEM;
-			rbo->placements[c++].flags = TTM_PL_FLAG_CACHED;
-		}
+		rbo->placements[c].fpfn = 0;
+		rbo->placements[c].mem_type = TTM_PL_SYSTEM;
+		rbo->placements[c++].flags = 0;
 	}
 	if (!c) {
 		rbo->placements[c].fpfn = 0;
 		rbo->placements[c].mem_type = TTM_PL_SYSTEM;
-		rbo->placements[c++].flags = TTM_PL_MASK_CACHING;
+		rbo->placements[c++].flags = 0;
 	}
 
 	rbo->placement.num_placement = c;
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 9b53a1d80632..d6f42fbc81f4 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -89,7 +89,7 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo,
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = TTM_PL_SYSTEM,
-		.flags = TTM_PL_MASK_CACHING
+		.flags = 0
 	};
 
 	struct radeon_bo *rbo;
@@ -225,17 +225,12 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
 	placements.fpfn = 0;
 	placements.lpfn = 0;
 	placements.mem_type = TTM_PL_TT;
-	placements.flags = TTM_PL_MASK_CACHING;
+	placements.flags = 0;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx);
 	if (unlikely(r)) {
 		return r;
 	}
 
-	r = ttm_tt_set_placement_caching(bo->ttm, tmp_mem.placement);
-	if (unlikely(r)) {
-		goto out_cleanup;
-	}
-
 	r = ttm_tt_populate(bo->bdev, bo->ttm, ctx);
 	if (unlikely(r)) {
 		goto out_cleanup;
@@ -275,7 +270,7 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
 	placements.fpfn = 0;
 	placements.lpfn = 0;
 	placements.mem_type = TTM_PL_TT;
-	placements.flags = TTM_PL_MASK_CACHING;
+	placements.flags = 0;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx);
 	if (unlikely(r)) {
 		return r;
@@ -389,12 +384,7 @@ static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 		 * Alpha: use bus.addr to hold the ioremap() return,
 		 * so we can modify bus.base below.
 		 */
-		if (mem->placement & TTM_PL_FLAG_WC)
-			mem->bus.addr =
-				ioremap_wc(mem->bus.offset, bus_size);
-		else
-			mem->bus.addr =
-				ioremap(mem->bus.offset, bus_size);
+		mem->bus.addr = ioremap_wc(mem->bus.offset, bus_size);
 		if (!mem->bus.addr)
 			return -ENOMEM;
 
diff --git a/drivers/gpu/drm/ttm/ttm_agp_backend.c b/drivers/gpu/drm/ttm/ttm_agp_backend.c
index a723062d37e7..4f76c9287159 100644
--- a/drivers/gpu/drm/ttm/ttm_agp_backend.c
+++ b/drivers/gpu/drm/ttm/ttm_agp_backend.c
@@ -54,7 +54,7 @@ int ttm_agp_bind(struct ttm_tt *ttm, struct ttm_resource *bo_mem)
 	struct page *dummy_read_page = ttm_bo_glob.dummy_read_page;
 	struct drm_mm_node *node = bo_mem->mm_node;
 	struct agp_memory *mem;
-	int ret, cached = (bo_mem->placement & TTM_PL_FLAG_CACHED);
+	int ret, cached = ttm->caching == ttm_cached;
 	unsigned i;
 
 	if (agp_be->mem)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 3732dcb58aad..b97ed6ca8765 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -252,10 +252,6 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 		if (ret)
 			goto out_err;
 
-		ret = ttm_tt_set_placement_caching(bo->ttm, mem->placement);
-		if (ret)
-			goto out_err;
-
 		if (mem->mem_type != TTM_PL_SYSTEM) {
 			ret = ttm_tt_populate(bdev, bo->ttm, ctx);
 			if (ret)
@@ -843,29 +839,6 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
 	return ttm_bo_add_move_fence(bo, man, mem, ctx->no_wait_gpu);
 }
 
-static uint32_t ttm_bo_select_caching(struct ttm_resource_manager *man,
-				      uint32_t cur_placement,
-				      uint32_t proposed_placement)
-{
-	uint32_t caching = proposed_placement & TTM_PL_MASK_CACHING;
-	uint32_t result = proposed_placement & ~TTM_PL_MASK_CACHING;
-
-	/**
-	 * Keep current caching if possible.
-	 */
-
-	if ((cur_placement & caching) != 0)
-		result |= (cur_placement & caching);
-	else if ((TTM_PL_FLAG_CACHED & caching) != 0)
-		result |= TTM_PL_FLAG_CACHED;
-	else if ((TTM_PL_FLAG_WC & caching) != 0)
-		result |= TTM_PL_FLAG_WC;
-	else if ((TTM_PL_FLAG_UNCACHED & caching) != 0)
-		result |= TTM_PL_FLAG_UNCACHED;
-
-	return result;
-}
-
 /**
  * ttm_bo_mem_placement - check if placement is compatible
  * @bo: BO to find memory for
@@ -884,18 +857,13 @@ static int ttm_bo_mem_placement(struct ttm_buffer_object *bo,
 {
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_resource_manager *man;
-	uint32_t cur_flags = 0;
 
 	man = ttm_manager_type(bdev, place->mem_type);
 	if (!man || !ttm_resource_manager_used(man))
 		return -EBUSY;
 
-	cur_flags = ttm_bo_select_caching(man, bo->mem.placement,
-					  place->flags);
-	cur_flags |= place->flags & ~TTM_PL_MASK_CACHING;
-
 	mem->mem_type = place->mem_type;
-	mem->placement = cur_flags;
+	mem->placement = place->flags;
 
 	spin_lock(&ttm_bo_glob.lru_lock);
 	ttm_bo_del_from_lru(bo);
@@ -1028,8 +996,7 @@ static bool ttm_bo_places_compat(const struct ttm_place *places,
 			continue;
 
