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_ttm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index e7b67dc330a4..db5f761f37ec 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -662,7 +662,7 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 
 	/* Can't move a pinned BO */
 	abo = ttm_to_amdgpu_bo(bo);
-	if (WARN_ON_ONCE(abo->pin_count > 0))
+	if (WARN_ON_ONCE(abo->tbo.pin_count > 0))
 		return -EINVAL;
 
 	adev = amdgpu_ttm_adev(bo->bdev);
-- 
cgit 


From 3a08446b31e3614f6e018c6ac297b1312689a73a Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Thu, 24 Sep 2020 15:18:03 +1000
Subject: drm/amdgpu/ttm: handle tt moves properly.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The core move code currently handles use_tt moves, for amdgpu
this was being handled also in the driver, but not using the same
paths.

If moving between TT/SYSTEM (all the use_tt paths on amdgpu) use
the core move function.

Eventually the core will be flipped over to calling the driver.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200924051845.397177-4-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index db5f761f37ec..d3bd2fd448be 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -671,14 +671,16 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 		ttm_bo_move_null(bo, new_mem);
 		return 0;
 	}
-	if ((old_mem->mem_type == TTM_PL_TT &&
-	     new_mem->mem_type == TTM_PL_SYSTEM) ||
-	    (old_mem->mem_type == TTM_PL_SYSTEM &&
-	     new_mem->mem_type == TTM_PL_TT)) {
-		/* bind is enough */
+	if (old_mem->mem_type == TTM_PL_SYSTEM &&
+	    new_mem->mem_type == TTM_PL_TT) {
 		ttm_bo_move_null(bo, new_mem);
 		return 0;
 	}
+
+	if (old_mem->mem_type == TTM_PL_TT &&
+	    new_mem->mem_type == TTM_PL_SYSTEM)
+		return ttm_bo_move_ttm(bo, ctx, new_mem);
+
 	if (old_mem->mem_type == AMDGPU_PL_GDS ||
 	    old_mem->mem_type == AMDGPU_PL_GWS ||
 	    old_mem->mem_type == AMDGPU_PL_OA ||
-- 
cgit 


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_ttm.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 


From 1b4ea4c5980ff3a64607166298269c30a9671d33 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Wed, 30 Sep 2020 10:38:48 +0200
Subject: drm/ttm: set the tt caching state at creation time
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

All drivers can determine the tt caching state at creation time,
no need to do this on the fly during every validation.

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/394253/
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c    |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    | 11 ++++++++--
 drivers/gpu/drm/drm_gem_vram_helper.c      |  2 +-
 drivers/gpu/drm/nouveau/nouveau_sgdma.c    | 13 +++++++++++-
 drivers/gpu/drm/qxl/qxl_ttm.c              |  2 +-
 drivers/gpu/drm/radeon/radeon_ttm.c        | 16 ++++++++++++--
 drivers/gpu/drm/ttm/ttm_agp_backend.c      |  2 +-
 drivers/gpu/drm/ttm/ttm_page_alloc.c       | 26 +++++++++++------------
 drivers/gpu/drm/ttm/ttm_page_alloc_dma.c   | 20 +++++++++---------
 drivers/gpu/drm/ttm/ttm_tt.c               | 33 +++++++++++++++--------------
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c |  6 ++++--
 include/drm/ttm/ttm_caching.h              | 34 ++++++++++++++++++++++++++++++
 include/drm/ttm/ttm_tt.h                   | 16 ++++++--------
 13 files changed, 123 insertions(+), 60 deletions(-)
 create mode 100644 include/drm/ttm/ttm_caching.h

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index 213ef090bb0e..3c5ad69eff19 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -124,7 +124,7 @@ uint64_t amdgpu_gmc_agp_addr(struct ttm_buffer_object *bo)
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
 	struct ttm_dma_tt *ttm;
 
-	if (bo->num_pages != 1 || bo->ttm->caching_state == tt_cached)
+	if (bo->num_pages != 1 || bo->ttm->caching == ttm_cached)
 		return AMDGPU_BO_INVALID_OFFSET;
 
 	ttm = container_of(bo->ttm, struct ttm_dma_tt, ttm);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 399961035ae6..7f41a47e7353 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1292,7 +1292,9 @@ static void amdgpu_ttm_backend_destroy(struct ttm_bo_device *bdev,
 static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
 					   uint32_t page_flags)
 {
+	struct amdgpu_bo *abo = ttm_to_amdgpu_bo(bo);
 	struct amdgpu_ttm_tt *gtt;
+	enum ttm_caching caching;
 
 	gtt = kzalloc(sizeof(struct amdgpu_ttm_tt), GFP_KERNEL);
 	if (gtt == NULL) {
@@ -1300,8 +1302,13 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
 	}
 	gtt->gobj = &bo->base;
 
+	if (abo->flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
+		caching = ttm_write_combined;
+	else
+		caching = ttm_cached;
+
 	/* allocate space for the uninitialized page entries */
-	if (ttm_sg_tt_init(&gtt->ttm, bo, page_flags)) {
+	if (ttm_sg_tt_init(&gtt->ttm, bo, page_flags, caching)) {
 		kfree(gtt);
 		return NULL;
 	}
@@ -1525,7 +1532,7 @@ uint64_t amdgpu_ttm_tt_pde_flags(struct ttm_tt *ttm, struct ttm_resource *mem)
 	if (mem && mem->mem_type == TTM_PL_TT) {
 		flags |= AMDGPU_PTE_SYSTEM;
 
-		if (ttm->caching_state == tt_cached)
+		if (ttm->caching == ttm_cached)
 			flags |= AMDGPU_PTE_SNOOPED;
 	}
 
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index 3213429f8444..ad58d0af5141 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -918,7 +918,7 @@ static struct ttm_tt *bo_driver_ttm_tt_create(struct ttm_buffer_object *bo,
 	if (!tt)
 		return NULL;
 
-	ret = ttm_tt_init(tt, bo, page_flags);
+	ret = ttm_tt_init(tt, bo, page_flags, ttm_cached);
 	if (ret < 0)
 		goto err_ttm_tt_init;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_sgdma.c b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
index 806d9ec310f5..cd6fdebae795 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sgdma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
@@ -5,6 +5,7 @@
 #include "nouveau_drv.h"
 #include "nouveau_mem.h"
 #include "nouveau_ttm.h"
+#include "nouveau_bo.h"
 
 struct nouveau_sgdma_be {
 	/* this has to be the first field so populate/unpopulated in
@@ -67,13 +68,23 @@ nouveau_sgdma_unbind(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 struct ttm_tt *
 nouveau_sgdma_create_ttm(struct ttm_buffer_object *bo, uint32_t page_flags)
 {
+	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
+	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	struct nouveau_sgdma_be *nvbe;
+	enum ttm_caching caching;
+
+	if (nvbo->force_coherent)
+		caching = ttm_uncached;
+	else if (drm->agp.bridge)
+		caching = ttm_write_combined;
+	else
+		caching = ttm_cached;
 
 	nvbe = kzalloc(sizeof(*nvbe), GFP_KERNEL);
 	if (!nvbe)
 		return NULL;
 
-	if (ttm_dma_tt_init(&nvbe->ttm, bo, page_flags)) {
+	if (ttm_dma_tt_init(&nvbe->ttm, bo, page_flags, caching)) {
 		kfree(nvbe);
 		return NULL;
 	}
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index 669bceb58205..f50863493f64 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -133,7 +133,7 @@ static struct ttm_tt *qxl_ttm_tt_create(struct ttm_buffer_object *bo,
 	ttm = kzalloc(sizeof(struct ttm_tt), GFP_KERNEL);
 	if (ttm == NULL)
 		return NULL;
-	if (ttm_tt_init(ttm, bo, page_flags)) {
+	if (ttm_tt_init(ttm, bo, page_flags, ttm_cached)) {
 		kfree(ttm);
 		return NULL;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 63e38b05a5bc..130a7cea35c3 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -546,7 +546,7 @@ static int radeon_ttm_backend_bind(struct ttm_bo_device *bdev,
 		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
 		     ttm->num_pages, bo_mem, ttm);
 	}
-	if (ttm->caching_state == tt_cached)
+	if (ttm->caching == ttm_cached)
 		flags |= RADEON_GART_PAGE_SNOOP;
 	r = radeon_gart_bind(rdev, gtt->offset, ttm->num_pages,
 			     ttm->pages, gtt->ttm.dma_address, flags);
@@ -590,6 +590,10 @@ static struct ttm_tt *radeon_ttm_tt_create(struct ttm_buffer_object *bo,
 {
 	struct radeon_device *rdev;
 	struct radeon_ttm_tt *gtt;
+	enum ttm_caching caching;
+	struct radeon_bo *rbo;
+
+	rbo = container_of(bo, struct radeon_bo, tbo);
 
 	rdev = radeon_get_rdev(bo->bdev);
 #if IS_ENABLED(CONFIG_AGP)
@@ -603,7 +607,15 @@ static struct ttm_tt *radeon_ttm_tt_create(struct ttm_buffer_object *bo,
 	if (gtt == NULL) {
 		return NULL;
 	}
-	if (ttm_dma_tt_init(&gtt->ttm, bo, page_flags)) {
+
+	if (rbo->flags & RADEON_GEM_GTT_UC)
+		caching = ttm_uncached;
+	else if (rbo->flags & RADEON_GEM_GTT_WC)
+		caching = ttm_write_combined;
+	else
+		caching = ttm_cached;
+
+	if (ttm_dma_tt_init(&gtt->ttm, bo, page_flags, caching)) {
 		kfree(gtt);
 		return NULL;
 	}
diff --git a/drivers/gpu/drm/ttm/ttm_agp_backend.c b/drivers/gpu/drm/ttm/ttm_agp_backend.c
index a98fd795b752..a723062d37e7 100644
--- a/drivers/gpu/drm/ttm/ttm_agp_backend.c
+++ b/drivers/gpu/drm/ttm/ttm_agp_backend.c
@@ -136,7 +136,7 @@ struct ttm_tt *ttm_agp_tt_create(struct ttm_buffer_object *bo,
 	agp_be->mem = NULL;
 	agp_be->bridge = bridge;
 
-	if (ttm_tt_init(&agp_be->ttm, bo, page_flags)) {
+	if (ttm_tt_init(&agp_be->ttm, bo, page_flags, ttm_write_combined)) {
 		kfree(agp_be);
 		return NULL;
 	}
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index 111031cbb6df..c8f6790962b9 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -220,14 +220,14 @@ static struct ttm_pool_manager *_manager;
 /**
  * Select the right pool or requested caching state and ttm flags. */
 static struct ttm_page_pool *ttm_get_pool(int flags, bool huge,
-					  enum ttm_caching_state cstate)
+					  enum ttm_caching cstate)
 {
 	int pool_index;
 
-	if (cstate == tt_cached)
+	if (cstate == ttm_cached)
 		return NULL;
 
-	if (cstate == tt_wc)
+	if (cstate == ttm_write_combined)
 		pool_index = 0x0;
 	else
 		pool_index = 0x1;
@@ -441,17 +441,17 @@ static void ttm_pool_mm_shrink_fini(struct ttm_pool_manager *manager)
 }
 
 static int ttm_set_pages_caching(struct page **pages,
-		enum ttm_caching_state cstate, unsigned cpages)
+		enum ttm_caching cstate, unsigned cpages)
 {
 	int r = 0;
 	/* Set page caching */
 	switch (cstate) {
-	case tt_uncached:
+	case ttm_uncached:
 		r = ttm_set_pages_array_uc(pages, cpages);
 		if (r)
 			pr_err("Failed to set %d pages to uc!\n", cpages);
 		break;
-	case tt_wc:
+	case ttm_write_combined:
 		r = ttm_set_pages_array_wc(pages, cpages);
 		if (r)
 			pr_err("Failed to set %d pages to wc!\n", cpages);
@@ -486,7 +486,7 @@ static void ttm_handle_caching_failure(struct page **failed_pages,
  * pages returned in pages array.
  */
 static int ttm_alloc_new_pages(struct list_head *pages, gfp_t gfp_flags,
-			       int ttm_flags, enum ttm_caching_state cstate,
+			       int ttm_flags, enum ttm_caching cstate,
 			       unsigned count, unsigned order)
 {
 	struct page **caching_array;
@@ -566,7 +566,7 @@ out:
  * pages is small.
  */
 static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool, int ttm_flags,
-				      enum ttm_caching_state cstate,
+				      enum ttm_caching cstate,
 				      unsigned count, unsigned long *irq_flags)
 {
 	struct page *p;
@@ -626,7 +626,7 @@ static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool, int ttm_flags,
 static int ttm_page_pool_get_pages(struct ttm_page_pool *pool,
 				   struct list_head *pages,
 				   int ttm_flags,
-				   enum ttm_caching_state cstate,
+				   enum ttm_caching cstate,
 				   unsigned count, unsigned order)
 {
 	unsigned long irq_flags;
@@ -703,7 +703,7 @@ out:
 
 /* Put all pages in pages list to correct pool to wait for reuse */
 static void ttm_put_pages(struct page **pages, unsigned npages, int flags,
-			  enum ttm_caching_state cstate)
+			  enum ttm_caching cstate)
 {
 	struct ttm_page_pool *pool = ttm_get_pool(flags, false, cstate);
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
@@ -821,7 +821,7 @@ static void ttm_put_pages(struct page **pages, unsigned npages, int flags,
  * cached pages.
  */
 static int ttm_get_pages(struct page **pages, unsigned npages, int flags,
-			 enum ttm_caching_state cstate)
+			 enum ttm_caching cstate)
 {
 	struct ttm_page_pool *pool = ttm_get_pool(flags, false, cstate);
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
@@ -1040,7 +1040,7 @@ ttm_pool_unpopulate_helper(struct ttm_tt *ttm, unsigned mem_count_update)
 
 put_pages:
 	ttm_put_pages(ttm->pages, ttm->num_pages, ttm->page_flags,
-		      ttm->caching_state);
+		      ttm->caching);
 	ttm_tt_set_unpopulated(ttm);
 }
 
@@ -1057,7 +1057,7 @@ int ttm_pool_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 		return -ENOMEM;
 
 	ret = ttm_get_pages(ttm->pages, ttm->num_pages, ttm->page_flags,
-			    ttm->caching_state);
+			    ttm->caching);
 	if (unlikely(ret != 0)) {
 		ttm_pool_unpopulate_helper(ttm, 0);
 		return ret;
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
index 1045a5c26ee3..6625b43f6256 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
@@ -325,15 +325,15 @@ static struct dma_page *__ttm_dma_alloc_page(struct dma_pool *pool)
 	}
 	return d_page;
 }
-static enum pool_type ttm_to_type(int flags, enum ttm_caching_state cstate)
+static enum pool_type ttm_to_type(int flags, enum ttm_caching cstate)
 {
 	enum pool_type type = IS_UNDEFINED;
 
 	if (flags & TTM_PAGE_FLAG_DMA32)
 		type |= IS_DMA32;
-	if (cstate == tt_cached)
+	if (cstate == ttm_cached)
 		type |= IS_CACHED;
-	else if (cstate == tt_uncached)
+	else if (cstate == ttm_uncached)
 		type |= IS_UC;
 	else
 		type |= IS_WC;
@@ -663,7 +663,7 @@ static struct dma_pool *ttm_dma_find_pool(struct device *dev,
  * are pages that have changed their caching state already put them to the
  * pool.
  */
-static void ttm_dma_handle_caching_state_failure(struct dma_pool *pool,
+static void ttm_dma_handle_caching_failure(struct dma_pool *pool,
 						 struct list_head *d_pages,
 						 struct page **failed_pages,
 						 unsigned cpages)
@@ -734,7 +734,7 @@ static int ttm_dma_pool_alloc_new_pages(struct dma_pool *pool,
 				r = ttm_set_pages_caching(pool, caching_array,
 							  cpages);
 				if (r)
-					ttm_dma_handle_caching_state_failure(
+					ttm_dma_handle_caching_failure(
 						pool, d_pages, caching_array,
 						cpages);
 			}
@@ -760,7 +760,7 @@ static int ttm_dma_pool_alloc_new_pages(struct dma_pool *pool,
 				r = ttm_set_pages_caching(pool, caching_array,
 							  cpages);
 				if (r) {
-					ttm_dma_handle_caching_state_failure(
+					ttm_dma_handle_caching_failure(
 					     pool, d_pages, caching_array,
 					     cpages);
 					goto out;
@@ -773,7 +773,7 @@ static int ttm_dma_pool_alloc_new_pages(struct dma_pool *pool,
 	if (cpages) {
 		r = ttm_set_pages_caching(pool, caching_array, cpages);
 		if (r)
-			ttm_dma_handle_caching_state_failure(pool, d_pages,
+			ttm_dma_handle_caching_failure(pool, d_pages,
 					caching_array, cpages);
 	}
 out:
@@ -904,7 +904,7 @@ int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev,
 	INIT_LIST_HEAD(&ttm_dma->pages_list);
 	i = 0;
 
-	type = ttm_to_type(ttm->page_flags, ttm->caching_state);
+	type = ttm_to_type(ttm->page_flags, ttm->caching);
 
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 	if (ttm->page_flags & TTM_PAGE_FLAG_DMA32)
@@ -1000,7 +1000,7 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 	unsigned count, i, npages = 0;
 	unsigned long irq_flags;
 
-	type = ttm_to_type(ttm->page_flags, ttm->caching_state);
+	type = ttm_to_type(ttm->page_flags, ttm->caching);
 
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 	pool = ttm_dma_find_pool(dev, type | IS_HUGE);
@@ -1032,7 +1032,7 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 		return;
 
 	is_cached = (ttm_dma_find_pool(pool->dev,
-		     ttm_to_type(ttm->page_flags, tt_cached)) == pool);
+		     ttm_to_type(ttm->page_flags, ttm_cached)) == pool);
 
 	/* make sure pages array match list and count number of pages */
 	count = 0;
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index 23e9604bc924..a465f51df027 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -114,31 +114,30 @@ 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_state c_state)
+static int ttm_tt_set_caching(struct ttm_tt *ttm, enum ttm_caching caching)
 {
-	if (ttm->caching_state == c_state)
+	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_state = c_state;
+	ttm->caching = caching;
 
 	return 0;
 }
 
 int ttm_tt_set_placement_caching(struct ttm_tt *ttm, uint32_t placement)
 {
-	enum ttm_caching_state state;
+	enum ttm_caching state;
 
 	if (placement & TTM_PL_FLAG_WC)
-		state = tt_wc;
+		state = ttm_write_combined;
 	else if (placement & TTM_PL_FLAG_UNCACHED)
-		state = tt_uncached;
+		state = ttm_uncached;
 	else
-		state = tt_cached;
+		state = ttm_cached;
 
 	return ttm_tt_set_caching(ttm, state);
 }
@@ -162,20 +161,22 @@ void ttm_tt_destroy(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 
 static void ttm_tt_init_fields(struct ttm_tt *ttm,
 			       struct ttm_buffer_object *bo,
-			       uint32_t page_flags)
+			       uint32_t page_flags,
+			       enum ttm_caching caching)
 {
 	ttm->num_pages = bo->num_pages;
-	ttm->caching_state = tt_cached;
+	ttm->caching = ttm_cached;
 	ttm->page_flags = page_flags;
 	ttm_tt_set_unpopulated(ttm);
 	ttm->swap_storage = NULL;
 	ttm->sg = bo->sg;
+	ttm->caching = caching;
 }
 
 int ttm_tt_init(struct ttm_tt *ttm, struct ttm_buffer_object *bo,
-		uint32_t page_flags)
+		uint32_t page_flags, enum ttm_caching caching)
 {
-	ttm_tt_init_fields(ttm, bo, page_flags);
+	ttm_tt_init_fields(ttm, bo, page_flags, caching);
 
 	if (ttm_tt_alloc_page_directory(ttm)) {
 		pr_err("Failed allocating page table\n");
@@ -193,11 +194,11 @@ void ttm_tt_fini(struct ttm_tt *ttm)
 EXPORT_SYMBOL(ttm_tt_fini);
 
 int ttm_dma_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
-		    uint32_t page_flags)
+		    uint32_t page_flags, enum ttm_caching caching)
 {
 	struct ttm_tt *ttm = &ttm_dma->ttm;
 
-	ttm_tt_init_fields(ttm, bo, page_flags);
+	ttm_tt_init_fields(ttm, bo, page_flags, caching);
 
 	INIT_LIST_HEAD(&ttm_dma->pages_list);
 	if (ttm_dma_tt_alloc_page_directory(ttm_dma)) {
@@ -209,12 +210,12 @@ int ttm_dma_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
 EXPORT_SYMBOL(ttm_dma_tt_init);
 
 int ttm_sg_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
-		   uint32_t page_flags)
+		   uint32_t page_flags, enum ttm_caching caching)
 {
 	struct ttm_tt *ttm = &ttm_dma->ttm;
 	int ret;
 
-	ttm_tt_init_fields(ttm, bo, page_flags);
+	ttm_tt_init_fields(ttm, bo, page_flags, caching);
 
 	INIT_LIST_HEAD(&ttm_dma->pages_list);
 	if (page_flags & TTM_PAGE_FLAG_SG)
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index 7b5fd5288870..1fa7f9438ec4 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -647,9 +647,11 @@ static struct ttm_tt *vmw_ttm_tt_create(struct ttm_buffer_object *bo,
 	vmw_be->mob = NULL;
 
 	if (vmw_be->dev_priv->map_mode == vmw_dma_alloc_coherent)
-		ret = ttm_dma_tt_init(&vmw_be->dma_ttm, bo, page_flags);
+		ret = ttm_dma_tt_init(&vmw_be->dma_ttm, bo, page_flags,
+				      ttm_cached);
 	else
-		ret = ttm_tt_init(&vmw_be->dma_ttm.ttm, bo, page_flags);
+		ret = ttm_tt_init(&vmw_be->dma_ttm.ttm, bo, page_flags,
+				  ttm_cached);
 	if (unlikely(ret != 0))
 		goto out_no_init;
 
diff --git a/include/drm/ttm/ttm_caching.h b/include/drm/ttm/ttm_caching.h
new file mode 100644
index 000000000000..161624dcf6be
--- /dev/null
+++ b/include/drm/ttm/ttm_caching.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Christian König
+ */
+
+#ifndef _TTM_CACHING_H_
+#define _TTM_CACHING_H_
+
+enum ttm_caching {
+	ttm_uncached,
+	ttm_write_combined,
+	ttm_cached
+};
+
+#endif
diff --git a/include/drm/ttm/ttm_tt.h b/include/drm/ttm/ttm_tt.h
index 5d1835d44084..c39c722d5184 100644
--- a/include/drm/ttm/ttm_tt.h
+++ b/include/drm/ttm/ttm_tt.h
@@ -28,6 +28,7 @@
 #define _TTM_TT_H_
 
 #include <linux/types.h>
+#include <drm/ttm/ttm_caching.h>
 
 struct ttm_tt;
 struct ttm_resource;
@@ -42,12 +43,6 @@ struct ttm_operation_ctx;
 
 #define TTM_PAGE_FLAG_PRIV_POPULATED  (1 << 31)
 
-enum ttm_caching_state {
-	tt_uncached,
-	tt_wc,
-	tt_cached
-};
-
 /**
  * struct ttm_tt
  *
@@ -69,7 +64,7 @@ struct ttm_tt {
 	unsigned long num_pages;
 	struct sg_table *sg; /* for SG objects via dma-buf */
 	struct file *swap_storage;
-	enum ttm_caching_state caching_state;
+	enum ttm_caching caching;
 };
 
 static inline bool ttm_tt_is_populated(struct ttm_tt *tt)
@@ -121,6 +116,7 @@ int ttm_tt_create(struct ttm_buffer_object *bo, bool zero_alloc);
  * @ttm: The struct ttm_tt.
  * @bo: The buffer object we create the ttm for.
  * @page_flags: Page flags as identified by TTM_PAGE_FLAG_XX flags.
+ * @caching: the desired caching state of the pages
  *
  * Create a struct ttm_tt to back data with system memory pages.
  * No pages are actually allocated.
@@ -128,11 +124,11 @@ int ttm_tt_create(struct ttm_buffer_object *bo, bool zero_alloc);
  * NULL: Out of memory.
  */
 int ttm_tt_init(struct ttm_tt *ttm, struct ttm_buffer_object *bo,
-		uint32_t page_flags);
+		uint32_t page_flags, enum ttm_caching caching);
 int ttm_dma_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
-		    uint32_t page_flags);
+		    uint32_t page_flags, enum ttm_caching caching);
 int ttm_sg_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
-		   uint32_t page_flags);
+		   uint32_t page_flags, enum ttm_caching caching);
 
 /**
  * ttm_tt_fini
-- 
cgit 


From 1cf65c45183a6c8b4703675d40e709f7ffed935c Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Wed, 30 Sep 2020 11:17:44 +0200
Subject: drm/ttm: add caching state to ttm_bus_placement
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

And implement setting it up correctly in the drivers.

This allows getting rid of the placement flags for this.

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/394254/
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    |  1 +
 drivers/gpu/drm/drm_gem_vram_helper.c      |  1 +
 drivers/gpu/drm/nouveau/nouveau_bo.c       | 11 +++++++++++
 drivers/gpu/drm/qxl/qxl_ttm.c              |  2 ++
 drivers/gpu/drm/radeon/radeon_ttm.c        |  2 ++
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c |  1 +
 include/drm/ttm/ttm_resource.h             |  8 +++++---
 7 files changed, 23 insertions(+), 3 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 7f41a47e7353..5b56a66063fd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -769,6 +769,7 @@ static int amdgpu_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 
 		mem->bus.offset += adev->gmc.aper_base;
 		mem->bus.is_iomem = true;
+		mem->bus.caching = ttm_write_combined;
 		break;
 	default:
 		return -EINVAL;
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index ad58d0af5141..b9e7ce1adf25 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -981,6 +981,7 @@ static int bo_driver_io_mem_reserve(struct ttm_bo_device *bdev,
 	case TTM_PL_VRAM:
 		mem->bus.offset = (mem->start << PAGE_SHIFT) + vmm->vram_base;
 		mem->bus.is_iomem = true;
+		mem->bus.caching = ttm_write_combined;
 		break;
 	default:
 		return -EINVAL;
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 0c0ca44a6802..cb878c0e8276 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1134,6 +1134,8 @@ nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_resource *reg)
 	struct nouveau_drm *drm = nouveau_bdev(bdev);
 	struct nvkm_device *device = nvxx_device(&drm->client.device);
 	struct nouveau_mem *mem = nouveau_mem(reg);
+	struct nvif_mmu *mmu = &drm->client.mmu;
+	const u8 type = mmu->type[drm->ttm.type_vram].type;
 	int ret;
 
 	mutex_lock(&drm->ttm.io_reserve_mutex);
@@ -1149,6 +1151,7 @@ retry:
 			reg->bus.offset = (reg->start << PAGE_SHIFT) +
 				drm->agp.base;
 			reg->bus.is_iomem = !drm->agp.cma;
+			reg->bus.caching = ttm_write_combined;
 		}
 #endif
 		if (drm->client.mem->oclass < NVIF_CLASS_MEM_NV50 ||
@@ -1162,6 +1165,14 @@ retry:
 		reg->bus.offset = (reg->start << PAGE_SHIFT) +
 			device->func->resource_addr(device, 1);
 		reg->bus.is_iomem = true;
+
+		/* Some BARs do not support being ioremapped WC */
+		if (drm->client.device.info.family >= NV_DEVICE_INFO_V0_TESLA &&
+		    type & NVIF_MEM_UNCACHED)
+			reg->bus.caching = ttm_uncached;
+		else
+			reg->bus.caching = ttm_write_combined;
+
 		if (drm->client.mem->oclass >= NVIF_CLASS_MEM_NV50) {
 			union {
 				struct nv50_mem_map_v0 nv50;
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index f50863493f64..61eb06dbbce8 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -83,11 +83,13 @@ int qxl_ttm_io_mem_reserve(struct ttm_bo_device *bdev,
 	case TTM_PL_VRAM:
 		mem->bus.is_iomem = true;
 		mem->bus.offset = (mem->start << PAGE_SHIFT) + qdev->vram_base;
+		mem->bus.caching = ttm_cached;
 		break;
 	case TTM_PL_PRIV:
 		mem->bus.is_iomem = true;
 		mem->bus.offset = (mem->start << PAGE_SHIFT) +
 			qdev->surfaceram_base;
+		mem->bus.caching = ttm_cached;
 		break;
 	default:
 		return -EINVAL;
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 130a7cea35c3..9b53a1d80632 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -372,6 +372,7 @@ static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 			mem->bus.offset = (mem->start << PAGE_SHIFT) +
 				rdev->mc.agp_base;
 			mem->bus.is_iomem = !rdev->ddev->agp->cant_use_aperture;
+			mem->bus.caching = ttm_write_combined;
 		}
 #endif
 		break;
@@ -382,6 +383,7 @@ static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 			return -EINVAL;
 		mem->bus.offset += rdev->mc.aper_base;
 		mem->bus.is_iomem = true;
+		mem->bus.caching = ttm_write_combined;
 #ifdef __alpha__
 		/*
 		 * Alpha: use bus.addr to hold the ioremap() return,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index 1fa7f9438ec4..fae88969a15a 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -688,6 +688,7 @@ static int vmw_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_resourc
 		mem->bus.offset = (mem->start << PAGE_SHIFT) +
 			dev_priv->vram_start;
 		mem->bus.is_iomem = true;
+		mem->bus.caching = ttm_cached;
 		break;
 	default:
 		return -EINVAL;
diff --git a/include/drm/ttm/ttm_resource.h b/include/drm/ttm/ttm_resource.h
index 1b2f56163c6c..f48a70d39ac5 100644
--- a/include/drm/ttm/ttm_resource.h
+++ b/include/drm/ttm/ttm_resource.h
@@ -29,6 +29,7 @@
 #include <linux/mutex.h>
 #include <linux/dma-fence.h>
 #include <drm/drm_print.h>
+#include <drm/ttm/ttm_caching.h>
 
 #define TTM_MAX_BO_PRIORITY	4U
 
@@ -148,9 +149,10 @@ struct ttm_resource_manager {
  * Structure indicating the bus placement of an object.
  */
 struct ttm_bus_placement {
-	void		*addr;
-	phys_addr_t	offset;
-	bool		is_iomem;
+	void			*addr;
+	phys_addr_t		offset;
+	bool			is_iomem;
+	enum ttm_caching	caching;
 };
 
 /**
-- 
cgit 


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_ttm.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 


From a07e32bda0265b62ed0f85fbdfcd316c1dfc98ca Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Mon, 19 Oct 2020 17:13:12 +1000
Subject: drm/ttm: use new move interface for known system->ttm moves
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

In all 3 drivers there is a case where the driver knows the
bo is in SYSTEM so don't call the api that checks that.

Reviewed-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201019071314.1671485-4-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 4 ++--
 drivers/gpu/drm/nouveau/nouveau_bo.c    | 3 ++-
 drivers/gpu/drm/radeon/radeon_ttm.c     | 3 ++-
 drivers/gpu/drm/ttm/ttm_bo_util.c       | 1 +
 4 files changed, 7 insertions(+), 4 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 8cdec58b9106..f8c9d66d3ef7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -601,11 +601,11 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo, bool evict,
 	}
 
 	/* move/bind old memory to GTT space */
-	r = ttm_bo_move_ttm(bo, ctx, &tmp_mem);
+	r = ttm_bo_move_to_new_tt_mem(bo, ctx, &tmp_mem);
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
-
+	ttm_bo_assign_mem(bo, &tmp_mem);
 	/* copy to VRAM */
 	r = amdgpu_move_blit(bo, evict, new_mem, old_mem);
 	if (unlikely(r)) {
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 4ccb3329014b..63baa9367851 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -927,10 +927,11 @@ nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict,
 	if (ret)
 		return ret;
 
-	ret = ttm_bo_move_ttm(bo, ctx, &tmp_reg);
+	ret = ttm_bo_move_to_new_tt_mem(bo, ctx, &tmp_reg);
 	if (ret)
 		goto out;
 
+	ttm_bo_assign_mem(bo, &tmp_reg);
 	ret = nouveau_bo_move_m2mf(bo, true, ctx, new_reg);
 	if (ret)
 		goto out;
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index d6f42fbc81f4..9437d29248fd 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -275,10 +275,11 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
 	if (unlikely(r)) {
 		return r;
 	}
-	r = ttm_bo_move_ttm(bo, ctx, &tmp_mem);
+	r = ttm_bo_move_to_new_tt_mem(bo, ctx, &tmp_mem);
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
+	ttm_bo_assign_mem(bo, &tmp_mem);
 	r = radeon_move_blit(bo, true, new_mem, old_mem);
 	if (unlikely(r)) {
 		goto out_cleanup;
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 4a461226b6ba..7ea0482c1aee 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -64,6 +64,7 @@ int ttm_bo_move_to_new_tt_mem(struct ttm_buffer_object *bo,
 
 	return 0;
 }
+EXPORT_SYMBOL(ttm_bo_move_to_new_tt_mem);
 
 static int ttm_bo_move_to_system(struct ttm_buffer_object *bo,
 				 struct ttm_operation_ctx *ctx)
-- 
cgit 


From c37d951cb42aa340513c0bc2df10b7324fa0d856 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Mon, 19 Oct 2020 17:13:13 +1000
Subject: drm/ttm: add move old to system to drivers.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Uninline ttm_bo_move_ttm. Eventually want to unhook the unbind out.

Reviewed-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201019071314.1671485-5-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    | 15 ++++++++++++---
 drivers/gpu/drm/nouveau/nouveau_bo.c       | 11 +++++++++--
 drivers/gpu/drm/radeon/radeon_ttm.c        | 15 ++++++++++++---
 drivers/gpu/drm/ttm/ttm_bo_util.c          |  5 +++--
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c |  7 ++++++-
 include/drm/ttm/ttm_bo_driver.h            |  2 ++
 6 files changed, 44 insertions(+), 11 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index f8c9d66d3ef7..00bec9924b5b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -562,7 +562,11 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, bool evict,
 	}
 
 	/* move BO (in tmp_mem) to new_mem */
-	r = ttm_bo_move_ttm(bo, ctx, new_mem);
+	r = ttm_bo_move_to_system(bo, ctx);
+	if (unlikely(r))
+		goto out_cleanup;
+
+	ttm_bo_assign_mem(bo, new_mem);
 out_cleanup:
 	ttm_resource_free(bo, &tmp_mem);
 	return r;
@@ -672,8 +676,13 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 	}
 
 	if (old_mem->mem_type == TTM_PL_TT &&
-	    new_mem->mem_type == TTM_PL_SYSTEM)
-		return ttm_bo_move_ttm(bo, ctx, new_mem);
+	    new_mem->mem_type == TTM_PL_SYSTEM) {
+		r = ttm_bo_move_to_system(bo, ctx);
+		if (r)
+			return r;
+		ttm_bo_assign_mem(bo, new_mem);
+		return 0;
+	}
 
 	if (old_mem->mem_type == AMDGPU_PL_GDS ||
 	    old_mem->mem_type == AMDGPU_PL_GWS ||
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 63baa9367851..ec79c3b251e8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -897,7 +897,11 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict,
 	if (ret)
 		goto out;
 
-	ret = ttm_bo_move_ttm(bo, ctx, new_reg);
+	ret = ttm_bo_move_to_system(bo, ctx);
+	if (ret)
+		goto out;
+
+	ttm_bo_assign_mem(bo, &tmp_reg);
 out:
 	ttm_resource_free(bo, &tmp_reg);
 	return ret;
@@ -1048,7 +1052,10 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
 
 	if (old_reg->mem_type == TTM_PL_TT &&
 	    new_reg->mem_type == TTM_PL_SYSTEM) {
-		ret = ttm_bo_move_ttm(bo, ctx, new_reg);
+		ret = ttm_bo_move_to_system(bo, ctx);
+		if (ret)
+			return ret;
+		ttm_bo_assign_mem(bo, new_reg);
 		goto out;
 	}
 
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 9437d29248fd..426643120e3a 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -244,7 +244,11 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
-	r = ttm_bo_move_ttm(bo, ctx, new_mem);
+	r = ttm_bo_move_to_system(bo, ctx);
+	if (unlikely(r))
+		goto out_cleanup;
+
+	ttm_bo_assign_mem(bo, new_mem);
 out_cleanup:
 	ttm_resource_free(bo, &tmp_mem);
 	return r;
@@ -319,9 +323,14 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, bool evict,
 	}
 
 	if (old_mem->mem_type == TTM_PL_TT &&
-	    new_mem->mem_type == TTM_PL_SYSTEM)
-		return ttm_bo_move_ttm(bo, ctx, new_mem);
+	    new_mem->mem_type == TTM_PL_SYSTEM) {
+		r = ttm_bo_move_to_system(bo, ctx);
+		if (r)
+			return r;
 
+		ttm_bo_assign_mem(bo, new_mem);
+		return 0;
+	}
 	if (!rdev->ring[radeon_copy_ring_index(rdev)].ready ||
 	    rdev->asic->copy.copy == NULL) {
 		/* use memcpy */
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 7ea0482c1aee..4a74a297e529 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -66,8 +66,8 @@ int ttm_bo_move_to_new_tt_mem(struct ttm_buffer_object *bo,
 }
 EXPORT_SYMBOL(ttm_bo_move_to_new_tt_mem);
 
-static int ttm_bo_move_to_system(struct ttm_buffer_object *bo,
-				 struct ttm_operation_ctx *ctx)
+int ttm_bo_move_to_system(struct ttm_buffer_object *bo,
+			  struct ttm_operation_ctx *ctx)
 {
 	struct ttm_resource *old_mem = &bo->mem;
 	int ret;
@@ -87,6 +87,7 @@ static int ttm_bo_move_to_system(struct ttm_buffer_object *bo,
 	old_mem->mem_type = TTM_PL_SYSTEM;
 	return 0;
 }
+EXPORT_SYMBOL(ttm_bo_move_to_system);
 
 int ttm_bo_move_ttm(struct ttm_buffer_object *bo,
 		   struct ttm_operation_ctx *ctx,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index 112253246f08..c5cf81c09971 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -735,13 +735,18 @@ static int vmw_move(struct ttm_buffer_object *bo,
 {
 	struct ttm_resource_manager *old_man = ttm_manager_type(bo->bdev, bo->mem.mem_type);
 	struct ttm_resource_manager *new_man = ttm_manager_type(bo->bdev, new_mem->mem_type);
+	int ret;
 
 	if (old_man->use_tt && new_man->use_tt) {
 		if (bo->mem.mem_type == TTM_PL_SYSTEM) {
 			ttm_bo_assign_mem(bo, new_mem);
 			return 0;
 		}
-		return ttm_bo_move_ttm(bo, ctx, new_mem);
+		ret = ttm_bo_move_to_system(bo, ctx);
+		if (ret)
+			return ret;
+		ttm_bo_assign_mem(bo, new_mem);
+		return 0;
 	} else {
 		return ttm_bo_move_memcpy(bo, ctx, new_mem);
 	}
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index 1f4d2b1febd0..9e60e6814ba1 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -596,6 +596,8 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo,
 int ttm_bo_move_to_new_tt_mem(struct ttm_buffer_object *bo,
 			      struct ttm_operation_ctx *ctx,
 			      struct ttm_resource *new_mem);
+int ttm_bo_move_to_system(struct ttm_buffer_object *bo,
+			  struct ttm_operation_ctx *ctx);
 
 /**
  * ttm_bo_move_memcpy
-- 
cgit 


From 9764c35348b4c0da722a234fd355119abb371196 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Tue, 20 Oct 2020 11:03:13 +1000
Subject: drm/ttm: move some move binds into the drivers
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This just gives the driver control over some of the bind paths.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201020010319.1692445-2-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c |  7 ++++++-
 drivers/gpu/drm/nouveau/nouveau_bo.c    | 10 +++++++---
 drivers/gpu/drm/radeon/radeon_ttm.c     | 11 ++++++++---
 drivers/gpu/drm/ttm/ttm_bo_util.c       |  1 -
 4 files changed, 21 insertions(+), 8 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 00bec9924b5b..11f464e0951d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -605,10 +605,15 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo, bool evict,
 	}
 
 	/* move/bind old memory to GTT space */
-	r = ttm_bo_move_to_new_tt_mem(bo, ctx, &tmp_mem);
+	r = ttm_tt_populate(bo->bdev, bo->ttm, ctx);
+	if (unlikely(r))
+		return r;
+
+	r = amdgpu_ttm_backend_bind(bo->bdev, bo->ttm, &tmp_mem);
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
+
 	ttm_bo_assign_mem(bo, &tmp_mem);
 	/* copy to VRAM */
 	r = amdgpu_move_blit(bo, evict, new_mem, old_mem);
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index ec79c3b251e8..526bbf4c399a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -931,9 +931,13 @@ nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict,
 	if (ret)
 		return ret;
 
-	ret = ttm_bo_move_to_new_tt_mem(bo, ctx, &tmp_reg);
-	if (ret)
-		goto out;
+	ret = ttm_tt_populate(bo->bdev, bo->ttm, ctx);
+	if (unlikely(ret != 0))
+		return ret;
+
+	ret = nouveau_ttm_tt_bind(bo->bdev, bo->ttm, &tmp_reg);
+	if (unlikely(ret != 0))
+		return ret;
 
 	ttm_bo_assign_mem(bo, &tmp_reg);
 	ret = nouveau_bo_move_m2mf(bo, true, ctx, new_reg);
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 426643120e3a..c1cdee33be11 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -279,10 +279,15 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
 	if (unlikely(r)) {
 		return r;
 	}
-	r = ttm_bo_move_to_new_tt_mem(bo, ctx, &tmp_mem);
-	if (unlikely(r)) {
+
+	r = ttm_tt_populate(bo->bdev, bo->ttm, ctx);
+	if (unlikely(r))
 		goto out_cleanup;
-	}
+
+	r = radeon_ttm_tt_bind(bo->bdev, bo->ttm, &tmp_mem);
+	if (unlikely(r))
+		goto out_cleanup;
+
 	ttm_bo_assign_mem(bo, &tmp_mem);
 	r = radeon_move_blit(bo, true, new_mem, old_mem);
 	if (unlikely(r)) {
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 558e78ad82aa..520b53e6e1d6 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -64,7 +64,6 @@ int ttm_bo_move_to_new_tt_mem(struct ttm_buffer_object *bo,
 
 	return 0;
 }
-EXPORT_SYMBOL(ttm_bo_move_to_new_tt_mem);
 
 int ttm_bo_move_to_system(struct ttm_buffer_object *bo,
 			  struct ttm_operation_ctx *ctx)
-- 
cgit 


From 29a1d482e4044ab76d0c0f6341212f1a51f48236 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Tue, 20 Oct 2020 11:03:15 +1000
Subject: drm/ttm: add move to system into drivers
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This moves the to system move into the drivers, and moves all
the unbinds in the move path under driver control

Note: radeon/nouveau already wait so don't duplicate it.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201020010319.1692445-4-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    | 12 +++++++++---
 drivers/gpu/drm/nouveau/nouveau_bo.c       | 10 ++++++----
 drivers/gpu/drm/radeon/radeon_ttm.c        | 12 +++++++-----
 drivers/gpu/drm/ttm/ttm_bo_util.c          | 21 ---------------------
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c |  5 ++++-
 include/drm/ttm/ttm_bo_driver.h            |  2 --
 6 files changed, 26 insertions(+), 36 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 11f464e0951d..56f16bdba936 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -66,6 +66,8 @@
 static int amdgpu_ttm_backend_bind(struct ttm_bo_device *bdev,
 				   struct ttm_tt *ttm,
 				   struct ttm_resource *bo_mem);
+static void amdgpu_ttm_backend_unbind(struct ttm_bo_device *bdev,
+				      struct ttm_tt *ttm);
 
 static int amdgpu_ttm_init_on_chip(struct amdgpu_device *adev,
 				    unsigned int type,
@@ -561,11 +563,12 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, bool evict,
 		goto out_cleanup;
 	}
 
-	/* move BO (in tmp_mem) to new_mem */
-	r = ttm_bo_move_to_system(bo, ctx);
+	r = ttm_bo_wait_ctx(bo, ctx);
 	if (unlikely(r))
 		goto out_cleanup;
 
+	amdgpu_ttm_backend_unbind(bo->bdev, bo->ttm);
+	ttm_resource_free(bo, &bo->mem);
 	ttm_bo_assign_mem(bo, new_mem);
 out_cleanup:
 	ttm_resource_free(bo, &tmp_mem);
@@ -682,9 +685,12 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 
 	if (old_mem->mem_type == TTM_PL_TT &&
 	    new_mem->mem_type == TTM_PL_SYSTEM) {
-		r = ttm_bo_move_to_system(bo, ctx);
+		r = ttm_bo_wait_ctx(bo, ctx);
 		if (r)
 			return r;
+
+		amdgpu_ttm_backend_unbind(bo->bdev, bo->ttm);
+		ttm_resource_free(bo, &bo->mem);
 		ttm_bo_assign_mem(bo, new_mem);
 		return 0;
 	}
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 526bbf4c399a..1ccd9ea6da95 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -46,6 +46,7 @@
 
 static int nouveau_ttm_tt_bind(struct ttm_bo_device *bdev, struct ttm_tt *ttm,
 			       struct ttm_resource *reg);
+static void nouveau_ttm_tt_unbind(struct ttm_bo_device *bdev, struct ttm_tt *ttm);
 
 /*
  * NV10-NV40 tiling helpers
@@ -897,10 +898,12 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict,
 	if (ret)
 		goto out;
 
-	ret = ttm_bo_move_to_system(bo, ctx);
+	ret = ttm_bo_wait_ctx(bo, ctx);
 	if (ret)
 		goto out;
 
+	nouveau_ttm_tt_unbind(bo->bdev, bo->ttm);
+	ttm_resource_free(bo, &bo->mem);
 	ttm_bo_assign_mem(bo, &tmp_reg);
 out:
 	ttm_resource_free(bo, &tmp_reg);
@@ -1056,9 +1059,8 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
 
 	if (old_reg->mem_type == TTM_PL_TT &&
 	    new_reg->mem_type == TTM_PL_SYSTEM) {
-		ret = ttm_bo_move_to_system(bo, ctx);
-		if (ret)
-			return ret;
+		nouveau_ttm_tt_unbind(bo->bdev, bo->ttm);
+		ttm_resource_free(bo, &bo->mem);
 		ttm_bo_assign_mem(bo, new_reg);
 		goto out;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index c1cdee33be11..42c85afe7955 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -59,6 +59,8 @@ static void radeon_ttm_debugfs_fini(struct radeon_device *rdev);
 static int radeon_ttm_tt_bind(struct ttm_bo_device *bdev,
 			      struct ttm_tt *ttm,
 			      struct ttm_resource *bo_mem);
+static void radeon_ttm_tt_unbind(struct ttm_bo_device *bdev,
+				 struct ttm_tt *ttm);
 
 struct radeon_device *radeon_get_rdev(struct ttm_bo_device *bdev)
 {
@@ -244,10 +246,12 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
-	r = ttm_bo_move_to_system(bo, ctx);
+	r = ttm_bo_wait_ctx(bo, ctx);
 	if (unlikely(r))
 		goto out_cleanup;
 
+	radeon_ttm_tt_unbind(bo->bdev, bo->ttm);
+	ttm_resource_free(bo, &bo->mem);
 	ttm_bo_assign_mem(bo, new_mem);
 out_cleanup:
 	ttm_resource_free(bo, &tmp_mem);
@@ -329,10 +333,8 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, bool evict,
 
 	if (old_mem->mem_type == TTM_PL_TT &&
 	    new_mem->mem_type == TTM_PL_SYSTEM) {
-		r = ttm_bo_move_to_system(bo, ctx);
-		if (r)
-			return r;
-
+		radeon_ttm_tt_unbind(bo->bdev, bo->ttm);
+		ttm_resource_free(bo, &bo->mem);
 		ttm_bo_assign_mem(bo, new_mem);
 		return 0;
 	}
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 5cfcb92a6d30..65f042d6da7c 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -65,27 +65,6 @@ int ttm_bo_move_to_new_tt_mem(struct ttm_buffer_object *bo,
 	return 0;
 }
 