 		*new_flags = heap->flags;
-		if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
-		    (mem->mem_type == heap->mem_type) &&
+		if ((mem->mem_type == heap->mem_type) &&
 		    (!(*new_flags & TTM_PL_FLAG_CONTIGUOUS) ||
 		     (mem->placement & TTM_PL_FLAG_CONTIGUOUS)))
 			return true;
@@ -1083,9 +1050,6 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 		ret = ttm_bo_move_buffer(bo, placement, ctx);
 		if (ret)
 			return ret;
-	} else {
-		bo->mem.placement &= TTM_PL_MASK_CACHING;
-		bo->mem.placement |= new_flags & ~TTM_PL_MASK_CACHING;
 	}
 	/*
 	 * We might need to add a TTM.
@@ -1153,7 +1117,7 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev,
 	bo->mem.bus.offset = 0;
 	bo->mem.bus.addr = NULL;
 	bo->moving = NULL;
-	bo->mem.placement = TTM_PL_FLAG_CACHED;
+	bo->mem.placement = 0;
 	bo->acc_size = acc_size;
 	bo->pin_count = 0;
 	bo->sg = sg;
@@ -1484,7 +1448,7 @@ int ttm_bo_swapout(struct ttm_operation_ctx *ctx)
 
 		evict_mem = bo->mem;
 		evict_mem.mm_node = NULL;
-		evict_mem.placement = TTM_PL_MASK_CACHING;
+		evict_mem.placement = 0;
 		evict_mem.mem_type = TTM_PL_SYSTEM;
 
 		ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, &ctx);
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 0542097dc419..ba7ab5ed85d0 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -72,10 +72,6 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo,
 		old_mem->mem_type = TTM_PL_SYSTEM;
 	}
 
-	ret = ttm_tt_set_placement_caching(ttm, new_mem->placement);
-	if (unlikely(ret != 0))
-		return ret;
-
 	if (new_mem->mem_type != TTM_PL_SYSTEM) {
 
 		ret = ttm_tt_populate(bo->bdev, ttm, ctx);
@@ -135,7 +131,7 @@ static int ttm_resource_ioremap(struct ttm_bo_device *bdev,
 	} else {
 		size_t bus_size = (size_t)mem->num_pages << PAGE_SHIFT;
 
-		if (mem->placement & TTM_PL_FLAG_WC)
+		if (mem->bus.caching == ttm_write_combined)
 			addr = ioremap_wc(mem->bus.offset, bus_size);
 		else
 			addr = ioremap(mem->bus.offset, bus_size);
@@ -427,7 +423,7 @@ static int ttm_bo_ioremap(struct ttm_buffer_object *bo,
 		map->virtual = (void *)(((u8 *)bo->mem.bus.addr) + offset);
 	} else {
 		map->bo_kmap_type = ttm_bo_map_iomap;
-		if (mem->placement & TTM_PL_FLAG_WC)
+		if (mem->bus.caching == ttm_write_combined)
 			map->virtual = ioremap_wc(bo->mem.bus.offset + offset,
 						  size);
 		else
@@ -457,7 +453,7 @@ static int ttm_bo_kmap_ttm(struct ttm_buffer_object *bo,
 	if (ret)
 		return ret;
 
-	if (num_pages == 1 && (mem->placement & TTM_PL_FLAG_CACHED)) {
+	if (num_pages == 1 && ttm->caching == ttm_cached) {
 		/*
 		 * We're mapping a single page, and the desired
 		 * page protection is consistent with the bo.
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index a465f51df027..3e5dd6271d4c 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -114,35 +114,6 @@ static int ttm_sg_tt_alloc_page_directory(struct ttm_dma_tt *ttm)
 	return 0;
 }
 
-static int ttm_tt_set_caching(struct ttm_tt *ttm, enum ttm_caching caching)
-{
-	if (ttm->caching == caching)
-		return 0;
-
-	/* Can't change the caching state after TT is populated */
-	if (WARN_ON_ONCE(ttm_tt_is_populated(ttm)))
-		return -EINVAL;
-
-	ttm->caching = caching;
-
-	return 0;
-}
-
-int ttm_tt_set_placement_caching(struct ttm_tt *ttm, uint32_t placement)
-{
-	enum ttm_caching state;
-
-	if (placement & TTM_PL_FLAG_WC)
-		state = ttm_write_combined;
-	else if (placement & TTM_PL_FLAG_UNCACHED)
-		state = ttm_uncached;
-	else
-		state = ttm_cached;
-
-	return ttm_tt_set_caching(ttm, state);
-}
-EXPORT_SYMBOL(ttm_tt_set_placement_caching);
-
 void ttm_tt_destroy_common(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
 	ttm_tt_unpopulate(bdev, ttm);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index fae88969a15a..112253246f08 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -34,28 +34,28 @@ static const struct ttm_place vram_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
 	.mem_type = TTM_PL_VRAM,
-	.flags = TTM_PL_FLAG_CACHED
+	.flags = 0
 };
 
 static const struct ttm_place sys_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
 	.mem_type = TTM_PL_SYSTEM,
-	.flags = TTM_PL_FLAG_CACHED
+	.flags = 0
 };
 
 static const struct ttm_place gmr_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
 	.mem_type = VMW_PL_GMR,
-	.flags = TTM_PL_FLAG_CACHED
+	.flags = 0
 };
 
 static const struct ttm_place mob_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
 	.mem_type = VMW_PL_MOB,
-	.flags = TTM_PL_FLAG_CACHED
+	.flags = 0
 };
 
 struct ttm_placement vmw_vram_placement = {
@@ -70,12 +70,12 @@ static const struct ttm_place vram_gmr_placement_flags[] = {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = TTM_PL_VRAM,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = VMW_PL_GMR,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}
 };
 
@@ -84,12 +84,12 @@ static const struct ttm_place gmr_vram_placement_flags[] = {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = VMW_PL_GMR,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = TTM_PL_VRAM,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}
 };
 
@@ -119,22 +119,22 @@ static const struct ttm_place evictable_placement_flags[] = {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = TTM_PL_SYSTEM,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = TTM_PL_VRAM,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = VMW_PL_GMR,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = VMW_PL_MOB,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}
 };
 
@@ -143,17 +143,17 @@ static const struct ttm_place nonfixed_placement_flags[] = {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = TTM_PL_SYSTEM,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = VMW_PL_GMR,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
 		.mem_type = VMW_PL_MOB,
-		.flags = TTM_PL_FLAG_CACHED
+		.flags = 0
 	}
 };
 
diff --git a/include/drm/ttm/ttm_placement.h b/include/drm/ttm/ttm_placement.h
index 50e72df48b8d..aa6ba4d0cf78 100644
--- a/include/drm/ttm/ttm_placement.h
+++ b/include/drm/ttm/ttm_placement.h
@@ -43,27 +43,13 @@
 #define TTM_PL_PRIV             3
 
 /*
- * Other flags that affects data placement.
- * TTM_PL_FLAG_CACHED indicates cache-coherent mappings
- * if available.
- * TTM_PL_FLAG_SHARED means that another application may
- * reference the buffer.
- * TTM_PL_FLAG_NO_EVICT means that the buffer may never
- * be evicted to make room for other buffers.
  * TTM_PL_FLAG_TOPDOWN requests to be placed from the
  * top of the memory area, instead of the bottom.
  */
 
-#define TTM_PL_FLAG_CACHED      (1 << 16)
-#define TTM_PL_FLAG_UNCACHED    (1 << 17)
-#define TTM_PL_FLAG_WC          (1 << 18)
 #define TTM_PL_FLAG_CONTIGUOUS  (1 << 19)
 #define TTM_PL_FLAG_TOPDOWN     (1 << 22)
 
-#define TTM_PL_MASK_CACHING     (TTM_PL_FLAG_CACHED | \
-				 TTM_PL_FLAG_UNCACHED | \
-				 TTM_PL_FLAG_WC)
-
 /**
  * struct ttm_place
  *
diff --git a/include/drm/ttm/ttm_tt.h b/include/drm/ttm/ttm_tt.h
index c39c722d5184..e042dec5e6c1 100644
--- a/include/drm/ttm/ttm_tt.h
+++ b/include/drm/ttm/ttm_tt.h
@@ -164,21 +164,6 @@ void ttm_tt_destroy_common(struct ttm_bo_device *bdev, struct ttm_tt *ttm);
  * Swap in a previously swap out ttm_tt.
  */
 int ttm_tt_swapin(struct ttm_tt *ttm);
-
-/**
- * ttm_tt_set_placement_caching:
- *
- * @ttm A struct ttm_tt the backing pages of which will change caching policy.
- * @placement: Flag indicating the desired caching policy.
- *
- * This function will change caching policy of any default kernel mappings of
- * the pages backing @ttm. If changing from cached to uncached or
- * write-combined,
- * all CPU caches will first be flushed to make sure the data of the pages
- * hit RAM. This function may be very costly as it involves global TLB
- * and cache flushes and potential page splitting / combining.
- */
-int ttm_tt_set_placement_caching(struct ttm_tt *ttm, uint32_t placement);
 int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm);
 
 /**
-- 
cgit v1.2.3


From ff72bc40317079b775a7a6d2ba94297b04b795e4 Mon Sep 17 00:00:00 2001
From: Mihir Bhogilal Patel <Mihir.Patel@amd.com>
Date: Thu, 8 Oct 2020 15:46:38 +0530
Subject: drm/amdgpu: Add debugfs entry for printing VM info
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Create new debugfs entry to print memory info using VM buffer
objects.

V2: Added Common function for printing BO info.
    Dump more VM lists for evicted, moved, relocated, invalidated.
    Removed dumping VM mapped BOs.
V3: Fixed coding style comments, renamed print API and variables.
V4: Fixed coding style comments.

Signed-off-by: Mihir Bhogilal Patel <Mihir.Patel@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c | 30 +++++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c     | 69 +++---------------------
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c  | 74 +++++++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.h  |  1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c      | 83 +++++++++++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h      |  4 ++
 6 files changed, 199 insertions(+), 62 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_object.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index 2d125b8b15ee..8ec4806c2ff4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -1335,11 +1335,41 @@ static int amdgpu_debugfs_evict_gtt(struct seq_file *m, void *data)
 	return 0;
 }
 
+static int amdgpu_debugfs_vm_info(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_file *file;
+	int r;
+
+	r = mutex_lock_interruptible(&dev->filelist_mutex);
+	if (r)
+		return r;
+
+	list_for_each_entry(file, &dev->filelist, lhead) {
+		struct amdgpu_fpriv *fpriv = file->driver_priv;
+		struct amdgpu_vm *vm = &fpriv->vm;
+
+		seq_printf(m, "pid:%d\tProcess:%s ----------\n",
+				vm->task_info.pid, vm->task_info.process_name);
+		r = amdgpu_bo_reserve(vm->root.base.bo, true);
+		if (r)
+			break;
+		amdgpu_debugfs_vm_bo_info(vm, m);
+		amdgpu_bo_unreserve(vm->root.base.bo);
+	}
+
+	mutex_unlock(&dev->filelist_mutex);
+
+	return r;
+}
+
 static const struct drm_info_list amdgpu_debugfs_list[] = {
 	{"amdgpu_vbios", amdgpu_debugfs_get_vbios_dump},
 	{"amdgpu_test_ib", &amdgpu_debugfs_test_ib},
 	{"amdgpu_evict_vram", &amdgpu_debugfs_evict_vram},
 	{"amdgpu_evict_gtt", &amdgpu_debugfs_evict_gtt},
+	{"amdgpu_vm_info", &amdgpu_debugfs_vm_info},
 };
 
 static void amdgpu_ib_preempt_fences_swap(struct amdgpu_ring *ring,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index aa7f230c71bf..5047989fe941 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -826,67 +826,6 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 }
 