-int ttm_bo_move_to_system(struct ttm_buffer_object *bo,
-			  struct ttm_operation_ctx *ctx)
-{
-	int ret;
-
-	if (bo->mem.mem_type == TTM_PL_SYSTEM)
-		return 0;
-
-	ret = ttm_bo_wait_ctx(bo, ctx);
-	if (unlikely(ret != 0)) {
-		if (ret != -ERESTARTSYS)
-			pr_err("Failed to expire sync object before unbinding TTM\n");
-		return ret;
-	}
-
-	ttm_bo_tt_unbind(bo);
-	ttm_resource_free(bo, &bo->mem);
-	return 0;
-}
-EXPORT_SYMBOL(ttm_bo_move_to_system);
-
 int ttm_mem_io_reserve(struct ttm_bo_device *bdev,
 		       struct ttm_resource *mem)
 {
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index c5cf81c09971..1d220a9794e6 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -742,9 +742,12 @@ static int vmw_move(struct ttm_buffer_object *bo,
 			ttm_bo_assign_mem(bo, new_mem);
 			return 0;
 		}
-		ret = ttm_bo_move_to_system(bo, ctx);
+		ret = ttm_bo_wait_ctx(bo, ctx);
 		if (ret)
 			return ret;
+
+		vmw_ttm_unbind(bo->bdev, bo->ttm);
+		ttm_resource_free(bo, &bo->mem);
 		ttm_bo_assign_mem(bo, new_mem);
 		return 0;
 	} else {
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index a89728cb9a23..81a1618b9535 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -574,8 +574,6 @@ void ttm_mem_io_free(struct ttm_bo_device *bdev,
 int ttm_bo_move_to_new_tt_mem(struct ttm_buffer_object *bo,
 			      struct ttm_operation_ctx *ctx,
 			      struct ttm_resource *new_mem);
-int ttm_bo_move_to_system(struct ttm_buffer_object *bo,
-			  struct ttm_operation_ctx *ctx);
 
 /**
  * ttm_bo_move_memcpy
-- 
cgit 


From f227ccc9612f5d2f6315874c93acf3945fd51dfb Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Tue, 20 Oct 2020 11:03:16 +1000
Subject: drm/ttm: drop unbind callback.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The drivers now control this, so drop unbinding.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201020010319.1692445-5-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    |  1 -
 drivers/gpu/drm/nouveau/nouveau_bo.c       |  1 -
 drivers/gpu/drm/qxl/qxl_ttm.c              |  7 -------
 drivers/gpu/drm/radeon/radeon_ttm.c        |  1 -
 drivers/gpu/drm/ttm/ttm_bo.c               |  5 -----
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c |  1 -
 include/drm/ttm/ttm_bo_driver.h            | 18 ------------------
 7 files changed, 34 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 56f16bdba936..d7a2f912055d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1722,7 +1722,6 @@ static struct ttm_bo_driver amdgpu_bo_driver = {
 	.ttm_tt_populate = &amdgpu_ttm_tt_populate,
 	.ttm_tt_unpopulate = &amdgpu_ttm_tt_unpopulate,
 	.ttm_tt_bind = &amdgpu_ttm_backend_bind,
-	.ttm_tt_unbind = &amdgpu_ttm_backend_unbind,
 	.ttm_tt_destroy = &amdgpu_ttm_backend_destroy,
 	.eviction_valuable = amdgpu_ttm_bo_eviction_valuable,
 	.evict_flags = &amdgpu_evict_flags,
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 1ccd9ea6da95..43e6e089319e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1394,7 +1394,6 @@ struct ttm_bo_driver nouveau_bo_driver = {
 	.ttm_tt_populate = &nouveau_ttm_tt_populate,
 	.ttm_tt_unpopulate = &nouveau_ttm_tt_unpopulate,
 	.ttm_tt_bind = &nouveau_ttm_tt_bind,
-	.ttm_tt_unbind = &nouveau_ttm_tt_unbind,
 	.ttm_tt_destroy = &nouveau_ttm_tt_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = nouveau_bo_evict_flags,
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index e3ed20215f18..95c4f2c7ab79 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -113,12 +113,6 @@ static int qxl_ttm_backend_bind(struct ttm_bo_device *bdev,
 	return -1;
 }
 
-static void qxl_ttm_backend_unbind(struct ttm_bo_device *bdev,
-				   struct ttm_tt *ttm)
-{
-	/* Not implemented */
-}
-
 static void qxl_ttm_backend_destroy(struct ttm_bo_device *bdev,
 				    struct ttm_tt *ttm)
 {
@@ -180,7 +174,6 @@ static struct ttm_bo_driver qxl_bo_driver = {
 	.ttm_tt_create = &qxl_ttm_tt_create,
 	.ttm_tt_bind = &qxl_ttm_backend_bind,
 	.ttm_tt_destroy = &qxl_ttm_backend_destroy,
-	.ttm_tt_unbind = &qxl_ttm_backend_unbind,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = &qxl_evict_flags,
 	.move = &qxl_bo_move,
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 42c85afe7955..914b2ca86936 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -817,7 +817,6 @@ static struct ttm_bo_driver radeon_bo_driver = {
 	.ttm_tt_populate = &radeon_ttm_tt_populate,
 	.ttm_tt_unpopulate = &radeon_ttm_tt_unpopulate,
 	.ttm_tt_bind = &radeon_ttm_tt_bind,
-	.ttm_tt_unbind = &radeon_ttm_tt_unbind,
 	.ttm_tt_destroy = &radeon_ttm_tt_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = &radeon_evict_flags,
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index cbc74a320db2..593e55879019 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1496,8 +1496,3 @@ int ttm_bo_tt_bind(struct ttm_buffer_object *bo, struct ttm_resource *mem)
 {
 	return bo->bdev->driver->ttm_tt_bind(bo->bdev, bo->ttm, mem);
 }
-
-void ttm_bo_tt_unbind(struct ttm_buffer_object *bo)
-{
-	bo->bdev->driver->ttm_tt_unbind(bo->bdev, bo->ttm);
-}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index 1d220a9794e6..6e07ea982961 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -760,7 +760,6 @@ struct ttm_bo_driver vmw_bo_driver = {
 	.ttm_tt_populate = &vmw_ttm_populate,
 	.ttm_tt_unpopulate = &vmw_ttm_unpopulate,
 	.ttm_tt_bind = &vmw_ttm_bind,
-	.ttm_tt_unbind = &vmw_ttm_unbind,
 	.ttm_tt_destroy = &vmw_ttm_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = vmw_evict_flags,
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index 81a1618b9535..fbbcf10670c1 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -104,17 +104,6 @@ struct ttm_bo_driver {
 	 */
 	int (*ttm_tt_bind)(struct ttm_bo_device *bdev, struct ttm_tt *ttm, struct ttm_resource *bo_mem);
 
-	/**
-	 * ttm_tt_unbind
-	 *
-	 * @bdev: Pointer to a ttm device
-	 * @ttm: Pointer to a struct ttm_tt.
-	 *
-	 * Unbind previously bound backend pages. This function should be
-	 * able to handle differences between aperture and system page sizes.
-	 */
-	void (*ttm_tt_unbind)(struct ttm_bo_device *bdev, struct ttm_tt *ttm);
-
 	/**
 	 * ttm_tt_destroy
 	 *
@@ -647,13 +636,6 @@ pgprot_t ttm_io_prot(struct ttm_buffer_object *bo, struct ttm_resource *res,
  */
 int ttm_bo_tt_bind(struct ttm_buffer_object *bo, struct ttm_resource *mem);
 
-/**
- * ttm_bo_tt_bind
- *
- * Unbind the object tt from a memory resource.
- */
-void ttm_bo_tt_unbind(struct ttm_buffer_object *bo);
-
 /**
  * ttm_bo_tt_destroy.
  */
-- 
cgit 


From 6d820003295977f865257f1845bcdebc5dab4fb5 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Tue, 20 Oct 2020 11:03:18 +1000
Subject: drm/ttm: drop move notify around move.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The drivers now do this in the move callback.

move_notify is still needed in the destroy path.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201020010319.1692445-7-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    | 13 +++++++--
 drivers/gpu/drm/drm_gem_vram_helper.c      | 11 +++++++-
 drivers/gpu/drm/nouveau/nouveau_bo.c       | 12 ++++++--
 drivers/gpu/drm/qxl/qxl_ttm.c              | 45 ++++++++++++++++++------------
 drivers/gpu/drm/radeon/radeon_ttm.c        | 11 ++++++--
 drivers/gpu/drm/ttm/ttm_bo.c               | 12 +-------
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c | 14 ++++++++--
 7 files changed, 78 insertions(+), 40 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index d7a2f912055d..fba8ada99b0e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -666,6 +666,8 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 	struct ttm_resource *old_mem = &bo->mem;
 	int r;
 
+	amdgpu_bo_move_notify(bo, evict, new_mem);
+
 	/* Can't move a pinned BO */
 	abo = ttm_to_amdgpu_bo(bo);
 	if (WARN_ON_ONCE(abo->tbo.pin_count > 0))
@@ -687,7 +689,7 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 	    new_mem->mem_type == TTM_PL_SYSTEM) {
 		r = ttm_bo_wait_ctx(bo, ctx);
 		if (r)
-			return r;
+			goto fail;
 
 		amdgpu_ttm_backend_unbind(bo->bdev, bo->ttm);
 		ttm_resource_free(bo, &bo->mem);
@@ -728,12 +730,12 @@ memcpy:
 		if (!amdgpu_mem_visible(adev, old_mem) ||
 		    !amdgpu_mem_visible(adev, new_mem)) {
 			pr_err("Move buffer fallback to memcpy unavailable\n");
-			return r;
+			goto fail;
 		}
 
 		r = ttm_bo_move_memcpy(bo, ctx, new_mem);
 		if (r)
-			return r;
+			goto fail;
 	}
 
 	if (bo->type == ttm_bo_type_device &&
@@ -748,6 +750,11 @@ memcpy:
 	/* update statistics */
 	atomic64_add((u64)bo->num_pages << PAGE_SHIFT, &adev->num_bytes_moved);
 	return 0;
+fail:
+	swap(*new_mem, bo->mem);
+	amdgpu_bo_move_notify(bo, false, new_mem);
+	swap(*new_mem, bo->mem);
+	return r;
 }
 
 /**
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index 7aeb5daf2805..19087b22bdbb 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -590,7 +590,16 @@ static int drm_gem_vram_bo_driver_move(struct drm_gem_vram_object *gbo,
 				       struct ttm_operation_ctx *ctx,
 				       struct ttm_resource *new_mem)
 {
-	return ttm_bo_move_memcpy(&gbo->bo, ctx, new_mem);
+	int ret;
+
+	drm_gem_vram_bo_driver_move_notify(gbo, evict, new_mem);
+	ret = ttm_bo_move_memcpy(&gbo->bo, ctx, new_mem);
+	if (ret) {
+		swap(*new_mem, gbo->bo.mem);
+		drm_gem_vram_bo_driver_move_notify(gbo, false, new_mem);
+		swap(*new_mem, gbo->bo.mem);
+	}
+	return ret;
 }
 
 /*
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 43e6e089319e..ecd16847fbd8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1032,9 +1032,10 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
 	struct nouveau_drm_tile *new_tile = NULL;
 	int ret = 0;
 
+	nouveau_bo_move_ntfy(bo, evict, new_reg);
 	ret = ttm_bo_wait_ctx(bo, ctx);
 	if (ret)
-		return ret;
+		goto out_ntfy;
 
 	if (nvbo->bo.pin_count)
 		NV_WARN(drm, "Moving pinned object %p!\n", nvbo);
@@ -1042,7 +1043,7 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
 	if (drm->client.device.info.family < NV_DEVICE_INFO_V0_TESLA) {
 		ret = nouveau_bo_vm_bind(bo, new_reg, &new_tile);
 		if (ret)
-			return ret;
+			goto out_ntfy;
 	}
 
 	/* Fake bo copy. */
@@ -1090,7 +1091,12 @@ out:
 		else
 			nouveau_bo_vm_cleanup(bo, new_tile, &nvbo->tile);
 	}
-
+out_ntfy:
+	if (ret) {
+		swap(*new_reg, bo->mem);
+		nouveau_bo_move_ntfy(bo, false, new_reg);
+		swap(*new_reg, bo->mem);
+	}
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index 95c4f2c7ab79..a6149e3cc3d2 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -136,24 +136,6 @@ static struct ttm_tt *qxl_ttm_tt_create(struct ttm_buffer_object *bo,
 	return ttm;
 }
 
-static int qxl_bo_move(struct ttm_buffer_object *bo, bool evict,
-		       struct ttm_operation_ctx *ctx,
-		       struct ttm_resource *new_mem)
-{
-	struct ttm_resource *old_mem = &bo->mem;
-	int ret;
-
-	ret = ttm_bo_wait_ctx(bo, ctx);
-	if (ret)
-		return ret;
-
-	if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
-		ttm_bo_move_null(bo, new_mem);
-		return 0;
-	}
-	return ttm_bo_move_memcpy(bo, ctx, new_mem);
-}
-
 static void qxl_bo_move_notify(struct ttm_buffer_object *bo,
 			       bool evict,
 			       struct ttm_resource *new_mem)
@@ -170,6 +152,33 @@ static void qxl_bo_move_notify(struct ttm_buffer_object *bo,
 		qxl_surface_evict(qdev, qbo, new_mem ? true : false);
 }
 
+static int qxl_bo_move(struct ttm_buffer_object *bo, bool evict,
+		       struct ttm_operation_ctx *ctx,
+		       struct ttm_resource *new_mem)
+{
+	struct ttm_resource *old_mem = &bo->mem;
+	int ret;
+
+	qxl_bo_move_notify(bo, evict, new_mem);
+
+	ret = ttm_bo_wait_ctx(bo, ctx);
+	if (ret)
+		goto out;
+
+	if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
+		ttm_bo_move_null(bo, new_mem);
+		return 0;
+	}
+	ret = ttm_bo_move_memcpy(bo, ctx, new_mem);
+out:
+	if (ret) {
+		swap(*new_mem, bo->mem);
+		qxl_bo_move_notify(bo, false, new_mem);
+		swap(*new_mem, bo->mem);
+	}
+	return ret;
+}
+
 static struct ttm_bo_driver qxl_bo_driver = {
 	.ttm_tt_create = &qxl_ttm_tt_create,
 	.ttm_tt_bind = &qxl_ttm_backend_bind,
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 914b2ca86936..e9c95ef0859b 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -311,9 +311,11 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, bool evict,
 	struct ttm_resource *old_mem = &bo->mem;
 	int r;
 
+	radeon_bo_move_notify(bo, evict, new_mem);
+
 	r = ttm_bo_wait_ctx(bo, ctx);
 	if (r)
-		return r;
+		goto fail;
 
 	/* Can't move a pinned BO */
 	rbo = container_of(bo, struct radeon_bo, tbo);
@@ -359,13 +361,18 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, bool evict,
 memcpy:
 		r = ttm_bo_move_memcpy(bo, ctx, new_mem);
 		if (r) {
-			return r;
+			goto fail;
 		}
 	}
 
 	/* update statistics */
 	atomic64_add((u64)bo->num_pages << PAGE_SHIFT, &rdev->num_bytes_moved);
 	return 0;
+fail:
+	swap(*new_mem, bo->mem);
+	radeon_bo_move_notify(bo, false, new_mem);
+	swap(*new_mem, bo->mem);
+	return r;
 }
 
 static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_resource *mem)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index a41be3dba23d..cd9c2bb183f4 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -263,19 +263,9 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 		}
 	}
 
-	if (bdev->driver->move_notify)
-		bdev->driver->move_notify(bo, evict, mem);
-
 	ret = bdev->driver->move(bo, evict, ctx, mem);
-	if (ret) {
-		if (bdev->driver->move_notify) {
-			swap(*mem, bo->mem);
-			bdev->driver->move_notify(bo, false, mem);
-			swap(*mem, bo->mem);
-		}
-
+	if (ret)
 		goto out_err;
-	}
 
 	ctx->bytes_moved += bo->num_pages << PAGE_SHIFT;
 	return 0;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index 6e07ea982961..fd82c9ba2d77 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -737,6 +737,8 @@ static int vmw_move(struct ttm_buffer_object *bo,
 	struct ttm_resource_manager *new_man = ttm_manager_type(bo->bdev, new_mem->mem_type);
 	int ret;
 
+	vmw_move_notify(bo, evict, new_mem);
+
 	if (old_man->use_tt && new_man->use_tt) {
 		if (bo->mem.mem_type == TTM_PL_SYSTEM) {
 			ttm_bo_assign_mem(bo, new_mem);
@@ -744,15 +746,23 @@ static int vmw_move(struct ttm_buffer_object *bo,
 		}
 		ret = ttm_bo_wait_ctx(bo, ctx);
 		if (ret)
-			return ret;
+			goto fail;
 
 		vmw_ttm_unbind(bo->bdev, bo->ttm);
 		ttm_resource_free(bo, &bo->mem);
 		ttm_bo_assign_mem(bo, new_mem);
 		return 0;
 	} else {
-		return ttm_bo_move_memcpy(bo, ctx, new_mem);
+		ret = ttm_bo_move_memcpy(bo, ctx, new_mem);
+		if (ret)
+			goto fail;
 	}
+	return 0;
+fail:
+	swap(*new_mem, bo->mem);
+	vmw_move_notify(bo, false, new_mem);
+	swap(*new_mem, bo->mem);
+	return ret;
 }
 
 struct ttm_bo_driver vmw_bo_driver = {
-- 
cgit 


From bfe5e585b44fb810e33fe54d4d822c6b2b40488c Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Tue, 20 Oct 2020 11:03:19 +1000
Subject: drm/ttm: move last binding into the drivers.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This moves the call to tt binding into the driver move,
and drops the driver callback.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201020010319.1692445-8-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    |  7 ++++++-
 drivers/gpu/drm/nouveau/nouveau_bo.c       |  7 ++++++-
 drivers/gpu/drm/qxl/qxl_ttm.c              | 14 --------------
 drivers/gpu/drm/radeon/radeon_ttm.c        |  6 +++++-
 drivers/gpu/drm/ttm/ttm_bo.c               |  4 ----
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c |  7 ++++++-
 include/drm/ttm/ttm_bo_driver.h            | 14 --------------
 7 files changed, 23 insertions(+), 36 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index fba8ada99b0e..87e10a212b8a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -666,6 +666,12 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 	struct ttm_resource *old_mem = &bo->mem;
 	int r;
 
+	if (new_mem->mem_type == TTM_PL_TT) {
+		r = amdgpu_ttm_backend_bind(bo->bdev, bo->ttm, new_mem);
+		if (r)
+			return r;
+	}
+
 	amdgpu_bo_move_notify(bo, evict, new_mem);
 
 	/* Can't move a pinned BO */
@@ -1728,7 +1734,6 @@ static struct ttm_bo_driver amdgpu_bo_driver = {
 	.ttm_tt_create = &amdgpu_ttm_tt_create,
 	.ttm_tt_populate = &amdgpu_ttm_tt_populate,
 	.ttm_tt_unpopulate = &amdgpu_ttm_tt_unpopulate,
-	.ttm_tt_bind = &amdgpu_ttm_backend_bind,
 	.ttm_tt_destroy = &amdgpu_ttm_backend_destroy,
 	.eviction_valuable = amdgpu_ttm_bo_eviction_valuable,
 	.evict_flags = &amdgpu_evict_flags,
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index ecd16847fbd8..70b6f3b1ae85 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1032,6 +1032,12 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
 	struct nouveau_drm_tile *new_tile = NULL;
 	int ret = 0;
 
+	if (new_reg->mem_type == TTM_PL_TT) {
+		ret = nouveau_ttm_tt_bind(bo->bdev, bo->ttm, new_reg);
+		if (ret)
+			return ret;
+	}
+
 	nouveau_bo_move_ntfy(bo, evict, new_reg);
 	ret = ttm_bo_wait_ctx(bo, ctx);
 	if (ret)
@@ -1399,7 +1405,6 @@ struct ttm_bo_driver nouveau_bo_driver = {
 	.ttm_tt_create = &nouveau_ttm_tt_create,
 	.ttm_tt_populate = &nouveau_ttm_tt_populate,
 	.ttm_tt_unpopulate = &nouveau_ttm_tt_unpopulate,
-	.ttm_tt_bind = &nouveau_ttm_tt_bind,
 	.ttm_tt_destroy = &nouveau_ttm_tt_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = nouveau_bo_evict_flags,
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index a6149e3cc3d2..1cc3c14bc684 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -100,19 +100,6 @@ int qxl_ttm_io_mem_reserve(struct ttm_bo_device *bdev,
 /*
  * TTM backend functions.
  */
-
-static int qxl_ttm_backend_bind(struct ttm_bo_device *bdev,
-				struct ttm_tt *ttm,
-				struct ttm_resource *bo_mem)
-{
-	if (!ttm->num_pages) {
-		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
-		     ttm->num_pages, bo_mem, ttm);
-	}
-	/* Not implemented */
-	return -1;
-}
-
 static void qxl_ttm_backend_destroy(struct ttm_bo_device *bdev,
 				    struct ttm_tt *ttm)
 {
@@ -181,7 +168,6 @@ out:
 
 static struct ttm_bo_driver qxl_bo_driver = {
 	.ttm_tt_create = &qxl_ttm_tt_create,
-	.ttm_tt_bind = &qxl_ttm_backend_bind,
 	.ttm_tt_destroy = &qxl_ttm_backend_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = &qxl_evict_flags,
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index e9c95ef0859b..cd454e5c802f 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -311,6 +311,11 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, bool evict,
 	struct ttm_resource *old_mem = &bo->mem;
 	int r;
 
+	if (new_mem->mem_type == TTM_PL_TT) {
+		r = radeon_ttm_tt_bind(bo->bdev, bo->ttm, new_mem);
+		if (r)
+			return r;
+	}
 	radeon_bo_move_notify(bo, evict, new_mem);
 
 	r = ttm_bo_wait_ctx(bo, ctx);
@@ -823,7 +828,6 @@ static struct ttm_bo_driver radeon_bo_driver = {
 	.ttm_tt_create = &radeon_ttm_tt_create,
 	.ttm_tt_populate = &radeon_ttm_tt_populate,
 	.ttm_tt_unpopulate = &radeon_ttm_tt_unpopulate,
-	.ttm_tt_bind = &radeon_ttm_tt_bind,
 	.ttm_tt_destroy = &radeon_ttm_tt_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = &radeon_evict_flags,
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index cd9c2bb183f4..2b578012cdef 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -256,10 +256,6 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 			ret = ttm_tt_populate(bo->bdev, bo->ttm, ctx);
 			if (ret)
 				goto out_err;
-
-			ret = bdev->driver->ttm_tt_bind(bo->bdev, bo->ttm, mem);
-			if (ret)
-				goto out_err;
 		}
 	}
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index fd82c9ba2d77..de25cf016be2 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -737,6 +737,12 @@ static int vmw_move(struct ttm_buffer_object *bo,
 	struct ttm_resource_manager *new_man = ttm_manager_type(bo->bdev, new_mem->mem_type);
 	int ret;
 
+	if (new_man->use_tt && new_mem->mem_type != TTM_PL_SYSTEM) {
+		ret = vmw_ttm_bind(bo->bdev, bo->ttm, new_mem);
+		if (ret)
+			return ret;
+	}
+
 	vmw_move_notify(bo, evict, new_mem);
 
 	if (old_man->use_tt && new_man->use_tt) {
@@ -769,7 +775,6 @@ struct ttm_bo_driver vmw_bo_driver = {
 	.ttm_tt_create = &vmw_ttm_tt_create,
 	.ttm_tt_populate = &vmw_ttm_populate,
 	.ttm_tt_unpopulate = &vmw_ttm_unpopulate,
-	.ttm_tt_bind = &vmw_ttm_bind,
 	.ttm_tt_destroy = &vmw_ttm_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = vmw_evict_flags,
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index 0c4efc169f46..72f106b335e9 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -90,20 +90,6 @@ struct ttm_bo_driver {
 	 */
 	void (*ttm_tt_unpopulate)(struct ttm_bo_device *bdev, struct ttm_tt *ttm);
 
-	/**
-	 * ttm_tt_bind
-	 *
-	 * @bdev: Pointer to a ttm device
-	 * @ttm: Pointer to a struct ttm_tt.
-	 * @bo_mem: Pointer to a struct ttm_resource describing the
-	 * memory type and location for binding.
-	 *
-	 * Bind the backend pages into the aperture in the location
-	 * indicated by @bo_mem. This function should be able to handle
-	 * differences between aperture and system page sizes.
-	 */
-	int (*ttm_tt_bind)(struct ttm_bo_device *bdev, struct ttm_tt *ttm, struct ttm_resource *bo_mem);
-
 	/**
 	 * ttm_tt_destroy
 	 *
-- 
cgit 


From 6a6e5988a2657cd0c91f6f1a3e7d194599248b6d Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Wed, 21 Oct 2020 14:40:29 +1000
Subject: drm/ttm: replace last move_notify with delete_mem_notify
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The move notify callback is only used in one place, this should
be removed in the future, but for now just rename it to the use
case which is to notify the driver that the GPU memory is to be
deleted.

Drivers can be cleaned up after this separately.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201021044031.1752624-2-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    |  8 +++++++-
 drivers/gpu/drm/drm_gem_vram_helper.c      |  8 +++-----
 drivers/gpu/drm/nouveau/nouveau_bo.c       |  8 +++++++-
 drivers/gpu/drm/qxl/qxl_ttm.c              |  7 ++++++-
 drivers/gpu/drm/radeon/radeon_ttm.c        |  8 +++++++-
 drivers/gpu/drm/ttm/ttm_bo.c               |  4 ++--
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c |  8 +++++++-
 include/drm/ttm/ttm_bo_driver.h            | 10 ++--------
 8 files changed, 41 insertions(+), 20 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 87e10a212b8a..62f9194b1dd1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1730,6 +1730,12 @@ static int amdgpu_ttm_access_memory(struct ttm_buffer_object *bo,
 	return ret;
 }
 
+static void
+amdgpu_bo_delete_mem_notify(struct ttm_buffer_object *bo)
+{
+	amdgpu_bo_move_notify(bo, false, NULL);
+}
+
 static struct ttm_bo_driver amdgpu_bo_driver = {
 	.ttm_tt_create = &amdgpu_ttm_tt_create,
 	.ttm_tt_populate = &amdgpu_ttm_tt_populate,
@@ -1739,7 +1745,7 @@ static struct ttm_bo_driver amdgpu_bo_driver = {
 	.evict_flags = &amdgpu_evict_flags,
 	.move = &amdgpu_bo_move,
 	.verify_access = &amdgpu_verify_access,
-	.move_notify = &amdgpu_bo_move_notify,
+	.delete_mem_notify = &amdgpu_bo_delete_mem_notify,
 	.release_notify = &amdgpu_bo_release_notify,
 	.io_mem_reserve = &amdgpu_ttm_io_mem_reserve,
 	.io_mem_pfn = amdgpu_ttm_io_mem_pfn,
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index 19087b22bdbb..9da823eb0edd 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -949,9 +949,7 @@ static void bo_driver_evict_flags(struct ttm_buffer_object *bo,
 	drm_gem_vram_bo_driver_evict_flags(gbo, placement);
 }
 
-static void bo_driver_move_notify(struct ttm_buffer_object *bo,
-				  bool evict,
-				  struct ttm_resource *new_mem)
+static void bo_driver_delete_mem_notify(struct ttm_buffer_object *bo)
 {
 	struct drm_gem_vram_object *gbo;
 
@@ -961,7 +959,7 @@ static void bo_driver_move_notify(struct ttm_buffer_object *bo,
 
 	gbo = drm_gem_vram_of_bo(bo);
 
-	drm_gem_vram_bo_driver_move_notify(gbo, evict, new_mem);
+	drm_gem_vram_bo_driver_move_notify(gbo, false, NULL);
 }
 
 static int bo_driver_move(struct ttm_buffer_object *bo,
@@ -1002,7 +1000,7 @@ static struct ttm_bo_driver bo_driver = {
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = bo_driver_evict_flags,
 	.move = bo_driver_move,
-	.move_notify = bo_driver_move_notify,
+	.delete_mem_notify = bo_driver_delete_mem_notify,
 	.io_mem_reserve = bo_driver_io_mem_reserve,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 70b6f3b1ae85..acff82afe260 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1401,6 +1401,12 @@ nouveau_bo_fence(struct nouveau_bo *nvbo, struct nouveau_fence *fence, bool excl
 		dma_resv_add_shared_fence(resv, &fence->base);
 }
 
+static void
+nouveau_bo_delete_mem_notify(struct ttm_buffer_object *bo)
+{
+	nouveau_bo_move_ntfy(bo, false, NULL);
+}
+
 struct ttm_bo_driver nouveau_bo_driver = {
 	.ttm_tt_create = &nouveau_ttm_tt_create,
 	.ttm_tt_populate = &nouveau_ttm_tt_populate,
@@ -1408,7 +1414,7 @@ struct ttm_bo_driver nouveau_bo_driver = {
 	.ttm_tt_destroy = &nouveau_ttm_tt_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = nouveau_bo_evict_flags,
-	.move_notify = nouveau_bo_move_ntfy,
+	.delete_mem_notify = nouveau_bo_delete_mem_notify,
 	.move = nouveau_bo_move,
 	.verify_access = nouveau_bo_verify_access,
 	.io_mem_reserve = &nouveau_ttm_io_mem_reserve,
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index 1cc3c14bc684..b52a4563b47b 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -166,6 +166,11 @@ out:
 	return ret;
 }
 
+static void qxl_bo_delete_mem_notify(struct ttm_buffer_object *bo)
+{
+	qxl_bo_move_notify(bo, false, NULL);
+}
+
 static struct ttm_bo_driver qxl_bo_driver = {
 	.ttm_tt_create = &qxl_ttm_tt_create,
 	.ttm_tt_destroy = &qxl_ttm_backend_destroy,
@@ -173,7 +178,7 @@ static struct ttm_bo_driver qxl_bo_driver = {
 	.evict_flags = &qxl_evict_flags,
 	.move = &qxl_bo_move,
 	.io_mem_reserve = &qxl_ttm_io_mem_reserve,
-	.move_notify = &qxl_bo_move_notify,
+	.delete_mem_notify = &qxl_bo_delete_mem_notify,
 };
 
 static int qxl_ttm_init_mem_type(struct qxl_device *qdev,
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index cd454e5c802f..321c09d20c6c 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -824,6 +824,12 @@ bool radeon_ttm_tt_is_readonly(struct radeon_device *rdev,
 	return !!(gtt->userflags & RADEON_GEM_USERPTR_READONLY);
 }
 
+static void
+radeon_bo_delete_mem_notify(struct ttm_buffer_object *bo)
+{
+	radeon_bo_move_notify(bo, false, NULL);
+}
+
 static struct ttm_bo_driver radeon_bo_driver = {
 	.ttm_tt_create = &radeon_ttm_tt_create,
 	.ttm_tt_populate = &radeon_ttm_tt_populate,
@@ -833,7 +839,7 @@ static struct ttm_bo_driver radeon_bo_driver = {
 	.evict_flags = &radeon_evict_flags,
 	.move = &radeon_bo_move,
 	.verify_access = &radeon_verify_access,
-	.move_notify = &radeon_bo_move_notify,
+	.delete_mem_notify = &radeon_bo_delete_mem_notify,
 	.io_mem_reserve = &radeon_ttm_io_mem_reserve,
 };
 
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 03b40ce7d2dc..5b411252a857 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -284,8 +284,8 @@ out_err:
 
 static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo)
 {
-	if (bo->bdev->driver->move_notify)
-		bo->bdev->driver->move_notify(bo, false, NULL);
+	if (bo->bdev->driver->delete_mem_notify)
+		bo->bdev->driver->delete_mem_notify(bo);
 
 	ttm_bo_tt_destroy(bo);
 	ttm_resource_free(bo, &bo->mem);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index de25cf016be2..88be48ad0344 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -771,6 +771,12 @@ fail:
 	return ret;
 }
 
+static void
+vmw_delete_mem_notify(struct ttm_buffer_object *bo)
+{
+	vmw_move_notify(bo, false, NULL);
+}
+
 struct ttm_bo_driver vmw_bo_driver = {
 	.ttm_tt_create = &vmw_ttm_tt_create,
 	.ttm_tt_populate = &vmw_ttm_populate,
@@ -780,7 +786,7 @@ struct ttm_bo_driver vmw_bo_driver = {
 	.evict_flags = vmw_evict_flags,
 	.move = vmw_move,
 	.verify_access = vmw_verify_access,
-	.move_notify = vmw_move_notify,
+	.delete_mem_notify = vmw_delete_mem_notify,
 	.swap_notify = vmw_swap_notify,
 	.io_mem_reserve = &vmw_ttm_io_mem_reserve,
 };
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index 72f106b335e9..29f6a1d1c853 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -156,15 +156,9 @@ struct ttm_bo_driver {
 			     struct file *filp);
 
 	/**
-	 * Hook to notify driver about a driver move so it
-	 * can do tiling things and book-keeping.
-	 *
-	 * @evict: whether this move is evicting the buffer from the graphics
-	 * address space
+	 * Hook to notify driver about a resource delete.
 	 */
-	void (*move_notify)(struct ttm_buffer_object *bo,
-			    bool evict,
-			    struct ttm_resource *new_mem);
+	void (*delete_mem_notify)(struct ttm_buffer_object *bo);
 
 	/**
 	 * notify the driver that we're about to swap out this bo
-- 
cgit 


From d836917da7e5ca9b33ef4d499972f1feeb519e00 Mon Sep 17 00:00:00 2001
From: xinhui pan <xinhui.pan@amd.com>
Date: Fri, 23 Oct 2020 13:41:12 +0800
Subject: drm/amdgpu: Fix size calculation when init onchip memory
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Size is page count here.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 4a85f8cedd77..11dd3d9eac15 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -65,11 +65,11 @@
 
 static int amdgpu_ttm_init_on_chip(struct amdgpu_device *adev,
 				    unsigned int type,
-				    uint64_t size)
+				    uint64_t size_in_page)
 {
 	return ttm_range_man_init(&adev->mman.bdev, type,
 				  TTM_PL_FLAG_UNCACHED, TTM_PL_FLAG_UNCACHED,
-				  false, size >> PAGE_SHIFT);
+				  false, size_in_page);
 }
 
 /**
-- 
cgit 


From d1cb1f254a5b1c07788eecb84b443d59ccdfb9e0 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Mon, 19 Oct 2020 18:49:27 +0200
Subject: drm/ttm: nuke ttm_tt_set_(un)populated again
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Neither page allocation backend nor the driver should mess with that.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Madhav Chauhan <madhav.chauhan@amd.com>
Link: https://patchwork.freedesktop.org/patch/396948/
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c  |  2 --
 drivers/gpu/drm/nouveau/nouveau_bo.c     |  1 -
 drivers/gpu/drm/radeon/radeon_ttm.c      |  2 --
 drivers/gpu/drm/ttm/ttm_page_alloc.c     |  2 --
 drivers/gpu/drm/ttm/ttm_page_alloc_dma.c |  2 --
 drivers/gpu/drm/ttm/ttm_tt.c             | 11 +++++++----
 include/drm/ttm/ttm_tt.h                 | 10 ----------
 7 files changed, 7 insertions(+), 23 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 62f9194b1dd1..169468572930 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1363,7 +1363,6 @@ static int amdgpu_ttm_tt_populate(struct ttm_bo_device *bdev,
 			return -ENOMEM;
 
 		ttm->page_flags |= TTM_PAGE_FLAG_SG;
-		ttm_tt_set_populated(ttm);
 		return 0;
 	}
 
@@ -1383,7 +1382,6 @@ static int amdgpu_ttm_tt_populate(struct ttm_bo_device *bdev,
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 gtt->ttm.dma_address,
 						 ttm->num_pages);
-		ttm_tt_set_populated(ttm);
 		return 0;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index acff82afe260..06a1f4c4e96e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1321,7 +1321,6 @@ nouveau_ttm_tt_populate(struct ttm_bo_device *bdev,
 		/* make userspace faulting work */
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 ttm_dma->dma_address, ttm->num_pages);
-		ttm_tt_set_populated(ttm);
 		return 0;
 	}
 
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 321c09d20c6c..75fa2f55186b 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -670,14 +670,12 @@ static int radeon_ttm_tt_populate(struct ttm_bo_device *bdev,
 			return -ENOMEM;
 
 		ttm->page_flags |= TTM_PAGE_FLAG_SG;
-		ttm_tt_set_populated(ttm);
 		return 0;
 	}
 
 	if (slave && ttm->sg) {
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 gtt->ttm.dma_address, ttm->num_pages);
-		ttm_tt_set_populated(ttm);
 		return 0;
 	}
 
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index c8f6790962b9..04099dddce16 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -1041,7 +1041,6 @@ ttm_pool_unpopulate_helper(struct ttm_tt *ttm, unsigned mem_count_update)
 put_pages:
 	ttm_put_pages(ttm->pages, ttm->num_pages, ttm->page_flags,
 		      ttm->caching);
-	ttm_tt_set_unpopulated(ttm);
 }
 
 int ttm_pool_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
@@ -1080,7 +1079,6 @@ int ttm_pool_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 		}
 	}
 
-	ttm_tt_set_populated(ttm);
 	return 0;
 }
 EXPORT_SYMBOL(ttm_pool_populate);
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
index 6625b43f6256..85dad69419f6 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
@@ -983,7 +983,6 @@ skip_huge:
 		}
 	}
 
-	ttm_tt_set_populated(ttm);
 	return 0;
 }
 EXPORT_SYMBOL_GPL(ttm_dma_populate);
@@ -1077,7 +1076,6 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 	/* shrink pool if necessary (only on !is_cached pools)*/
 	if (npages)
 		ttm_dma_page_pool_free(pool, npages, false);
-	ttm_tt_set_unpopulated(ttm);
 }
 EXPORT_SYMBOL_GPL(ttm_dma_unpopulate);
 
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index 3e5dd6271d4c..beb1e878fed9 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -138,7 +138,6 @@ static void ttm_tt_init_fields(struct ttm_tt *ttm,
 	ttm->num_pages = bo->num_pages;
 	ttm->caching = ttm_cached;
 	ttm->page_flags = page_flags;
-	ttm_tt_set_unpopulated(ttm);
 	ttm->swap_storage = NULL;
 	ttm->sg = bo->sg;
 	ttm->caching = caching;
@@ -334,9 +333,12 @@ int ttm_tt_populate(struct ttm_bo_device *bdev,
 		ret = bdev->driver->ttm_tt_populate(bdev, ttm, ctx);
 	else
 		ret = ttm_pool_populate(ttm, ctx);
-	if (!ret)
-		ttm_tt_add_mapping(bdev, ttm);
-	return ret;
+	if (ret)
+		return ret;
+
+	ttm_tt_add_mapping(bdev, ttm);
+	ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED;
+	return 0;
 }
 EXPORT_SYMBOL(ttm_tt_populate);
 
@@ -365,4 +367,5 @@ void ttm_tt_unpopulate(struct ttm_bo_device *bdev,
 		bdev->driver->ttm_tt_unpopulate(bdev, ttm);
 	else
 		ttm_pool_unpopulate(ttm);
+	ttm->page_flags &= ~TTM_PAGE_FLAG_PRIV_POPULATED;
 }
diff --git a/include/drm/ttm/ttm_tt.h b/include/drm/ttm/ttm_tt.h
index e042dec5e6c1..e3e60c1da754 100644
--- a/include/drm/ttm/ttm_tt.h
+++ b/include/drm/ttm/ttm_tt.h
@@ -72,16 +72,6 @@ static inline bool ttm_tt_is_populated(struct ttm_tt *tt)
 	return tt->page_flags & TTM_PAGE_FLAG_PRIV_POPULATED;
 }
 
-static inline void ttm_tt_set_unpopulated(struct ttm_tt *tt)
-{
-	tt->page_flags &= ~TTM_PAGE_FLAG_PRIV_POPULATED;
-}
-
-static inline void ttm_tt_set_populated(struct ttm_tt *tt)
-{
-	tt->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED;
-}
-
 /**
  * struct ttm_dma_tt
  *
-- 
cgit 


From 230c079fdcf45efacd316a76c3132b9f42cd3565 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Tue, 20 Oct 2020 20:10:39 +0200
Subject: drm/ttm: make num_pages uint32_t
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

We can still allocate 16TiB with that.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/396946/
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 10 +++++-----
 drivers/gpu/drm/nouveau/nouveau_ttm.c   |  2 +-
 drivers/gpu/drm/radeon/radeon_ttm.c     |  4 ++--
 include/drm/ttm/ttm_tt.h                |  2 +-
 4 files changed, 9 insertions(+), 9 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 169468572930..0a3270f25e40 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -973,7 +973,7 @@ bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
 	if (!gtt || !gtt->userptr)
 		return false;
 
-	DRM_DEBUG_DRIVER("user_pages_done 0x%llx pages 0x%lx\n",
+	DRM_DEBUG_DRIVER("user_pages_done 0x%llx pages 0x%x\n",
 		gtt->userptr, ttm->num_pages);
 
 	WARN_ONCE(!gtt->range || !gtt->range->hmm_pfns,
@@ -1124,7 +1124,7 @@ static int amdgpu_ttm_gart_bind(struct amdgpu_device *adev,
 
 gart_bind_fail:
 	if (r)
-		DRM_ERROR("failed to bind %lu pages at 0x%08llX\n",
+		DRM_ERROR("failed to bind %u pages at 0x%08llX\n",
 			  ttm->num_pages, gtt->offset);
 
 	return r;
@@ -1159,7 +1159,7 @@ static int amdgpu_ttm_backend_bind(struct ttm_bo_device *bdev,
 		}
 	}
 	if (!ttm->num_pages) {
-		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
+		WARN(1, "nothing to bind %u pages for mreg %p back %p!\n",
 		     ttm->num_pages, bo_mem, ttm);
 	}
 
@@ -1182,7 +1182,7 @@ static int amdgpu_ttm_backend_bind(struct ttm_bo_device *bdev,
 		ttm->pages, gtt->ttm.dma_address, flags);
 
 	if (r)
-		DRM_ERROR("failed to bind %lu pages at 0x%08llX\n",
+		DRM_ERROR("failed to bind %u pages at 0x%08llX\n",
 			  ttm->num_pages, gtt->offset);
 	gtt->bound = true;
 	return r;
@@ -1291,7 +1291,7 @@ static void amdgpu_ttm_backend_unbind(struct ttm_bo_device *bdev,
 	/* unbind shouldn't be done for GDS/GWS/OA in ttm_bo_clean_mm */
 	r = amdgpu_gart_unbind(adev, gtt->offset, ttm->num_pages);
 	if (r)
-		DRM_ERROR("failed to unbind %lu pages at 0x%08llX\n",
+		DRM_ERROR("failed to unbind %u pages at 0x%08llX\n",
 			  gtt->ttm.ttm.num_pages, gtt->offset);
 	gtt->bound = false;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index 04b95277c73a..0592ed6eaad1 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -108,7 +108,7 @@ nv04_gart_manager_new(struct ttm_resource_manager *man,
 		return ret;
 
 	ret = nvif_vmm_get(&mem->cli->vmm.vmm, PTES, false, 12, 0,
-			   reg->num_pages << PAGE_SHIFT, &mem->vma[0]);
+			   (long)reg->num_pages << PAGE_SHIFT, &mem->vma[0]);
 	if (ret) {
 		nouveau_mem_del(reg);
 		return ret;
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 75fa2f55186b..c51dcbc818ef 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -564,7 +564,7 @@ static int radeon_ttm_backend_bind(struct ttm_bo_device *bdev,
 
 	gtt->offset = (unsigned long)(bo_mem->start << PAGE_SHIFT);
 	if (!ttm->num_pages) {
-		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
+		WARN(1, "nothing to bind %u pages for mreg %p back %p!\n",
 		     ttm->num_pages, bo_mem, ttm);
 	}
 	if (ttm->caching == ttm_cached)
@@ -572,7 +572,7 @@ static int radeon_ttm_backend_bind(struct ttm_bo_device *bdev,
 	r = radeon_gart_bind(rdev, gtt->offset, ttm->num_pages,
 			     ttm->pages, gtt->ttm.dma_address, flags);
 	if (r) {
-		DRM_ERROR("failed to bind %lu pages at 0x%08X\n",
+		DRM_ERROR("failed to bind %u pages at 0x%08X\n",
 			  ttm->num_pages, (unsigned)gtt->offset);
 		return r;
 	}
diff --git a/include/drm/ttm/ttm_tt.h b/include/drm/ttm/ttm_tt.h
index e3e60c1da754..931a31355870 100644
--- a/include/drm/ttm/ttm_tt.h
+++ b/include/drm/ttm/ttm_tt.h
@@ -61,7 +61,7 @@ struct ttm_operation_ctx;
 struct ttm_tt {
 	struct page **pages;
 	uint32_t page_flags;
-	unsigned long num_pages;
+	uint32_t num_pages;
 	struct sg_table *sg; /* for SG objects via dma-buf */
 	struct file *swap_storage;
 	enum ttm_caching caching;
-- 
cgit 


From e34b8feeaa4b65725b25f49c9b08a0f8707e8e86 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Wed, 21 Oct 2020 14:06:49 +0200
Subject: drm/ttm: merge ttm_dma_tt back into ttm_tt
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

It makes no difference to kmalloc if the structure
is 48 or 64 bytes in size.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/396950/
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c    | 10 ++----
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    | 14 ++++----
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c     |  7 ++--
 drivers/gpu/drm/nouveau/nouveau_bo.c       | 12 +++----
 drivers/gpu/drm/nouveau/nouveau_mem.c      |  8 +++--
 drivers/gpu/drm/nouveau/nouveau_mem.h      |  4 +--
 drivers/gpu/drm/nouveau/nouveau_sgdma.c    |  6 ++--
 drivers/gpu/drm/qxl/qxl_ttm.c              |  2 +-
 drivers/gpu/drm/radeon/radeon_ttm.c        |  8 ++---
 drivers/gpu/drm/ttm/ttm_bo.c               |  2 +-
 drivers/gpu/drm/ttm/ttm_page_alloc.c       | 30 ++++++++--------
 drivers/gpu/drm/ttm/ttm_page_alloc_dma.c   | 44 +++++++++++-------------
 drivers/gpu/drm/ttm/ttm_tt.c               | 55 ++++++++++++------------------
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c | 26 +++++++-------
 include/drm/ttm/ttm_page_alloc.h           | 12 +++----
 include/drm/ttm/ttm_tt.h                   | 35 ++++++-------------
 16 files changed, 119 insertions(+), 156 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index 3c5ad69eff19..0e35023b5703 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -45,12 +45,10 @@ void amdgpu_gmc_get_pde_for_bo(struct amdgpu_bo *bo, int level,
 			       uint64_t *addr, uint64_t *flags)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	struct ttm_dma_tt *ttm;
 
 	switch (bo->tbo.mem.mem_type) {
 	case TTM_PL_TT:
-		ttm = container_of(bo->tbo.ttm, struct ttm_dma_tt, ttm);
-		*addr = ttm->dma_address[0];
+		*addr = bo->tbo.ttm->dma_address[0];
 		break;
 	case TTM_PL_VRAM:
 		*addr = amdgpu_bo_gpu_offset(bo);
@@ -122,16 +120,14 @@ int amdgpu_gmc_set_pte_pde(struct amdgpu_device *adev, void *cpu_pt_addr,
 uint64_t amdgpu_gmc_agp_addr(struct ttm_buffer_object *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
-	struct ttm_dma_tt *ttm;
 
 	if (bo->num_pages != 1 || bo->ttm->caching == ttm_cached)
 		return AMDGPU_BO_INVALID_OFFSET;
 