 #if defined(CONFIG_DEBUG_FS)
-
-#define amdgpu_debugfs_gem_bo_print_flag(m, bo, flag)	\
-	if (bo->flags & (AMDGPU_GEM_CREATE_ ## flag)) {	\
-		seq_printf((m), " " #flag);		\
-	}
-
-static int amdgpu_debugfs_gem_bo_info(int id, void *ptr, void *data)
-{
-	struct drm_gem_object *gobj = ptr;
-	struct amdgpu_bo *bo = gem_to_amdgpu_bo(gobj);
-	struct seq_file *m = data;
-
-	struct dma_buf_attachment *attachment;
-	struct dma_buf *dma_buf;
-	unsigned domain;
-	const char *placement;
-	unsigned pin_count;
-
-	domain = amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type);
-	switch (domain) {
-	case AMDGPU_GEM_DOMAIN_VRAM:
-		placement = "VRAM";
-		break;
-	case AMDGPU_GEM_DOMAIN_GTT:
-		placement = " GTT";
-		break;
-	case AMDGPU_GEM_DOMAIN_CPU:
-	default:
-		placement = " CPU";
-		break;
-	}
-	seq_printf(m, "\t0x%08x: %12ld byte %s",
-		   id, amdgpu_bo_size(bo), placement);
-
-	pin_count = READ_ONCE(bo->pin_count);
-	if (pin_count)
-		seq_printf(m, " pin count %d", pin_count);
-
-	dma_buf = READ_ONCE(bo->tbo.base.dma_buf);
-	attachment = READ_ONCE(bo->tbo.base.import_attach);
-
-	if (attachment)
-		seq_printf(m, " imported from %p%s", dma_buf,
-			   attachment->peer2peer ? " P2P" : "");
-	else if (dma_buf)
-		seq_printf(m, " exported as %p", dma_buf);
-
-	amdgpu_debugfs_gem_bo_print_flag(m, bo, CPU_ACCESS_REQUIRED);
-	amdgpu_debugfs_gem_bo_print_flag(m, bo, NO_CPU_ACCESS);
-	amdgpu_debugfs_gem_bo_print_flag(m, bo, CPU_GTT_USWC);
-	amdgpu_debugfs_gem_bo_print_flag(m, bo, VRAM_CLEARED);
-	amdgpu_debugfs_gem_bo_print_flag(m, bo, SHADOW);
-	amdgpu_debugfs_gem_bo_print_flag(m, bo, VRAM_CONTIGUOUS);
-	amdgpu_debugfs_gem_bo_print_flag(m, bo, VM_ALWAYS_VALID);
-	amdgpu_debugfs_gem_bo_print_flag(m, bo, EXPLICIT_SYNC);
-
-	seq_printf(m, "\n");
-
-	return 0;
-}
-
 static int amdgpu_debugfs_gem_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = (struct drm_info_node *)m->private;
@@ -900,6 +839,8 @@ static int amdgpu_debugfs_gem_info(struct seq_file *m, void *data)
 
 	list_for_each_entry(file, &dev->filelist, lhead) {
 		struct task_struct *task;
+		struct drm_gem_object *gobj;
+		int id;
 
 		/*
 		 * Although we have a valid reference on file->pid, that does
@@ -914,7 +855,11 @@ static int amdgpu_debugfs_gem_info(struct seq_file *m, void *data)
 		rcu_read_unlock();
 
 		spin_lock(&file->table_lock);
-		idr_for_each(&file->object_idr, amdgpu_debugfs_gem_bo_info, m);
+		idr_for_each_entry(&file->object_idr, gobj, id) {
+			struct amdgpu_bo *bo = gem_to_amdgpu_bo(gobj);
+
+			amdgpu_bo_print_info(id, bo, m);
+		}
 		spin_unlock(&file->table_lock);
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 07e81aca9c34..b191701098f0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -1528,3 +1528,77 @@ uint32_t amdgpu_bo_get_preferred_pin_domain(struct amdgpu_device *adev,
 	}
 	return domain;
 }
+
+#if defined(CONFIG_DEBUG_FS)
+#define amdgpu_bo_print_flag(m, bo, flag)		        \
+	do {							\
+		if (bo->flags & (AMDGPU_GEM_CREATE_ ## flag)) {	\
+			seq_printf((m), " " #flag);		\
+		}						\
+	} while (0)
+
+/**
+ * amdgpu_debugfs_print_bo_info - print BO info in debugfs file
+ *
+ * @id: Index or Id of the BO
+ * @bo: Requested BO for printing info
+ * @m: debugfs file
+ *
+ * Print BO information in debugfs file
+ *
+ * Returns:
+ * Size of the BO in bytes.
+ */
+u64 amdgpu_bo_print_info(int id, struct amdgpu_bo *bo, struct seq_file *m)
+{
+	struct dma_buf_attachment *attachment;
+	struct dma_buf *dma_buf;
+	unsigned int domain;
+	const char *placement;
+	unsigned int pin_count;
+	u64 size;
+
+	domain = amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type);
+	switch (domain) {
+	case AMDGPU_GEM_DOMAIN_VRAM:
+		placement = "VRAM";
+		break;
+	case AMDGPU_GEM_DOMAIN_GTT:
+		placement = " GTT";
+		break;
+	case AMDGPU_GEM_DOMAIN_CPU:
+	default:
+		placement = " CPU";
+		break;
+	}
+
+	size = amdgpu_bo_size(bo);
+	seq_printf(m, "\t\t0x%08x: %12lld byte %s",
+			id, size, placement);
+
+	pin_count = READ_ONCE(bo->pin_count);
+	if (pin_count)
+		seq_printf(m, " pin count %d", pin_count);
+
+	dma_buf = READ_ONCE(bo->tbo.base.dma_buf);
+	attachment = READ_ONCE(bo->tbo.base.import_attach);
+
+	if (attachment)
+		seq_printf(m, " imported from %p", dma_buf);
+	else if (dma_buf)
+		seq_printf(m, " exported as %p", dma_buf);
+
+	amdgpu_bo_print_flag(m, bo, CPU_ACCESS_REQUIRED);
+	amdgpu_bo_print_flag(m, bo, NO_CPU_ACCESS);
+	amdgpu_bo_print_flag(m, bo, CPU_GTT_USWC);
+	amdgpu_bo_print_flag(m, bo, VRAM_CLEARED);
+	amdgpu_bo_print_flag(m, bo, SHADOW);
+	amdgpu_bo_print_flag(m, bo, VRAM_CONTIGUOUS);
+	amdgpu_bo_print_flag(m, bo, VM_ALWAYS_VALID);
+	amdgpu_bo_print_flag(m, bo, EXPLICIT_SYNC);
+
+	seq_puts(m, "\n");
+
+	return size;
+}
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
index 5ddb6cf96030..621c0bfee6e3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
@@ -330,6 +330,7 @@ void amdgpu_sa_bo_free(struct amdgpu_device *adev,
 #if defined(CONFIG_DEBUG_FS)
 void amdgpu_sa_bo_dump_debug_info(struct amdgpu_sa_manager *sa_manager,
 					 struct seq_file *m);
+u64 amdgpu_bo_print_info(int id, struct amdgpu_bo *bo, struct seq_file *m);
 #endif
 int amdgpu_debugfs_sa_init(struct amdgpu_device *adev);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 2b65e83c808b..461fcde19c8a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -3392,3 +3392,86 @@ error_unref:
 
 	return false;
 }
+
+#if defined(CONFIG_DEBUG_FS)
+/**
+ * amdgpu_debugfs_vm_bo_info  - print BO info for the VM
+ *
+ * @vm: Requested VM for printing BO info
+ * @m: debugfs file
+ *
+ * Print BO information in debugfs file for the VM
+ */
+void amdgpu_debugfs_vm_bo_info(struct amdgpu_vm *vm, struct seq_file *m)
+{
+	struct amdgpu_bo_va *bo_va, *tmp;
+	u64 total_idle = 0;
+	u64 total_evicted = 0;
+	u64 total_relocated = 0;
+	u64 total_moved = 0;
+	u64 total_invalidated = 0;
+	unsigned int total_idle_objs = 0;
+	unsigned int total_evicted_objs = 0;
+	unsigned int total_relocated_objs = 0;
+	unsigned int total_moved_objs = 0;
+	unsigned int total_invalidated_objs = 0;
+	unsigned int id = 0;
+
+	seq_puts(m, "\tIdle BOs:\n");
+	list_for_each_entry_safe(bo_va, tmp, &vm->idle, base.vm_status) {
+		if (!bo_va->base.bo)
+			continue;
+		total_idle += amdgpu_bo_print_info(id++, bo_va->base.bo, m);
+	}
+	total_idle_objs = id;
+	id = 0;
+
+	seq_puts(m, "\tEvicted BOs:\n");
+	list_for_each_entry_safe(bo_va, tmp, &vm->evicted, base.vm_status) {
+		if (!bo_va->base.bo)
+			continue;
+		total_evicted += amdgpu_bo_print_info(id++, bo_va->base.bo, m);
+	}
+	total_evicted_objs = id;
+	id = 0;
+
+	seq_puts(m, "\tRelocated BOs:\n");
+	list_for_each_entry_safe(bo_va, tmp, &vm->relocated, base.vm_status) {
+		if (!bo_va->base.bo)
+			continue;
+		total_relocated += amdgpu_bo_print_info(id++, bo_va->base.bo, m);
+	}
+	total_relocated_objs = id;
+	id = 0;
+
+	seq_puts(m, "\tMoved BOs:\n");
+	list_for_each_entry_safe(bo_va, tmp, &vm->moved, base.vm_status) {
+		if (!bo_va->base.bo)
+			continue;
+		total_moved += amdgpu_bo_print_info(id++, bo_va->base.bo, m);
+	}
+	total_moved_objs = id;
+	id = 0;
+
+	seq_puts(m, "\tInvalidated BOs:\n");
+	spin_lock(&vm->invalidated_lock);
+	list_for_each_entry_safe(bo_va, tmp, &vm->invalidated, base.vm_status) {
+		if (!bo_va->base.bo)
+			continue;
+		total_invalidated += amdgpu_bo_print_info(id++,	bo_va->base.bo, m);
+	}
+	spin_unlock(&vm->invalidated_lock);
+	total_invalidated_objs = id;
+
+	seq_printf(m, "\tTotal idle size:        %12lld\tobjs:\t%d\n", total_idle,
+		   total_idle_objs);
+	seq_printf(m, "\tTotal evicted size:     %12lld\tobjs:\t%d\n", total_evicted,
+		   total_evicted_objs);
+	seq_printf(m, "\tTotal relocated size:   %12lld\tobjs:\t%d\n", total_relocated,
+		   total_relocated_objs);
+	seq_printf(m, "\tTotal moved size:       %12lld\tobjs:\t%d\n", total_moved,
+		   total_moved_objs);
+	seq_printf(m, "\tTotal invalidated size: %12lld\tobjs:\t%d\n", total_invalidated,
+		   total_invalidated_objs);
+}
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 7c46937c1c0e..74cc14179c41 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -441,4 +441,8 @@ void amdgpu_vm_move_to_lru_tail(struct amdgpu_device *adev,
 				struct amdgpu_vm *vm);
 void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo);
 
+#if defined(CONFIG_DEBUG_FS)
+void amdgpu_debugfs_vm_bo_info(struct amdgpu_vm *vm, struct seq_file *m);
+#endif
+
 #endif
-- 
cgit v1.2.3


From 586052b0a6062e2fa98189d7f24d8cb9ccf4258b Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Mon, 2 Nov 2020 13:01:53 +0100
Subject: drm/ttm: rework no_retry handling v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

During eviction we do want to trigger the OOM killer.