-	ttm = container_of(bo->ttm, struct ttm_dma_tt, ttm);
-	if (ttm->dma_address[0] + PAGE_SIZE >= adev->gmc.agp_size)
+	if (bo->ttm->dma_address[0] + PAGE_SIZE >= adev->gmc.agp_size)
 		return AMDGPU_BO_INVALID_OFFSET;
 
-	return adev->gmc.agp_start + ttm->dma_address[0];
+	return adev->gmc.agp_start + bo->ttm->dma_address[0];
 }
 
 /**
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 0a3270f25e40..beacd00221d8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -294,11 +294,9 @@ static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo,
 	cpu_addr = &job->ibs[0].ptr[num_dw];
 
 	if (mem->mem_type == TTM_PL_TT) {
-		struct ttm_dma_tt *dma;
 		dma_addr_t *dma_address;
 
-		dma = container_of(bo->ttm, struct ttm_dma_tt, ttm);
-		dma_address = &dma->dma_address[offset >> PAGE_SHIFT];
+		dma_address = &bo->ttm->dma_address[offset >> PAGE_SHIFT];
 		r = amdgpu_gart_map(adev, 0, num_pages, dma_address, flags,
 				    cpu_addr);
 		if (r)
@@ -841,7 +839,7 @@ uint64_t amdgpu_ttm_domain_start(struct amdgpu_device *adev, uint32_t type)
  * TTM backend functions.
  */
 struct amdgpu_ttm_tt {
-	struct ttm_dma_tt	ttm;
+	struct ttm_tt	ttm;
 	struct drm_gem_object	*gobj;
 	u64			offset;
 	uint64_t		userptr;
@@ -1292,7 +1290,7 @@ static void amdgpu_ttm_backend_unbind(struct ttm_bo_device *bdev,
 	r = amdgpu_gart_unbind(adev, gtt->offset, ttm->num_pages);
 	if (r)
 		DRM_ERROR("failed to unbind %u pages at 0x%08llX\n",
-			  gtt->ttm.ttm.num_pages, gtt->offset);
+			  gtt->ttm.num_pages, gtt->offset);
 	gtt->bound = false;
 }
 
@@ -1306,7 +1304,7 @@ static void amdgpu_ttm_backend_destroy(struct ttm_bo_device *bdev,
 	if (gtt->usertask)
 		put_task_struct(gtt->usertask);
 
-	ttm_dma_tt_fini(&gtt->ttm);
+	ttm_tt_fini(&gtt->ttm);
 	kfree(gtt);
 }
 
@@ -1340,7 +1338,7 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
 		kfree(gtt);
 		return NULL;
 	}
-	return &gtt->ttm.ttm;
+	return &gtt->ttm;
 }
 
 /**
@@ -1507,7 +1505,7 @@ bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start,
 	/* Return false if no part of the ttm_tt object lies within
 	 * the range
 	 */
-	size = (unsigned long)gtt->ttm.ttm.num_pages * PAGE_SIZE;
+	size = (unsigned long)gtt->ttm.num_pages * PAGE_SIZE;
 	if (gtt->userptr > end || gtt->userptr + size <= start)
 		return false;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 3e6243623082..f0e6fafd0938 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -1781,7 +1781,6 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va,
 		resv = vm->root.base.bo->tbo.base.resv;
 	} else {
 		struct drm_gem_object *obj = &bo->tbo.base;
-		struct ttm_dma_tt *ttm;
 
 		resv = bo->tbo.base.resv;
 		if (obj->import_attach && bo_va->is_xgmi) {
@@ -1794,10 +1793,8 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va,
 		}
 		mem = &bo->tbo.mem;
 		nodes = mem->mm_node;
-		if (mem->mem_type == TTM_PL_TT) {
-			ttm = container_of(bo->tbo.ttm, struct ttm_dma_tt, ttm);
-			pages_addr = ttm->dma_address;
-		}
+		if (mem->mem_type == TTM_PL_TT)
+			pages_addr = bo->tbo.ttm->dma_address;
 	}
 
 	if (bo) {
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 06a1f4c4e96e..75fddbcd7832 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -547,7 +547,7 @@ void
 nouveau_bo_sync_for_device(struct nouveau_bo *nvbo)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
-	struct ttm_dma_tt *ttm_dma = (struct ttm_dma_tt *)nvbo->bo.ttm;
+	struct ttm_tt *ttm_dma = (struct ttm_tt *)nvbo->bo.ttm;
 	int i;
 
 	if (!ttm_dma)
@@ -557,7 +557,7 @@ nouveau_bo_sync_for_device(struct nouveau_bo *nvbo)
 	if (nvbo->force_coherent)
 		return;
 
-	for (i = 0; i < ttm_dma->ttm.num_pages; i++)
+	for (i = 0; i < ttm_dma->num_pages; i++)
 		dma_sync_single_for_device(drm->dev->dev,
 					   ttm_dma->dma_address[i],
 					   PAGE_SIZE, DMA_TO_DEVICE);
@@ -567,7 +567,7 @@ void
 nouveau_bo_sync_for_cpu(struct nouveau_bo *nvbo)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
-	struct ttm_dma_tt *ttm_dma = (struct ttm_dma_tt *)nvbo->bo.ttm;
+	struct ttm_tt *ttm_dma = (struct ttm_tt *)nvbo->bo.ttm;
 	int i;
 
 	if (!ttm_dma)
@@ -577,7 +577,7 @@ nouveau_bo_sync_for_cpu(struct nouveau_bo *nvbo)
 	if (nvbo->force_coherent)
 		return;
 
-	for (i = 0; i < ttm_dma->ttm.num_pages; i++)
+	for (i = 0; i < ttm_dma->num_pages; i++)
 		dma_sync_single_for_cpu(drm->dev->dev, ttm_dma->dma_address[i],
 					PAGE_SIZE, DMA_FROM_DEVICE);
 }
@@ -1309,7 +1309,7 @@ static int
 nouveau_ttm_tt_populate(struct ttm_bo_device *bdev,
 			struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 {
-	struct ttm_dma_tt *ttm_dma = (void *)ttm;
+	struct ttm_tt *ttm_dma = (void *)ttm;
 	struct nouveau_drm *drm;
 	struct device *dev;
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
@@ -1345,7 +1345,7 @@ static void
 nouveau_ttm_tt_unpopulate(struct ttm_bo_device *bdev,
 			  struct ttm_tt *ttm)
 {
-	struct ttm_dma_tt *ttm_dma = (void *)ttm;
+	struct ttm_tt *ttm_dma = (void *)ttm;
 	struct nouveau_drm *drm;
 	struct device *dev;
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c
index 269d8707acc3..0a4c04aaad30 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mem.c
@@ -92,7 +92,7 @@ nouveau_mem_fini(struct nouveau_mem *mem)
 }
 
 int
-nouveau_mem_host(struct ttm_resource *reg, struct ttm_dma_tt *tt)
+nouveau_mem_host(struct ttm_resource *reg, struct ttm_tt *tt)
 {
 	struct nouveau_mem *mem = nouveau_mem(reg);
 	struct nouveau_cli *cli = mem->cli;
@@ -116,8 +116,10 @@ nouveau_mem_host(struct ttm_resource *reg, struct ttm_dma_tt *tt)
 		mem->comp = 0;
 	}
 
-	if (tt->ttm.sg) args.sgl = tt->ttm.sg->sgl;
-	else            args.dma = tt->dma_address;
+	if (tt->sg)
+		args.sgl = tt->sg->sgl;
+	else
+		args.dma = tt->dma_address;
 
 	mutex_lock(&drm->master.lock);
 	cli->base.super = true;
diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.h b/drivers/gpu/drm/nouveau/nouveau_mem.h
index 3fe1cfed57a1..7df3848e85aa 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mem.h
+++ b/drivers/gpu/drm/nouveau/nouveau_mem.h
@@ -1,7 +1,7 @@
 #ifndef __NOUVEAU_MEM_H__
 #define __NOUVEAU_MEM_H__
 #include <drm/ttm/ttm_bo_api.h>
-struct ttm_dma_tt;
+struct ttm_tt;
 
 #include <nvif/mem.h>
 #include <nvif/vmm.h>
@@ -24,7 +24,7 @@ int nouveau_mem_new(struct nouveau_cli *, u8 kind, u8 comp,
 		    struct ttm_resource *);
 void nouveau_mem_del(struct ttm_resource *);
 int nouveau_mem_vram(struct ttm_resource *, bool contig, u8 page);
-int nouveau_mem_host(struct ttm_resource *, struct ttm_dma_tt *);
+int nouveau_mem_host(struct ttm_resource *, struct ttm_tt *);
 void nouveau_mem_fini(struct nouveau_mem *);
 int nouveau_mem_map(struct nouveau_mem *, struct nvif_vmm *, struct nvif_vma *);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_sgdma.c b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
index cd6fdebae795..a2e23fd4906c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sgdma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
@@ -11,7 +11,7 @@ struct nouveau_sgdma_be {
 	/* this has to be the first field so populate/unpopulated in
 	 * nouve_bo.c works properly, otherwise have to move them here
 	 */
-	struct ttm_dma_tt ttm;
+	struct ttm_tt ttm;
 	struct nouveau_mem *mem;
 };
 
@@ -23,7 +23,7 @@ nouveau_sgdma_destroy(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 	if (ttm) {
 		nouveau_sgdma_unbind(bdev, ttm);
 		ttm_tt_destroy_common(bdev, ttm);
-		ttm_dma_tt_fini(&nvbe->ttm);
+		ttm_tt_fini(&nvbe->ttm);
 		kfree(nvbe);
 	}
 }
@@ -88,5 +88,5 @@ nouveau_sgdma_create_ttm(struct ttm_buffer_object *bo, uint32_t page_flags)
 		kfree(nvbe);
 		return NULL;
 	}
-	return &nvbe->ttm.ttm;
+	return &nvbe->ttm;
 }
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index b52a4563b47b..9609eeb52821 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -116,7 +116,7 @@ static struct ttm_tt *qxl_ttm_tt_create(struct ttm_buffer_object *bo,
 	ttm = kzalloc(sizeof(struct ttm_tt), GFP_KERNEL);
 	if (ttm == NULL)
 		return NULL;
-	if (ttm_tt_init(ttm, bo, page_flags, ttm_cached)) {
+	if (ttm_dma_tt_init(ttm, bo, page_flags, ttm_cached)) {
 		kfree(ttm);
 		return NULL;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index c51dcbc818ef..0a6d7ea847db 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -437,7 +437,7 @@ static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
  * TTM backend functions.
  */
 struct radeon_ttm_tt {
-	struct ttm_dma_tt		ttm;
+	struct ttm_tt		ttm;
 	u64				offset;
 
 	uint64_t			userptr;
@@ -602,7 +602,7 @@ static void radeon_ttm_backend_destroy(struct ttm_bo_device *bdev, struct ttm_tt
 	radeon_ttm_backend_unbind(bdev, ttm);
 	ttm_tt_destroy_common(bdev, ttm);
 
-	ttm_dma_tt_fini(&gtt->ttm);
+	ttm_tt_fini(&gtt->ttm);
 	kfree(gtt);
 }
 
@@ -640,7 +640,7 @@ static struct ttm_tt *radeon_ttm_tt_create(struct ttm_buffer_object *bo,
 		kfree(gtt);
 		return NULL;
 	}
-	return &gtt->ttm.ttm;
+	return &gtt->ttm;
 }
 
 static struct radeon_ttm_tt *radeon_ttm_tt_to_gtt(struct radeon_device *rdev,
@@ -653,7 +653,7 @@ static struct radeon_ttm_tt *radeon_ttm_tt_to_gtt(struct radeon_device *rdev,
 
 	if (!ttm)
 		return NULL;
-	return container_of(ttm, struct radeon_ttm_tt, ttm.ttm);
+	return container_of(ttm, struct radeon_ttm_tt, ttm);
 }
 
 static int radeon_ttm_tt_populate(struct ttm_bo_device *bdev,
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 5b411252a857..40c72a0f9325 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1192,7 +1192,7 @@ size_t ttm_bo_dma_acc_size(struct ttm_bo_device *bdev,
 
 	size += ttm_round_pot(struct_size);
 	size += ttm_round_pot(npages * (2*sizeof(void *) + sizeof(dma_addr_t)));
-	size += ttm_round_pot(sizeof(struct ttm_dma_tt));
+	size += ttm_round_pot(sizeof(struct ttm_tt));
 	return size;
 }
 EXPORT_SYMBOL(ttm_bo_dma_acc_size);
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index 661b75d19cad..29e6c29ad60e 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -1081,28 +1081,28 @@ void ttm_pool_unpopulate(struct ttm_tt *ttm)
 }
 EXPORT_SYMBOL(ttm_pool_unpopulate);
 
-int ttm_populate_and_map_pages(struct device *dev, struct ttm_dma_tt *tt,
+int ttm_populate_and_map_pages(struct device *dev, struct ttm_tt *tt,
 					struct ttm_operation_ctx *ctx)
 {
 	unsigned i, j;
 	int r;
 
-	r = ttm_pool_populate(&tt->ttm, ctx);
+	r = ttm_pool_populate(tt, ctx);
 	if (r)
 		return r;
 
-	for (i = 0; i < tt->ttm.num_pages; ++i) {
-		struct page *p = tt->ttm.pages[i];
+	for (i = 0; i < tt->num_pages; ++i) {
+		struct page *p = tt->pages[i];
 		size_t num_pages = 1;
 
-		for (j = i + 1; j < tt->ttm.num_pages; ++j) {
-			if (++p != tt->ttm.pages[j])
+		for (j = i + 1; j < tt->num_pages; ++j) {
+			if (++p != tt->pages[j])
 				break;
 
 			++num_pages;
 		}
 
-		tt->dma_address[i] = dma_map_page(dev, tt->ttm.pages[i],
+		tt->dma_address[i] = dma_map_page(dev, tt->pages[i],
 						  0, num_pages * PAGE_SIZE,
 						  DMA_BIDIRECTIONAL);
 		if (dma_mapping_error(dev, tt->dma_address[i])) {
@@ -1111,7 +1111,7 @@ int ttm_populate_and_map_pages(struct device *dev, struct ttm_dma_tt *tt,
 					       PAGE_SIZE, DMA_BIDIRECTIONAL);
 				tt->dma_address[i] = 0;
 			}
-			ttm_pool_unpopulate(&tt->ttm);
+			ttm_pool_unpopulate(tt);
 			return -EFAULT;
 		}
 
@@ -1124,21 +1124,21 @@ int ttm_populate_and_map_pages(struct device *dev, struct ttm_dma_tt *tt,
 }
 EXPORT_SYMBOL(ttm_populate_and_map_pages);
 
-void ttm_unmap_and_unpopulate_pages(struct device *dev, struct ttm_dma_tt *tt)
+void ttm_unmap_and_unpopulate_pages(struct device *dev, struct ttm_tt *tt)
 {
 	unsigned i, j;
 
-	for (i = 0; i < tt->ttm.num_pages;) {
-		struct page *p = tt->ttm.pages[i];
+	for (i = 0; i < tt->num_pages;) {
+		struct page *p = tt->pages[i];
 		size_t num_pages = 1;
 
-		if (!tt->dma_address[i] || !tt->ttm.pages[i]) {
+		if (!tt->dma_address[i] || !tt->pages[i]) {
 			++i;
 			continue;
 		}
 
-		for (j = i + 1; j < tt->ttm.num_pages; ++j) {
-			if (++p != tt->ttm.pages[j])
+		for (j = i + 1; j < tt->num_pages; ++j) {
+			if (++p != tt->pages[j])
 				break;
 
 			++num_pages;
@@ -1149,7 +1149,7 @@ void ttm_unmap_and_unpopulate_pages(struct device *dev, struct ttm_dma_tt *tt)
 
 		i += num_pages;
 	}
-	ttm_pool_unpopulate(&tt->ttm);
+	ttm_pool_unpopulate(tt);
 }
 EXPORT_SYMBOL(ttm_unmap_and_unpopulate_pages);
 
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
index a9aaed7e618a..c0353c25efd6 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
@@ -832,11 +832,10 @@ static int ttm_dma_page_pool_fill_locked(struct dma_pool *pool,
  * return dma_page pointer if success, otherwise NULL.
  */
 static struct dma_page *ttm_dma_pool_get_pages(struct dma_pool *pool,
-				  struct ttm_dma_tt *ttm_dma,
+				  struct ttm_tt *ttm,
 				  unsigned index)
 {
 	struct dma_page *d_page = NULL;
-	struct ttm_tt *ttm = &ttm_dma->ttm;
 	unsigned long irq_flags;
 	int count;
 
@@ -845,8 +844,8 @@ static struct dma_page *ttm_dma_pool_get_pages(struct dma_pool *pool,
 	if (count) {
 		d_page = list_first_entry(&pool->free_list, struct dma_page, page_list);
 		ttm->pages[index] = d_page->p;
-		ttm_dma->dma_address[index] = d_page->dma;
-		list_move_tail(&d_page->page_list, &ttm_dma->pages_list);
+		ttm->dma_address[index] = d_page->dma;
+		list_move_tail(&d_page->page_list, &ttm->pages_list);
 		pool->npages_in_use += 1;
 		pool->npages_free -= 1;
 	}
@@ -854,9 +853,8 @@ static struct dma_page *ttm_dma_pool_get_pages(struct dma_pool *pool,
 	return d_page;
 }
 
-static gfp_t ttm_dma_pool_gfp_flags(struct ttm_dma_tt *ttm_dma, bool huge)
+static gfp_t ttm_dma_pool_gfp_flags(struct ttm_tt *ttm, bool huge)
 {
-	struct ttm_tt *ttm = &ttm_dma->ttm;
 	gfp_t gfp_flags;
 
 	if (ttm->page_flags & TTM_PAGE_FLAG_DMA32)
@@ -883,11 +881,10 @@ static gfp_t ttm_dma_pool_gfp_flags(struct ttm_dma_tt *ttm_dma, bool huge)
  * On success pages list will hold count number of correctly
  * cached pages. On failure will hold the negative return value (-ENOMEM, etc).
  */
-int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev,
+int ttm_dma_populate(struct ttm_tt *ttm, struct device *dev,
 			struct ttm_operation_ctx *ctx)
 {
 	struct ttm_mem_global *mem_glob = &ttm_mem_glob;
-	struct ttm_tt *ttm = &ttm_dma->ttm;
 	unsigned long num_pages = ttm->num_pages;
 	struct dma_pool *pool;
 	struct dma_page *d_page;
@@ -901,7 +898,7 @@ int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev,
 	if (ttm_check_under_lowerlimit(mem_glob, num_pages, ctx))
 		return -ENOMEM;
 
-	INIT_LIST_HEAD(&ttm_dma->pages_list);
+	INIT_LIST_HEAD(&ttm->pages_list);
 	i = 0;
 
 	type = ttm_to_type(ttm->page_flags, ttm->caching);
@@ -912,7 +909,7 @@ int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev,
 
 	pool = ttm_dma_find_pool(dev, type | IS_HUGE);
 	if (!pool) {
-		gfp_t gfp_flags = ttm_dma_pool_gfp_flags(ttm_dma, true);
+		gfp_t gfp_flags = ttm_dma_pool_gfp_flags(ttm, true);
 
 		pool = ttm_dma_pool_init(dev, gfp_flags, type | IS_HUGE);
 		if (IS_ERR_OR_NULL(pool))
@@ -922,21 +919,21 @@ int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev,
 	while (num_pages >= HPAGE_PMD_NR) {
 		unsigned j;
 
-		d_page = ttm_dma_pool_get_pages(pool, ttm_dma, i);
+		d_page = ttm_dma_pool_get_pages(pool, ttm, i);
 		if (!d_page)
 			break;
 
 		ret = ttm_mem_global_alloc_page(mem_glob, ttm->pages[i],
 						pool->size, ctx);
 		if (unlikely(ret != 0)) {
-			ttm_dma_unpopulate(ttm_dma, dev);
+			ttm_dma_unpopulate(ttm, dev);
 			return -ENOMEM;
 		}
 
 		d_page->vaddr |= VADDR_FLAG_UPDATED_COUNT;
 		for (j = i + 1; j < (i + HPAGE_PMD_NR); ++j) {
 			ttm->pages[j] = ttm->pages[j - 1] + 1;
-			ttm_dma->dma_address[j] = ttm_dma->dma_address[j - 1] +
+			ttm->dma_address[j] = ttm->dma_address[j - 1] +
 				PAGE_SIZE;
 		}
 
@@ -949,7 +946,7 @@ skip_huge:
 
 	pool = ttm_dma_find_pool(dev, type);
 	if (!pool) {
-		gfp_t gfp_flags = ttm_dma_pool_gfp_flags(ttm_dma, false);
+		gfp_t gfp_flags = ttm_dma_pool_gfp_flags(ttm, false);
 
 		pool = ttm_dma_pool_init(dev, gfp_flags, type);
 		if (IS_ERR_OR_NULL(pool))
@@ -957,16 +954,16 @@ skip_huge:
 	}
 
 	while (num_pages) {
-		d_page = ttm_dma_pool_get_pages(pool, ttm_dma, i);
+		d_page = ttm_dma_pool_get_pages(pool, ttm, i);
 		if (!d_page) {
-			ttm_dma_unpopulate(ttm_dma, dev);
+			ttm_dma_unpopulate(ttm, dev);
 			return -ENOMEM;
 		}
 
 		ret = ttm_mem_global_alloc_page(mem_glob, ttm->pages[i],
 						pool->size, ctx);
 		if (unlikely(ret != 0)) {
-			ttm_dma_unpopulate(ttm_dma, dev);
+			ttm_dma_unpopulate(ttm, dev);
 			return -ENOMEM;
 		}
 
@@ -980,10 +977,9 @@ skip_huge:
 EXPORT_SYMBOL_GPL(ttm_dma_populate);
 
 /* Put all pages in pages list to correct pool to wait for reuse */
-void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
+void ttm_dma_unpopulate(struct ttm_tt *ttm, struct device *dev)
 {
 	struct ttm_mem_global *mem_glob = &ttm_mem_glob;
-	struct ttm_tt *ttm = &ttm_dma->ttm;
 	struct dma_pool *pool;
 	struct dma_page *d_page, *next;
 	enum pool_type type;
@@ -997,7 +993,7 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 	pool = ttm_dma_find_pool(dev, type | IS_HUGE);
 	if (pool) {
 		count = 0;
-		list_for_each_entry_safe(d_page, next, &ttm_dma->pages_list,
+		list_for_each_entry_safe(d_page, next, &ttm->pages_list,
 					 page_list) {
 			if (!(d_page->vaddr & VADDR_FLAG_HUGE_POOL))
 				continue;
@@ -1027,7 +1023,7 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 
 	/* make sure pages array match list and count number of pages */
 	count = 0;
-	list_for_each_entry_safe(d_page, next, &ttm_dma->pages_list,
+	list_for_each_entry_safe(d_page, next, &ttm->pages_list,
 				 page_list) {
 		ttm->pages[count] = d_page->p;
 		count++;
@@ -1048,7 +1044,7 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 		pool->nfrees += count;
 	} else {
 		pool->npages_free += count;
-		list_splice(&ttm_dma->pages_list, &pool->free_list);
+		list_splice(&ttm->pages_list, &pool->free_list);
 		/*
 		 * Wait to have at at least NUM_PAGES_TO_ALLOC number of pages
 		 * to free in order to minimize calls to set_memory_wb().
@@ -1059,10 +1055,10 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 	}
 	spin_unlock_irqrestore(&pool->lock, irq_flags);
 
-	INIT_LIST_HEAD(&ttm_dma->pages_list);
+	INIT_LIST_HEAD(&ttm->pages_list);
 	for (i = 0; i < ttm->num_pages; i++) {
 		ttm->pages[i] = NULL;
-		ttm_dma->dma_address[i] = 0;
+		ttm->dma_address[i] = 0;
 	}
 