Only while doing new allocations we should try to avoid that and
return -ENOMEM to the application.

v2: rename the flag to gfp_retry_mayfail.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/398685/
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 2 ++
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    | 3 ---
 drivers/gpu/drm/ttm/ttm_pool.c             | 2 +-
 drivers/gpu/drm/ttm/ttm_tt.c               | 7 -------
 include/drm/ttm/ttm_bo_api.h               | 2 ++
 include/drm/ttm/ttm_bo_driver.h            | 3 ---
 6 files changed, 5 insertions(+), 14 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_object.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 1aa516429c80..4e9dfbea31c6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -516,6 +516,8 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
 	struct ttm_operation_ctx ctx = {
 		.interruptible = (bp->type != ttm_bo_type_kernel),
 		.no_wait_gpu = bp->no_wait_gpu,
+		/* We opt to avoid OOM on system pages allocations */
+		.gfp_retry_mayfail = true,
 		.resv = bp->resv,
 		.flags = bp->type != ttm_bo_type_kernel ?
 			TTM_OPT_FLAG_ALLOW_RES_EVICT : 0
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index bd6e6641c3fc..c01c060e4ac5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1914,9 +1914,6 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	}
 	adev->mman.initialized = true;
 
-	/* We opt to avoid OOM on system pages allocations */
-	adev->mman.bdev.no_retry = true;
-
 	/* Initialize VRAM pool with all of VRAM divided into pages */
 	r = amdgpu_vram_mgr_init(adev);
 	if (r) {
diff --git a/drivers/gpu/drm/ttm/ttm_pool.c b/drivers/gpu/drm/ttm/ttm_pool.c
index 1e50deefb5d5..44ec41aa78d6 100644
--- a/drivers/gpu/drm/ttm/ttm_pool.c
+++ b/drivers/gpu/drm/ttm/ttm_pool.c
@@ -367,7 +367,7 @@ int ttm_pool_alloc(struct ttm_pool *pool, struct ttm_tt *tt,
 	if (tt->page_flags & TTM_PAGE_FLAG_ZERO_ALLOC)
 		gfp_flags |= __GFP_ZERO;
 
-	if (tt->page_flags & TTM_PAGE_FLAG_NO_RETRY)
+	if (ctx->gfp_retry_mayfail)
 		gfp_flags |= __GFP_RETRY_MAYFAIL;
 
 	if (pool->use_dma32)
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index 8861a74ac335..cfd633c7e764 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -51,9 +51,6 @@ int ttm_tt_create(struct ttm_buffer_object *bo, bool zero_alloc)
 	if (bo->ttm)
 		return 0;
 
-	if (bdev->no_retry)
-		page_flags |= TTM_PAGE_FLAG_NO_RETRY;
-
 	switch (bo->type) {
 	case ttm_bo_type_device:
 		if (zero_alloc)
@@ -211,8 +208,6 @@ int ttm_tt_swapin(struct ttm_tt *ttm)
 
 	swap_space = swap_storage->f_mapping;
 	gfp_mask = mapping_gfp_mask(swap_space);
-	if (ttm->page_flags & TTM_PAGE_FLAG_NO_RETRY)
-		gfp_mask |= __GFP_RETRY_MAYFAIL;
 
 	for (i = 0; i < ttm->num_pages; ++i) {
 		from_page = shmem_read_mapping_page_gfp(swap_space, i,
@@ -260,8 +255,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 
 	swap_space = swap_storage->f_mapping;
 	gfp_mask = mapping_gfp_mask(swap_space);
-	if (ttm->page_flags & TTM_PAGE_FLAG_NO_RETRY)
-		gfp_mask |= __GFP_RETRY_MAYFAIL;
 
 	for (i = 0; i < ttm->num_pages; ++i) {
 		from_page = ttm->pages[i];
diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h
index 37102e45e496..4637357ba856 100644
--- a/include/drm/ttm/ttm_bo_api.h
+++ b/include/drm/ttm/ttm_bo_api.h
@@ -195,6 +195,7 @@ struct ttm_bo_kmap_obj {
  *
  * @interruptible: Sleep interruptible if sleeping.
  * @no_wait_gpu: Return immediately if the GPU is busy.
+ * @gfp_retry_mayfail: Set the __GFP_RETRY_MAYFAIL when allocation pages.
  * @resv: Reservation object to allow reserved evictions with.
  * @flags: Including the following flags
  *
@@ -204,6 +205,7 @@ struct ttm_bo_kmap_obj {
 struct ttm_operation_ctx {
 	bool interruptible;
 	bool no_wait_gpu;
+	bool gfp_retry_mayfail;
 	struct dma_resv *resv;
 	uint64_t bytes_moved;
 	uint32_t flags;
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index e9f683fa72dc..da8208f43378 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -276,7 +276,6 @@ extern struct ttm_bo_global {
  * @dev_mapping: A pointer to the struct address_space representing the
  * device address space.
  * @wq: Work queue structure for the delayed delete workqueue.
- * @no_retry: Don't retry allocation if it fails
  *
  */
 
@@ -314,8 +313,6 @@ struct ttm_bo_device {
 	 */
 
 	struct delayed_work wq;
-
-	bool no_retry;
 };
 
 static inline struct ttm_resource_manager *ttm_manager_type(struct ttm_bo_device *bdev,
-- 
cgit v1.2.3


From c44dfe4de053914cae61ed0a36421b3017f428bd Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Mon, 2 Nov 2020 13:16:13 +0100
Subject: drm/ttm: replace context flags with bools v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The ttm_operation_ctx structure has a mixture of flags and bools. Drop the
flags and replace them with bools as well.

v2: fix typos, improve comments

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/398686/
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c     |  3 +--
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c |  5 ++---
 drivers/gpu/drm/ttm/ttm_bo.c               |  2 +-
 drivers/gpu/drm/ttm/ttm_bo_vm.c            |  3 +--
 drivers/gpu/drm/ttm/ttm_memory.c           |  3 ++-
 drivers/gpu/drm/ttm/ttm_resource.c         |  2 +-
 include/drm/ttm/ttm_bo_api.h               | 13 ++++++-------
 7 files changed, 14 insertions(+), 17 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_object.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index d50b63a93d37..8466558d0d93 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -404,8 +404,7 @@ static int amdgpu_cs_bo_validate(struct amdgpu_cs_parser *p,
 	struct ttm_operation_ctx ctx = {
 		.interruptible = true,
 		.no_wait_gpu = false,
-		.resv = bo->tbo.base.resv,
-		.flags = 0
+		.resv = bo->tbo.base.resv
 	};
 	uint32_t domain;
 	int r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 4e9dfbea31c6..e1f64ef8c765 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -518,9 +518,8 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
 		.no_wait_gpu = bp->no_wait_gpu,
 		/* We opt to avoid OOM on system pages allocations */
 		.gfp_retry_mayfail = true,
-		.resv = bp->resv,
-		.flags = bp->type != ttm_bo_type_kernel ?
-			TTM_OPT_FLAG_ALLOW_RES_EVICT : 0
+		.allow_res_evict = bp->type != ttm_bo_type_kernel,
+		.resv = bp->resv
 	};
 	struct amdgpu_bo *bo;
 	unsigned long page_align, size = bp->size;
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index c63b7ea1cd5d..e2a124b3affb 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -637,7 +637,7 @@ static bool ttm_bo_evict_swapout_allowable(struct ttm_buffer_object *bo,
 
 	if (bo->base.resv == ctx->resv) {
 		dma_resv_assert_held(bo->base.resv);
-		if (ctx->flags & TTM_OPT_FLAG_ALLOW_RES_EVICT)
+		if (ctx->allow_res_evict)
 			ret = true;
 		*locked = false;
 		if (busy)
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index eeaca5d1efe3..2944fa0af493 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -315,8 +315,7 @@ vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 		struct ttm_operation_ctx ctx = {
 			.interruptible = false,
 			.no_wait_gpu = false,
-			.flags = TTM_OPT_FLAG_FORCE_ALLOC
-
+			.force_alloc = true
 		};
 
 		ttm = bo->ttm;
diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c
index f9a90bfaa3c1..5ed1fc8f2ace 100644
--- a/drivers/gpu/drm/ttm/ttm_memory.c
+++ b/drivers/gpu/drm/ttm/ttm_memory.c
@@ -542,7 +542,8 @@ ttm_check_under_lowerlimit(struct ttm_mem_global *glob,
 {
 	int64_t available;
 
-	if (ctx->flags & TTM_OPT_FLAG_FORCE_ALLOC)
+	/* We allow over commit during suspend */
+	if (ctx->force_alloc)
 		return false;
 
 	available = get_nr_swap_pages() + si_mem_available();
diff --git a/drivers/gpu/drm/ttm/ttm_resource.c b/drivers/gpu/drm/ttm/ttm_resource.c
index 4ebc043e2867..b60699bf4816 100644
--- a/drivers/gpu/drm/ttm/ttm_resource.c
+++ b/drivers/gpu/drm/ttm/ttm_resource.c
@@ -89,7 +89,7 @@ int ttm_resource_manager_evict_all(struct ttm_bo_device *bdev,
 	struct ttm_operation_ctx ctx = {
 		.interruptible = false,
 		.no_wait_gpu = false,
-		.flags = TTM_OPT_FLAG_FORCE_ALLOC
+		.force_alloc = true
 	};
 	struct ttm_bo_global *glob = &ttm_bo_glob;
 	struct dma_fence *fence;
diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h
index 4637357ba856..5ddad88ae6ed 100644
--- a/include/drm/ttm/ttm_bo_api.h
+++ b/include/drm/ttm/ttm_bo_api.h
@@ -196,8 +196,11 @@ struct ttm_bo_kmap_obj {
  * @interruptible: Sleep interruptible if sleeping.
  * @no_wait_gpu: Return immediately if the GPU is busy.
  * @gfp_retry_mayfail: Set the __GFP_RETRY_MAYFAIL when allocation pages.
+ * @allow_res_evict: Allow eviction of reserved BOs. Can be used when multiple
+ * BOs share the same reservation object.
+ * @force_alloc: Don't check the memory account during suspend or CPU page
+ * faults. Should only be used by TTM internally.
  * @resv: Reservation object to allow reserved evictions with.
- * @flags: Including the following flags
  *
  * Context for TTM operations like changing buffer placement or general memory
  * allocation.
@@ -206,16 +209,12 @@ struct ttm_operation_ctx {
 	bool interruptible;
 	bool no_wait_gpu;
 	bool gfp_retry_mayfail;
+	bool allow_res_evict;
+	bool force_alloc;
 	struct dma_resv *resv;
 	uint64_t bytes_moved;
-	uint32_t flags;
 };
 