 	/* shrink pool if necessary (only on !is_cached pools)*/
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index dc1dad982f28..65c4254eea5c 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -92,21 +92,22 @@ static int ttm_tt_alloc_page_directory(struct ttm_tt *ttm)
 	return 0;
 }
 
-static int ttm_dma_tt_alloc_page_directory(struct ttm_dma_tt *ttm)
+static int ttm_dma_tt_alloc_page_directory(struct ttm_tt *ttm)
 {
-	ttm->ttm.pages = kvmalloc_array(ttm->ttm.num_pages,
-					  sizeof(*ttm->ttm.pages) +
-					  sizeof(*ttm->dma_address),
-					  GFP_KERNEL | __GFP_ZERO);
-	if (!ttm->ttm.pages)
+	ttm->pages = kvmalloc_array(ttm->num_pages,
+				    sizeof(*ttm->pages) +
+				    sizeof(*ttm->dma_address),
+				    GFP_KERNEL | __GFP_ZERO);
+	if (!ttm->pages)
 		return -ENOMEM;
-	ttm->dma_address = (void *) (ttm->ttm.pages + ttm->ttm.num_pages);
+
+	ttm->dma_address = (void *)(ttm->pages + ttm->num_pages);
 	return 0;
 }
 
-static int ttm_sg_tt_alloc_page_directory(struct ttm_dma_tt *ttm)
+static int ttm_sg_tt_alloc_page_directory(struct ttm_tt *ttm)
 {
-	ttm->dma_address = kvmalloc_array(ttm->ttm.num_pages,
+	ttm->dma_address = kvmalloc_array(ttm->num_pages,
 					  sizeof(*ttm->dma_address),
 					  GFP_KERNEL | __GFP_ZERO);
 	if (!ttm->dma_address)
@@ -138,8 +139,10 @@ static void ttm_tt_init_fields(struct ttm_tt *ttm,
 	ttm->num_pages = bo->num_pages;
 	ttm->caching = ttm_cached;
 	ttm->page_flags = page_flags;
+	ttm->dma_address = NULL;
 	ttm->swap_storage = NULL;
 	ttm->sg = bo->sg;
+	INIT_LIST_HEAD(&ttm->pages_list);
 	ttm->caching = caching;
 }
 
@@ -158,20 +161,21 @@ EXPORT_SYMBOL(ttm_tt_init);
 
 void ttm_tt_fini(struct ttm_tt *ttm)
 {
-	kvfree(ttm->pages);
+	if (ttm->pages)
+		kvfree(ttm->pages);
+	else
+		kvfree(ttm->dma_address);
 	ttm->pages = NULL;
+	ttm->dma_address = NULL;
 }
 EXPORT_SYMBOL(ttm_tt_fini);
 
-int ttm_dma_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
+int ttm_dma_tt_init(struct ttm_tt *ttm, struct ttm_buffer_object *bo,
 		    uint32_t page_flags, enum ttm_caching caching)
 {
-	struct ttm_tt *ttm = &ttm_dma->ttm;
-
 	ttm_tt_init_fields(ttm, bo, page_flags, caching);
 
-	INIT_LIST_HEAD(&ttm_dma->pages_list);
-	if (ttm_dma_tt_alloc_page_directory(ttm_dma)) {
+	if (ttm_dma_tt_alloc_page_directory(ttm)) {
 		pr_err("Failed allocating page table\n");
 		return -ENOMEM;
 	}
@@ -179,19 +183,17 @@ int ttm_dma_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
 }
 EXPORT_SYMBOL(ttm_dma_tt_init);
 
-int ttm_sg_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
+int ttm_sg_tt_init(struct ttm_tt *ttm, struct ttm_buffer_object *bo,
 		   uint32_t page_flags, enum ttm_caching caching)
 {
-	struct ttm_tt *ttm = &ttm_dma->ttm;
 	int ret;
 
 	ttm_tt_init_fields(ttm, bo, page_flags, caching);
 
-	INIT_LIST_HEAD(&ttm_dma->pages_list);
 	if (page_flags & TTM_PAGE_FLAG_SG)
-		ret = ttm_sg_tt_alloc_page_directory(ttm_dma);
+		ret = ttm_sg_tt_alloc_page_directory(ttm);
 	else
-		ret = ttm_dma_tt_alloc_page_directory(ttm_dma);
+		ret = ttm_dma_tt_alloc_page_directory(ttm);
 	if (ret) {
 		pr_err("Failed allocating page table\n");
 		return -ENOMEM;
@@ -200,19 +202,6 @@ int ttm_sg_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
 }
 EXPORT_SYMBOL(ttm_sg_tt_init);
 
-void ttm_dma_tt_fini(struct ttm_dma_tt *ttm_dma)
-{
-	struct ttm_tt *ttm = &ttm_dma->ttm;
-
-	if (ttm->pages)
-		kvfree(ttm->pages);
-	else
-		kvfree(ttm_dma->dma_address);
-	ttm->pages = NULL;
-	ttm_dma->dma_address = NULL;
-}
-EXPORT_SYMBOL(ttm_dma_tt_fini);
-
 int ttm_tt_swapin(struct ttm_tt *ttm)
 {
 	struct address_space *swap_space;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index 88be48ad0344..92a5d245ff4d 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -186,7 +186,7 @@ struct ttm_placement vmw_nonfixed_placement = {
 };
 
 struct vmw_ttm_tt {
-	struct ttm_dma_tt dma_ttm;
+	struct ttm_tt dma_ttm;
 	struct vmw_private *dev_priv;
 	int gmr_id;
 	struct vmw_mob *mob;
@@ -374,8 +374,8 @@ static int vmw_ttm_map_dma(struct vmw_ttm_tt *vmw_tt)
 		return 0;
 
 	vsgt->mode = dev_priv->map_mode;
-	vsgt->pages = vmw_tt->dma_ttm.ttm.pages;
-	vsgt->num_pages = vmw_tt->dma_ttm.ttm.num_pages;
+	vsgt->pages = vmw_tt->dma_ttm.pages;
+	vsgt->num_pages = vmw_tt->dma_ttm.num_pages;
 	vsgt->addrs = vmw_tt->dma_ttm.dma_address;
 	vsgt->sgt = &vmw_tt->sgt;
 
@@ -483,7 +483,7 @@ static void vmw_ttm_unmap_dma(struct vmw_ttm_tt *vmw_tt)
 const struct vmw_sg_table *vmw_bo_sg_table(struct ttm_buffer_object *bo)
 {
 	struct vmw_ttm_tt *vmw_tt =
-		container_of(bo->ttm, struct vmw_ttm_tt, dma_ttm.ttm);
+		container_of(bo->ttm, struct vmw_ttm_tt, dma_ttm);
 
 	return &vmw_tt->vsgt;
 }
@@ -493,7 +493,7 @@ static int vmw_ttm_bind(struct ttm_bo_device *bdev,
 			struct ttm_tt *ttm, struct ttm_resource *bo_mem)
 {
 	struct vmw_ttm_tt *vmw_be =
-		container_of(ttm, struct vmw_ttm_tt, dma_ttm.ttm);
+		container_of(ttm, struct vmw_ttm_tt, dma_ttm);
 	int ret = 0;
 
 	if (!bo_mem)
@@ -537,7 +537,7 @@ static void vmw_ttm_unbind(struct ttm_bo_device *bdev,
 			   struct ttm_tt *ttm)
 {
 	struct vmw_ttm_tt *vmw_be =
-		container_of(ttm, struct vmw_ttm_tt, dma_ttm.ttm);
+		container_of(ttm, struct vmw_ttm_tt, dma_ttm);
 
 	if (!vmw_be->bound)
 		return;
@@ -562,13 +562,13 @@ static void vmw_ttm_unbind(struct ttm_bo_device *bdev,
 static void vmw_ttm_destroy(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
 	struct vmw_ttm_tt *vmw_be =
-		container_of(ttm, struct vmw_ttm_tt, dma_ttm.ttm);
+		container_of(ttm, struct vmw_ttm_tt, dma_ttm);
 
 	vmw_ttm_unbind(bdev, ttm);
 	ttm_tt_destroy_common(bdev, ttm);
 	vmw_ttm_unmap_dma(vmw_be);
 	if (vmw_be->dev_priv->map_mode == vmw_dma_alloc_coherent)
-		ttm_dma_tt_fini(&vmw_be->dma_ttm);
+		ttm_tt_fini(&vmw_be->dma_ttm);
 	else
 		ttm_tt_fini(ttm);
 
@@ -583,7 +583,7 @@ static int vmw_ttm_populate(struct ttm_bo_device *bdev,
 			    struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 {
 	struct vmw_ttm_tt *vmw_tt =
-		container_of(ttm, struct vmw_ttm_tt, dma_ttm.ttm);
+		container_of(ttm, struct vmw_ttm_tt, dma_ttm);
 	struct vmw_private *dev_priv = vmw_tt->dev_priv;
 	struct ttm_mem_global *glob = vmw_mem_glob(dev_priv);
 	int ret;
@@ -612,7 +612,7 @@ static void vmw_ttm_unpopulate(struct ttm_bo_device *bdev,
 			       struct ttm_tt *ttm)
 {
 	struct vmw_ttm_tt *vmw_tt = container_of(ttm, struct vmw_ttm_tt,
-						 dma_ttm.ttm);
+						 dma_ttm);
 	struct vmw_private *dev_priv = vmw_tt->dev_priv;
 	struct ttm_mem_global *glob = vmw_mem_glob(dev_priv);
 
@@ -650,12 +650,12 @@ static struct ttm_tt *vmw_ttm_tt_create(struct ttm_buffer_object *bo,
 		ret = ttm_dma_tt_init(&vmw_be->dma_ttm, bo, page_flags,
 				      ttm_cached);
 	else
-		ret = ttm_tt_init(&vmw_be->dma_ttm.ttm, bo, page_flags,
+		ret = ttm_tt_init(&vmw_be->dma_ttm, bo, page_flags,
 				  ttm_cached);
 	if (unlikely(ret != 0))
 		goto out_no_init;
 
-	return &vmw_be->dma_ttm.ttm;
+	return &vmw_be->dma_ttm;
 out_no_init:
 	kfree(vmw_be);
 	return NULL;
@@ -813,7 +813,7 @@ int vmw_bo_create_and_populate(struct vmw_private *dev_priv,
 	ret = vmw_ttm_populate(bo->bdev, bo->ttm, &ctx);
 	if (likely(ret == 0)) {
 		struct vmw_ttm_tt *vmw_tt =
-			container_of(bo->ttm, struct vmw_ttm_tt, dma_ttm.ttm);
+			container_of(bo->ttm, struct vmw_ttm_tt, dma_ttm);
 		ret = vmw_ttm_map_dma(vmw_tt);
 	}
 
diff --git a/include/drm/ttm/ttm_page_alloc.h b/include/drm/ttm/ttm_page_alloc.h
index a6b6ef5f9bf4..8fa1e7df6213 100644
--- a/include/drm/ttm/ttm_page_alloc.h
+++ b/include/drm/ttm/ttm_page_alloc.h
@@ -61,13 +61,13 @@ void ttm_pool_unpopulate(struct ttm_tt *ttm);
 /**
  * Populates and DMA maps pages to fullfil a ttm_dma_populate() request
  */
-int ttm_populate_and_map_pages(struct device *dev, struct ttm_dma_tt *tt,
+int ttm_populate_and_map_pages(struct device *dev, struct ttm_tt *tt,
 				struct ttm_operation_ctx *ctx);
 
 /**
  * Unpopulates and DMA unmaps pages as part of a
  * ttm_dma_unpopulate() request */
-void ttm_unmap_and_unpopulate_pages(struct device *dev, struct ttm_dma_tt *tt);
+void ttm_unmap_and_unpopulate_pages(struct device *dev, struct ttm_tt *tt);
 
 /**
  * Output the state of pools to debugfs file
@@ -90,9 +90,9 @@ void ttm_dma_page_alloc_fini(void);
  */
 int ttm_dma_page_alloc_debugfs(struct seq_file *m, void *data);
 
-int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev,
+int ttm_dma_populate(struct ttm_tt *ttm_dma, struct device *dev,
 			struct ttm_operation_ctx *ctx);
-void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev);
+void ttm_dma_unpopulate(struct ttm_tt *ttm_dma, struct device *dev);
 
 #else
 static inline int ttm_dma_page_alloc_init(struct ttm_mem_global *glob,
@@ -107,13 +107,13 @@ static inline int ttm_dma_page_alloc_debugfs(struct seq_file *m, void *data)
 {
 	return 0;
 }
-static inline int ttm_dma_populate(struct ttm_dma_tt *ttm_dma,
+static inline int ttm_dma_populate(struct ttm_tt *ttm_dma,
 				struct device *dev,
 				struct ttm_operation_ctx *ctx)
 {
 	return -ENOMEM;
 }
-static inline void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma,
+static inline void ttm_dma_unpopulate(struct ttm_tt *ttm_dma,
 				      struct device *dev)
 {
 }
diff --git a/include/drm/ttm/ttm_tt.h b/include/drm/ttm/ttm_tt.h
index 931a31355870..df9a80650feb 100644
--- a/include/drm/ttm/ttm_tt.h
+++ b/include/drm/ttm/ttm_tt.h
@@ -47,12 +47,13 @@ struct ttm_operation_ctx;
  * struct ttm_tt
  *
  * @pages: Array of pages backing the data.
+ * @page_flags: see TTM_PAGE_FLAG_*
  * @num_pages: Number of pages in the page array.
- * @bdev: Pointer to the current struct ttm_bo_device.
- * @be: Pointer to the ttm backend.
+ * @sg: for SG objects via dma-buf
+ * @dma_address: The DMA (bus) addresses of the pages
  * @swap_storage: Pointer to shmem struct file for swap storage.
- * @caching_state: The current caching state of the pages.
- * @state: The current binding state of the pages.
+ * @pages_list: used by some page allocation backend
+ * @caching: The current caching state of the pages.
  *
  * This is a structure holding the pages, caching- and aperture binding
  * status for a buffer object that isn't backed by fixed (VRAM / AGP)
@@ -62,8 +63,10 @@ struct ttm_tt {
 	struct page **pages;
 	uint32_t page_flags;
 	uint32_t num_pages;
-	struct sg_table *sg; /* for SG objects via dma-buf */
+	struct sg_table *sg;
+	dma_addr_t *dma_address;
 	struct file *swap_storage;
+	struct list_head pages_list;
 	enum ttm_caching caching;
 };
 
@@ -72,23 +75,6 @@ static inline bool ttm_tt_is_populated(struct ttm_tt *tt)
 	return tt->page_flags & TTM_PAGE_FLAG_PRIV_POPULATED;
 }
 
-/**
- * struct ttm_dma_tt
- *
- * @ttm: Base ttm_tt struct.
- * @dma_address: The DMA (bus) addresses of the pages
- * @pages_list: used by some page allocation backend
- *
- * This is a structure holding the pages, caching- and aperture binding
- * status for a buffer object that isn't backed by fixed (VRAM / AGP)
- * memory.
- */
-struct ttm_dma_tt {
-	struct ttm_tt ttm;
-	dma_addr_t *dma_address;
-	struct list_head pages_list;
-};
-
 /**
  * ttm_tt_create
  *
@@ -115,9 +101,9 @@ int ttm_tt_create(struct ttm_buffer_object *bo, bool zero_alloc);
  */
 int ttm_tt_init(struct ttm_tt *ttm, struct ttm_buffer_object *bo,
 		uint32_t page_flags, enum ttm_caching caching);
-int ttm_dma_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
+int ttm_dma_tt_init(struct ttm_tt *ttm_dma, struct ttm_buffer_object *bo,
 		    uint32_t page_flags, enum ttm_caching caching);
-int ttm_sg_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
+int ttm_sg_tt_init(struct ttm_tt *ttm_dma, struct ttm_buffer_object *bo,
 		   uint32_t page_flags, enum ttm_caching caching);
 
 /**
@@ -128,7 +114,6 @@ int ttm_sg_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_buffer_object *bo,
  * Free memory of ttm_tt structure
  */
 void ttm_tt_fini(struct ttm_tt *ttm);
-void ttm_dma_tt_fini(struct ttm_dma_tt *ttm_dma);
 
 /**
  * ttm_ttm_destroy:
-- 
cgit 


From ee5d2a8e549e90325fcc31825269f89647cd6fac Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Sat, 24 Oct 2020 13:10:28 +0200
Subject: drm/ttm: wire up the new pool as default one v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Provide the necessary parameters by all drivers and use the new pool alloc
when no driver specific function is provided.

v2: fix the GEM VRAM helpers

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Madhav Chauhan <madhav.chauhan@amd.com>
Tested-by: Huang Rui <ray.huang@amd.com>
Link: https://patchwork.freedesktop.org/patch/397081/?series=83051&rev=1
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c |  4 ++--
 drivers/gpu/drm/drm_gem_vram_helper.c   |  4 ++--
 drivers/gpu/drm/nouveau/nouveau_ttm.c   | 14 +++++++++-----
 drivers/gpu/drm/qxl/qxl_ttm.c           |  5 ++---
 drivers/gpu/drm/radeon/radeon_ttm.c     |  4 ++--
 drivers/gpu/drm/ttm/ttm_bo.c            |  8 ++++++--
 drivers/gpu/drm/ttm/ttm_memory.c        |  2 +-
 drivers/gpu/drm/ttm/ttm_tt.c            |  5 ++---
 drivers/gpu/drm/vmwgfx/vmwgfx_drv.c     |  5 +++--
 include/drm/ttm/ttm_bo_driver.h         | 11 +++++++----
 10 files changed, 36 insertions(+), 26 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index beacd00221d8..34944927838e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1914,10 +1914,10 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	mutex_init(&adev->mman.gtt_window_lock);
 
 	/* No others user of address space so set it to 0 */
-	r = ttm_bo_device_init(&adev->mman.bdev,
-			       &amdgpu_bo_driver,
+	r = ttm_bo_device_init(&adev->mman.bdev, &amdgpu_bo_driver, adev->dev,
 			       adev_to_drm(adev)->anon_inode->i_mapping,
 			       adev_to_drm(adev)->vma_offset_manager,
+			       adev->need_swiotlb,
 			       dma_addressing_limited(adev->dev));
 	if (r) {
 		DRM_ERROR("failed initializing buffer object driver(%d).\n", r);
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index 9da823eb0edd..683762b2f9a8 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -1045,10 +1045,10 @@ static int drm_vram_mm_init(struct drm_vram_mm *vmm, struct drm_device *dev,
 	vmm->vram_base = vram_base;
 	vmm->vram_size = vram_size;
 
-	ret = ttm_bo_device_init(&vmm->bdev, &bo_driver,
+	ret = ttm_bo_device_init(&vmm->bdev, &bo_driver, dev->dev,
 				 dev->anon_inode->i_mapping,
 				 dev->vma_offset_manager,
-				 true);
+				 false, true);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index 0592ed6eaad1..d696d882c9eb 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -281,6 +281,7 @@ nouveau_ttm_init(struct nouveau_drm *drm)
 	struct nvkm_pci *pci = device->pci;
 	struct nvif_mmu *mmu = &drm->client.mmu;
 	struct drm_device *dev = drm->dev;
+	bool need_swiotlb = false;
 	int typei, ret;
 
 	ret = nouveau_ttm_init_host(drm, 0);
@@ -315,11 +316,14 @@ nouveau_ttm_init(struct nouveau_drm *drm)
 		drm->agp.cma = pci->agp.cma;
 	}
 
-	ret = ttm_bo_device_init(&drm->ttm.bdev,
-				  &nouveau_bo_driver,
-				  dev->anon_inode->i_mapping,
-				  dev->vma_offset_manager,
-				  drm->client.mmu.dmabits <= 32 ? true : false);
+#if IS_ENABLED(CONFIG_SWIOTLB) && IS_ENABLED(CONFIG_X86)
+	need_swiotlb = !!swiotlb_nr_tbl();
+#endif
+
+	ret = ttm_bo_device_init(&drm->ttm.bdev, &nouveau_bo_driver,
+				 drm->dev->dev, dev->anon_inode->i_mapping,
+				 dev->vma_offset_manager, need_swiotlb,
+				 drm->client.mmu.dmabits <= 32);
 	if (ret) {
 		NV_ERROR(drm, "error initialising bo driver, %d\n", ret);
 		return ret;
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index 9609eeb52821..d8ecfb8b3193 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -194,11 +194,10 @@ int qxl_ttm_init(struct qxl_device *qdev)
 	int num_io_pages; /* != rom->num_io_pages, we include surface0 */
 
 	/* No others user of address space so set it to 0 */
-	r = ttm_bo_device_init(&qdev->mman.bdev,
-			       &qxl_bo_driver,
+	r = ttm_bo_device_init(&qdev->mman.bdev, &qxl_bo_driver, NULL,
 			       qdev->ddev.anon_inode->i_mapping,
 			       qdev->ddev.vma_offset_manager,
-			       false);
+			       false, false);
 	if (r) {
 		DRM_ERROR("failed initializing buffer object driver(%d).\n", r);
 		return r;
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 0a6d7ea847db..1add3918519c 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -846,10 +846,10 @@ int radeon_ttm_init(struct radeon_device *rdev)
 	int r;
 
 	/* No others user of address space so set it to 0 */
-	r = ttm_bo_device_init(&rdev->mman.bdev,
-			       &radeon_bo_driver,
+	r = ttm_bo_device_init(&rdev->mman.bdev, &radeon_bo_driver, rdev->dev,
 			       rdev->ddev->anon_inode->i_mapping,
 			       rdev->ddev->vma_offset_manager,
+			       rdev->need_swiotlb,
 			       dma_addressing_limited(&rdev->pdev->dev));
 	if (r) {
 		DRM_ERROR("failed initializing buffer object driver(%d).\n", r);
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 40c72a0f9325..e9f91cfce5e9 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1283,6 +1283,8 @@ int ttm_bo_device_release(struct ttm_bo_device *bdev)
 			pr_debug("Swap list %d was clean\n", i);
 	spin_unlock(&glob->lru_lock);
 
+	ttm_pool_fini(&bdev->pool);
+
 	if (!ret)
 		ttm_bo_global_release();
 
@@ -1307,9 +1309,10 @@ static void ttm_bo_init_sysman(struct ttm_bo_device *bdev)
 
 int ttm_bo_device_init(struct ttm_bo_device *bdev,
 		       struct ttm_bo_driver *driver,
+		       struct device *dev,
 		       struct address_space *mapping,
 		       struct drm_vma_offset_manager *vma_manager,
-		       bool need_dma32)
+		       bool use_dma_alloc, bool use_dma32)
 {
 	struct ttm_bo_global *glob = &ttm_bo_glob;
 	int ret;
@@ -1324,12 +1327,13 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev,
 	bdev->driver = driver;
 
 	ttm_bo_init_sysman(bdev);
+	ttm_pool_init(&bdev->pool, dev, use_dma_alloc, use_dma32);
 
 	bdev->vma_manager = vma_manager;
 	INIT_DELAYED_WORK(&bdev->wq, ttm_bo_delayed_workqueue);
 	INIT_LIST_HEAD(&bdev->ddestroy);
 	bdev->dev_mapping = mapping;
-	bdev->need_dma32 = need_dma32;
+	bdev->need_dma32 = use_dma32;
 	mutex_lock(&ttm_global_mutex);
 	list_add_tail(&bdev->device_list, &glob->device_list);
 	mutex_unlock(&ttm_global_mutex);
diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c
index 3012d0388c51..b15a91c90271 100644
--- a/drivers/gpu/drm/ttm/ttm_memory.c
+++ b/drivers/gpu/drm/ttm/ttm_memory.c
@@ -454,7 +454,7 @@ int ttm_mem_global_init(struct ttm_mem_global *glob)
 	}
 	ttm_page_alloc_init(glob, glob->zone_kernel->max_mem/(2*PAGE_SIZE));
 	ttm_dma_page_alloc_init(glob, glob->zone_kernel->max_mem/(2*PAGE_SIZE));
-	ttm_pool_mgr_init(glob->zone_kernel->max_mem/(2*PAGE_SIZE));
+	ttm_pool_mgr_init(glob->zone_kernel->max_mem / (2 * PAGE_SIZE));
 	return 0;
 out_no_zone:
 	ttm_mem_global_release(glob);
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index 65c4254eea5c..90054280cd8f 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -37,7 +37,6 @@
 #include <linux/file.h>
 #include <drm/drm_cache.h>
 #include <drm/ttm/ttm_bo_driver.h>
-#include <drm/ttm/ttm_page_alloc.h>
 