-/* Allow eviction of reserved BOs */
-#define TTM_OPT_FLAG_ALLOW_RES_EVICT		0x1
-/* when serving page fault or suspend, allow alloc anyway */
-#define TTM_OPT_FLAG_FORCE_ALLOC		0x2
-
 /**
  * ttm_bo_get - reference a struct ttm_buffer_object
  *
-- 
cgit v1.2.3


From 25dd7a4469ae42369c2d90b1c73cde1e27bf4b54 Mon Sep 17 00:00:00 2001
From: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Date: Wed, 2 Dec 2020 09:27:15 +0100
Subject: drm: amdgpu: fix a kernel-doc markup
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The function name at kernel-doc markup doesn't match the name
of the function:

	drivers/gpu/drm/amd/amdgpu/amdgpu_object.c:1534: warning: expecting prototype for amdgpu_debugfs_print_bo_info(). Prototype was for amdgpu_bo_print_info() instead

Fix it.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_object.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index c6c9723d3d8a..fd7a93c32235 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -1518,7 +1518,7 @@ uint32_t amdgpu_bo_get_preferred_pin_domain(struct amdgpu_device *adev,
 	} while (0)
 
 /**
- * amdgpu_debugfs_print_bo_info - print BO info in debugfs file
+ * amdgpu_bo_print_info - print BO info in debugfs file
  *
  * @id: Index or Id of the BO
  * @bo: Requested BO for printing info
-- 
cgit v1.2.3


From 64f2c15892ae9dd8f0073f576601b1c9582d823b Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 4 Dec 2020 11:56:46 -0500
Subject: drm/amdgpu: remove amdgpu_ttm_late_init  and amdgpu_bo_late_init
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

No longer used.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 17 -----------------
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.h |  1 -
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    |  7 -------
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h    |  1 -
 drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c     |  2 --
 drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c      |  2 --
 drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c      |  2 --
 drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c      |  2 --
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c      |  2 --
 9 files changed, 36 deletions(-)

(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_object.c')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index fd7a93c32235..25ec4d57333f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -1074,23 +1074,6 @@ int amdgpu_bo_init(struct amdgpu_device *adev)
 	return amdgpu_ttm_init(adev);
 }
 
-/**
- * amdgpu_bo_late_init - late init
- * @adev: amdgpu device object
- *
- * Calls amdgpu_ttm_late_init() to free resources used earlier during
- * initialization.
- *
- * Returns:
- * 0 for success or a negative error code on failure.
- */
-int amdgpu_bo_late_init(struct amdgpu_device *adev)
-{
-	amdgpu_ttm_late_init(adev);
-
-	return 0;
-}
-
 /**
  * amdgpu_bo_fini - tear down memory manager
  * @adev: amdgpu device object
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
index a5a7992b5773..2bfef286fda4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
@@ -269,7 +269,6 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 void amdgpu_bo_unpin(struct amdgpu_bo *bo);
 int amdgpu_bo_evict_vram(struct amdgpu_device *adev);
 int amdgpu_bo_init(struct amdgpu_device *adev);
-int amdgpu_bo_late_init(struct amdgpu_device *adev);
 void amdgpu_bo_fini(struct amdgpu_device *adev);
 int amdgpu_bo_fbdev_mmap(struct amdgpu_bo *bo,
 				struct vm_area_struct *vma);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 6be131df21f5..b719b4281cc9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -2021,13 +2021,6 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	return 0;
 }
 
-/*
- * amdgpu_ttm_late_init - Handle any late initialization for amdgpu_ttm
- */
-void amdgpu_ttm_late_init(struct amdgpu_device *adev)
-{
-}
-
 /*
  * amdgpu_ttm_fini - De-initialize the TTM memory pools
  */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
index bdca2970173e..d2987536d7cd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
@@ -131,7 +131,6 @@ int amdgpu_vram_mgr_query_page_status(struct ttm_resource_manager *man,
 				      uint64_t start);
 
 int amdgpu_ttm_init(struct amdgpu_device *adev);
-void amdgpu_ttm_late_init(struct amdgpu_device *adev);
 void amdgpu_ttm_fini(struct amdgpu_device *adev);
 void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev,
 					bool enable);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index 11fedcf86fae..5648c48be77f 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -717,8 +717,6 @@ static int gmc_v10_0_late_init(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int r;
 
-	amdgpu_bo_late_init(adev);
-
 	r = amdgpu_gmc_allocate_vm_inv_eng(adev);
 	if (r)
 		return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
index 998d6c8fde79..f5b69484c45a 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
@@ -791,8 +791,6 @@ static int gmc_v6_0_late_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	amdgpu_bo_late_init(adev);
-
 	if (amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS)
 		return amdgpu_irq_get(adev, &adev->gmc.vm_fault, 0);
 	else
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
index d06e3f5e38ae..dee2b34effb6 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
@@ -961,8 +961,6 @@ static int gmc_v7_0_late_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	amdgpu_bo_late_init(adev);
-
 	if (amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS)
 		return amdgpu_irq_get(adev, &adev->gmc.vm_fault, 0);
 	else
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
index e8d07109c628..2d832fc23119 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
@@ -1062,8 +1062,6 @@ static int gmc_v8_0_late_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	amdgpu_bo_late_init(adev);
-
 	if (amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS)
 		return amdgpu_irq_get(adev, &adev->gmc.vm_fault, 0);
 	else
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 07126b2b46c2..e1531d97f486 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -1177,8 +1177,6 @@ static int gmc_v9_0_late_init(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int r;
 
-	amdgpu_bo_late_init(adev);
-
 	r = amdgpu_gmc_allocate_vm_inv_eng(adev);
 	if (r)
 		return r;
-- 
cgit v1.2.3