 /**
  * Allocates a ttm structure for the given BO.
@@ -321,7 +320,7 @@ int ttm_tt_populate(struct ttm_bo_device *bdev,
 	if (bdev->driver->ttm_tt_populate)
 		ret = bdev->driver->ttm_tt_populate(bdev, ttm, ctx);
 	else
-		ret = ttm_pool_populate(ttm, ctx);
+		ret = ttm_pool_alloc(&bdev->pool, ttm, ctx);
 	if (ret)
 		return ret;
 
@@ -363,6 +362,6 @@ void ttm_tt_unpopulate(struct ttm_bo_device *bdev,
 	if (bdev->driver->ttm_tt_unpopulate)
 		bdev->driver->ttm_tt_unpopulate(bdev, ttm);
 	else
-		ttm_pool_unpopulate(ttm);
+		ttm_pool_free(&bdev->pool, ttm);
 	ttm->page_flags &= ~TTM_PAGE_FLAG_PRIV_POPULATED;
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 4860370740e0..7bd1be26afe4 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -878,10 +878,11 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 	drm_vma_offset_manager_init(&dev_priv->vma_manager,
 				    DRM_FILE_PAGE_OFFSET_START,
 				    DRM_FILE_PAGE_OFFSET_SIZE);
-	ret = ttm_bo_device_init(&dev_priv->bdev,
-				 &vmw_bo_driver,
+	ret = ttm_bo_device_init(&dev_priv->bdev, &vmw_bo_driver,
+				 dev_priv->dev->dev,
 				 dev->anon_inode->i_mapping,
 				 &dev_priv->vma_manager,
+				 dev_priv->map_mode == vmw_dma_alloc_coherent,
 				 false);
 	if (unlikely(ret != 0)) {
 		DRM_ERROR("Failed initializing TTM buffer object driver.\n");
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index 29f6a1d1c853..45ae87640909 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -42,6 +42,7 @@
 #include "ttm_module.h"
 #include "ttm_placement.h"
 #include "ttm_tt.h"
+#include "ttm_pool.h"
 
 /**
  * struct ttm_bo_driver
@@ -295,6 +296,7 @@ struct ttm_bo_device {
 	 * Protected by internal locks.
 	 */
 	struct drm_vma_offset_manager *vma_manager;
+	struct ttm_pool pool;
 
 	/*
 	 * Protected by the global:lru lock.
@@ -395,11 +397,11 @@ int ttm_bo_device_release(struct ttm_bo_device *bdev);
  * @bdev: A pointer to a struct ttm_bo_device to initialize.
  * @glob: A pointer to an initialized struct ttm_bo_global.
  * @driver: A pointer to a struct ttm_bo_driver set up by the caller.
+ * @dev: The core kernel device pointer for DMA mappings and allocations.
  * @mapping: The address space to use for this bo.
  * @vma_manager: A pointer to a vma manager.
- * @file_page_offset: Offset into the device address space that is available
- * for buffer data. This ensures compatibility with other users of the
- * address space.
+ * @use_dma_alloc: If coherent DMA allocation API should be used.
+ * @use_dma32: If we should use GFP_DMA32 for device memory allocations.
  *
  * Initializes a struct ttm_bo_device:
  * Returns:
@@ -407,9 +409,10 @@ int ttm_bo_device_release(struct ttm_bo_device *bdev);
  */
 int ttm_bo_device_init(struct ttm_bo_device *bdev,
 		       struct ttm_bo_driver *driver,
+		       struct device *dev,
 		       struct address_space *mapping,
 		       struct drm_vma_offset_manager *vma_manager,
-		       bool need_dma32);
+		       bool use_dma_alloc, bool use_dma32);
 
 /**
  * ttm_bo_unmap_virtual
-- 
cgit 


From e93b2da9799e5cb97760969f3e1f02a5bdac29fe Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Sat, 24 Oct 2020 13:11:29 +0200
Subject: drm/amdgpu: switch to new allocator v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

It should be able to handle all cases here.

v2: fix debugfs as well

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Madhav Chauhan <madhav.chauhan@amd.com>
Tested-by: Huang Rui <ray.huang@amd.com>
Link: https://patchwork.freedesktop.org/patch/397086/?series=83051&rev=1
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 45 ++++++++++-----------------------
 1 file changed, 14 insertions(+), 31 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 34944927838e..f3416cc1bd04 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -47,7 +47,6 @@
 #include <drm/ttm/ttm_bo_driver.h>
 #include <drm/ttm/ttm_placement.h>
 #include <drm/ttm/ttm_module.h>
-#include <drm/ttm/ttm_page_alloc.h>
 
 #include <drm/drm_debugfs.h>
 #include <drm/amdgpu_drm.h>
@@ -1383,15 +1382,7 @@ static int amdgpu_ttm_tt_populate(struct ttm_bo_device *bdev,
 		return 0;
 	}
 
-#ifdef CONFIG_SWIOTLB
-	if (adev->need_swiotlb && swiotlb_nr_tbl()) {
-		return ttm_dma_populate(&gtt->ttm, adev->dev, ctx);
-	}
-#endif
-
-	/* fall back to generic helper to populate the page array
-	 * and map them to the device */
-	return ttm_populate_and_map_pages(adev->dev, &gtt->ttm, ctx);
+	return ttm_pool_alloc(&adev->mman.bdev.pool, ttm, ctx);
 }
 
 /**
@@ -1400,7 +1391,8 @@ static int amdgpu_ttm_tt_populate(struct ttm_bo_device *bdev,
  * Unmaps pages of a ttm_tt object from the device address space and
  * unpopulates the page array backing it.
  */
-static void amdgpu_ttm_tt_unpopulate(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
+static void amdgpu_ttm_tt_unpopulate(struct ttm_bo_device *bdev,
+				     struct ttm_tt *ttm)
 {
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
 	struct amdgpu_device *adev;
@@ -1425,16 +1417,7 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_bo_device *bdev, struct ttm_tt *
 		return;
 
 	adev = amdgpu_ttm_adev(bdev);
-
-#ifdef CONFIG_SWIOTLB
-	if (adev->need_swiotlb && swiotlb_nr_tbl()) {
-		ttm_dma_unpopulate(&gtt->ttm, adev->dev);
-		return;
-	}
-#endif
-
-	/* fall back to generic helper to unmap and unpopulate array */
-	ttm_unmap_and_unpopulate_pages(adev->dev, &gtt->ttm);
+	return ttm_pool_free(&adev->mman.bdev.pool, ttm);
 }
 
 /**
@@ -2347,16 +2330,22 @@ static int amdgpu_mm_dump_table(struct seq_file *m, void *data)
 	return 0;
 }
 
+static int amdgpu_ttm_pool_debugfs(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);
+
+	return ttm_pool_debugfs(&adev->mman.bdev.pool, m);
+}
+
 static const struct drm_info_list amdgpu_ttm_debugfs_list[] = {
 	{"amdgpu_vram_mm", amdgpu_mm_dump_table, 0, (void *)TTM_PL_VRAM},
 	{"amdgpu_gtt_mm", amdgpu_mm_dump_table, 0, (void *)TTM_PL_TT},
 	{"amdgpu_gds_mm", amdgpu_mm_dump_table, 0, (void *)AMDGPU_PL_GDS},
 	{"amdgpu_gws_mm", amdgpu_mm_dump_table, 0, (void *)AMDGPU_PL_GWS},
 	{"amdgpu_oa_mm", amdgpu_mm_dump_table, 0, (void *)AMDGPU_PL_OA},
-	{"ttm_page_pool", ttm_page_alloc_debugfs, 0, NULL},
-#ifdef CONFIG_SWIOTLB
-	{"ttm_dma_page_pool", ttm_dma_page_alloc_debugfs, 0, NULL}
-#endif
+	{"ttm_page_pool", amdgpu_ttm_pool_debugfs, 0, NULL},
 };
 
 /**
@@ -2649,12 +2638,6 @@ int amdgpu_ttm_debugfs_init(struct amdgpu_device *adev)
 	}
 
 	count = ARRAY_SIZE(amdgpu_ttm_debugfs_list);
-
-#ifdef CONFIG_SWIOTLB
-	if (!(adev->need_swiotlb && swiotlb_nr_tbl()))
-		--count;
-#endif
-
 	return amdgpu_debugfs_add_files(adev, amdgpu_ttm_debugfs_list, count);
 #else
 	return 0;
-- 
cgit 


From f3729f7b1ab3c47d1fb9b445af30854d4b0949f8 Mon Sep 17 00:00:00 2001
From: Deepak R Varma <mh12gx2825@gmail.com>
Date: Mon, 2 Nov 2020 22:50:19 +0530
Subject: drm/amdgpu/amdgpu: improve code indentation and alignment

General code indentation and alignment changes such as replace spaces
by tabs or align function arguments as per the coding style
guidelines. The patch corrects issues for various amdgpu_*.c files
for this driver. Issue reported by checkpatch script.

Signed-off-by: Deepak R Varma <mh12gx2825@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c          | 22 +++++++++++-----------
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 20 ++++++++++----------
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c           |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c      | 10 +++++-----
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c          |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c          | 18 +++++++++---------
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c          | 10 +++++-----
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c   | 14 +++++++-------
 drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c           |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c         |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c          |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c          | 18 +++++++++---------
 drivers/gpu/drm/amd/amdgpu/amdgpu_vf_error.c     |  2 +-
 13 files changed, 62 insertions(+), 62 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
index d3e51d361179..1400957034a1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
@@ -36,17 +36,17 @@
 
 #include "acp_gfx_if.h"
 
-#define ACP_TILE_ON_MASK                	0x03
-#define ACP_TILE_OFF_MASK               	0x02
-#define ACP_TILE_ON_RETAIN_REG_MASK     	0x1f
-#define ACP_TILE_OFF_RETAIN_REG_MASK    	0x20
+#define ACP_TILE_ON_MASK			0x03
+#define ACP_TILE_OFF_MASK			0x02
+#define ACP_TILE_ON_RETAIN_REG_MASK		0x1f
+#define ACP_TILE_OFF_RETAIN_REG_MASK		0x20
 
-#define ACP_TILE_P1_MASK                	0x3e
-#define ACP_TILE_P2_MASK                	0x3d
-#define ACP_TILE_DSP0_MASK              	0x3b
-#define ACP_TILE_DSP1_MASK              	0x37
+#define ACP_TILE_P1_MASK			0x3e
+#define ACP_TILE_P2_MASK			0x3d
+#define ACP_TILE_DSP0_MASK			0x3b
+#define ACP_TILE_DSP1_MASK			0x37
 
-#define ACP_TILE_DSP2_MASK              	0x2f
+#define ACP_TILE_DSP2_MASK			0x2f
 
 #define ACP_DMA_REGS_END			0x146c0
 #define ACP_I2S_PLAY_REGS_START			0x14840
@@ -75,8 +75,8 @@
 #define mmACP_CONTROL				0x5131
 #define mmACP_STATUS				0x5133
 #define mmACP_SOFT_RESET			0x5134
-#define ACP_CONTROL__ClkEn_MASK 		0x1
-#define ACP_SOFT_RESET__SoftResetAud_MASK 	0x100
+#define ACP_CONTROL__ClkEn_MASK			0x1
+#define ACP_SOFT_RESET__SoftResetAud_MASK	0x100
 #define ACP_SOFT_RESET__SoftResetAudDone_MASK	0x1000000
 #define ACP_CLOCK_EN_TIME_OUT_VALUE		0x000000FF
 #define ACP_SOFT_RESET_DONE_TIME_OUT_VALUE	0x000000FF
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 4324479d0f75..703cd5a7b8f7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -1114,19 +1114,19 @@ void amdgpu_amdkfd_gpuvm_destroy_process_vm(struct kgd_dev *kgd, void *vm)
 void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *vm)
 {
 	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-        struct amdgpu_vm *avm = (struct amdgpu_vm *)vm;
+	struct amdgpu_vm *avm = (struct amdgpu_vm *)vm;
 
 	if (WARN_ON(!kgd || !vm))
-                return;
+		return;
 
-        pr_debug("Releasing process vm %p\n", vm);
+	pr_debug("Releasing process vm %p\n", vm);
 
-        /* The original pasid of amdgpu vm has already been
-         * released during making a amdgpu vm to a compute vm
-         * The current pasid is managed by kfd and will be
-         * released on kfd process destroy. Set amdgpu pasid
-         * to 0 to avoid duplicate release.
-         */
+	/* The original pasid of amdgpu vm has already been
+	 * released during making a amdgpu vm to a compute vm
+	 * The current pasid is managed by kfd and will be
+	 * released on kfd process destroy. Set amdgpu pasid
+	 * to 0 to avoid duplicate release.
+	 */
 	amdgpu_vm_release_compute(adev, avm);
 }
 
@@ -1693,7 +1693,7 @@ int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd,
 
 	INIT_LIST_HEAD(&(*mem)->bo_va_list);
 	mutex_init(&(*mem)->lock);
-	
+
 	(*mem)->alloc_flags =
 		((bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM) ?
 		KFD_IOC_ALLOC_MEM_FLAGS_VRAM : KFD_IOC_ALLOC_MEM_FLAGS_GTT)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 12598a4b5c78..309464a8c06b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -326,7 +326,7 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev,
 	increment_us = time_us - adev->mm_stats.last_update_us;
 	adev->mm_stats.last_update_us = time_us;
 	adev->mm_stats.accum_us = min(adev->mm_stats.accum_us + increment_us,
-                                      us_upper_bound);
+				      us_upper_bound);
 
 	/* This prevents the short period of low performance when the VRAM
 	 * usage is low and the driver is in debt or doesn't have enough
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index 58116ca482f8..41ca13f0acd5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -170,14 +170,14 @@ static void amdgpu_debugfs_autodump_init(struct amdgpu_device *adev)
  *
  * Bit 62:  Indicates a GRBM bank switch is needed
  * Bit 61:  Indicates a SRBM bank switch is needed (implies bit 62 is
- * 			zero)
+ * 	    zero)
  * Bits 24..33: The SE or ME selector if needed
  * Bits 34..43: The SH (or SA) or PIPE selector if needed
  * Bits 44..53: The INSTANCE (or CU/WGP) or QUEUE selector if needed
  *
  * Bit 23:  Indicates that the PM power gating lock should be held
- * 			This is necessary to read registers that might be
- * 			unreliable during a power gating transistion.
+ * 	    This is necessary to read registers that might be
+ * 	    unreliable during a power gating transistion.
  *
  * The lower bits are the BYTE offset of the register to read.  This
  * allows reading multiple registers in a single call and having
@@ -865,7 +865,7 @@ static ssize_t amdgpu_debugfs_wave_read(struct file *f, char __user *buf,
 {
 	struct amdgpu_device *adev = f->f_inode->i_private;
 	int r, x;
-	ssize_t result=0;
+	ssize_t result = 0;
 	uint32_t offset, se, sh, cu, wave, simd, data[32];
 
 	if (size & 3 || *pos & 3)
@@ -1211,7 +1211,7 @@ static const char *debugfs_regs_names[] = {
 
 /**
  * amdgpu_debugfs_regs_init -	Initialize debugfs entries that provide
- * 								register access.
+ * 				register access.
  *
  * @adev: The device to attach the debugfs entries to
  */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 4933f6599e16..3d0ea78e444b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -1514,7 +1514,7 @@ static const struct file_operations amdgpu_driver_kms_fops = {
 
 int amdgpu_file_to_fpriv(struct file *filp, struct amdgpu_fpriv **fpriv)
 {
-        struct drm_file *file;
+	struct drm_file *file;
 
 	if (!filp)
 		return -EINVAL;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 487961ed2600..2caea7e340ea 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -656,14 +656,14 @@ static void psp_prep_ta_load_cmd_buf(struct psp_gfx_cmd_resp *cmd,
 				     uint64_t ta_shared_mc,
 				     uint32_t ta_shared_size)
 {
-	cmd->cmd_id 				= GFX_CMD_ID_LOAD_TA;
+	cmd->cmd_id				= GFX_CMD_ID_LOAD_TA;
 	cmd->cmd.cmd_load_ta.app_phy_addr_lo 	= lower_32_bits(ta_bin_mc);
-	cmd->cmd.cmd_load_ta.app_phy_addr_hi 	= upper_32_bits(ta_bin_mc);
-	cmd->cmd.cmd_load_ta.app_len 		= ta_bin_size;
+	cmd->cmd.cmd_load_ta.app_phy_addr_hi	= upper_32_bits(ta_bin_mc);
+	cmd->cmd.cmd_load_ta.app_len		= ta_bin_size;
 
 	cmd->cmd.cmd_load_ta.cmd_buf_phy_addr_lo = lower_32_bits(ta_shared_mc);
 	cmd->cmd.cmd_load_ta.cmd_buf_phy_addr_hi = upper_32_bits(ta_shared_mc);
-	cmd->cmd.cmd_load_ta.cmd_buf_len 	 = ta_shared_size;
+	cmd->cmd.cmd_load_ta.cmd_buf_len	 = ta_shared_size;
 }
 
 static int psp_xgmi_init_shared_buf(struct psp_context *psp)
@@ -687,9 +687,9 @@ static void psp_prep_ta_invoke_cmd_buf(struct psp_gfx_cmd_resp *cmd,
 				       uint32_t ta_cmd_id,
 				       uint32_t session_id)
 {
-	cmd->cmd_id 				= GFX_CMD_ID_INVOKE_CMD;
-	cmd->cmd.cmd_invoke_cmd.session_id 	= session_id;
-	cmd->cmd.cmd_invoke_cmd.ta_cmd_id 	= ta_cmd_id;
+	cmd->cmd_id				= GFX_CMD_ID_INVOKE_CMD;
+	cmd->cmd.cmd_invoke_cmd.session_id	= session_id;
+	cmd->cmd.cmd_invoke_cmd.ta_cmd_id	= ta_cmd_id;
 }
 
 static int psp_ta_invoke(struct psp_context *psp,
@@ -2588,9 +2588,9 @@ int parse_ta_bin_descriptor(struct psp_context *psp,
 
 	switch (desc->fw_type) {
 	case TA_FW_TYPE_PSP_ASD:
-		psp->asd_fw_version 	   = le32_to_cpu(desc->fw_version);
+		psp->asd_fw_version	   = le32_to_cpu(desc->fw_version);
 		psp->asd_feature_version   = le32_to_cpu(desc->fw_version);
-		psp->asd_ucode_size 	   = le32_to_cpu(desc->size_bytes);
+		psp->asd_ucode_size	   = le32_to_cpu(desc->size_bytes);
 		psp->asd_start_addr 	   = ucode_start_addr;
 		break;
 	case TA_FW_TYPE_PSP_XGMI:
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index cf7a79a48221..b96267068a72 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -518,9 +518,9 @@ struct ras_manager *amdgpu_ras_find_obj(struct amdgpu_device *adev,
 /* obj end */
 
 static void amdgpu_ras_parse_status_code(struct amdgpu_device *adev,
-				  const char* 		invoke_type,
-				  const char* 		block_name,
-				  enum ta_ras_status 	ret)
+					 const char* invoke_type,
+					 const char* block_name,
+					 enum ta_ras_status ret)
 {
 	switch (ret) {
 	case TA_RAS_STATUS__SUCCESS:
@@ -609,7 +609,7 @@ int amdgpu_ras_feature_enable(struct amdgpu_device *adev,
 	if (!con)
 		return -EINVAL;
 
-        info = kzalloc(sizeof(union ta_ras_cmd_input), GFP_KERNEL);
+	info = kzalloc(sizeof(union ta_ras_cmd_input), GFP_KERNEL);
 	if (!info)
 		return -ENOMEM;
 
@@ -2081,7 +2081,7 @@ void amdgpu_ras_late_fini(struct amdgpu_device *adev,
 
 	amdgpu_ras_sysfs_remove(adev, ras_block);
 	if (ih_info->cb)
-                amdgpu_ras_interrupt_remove_handler(adev, ih_info);
+		amdgpu_ras_interrupt_remove_handler(adev, ih_info);
 	amdgpu_ras_feature_enable(adev, ras_block, 0);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
index c3710c591b55..1dd040166c63 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
@@ -27,9 +27,9 @@
 #include <linux/bits.h>
 #include "atom.h"
 
-#define EEPROM_I2C_TARGET_ADDR_VEGA20    	0xA0
-#define EEPROM_I2C_TARGET_ADDR_ARCTURUS  	0xA8
-#define EEPROM_I2C_TARGET_ADDR_ARCTURUS_D342  	0xA0
+#define EEPROM_I2C_TARGET_ADDR_VEGA20		0xA0
+#define EEPROM_I2C_TARGET_ADDR_ARCTURUS		0xA8
+#define EEPROM_I2C_TARGET_ADDR_ARCTURUS_D342	0xA0
 
 /*
  * The 2 macros bellow represent the actual size in bytes that
@@ -124,11 +124,11 @@ static void __decode_table_header_from_buff(struct amdgpu_ras_eeprom_table_heade
 {
 	uint32_t *pp = (uint32_t *)buff;
 
-	hdr->header 	      = le32_to_cpu(pp[0]);
-	hdr->version 	      = le32_to_cpu(pp[1]);
+	hdr->header	      = le32_to_cpu(pp[0]);
+	hdr->version	      = le32_to_cpu(pp[1]);
 	hdr->first_rec_offset = le32_to_cpu(pp[2]);
-	hdr->tbl_size 	      = le32_to_cpu(pp[3]);
-	hdr->checksum 	      = le32_to_cpu(pp[4]);
+	hdr->tbl_size	      = le32_to_cpu(pp[3]);
+	hdr->checksum	      = le32_to_cpu(pp[4]);
 }
 
 static int __update_table_header(struct amdgpu_ras_eeprom_control *control,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
index 0bd1d4ffc19e..524d10b21041 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
@@ -75,7 +75,7 @@ int amdgpu_sa_bo_manager_init(struct amdgpu_device *adev,
 }
 
 void amdgpu_sa_bo_manager_fini(struct amdgpu_device *adev,
-                              struct amdgpu_sa_manager *sa_manager)
+			       struct amdgpu_sa_manager *sa_manager)
 {
 	struct amdgpu_sa_bo *sa_bo, *tmp;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
index 250a309e4dee..de91d29c9d96 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
@@ -126,7 +126,7 @@ int amdgpu_sdma_ras_late_init(struct amdgpu_device *adev,
 		goto free;
 	}
 
-        return 0;
+	return 0;
 
 late_fini:
 	amdgpu_ras_late_fini(adev, adev->sdma.ras_if, ih_info);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 11dd3d9eac15..7daded325c3d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -987,7 +987,7 @@ void amdgpu_ttm_tt_set_user_pages(struct ttm_tt *ttm, struct page **pages)
 }
 
 /**
- * amdgpu_ttm_tt_pin_userptr - 	prepare the sg table with the user pages
+ * amdgpu_ttm_tt_pin_userptr - prepare the sg table with the user pages
  *
  * Called by amdgpu_ttm_backend_bind()
  **/
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index 62d4614d48eb..0145504e408b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -37,16 +37,16 @@
 #define FIRMWARE_RAVEN		"amdgpu/raven_vcn.bin"
 #define FIRMWARE_PICASSO	"amdgpu/picasso_vcn.bin"
 #define FIRMWARE_RAVEN2		"amdgpu/raven2_vcn.bin"
-#define FIRMWARE_ARCTURUS 	"amdgpu/arcturus_vcn.bin"
-#define FIRMWARE_RENOIR 	"amdgpu/renoir_vcn.bin"
-#define FIRMWARE_GREEN_SARDINE 	"amdgpu/green_sardine_vcn.bin"
-#define FIRMWARE_NAVI10 	"amdgpu/navi10_vcn.bin"
-#define FIRMWARE_NAVI14 	"amdgpu/navi14_vcn.bin"
-#define FIRMWARE_NAVI12 	"amdgpu/navi12_vcn.bin"
-#define FIRMWARE_SIENNA_CICHLID 	"amdgpu/sienna_cichlid_vcn.bin"
-#define FIRMWARE_NAVY_FLOUNDER 	"amdgpu/navy_flounder_vcn.bin"
+#define FIRMWARE_ARCTURUS	"amdgpu/arcturus_vcn.bin"
+#define FIRMWARE_RENOIR		"amdgpu/renoir_vcn.bin"
+#define FIRMWARE_GREEN_SARDINE	"amdgpu/green_sardine_vcn.bin"
+#define FIRMWARE_NAVI10		"amdgpu/navi10_vcn.bin"
+#define FIRMWARE_NAVI14		"amdgpu/navi14_vcn.bin"
+#define FIRMWARE_NAVI12		"amdgpu/navi12_vcn.bin"
+#define FIRMWARE_SIENNA_CICHLID	"amdgpu/sienna_cichlid_vcn.bin"
+#define FIRMWARE_NAVY_FLOUNDER	"amdgpu/navy_flounder_vcn.bin"
 #define FIRMWARE_VANGOGH	"amdgpu/vangogh_vcn.bin"
-#define FIRMWARE_DIMGREY_CAVEFISH 	"amdgpu/dimgrey_cavefish_vcn.bin"
+#define FIRMWARE_DIMGREY_CAVEFISH	"amdgpu/dimgrey_cavefish_vcn.bin"
 
 MODULE_FIRMWARE(FIRMWARE_RAVEN);
 MODULE_FIRMWARE(FIRMWARE_PICASSO);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vf_error.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vf_error.c
index 7f7097931c6f..f9d3d79f68b1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vf_error.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vf_error.c
@@ -59,7 +59,7 @@ void amdgpu_vf_error_trans_all(struct amdgpu_device *adev)
 		return;
 	}
 /*
- 	TODO: Enable these code when pv2vf_info is merged
+	TODO: Enable these code when pv2vf_info is merged
 	AMDGPU_FW_VRAM_PF2VF_READ (adev, feature_flags, &pf2vf_flags);
 	if (!(pf2vf_flags & AMDGIM_FEATURE_ERROR_LOG_COLLECT)) {
 		return;
-- 
cgit 


From c4c5ae67d17976d058341b24908d5f562f3ce933 Mon Sep 17 00:00:00 2001
From: Deepak R Varma <mh12gx2825@gmail.com>
Date: Tue, 3 Nov 2020 01:07:12 +0530
Subject: drm/amdgpu/amdgpu: use "*" adjacent to data name

When declaring pointer data, the "*" symbol should be used adjacent to
the data name as per the coding standards. This resolves following
issues reported by checkpatch script:
	ERROR: "foo *   bar" should be "foo *bar"
	ERROR: "foo * bar" should be "foo *bar"
	ERROR: "foo*            bar" should be "foo *bar"
	ERROR: "(foo*)" should be "(foo *)"

Signed-off-by: Deepak R Varma <mh12gx2825@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c     |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c           |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c          |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c          | 14 +++++++-------
 drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c         |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_test.c         |  8 ++++----
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c          |  6 +++---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c        |  4 ++--
 9 files changed, 21 insertions(+), 21 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
index 469352e2d6ec..86add0f4ea4d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
@@ -1401,7 +1401,7 @@ static ATOM_VOLTAGE_OBJECT_V3 *amdgpu_atombios_lookup_voltage_object_v3(ATOM_VOL
 {
 	u32 size = le16_to_cpu(v3->sHeader.usStructureSize);
 	u32 offset = offsetof(ATOM_VOLTAGE_OBJECT_INFO_V3_1, asVoltageObj[0]);
-	u8 *start = (u8*)v3;
+	u8 *start = (u8 *)v3;
 
 	while (offset < size) {
 		ATOM_VOLTAGE_OBJECT_V3 *vo = (ATOM_VOLTAGE_OBJECT_V3 *)(start + offset);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index 09cd2d228fdf..306077884a67 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -70,7 +70,7 @@ int amdgpu_atomfirmware_allocate_fb_scratch(struct amdgpu_device *adev)
 	struct atom_context *ctx = adev->mode_info.atom_context;
 	int index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
 						vram_usagebyfirmware);
-	struct vram_usagebyfirmware_v2_1 *	firmware_usage;
+	struct vram_usagebyfirmware_v2_1 *firmware_usage;
 	uint32_t start_addr, size;
 	uint16_t data_offset;
 	int usage_bytes = 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 309464a8c06b..79342976fa76 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -1461,7 +1461,7 @@ int amdgpu_cs_fence_to_handle_ioctl(struct drm_device *dev, void *data,
 		dma_fence_put(fence);
 		if (r)
 			return r;
-		r = drm_syncobj_get_fd(syncobj, (int*)&info->out.handle);
+		r = drm_syncobj_get_fd(syncobj, (int *)&info->out.handle);
 		drm_syncobj_put(syncobj);
 		return r;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
index a03398c87344..0350205c4897 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
@@ -450,7 +450,7 @@ int amdgpu_ctx_put(struct amdgpu_ctx *ctx)
 
 void amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx,
 			  struct drm_sched_entity *entity,
-			  struct dma_fence *fence, uint64_t* handle)
+			  struct dma_fence *fence, uint64_t *handle)
 {
 	struct amdgpu_ctx_entity *centity = to_amdgpu_ctx_entity(entity);
 	uint64_t seq = centity->sequence;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 2caea7e340ea..c3457d675edb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -838,7 +838,7 @@ int psp_xgmi_get_hive_id(struct psp_context *psp, uint64_t *hive_id)
 	struct ta_xgmi_shared_memory *xgmi_cmd;
 	int ret;
 
-	xgmi_cmd = (struct ta_xgmi_shared_memory*)psp->xgmi_context.xgmi_shared_buf;
+	xgmi_cmd = (struct ta_xgmi_shared_memory *)psp->xgmi_context.xgmi_shared_buf;
 	memset(xgmi_cmd, 0, sizeof(struct ta_xgmi_shared_memory));
 
 	xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_HIVE_ID;
@@ -858,7 +858,7 @@ int psp_xgmi_get_node_id(struct psp_context *psp, uint64_t *node_id)
 	struct ta_xgmi_shared_memory *xgmi_cmd;
 	int ret;
 
-	xgmi_cmd = (struct ta_xgmi_shared_memory*)psp->xgmi_context.xgmi_shared_buf;
+	xgmi_cmd = (struct ta_xgmi_shared_memory *)psp->xgmi_context.xgmi_shared_buf;
 	memset(xgmi_cmd, 0, sizeof(struct ta_xgmi_shared_memory));
 
 	xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_NODE_ID;
@@ -886,7 +886,7 @@ int psp_xgmi_get_topology_info(struct psp_context *psp,
 	if (!topology || topology->num_nodes > TA_XGMI__MAX_CONNECTED_NODES)
 		return -EINVAL;
 
-	xgmi_cmd = (struct ta_xgmi_shared_memory*)psp->xgmi_context.xgmi_shared_buf;
+	xgmi_cmd = (struct ta_xgmi_shared_memory *)psp->xgmi_context.xgmi_shared_buf;
 	memset(xgmi_cmd, 0, sizeof(struct ta_xgmi_shared_memory));
 
 	/* Fill in the shared memory with topology information as input */
@@ -930,7 +930,7 @@ int psp_xgmi_set_topology_info(struct psp_context *psp,
 	if (!topology || topology->num_nodes > TA_XGMI__MAX_CONNECTED_NODES)
 		return -EINVAL;
 
-	xgmi_cmd = (struct ta_xgmi_shared_memory*)psp->xgmi_context.xgmi_shared_buf;
+	xgmi_cmd = (struct ta_xgmi_shared_memory *)psp->xgmi_context.xgmi_shared_buf;
 	memset(xgmi_cmd, 0, sizeof(struct ta_xgmi_shared_memory));
 
 	topology_info_input = &xgmi_cmd->xgmi_in_message.get_topology_info;
@@ -994,7 +994,7 @@ static int psp_ras_load(struct psp_context *psp)
 	ret = psp_cmd_submit_buf(psp, NULL, cmd,
 			psp->fence_buf_mc_addr);
 
-	ras_cmd = (struct ta_ras_shared_memory*)psp->ras.ras_shared_buf;
+	ras_cmd = (struct ta_ras_shared_memory *)psp->ras.ras_shared_buf;
 
 	if (!ret) {
 		psp->ras.session_id = cmd->resp.session_id;
@@ -1916,7 +1916,7 @@ static int psp_execute_np_fw_load(struct psp_context *psp,
 static int psp_load_smu_fw(struct psp_context *psp)
 {
 	int ret;
-	struct amdgpu_device* adev = psp->adev;
+	struct amdgpu_device *adev = psp->adev;
 	struct amdgpu_firmware_info *ucode =
 			&adev->firmware.ucode[AMDGPU_UCODE_ID_SMC];
 	struct amdgpu_ras *ras = psp->ras.ras;
@@ -1982,7 +1982,7 @@ static int psp_np_fw_load(struct psp_context *psp)
 {
 	int i, ret;
 	struct amdgpu_firmware_info *ucode;
-	struct amdgpu_device* adev = psp->adev;
+	struct amdgpu_device *adev = psp->adev;
 
 	if (psp->autoload_supported &&
 	    !psp->pmfw_centralized_cstate_management) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
index 15ee13c3bd9e..2697b250dc32 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
@@ -396,7 +396,7 @@ static ssize_t amdgpu_debugfs_ring_read(struct file *f, char __user *buf,
 			return result;
 
 		value = ring->ring[(*pos - 12)/4];
-		r = put_user(value, (uint32_t*)buf);
+		r = put_user(value, (uint32_t *)buf);
 		if (r)
 			return r;
 		buf += 4;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c
index 2f4d5ca9894f..6042b3b81a4c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c
@@ -157,10 +157,10 @@ static void amdgpu_do_test_moves(struct amdgpu_device *adev)
 					  i, *vram_start, gart_start,
 					  (unsigned long long)
 					  (gart_addr - adev->gmc.gart_start +
-					   (void*)gart_start - gtt_map),
+					   (void *)gart_start - gtt_map),
 					  (unsigned long long)
 					  (vram_addr - adev->gmc.vram_start +
-					   (void*)gart_start - gtt_map));
+					   (void *)gart_start - gtt_map));
 				amdgpu_bo_kunmap(vram_obj);
 				goto out_lclean_unpin;
 			}
@@ -203,10 +203,10 @@ static void amdgpu_do_test_moves(struct amdgpu_device *adev)
 					  i, *gart_start, vram_start,
 					  (unsigned long long)
 					  (vram_addr - adev->gmc.vram_start +
-					   (void*)vram_start - vram_map),
+					   (void *)vram_start - vram_map),
 					  (unsigned long long)
 					  (gart_addr - adev->gmc.gart_start +
-					   (void*)vram_start - vram_map));
+					   (void *)vram_start - vram_map));
 				amdgpu_bo_kunmap(gtt_obj[i]);
 				goto out_lclean_unpin;
 			}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 7daded325c3d..69ebd07f3eee 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1115,7 +1115,7 @@ static int amdgpu_ttm_backend_bind(struct ttm_tt *ttm,
 				   struct ttm_resource *bo_mem)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
-	struct amdgpu_ttm_tt *gtt = (void*)ttm;
+	struct amdgpu_ttm_tt *gtt = (void *)ttm;
 	uint64_t flags;
 	int r = 0;
 
@@ -1167,7 +1167,7 @@ int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
 	struct ttm_operation_ctx ctx = { false, false };
-	struct amdgpu_ttm_tt *gtt = (void*)bo->ttm;
+	struct amdgpu_ttm_tt *gtt = (void *)bo->ttm;
 	struct ttm_resource tmp;
 	struct ttm_placement placement;
 	struct ttm_place placements;
@@ -1427,7 +1427,7 @@ int amdgpu_ttm_tt_set_userptr(struct ttm_buffer_object *bo,
 			return -ENOMEM;
 	}
 
-	gtt = (void*)bo->ttm;
+	gtt = (void *)bo->ttm;
 	gtt->userptr = addr;
 	gtt->userflags = flags;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
index 1c9476a060d7..2b7c90b7a712 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
@@ -588,8 +588,8 @@ static int amdgpu_ucode_patch_jt(struct amdgpu_firmware_info *ucode,
 {
 	const struct gfx_firmware_header_v1_0 *header = NULL;
 	const struct common_firmware_header *comm_hdr = NULL;
-	uint8_t* src_addr = NULL;
-	uint8_t* dst_addr = NULL;
+	uint8_t *src_addr = NULL;
+	uint8_t *dst_addr = NULL;
 
 	if (NULL == ucode->fw)
 		return 0;
-- 
cgit 


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_ttm.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 


From ebdf565169af006ee3be8c40eecbfc77d28a3b84 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Thu, 29 Oct 2020 13:58:52 +1000
Subject: drm/ttm: add multihop infrastrucutre (v3)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Currently drivers get called to move a buffer, but if they have to
move it temporarily through another space (SYSTEM->VRAM via TT)
then they can end up with a lot of ttm->driver->ttm call stacks,
if the temprorary space moves requires eviction.

Instead of letting the driver do all the placement/space for the
temporary, allow it to report back (-EMULTIHOP) and a placement (hop)
to the move code, which will then do the temporary move, and the
correct placement move afterwards.

This removes a lot of code from drivers, at the expense of
adding some midlayering. I've some further ideas on how to turn
it inside out, but I think this is a good solution to the call
stack problems.

v2: separate out the driver patches, add WARN for getting
MULTHOP in paths we shouldn't (Daniel)
v3: use memset (Christian)

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: hristian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201109005432.861936-2-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    |  3 +-
 drivers/gpu/drm/drm_gem_vram_helper.c      |  3 +-
 drivers/gpu/drm/nouveau/nouveau_bo.c       |  3 +-
 drivers/gpu/drm/qxl/qxl_ttm.c              |  3 +-
 drivers/gpu/drm/radeon/radeon_ttm.c        |  3 +-
 drivers/gpu/drm/ttm/ttm_bo.c               | 74 ++++++++++++++++++++++++++----
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c |  3 +-
 include/drm/ttm/ttm_bo_driver.h            |  7 ++-
 8 files changed, 83 insertions(+), 16 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 12aa35ac0eb4..fad7ca959316 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -656,7 +656,8 @@ static bool amdgpu_mem_visible(struct amdgpu_device *adev,
  */
 static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 			  struct ttm_operation_ctx *ctx,
-			  struct ttm_resource *new_mem)
+			  struct ttm_resource *new_mem,
+			  struct ttm_place *hop)
 {
 	struct amdgpu_device *adev;
 	struct amdgpu_bo *abo;
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index 5e1b09ba5f2b..889a06696f7e 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -955,7 +955,8 @@ static void bo_driver_delete_mem_notify(struct ttm_buffer_object *bo)
 static int bo_driver_move(struct ttm_buffer_object *bo,
 			  bool evict,
 			  struct ttm_operation_ctx *ctx,
-			  struct ttm_resource *new_mem)
+			  struct ttm_resource *new_mem,
+			  struct ttm_place *hop)
 {
 	struct drm_gem_vram_object *gbo;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 8133377d865d..fee07b9d19ed 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -1023,7 +1023,8 @@ nouveau_bo_vm_cleanup(struct ttm_buffer_object *bo,
 static int
 nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
 		struct ttm_operation_ctx *ctx,
-		struct ttm_resource *new_reg)
+		struct ttm_resource *new_reg,
+		struct ttm_place *hop)
 {
 	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index a80d59634143..128c38c8a837 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -140,7 +140,8 @@ static void qxl_bo_move_notify(struct ttm_buffer_object *bo,
 
 static int qxl_bo_move(struct ttm_buffer_object *bo, bool evict,
 		       struct ttm_operation_ctx *ctx,
-		       struct ttm_resource *new_mem)
+		       struct ttm_resource *new_mem,
+		       struct ttm_place *hop)
 {
 	struct ttm_resource *old_mem = &bo->mem;
 	int ret;
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 95038ac3382e..29062dbea299 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -303,7 +303,8 @@ out_cleanup:
 
 static int radeon_bo_move(struct ttm_buffer_object *bo, bool evict,
 			  struct ttm_operation_ctx *ctx,
-			  struct ttm_resource *new_mem)
+			  struct ttm_resource *new_mem,
+			  struct ttm_place *hop)
 {
 	struct radeon_device *rdev;
 	struct radeon_bo *rbo;
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index e2a124b3affb..e6bcbfe530ec 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -231,7 +231,8 @@ EXPORT_SYMBOL(ttm_bo_bulk_move_lru_tail);
 
 static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 				  struct ttm_resource *mem, bool evict,
-				  struct ttm_operation_ctx *ctx)
+				  struct ttm_operation_ctx *ctx,
+				  struct ttm_place *hop)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_resource_manager *old_man = ttm_manager_type(bdev, bo->mem.mem_type);
@@ -259,9 +260,12 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 		}
 	}
 
-	ret = bdev->driver->move(bo, evict, ctx, mem);
-	if (ret)
+	ret = bdev->driver->move(bo, evict, ctx, mem, hop);
+	if (ret) {
+		if (ret == -EMULTIHOP)
+			return ret;
 		goto out_err;
+	}
 
 	ctx->bytes_moved += bo->num_pages << PAGE_SHIFT;
 	return 0;
@@ -566,8 +570,11 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo,
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_resource evict_mem;
 	struct ttm_placement placement;
+	struct ttm_place hop;
 	int ret = 0;
 
+	memset(&hop, 0, sizeof(hop));
+
 	dma_resv_assert_held(bo->base.resv);
 
 	placement.num_placement = 0;
@@ -596,8 +603,9 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo,
 		goto out;
 	}
 
-	ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, ctx);
+	ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, ctx, &hop);
 	if (unlikely(ret)) {
+		WARN(ret == -EMULTIHOP, "Unexpected multihop in eviction - likely driver bug\n");
 		if (ret != -ERESTARTSYS)
 			pr_err("Buffer eviction failed\n");
 		ttm_resource_free(bo, &evict_mem);
@@ -936,15 +944,45 @@ error:
 }
 EXPORT_SYMBOL(ttm_bo_mem_space);
 
+static int ttm_bo_bounce_temp_buffer(struct ttm_buffer_object *bo,
+				     struct ttm_resource *mem,
+				     struct ttm_operation_ctx *ctx,
+				     struct ttm_place *hop)
+{
+	struct ttm_placement hop_placement;
+	int ret;
+	struct ttm_resource hop_mem = *mem;
+
+	hop_mem.mm_node = NULL;
+	hop_mem.mem_type = TTM_PL_SYSTEM;
+	hop_mem.placement = 0;
+
+	hop_placement.num_placement = hop_placement.num_busy_placement = 1;
+	hop_placement.placement = hop_placement.busy_placement = hop;
+
+	/* find space in the bounce domain */
+	ret = ttm_bo_mem_space(bo, &hop_placement, &hop_mem, ctx);
+	if (ret)
+		return ret;
+	/* move to the bounce domain */
+	ret = ttm_bo_handle_move_mem(bo, &hop_mem, false, ctx, NULL);
+	if (ret)
+		return ret;
+	return 0;
+}
+
 static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 			      struct ttm_placement *placement,
 			      struct ttm_operation_ctx *ctx)
 {
 	int ret = 0;
+	struct ttm_place hop;
 	struct ttm_resource mem;
 
 	dma_resv_assert_held(bo->base.resv);
 
+	memset(&hop, 0, sizeof(hop));
+
 	mem.num_pages = bo->num_pages;
 	mem.size = mem.num_pages << PAGE_SHIFT;
 	mem.page_alignment = bo->mem.page_alignment;
@@ -954,12 +992,25 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 
 	/*
 	 * Determine where to move the buffer.
+	 *
+	 * If driver determines move is going to need
+	 * an extra step then it will return -EMULTIHOP
+	 * and the buffer will be moved to the temporary
+	 * stop and the driver will be called to make
+	 * the second hop.
 	 */
+bounce:
 	ret = ttm_bo_mem_space(bo, placement, &mem, ctx);
 	if (ret)
-		goto out_unlock;
-	ret = ttm_bo_handle_move_mem(bo, &mem, false, ctx);
-out_unlock:
+		return ret;
+	ret = ttm_bo_handle_move_mem(bo, &mem, false, ctx, &hop);
+	if (ret == -EMULTIHOP) {
+		ret = ttm_bo_bounce_temp_buffer(bo, &mem, ctx, &hop);
+		if (ret)
+			return ret;
+		/* try and move to final place now. */
+		goto bounce;
+	}
 	if (ret)
 		ttm_resource_free(bo, &mem);
 	return ret;
@@ -1432,15 +1483,20 @@ int ttm_bo_swapout(struct ttm_operation_ctx *ctx)
 	if (bo->mem.mem_type != TTM_PL_SYSTEM) {
 		struct ttm_operation_ctx ctx = { false, false };
 		struct ttm_resource evict_mem;
+		struct ttm_place hop;
+
+		memset(&hop, 0, sizeof(hop));
 
 		evict_mem = bo->mem;
 		evict_mem.mm_node = NULL;
 		evict_mem.placement = 0;
 		evict_mem.mem_type = TTM_PL_SYSTEM;
 
-		ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, &ctx);
-		if (unlikely(ret != 0))
+		ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, &ctx, &hop);
+		if (unlikely(ret != 0)) {
+			WARN(ret == -EMULTIHOP, "Unexpected multihop in swaput - likely driver bug.\n");
 			goto out;
+		}
 	}
 
 	/**
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index 51f70bea41cc..6a04261ce760 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -695,7 +695,8 @@ static void vmw_swap_notify(struct ttm_buffer_object *bo)
 static int vmw_move(struct ttm_buffer_object *bo,
 		    bool evict,
 		    struct ttm_operation_ctx *ctx,
-		    struct ttm_resource *new_mem)
+		    struct ttm_resource *new_mem,
+		    struct ttm_place *hop)
 {
 	struct ttm_resource_manager *old_man = ttm_manager_type(bo->bdev, bo->mem.mem_type);
 	struct ttm_resource_manager *new_man = ttm_manager_type(bo->bdev, new_mem->mem_type);
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index da8208f43378..f02f7cf9ae90 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -121,6 +121,8 @@ struct ttm_bo_driver {
 	 * Return the bo flags for a buffer which is not mapped to the hardware.
 	 * These will be placed in proposed_flags so that when the move is
 	 * finished, they'll end up in bo->mem.flags
+	 * This should not cause multihop evictions, and the core will warn
+	 * if one is proposed.
 	 */
 
 	void (*evict_flags)(struct ttm_buffer_object *bo,
@@ -134,12 +136,15 @@ struct ttm_bo_driver {
 	 * the graphics address space
 	 * @ctx: context for this move with parameters
 	 * @new_mem: the new memory region receiving the buffer
+	 @ @hop: placement for driver directed intermediate hop
 	 *
 	 * Move a buffer between two memory regions.
+	 * Returns errno -EMULTIHOP if driver requests a hop
 	 */
 	int (*move)(struct ttm_buffer_object *bo, bool evict,
 		    struct ttm_operation_ctx *ctx,
-		    struct ttm_resource *new_mem);
+		    struct ttm_resource *new_mem,
+		    struct ttm_place *hop);
 
 	/**
 	 * struct ttm_bo_driver_member verify_access
-- 
cgit 


From f5a89a5cae812a39993be32e74c8ed7856b1e2b2 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Thu, 29 Oct 2020 13:59:08 +1000
Subject: drm/amdgpu/ttm: use multihop
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This removes the code to move resources directly between
SYSTEM and VRAM in favour of using the core ttm mulithop code.

Signed-off-by: Dave Airlie <airlied@redhat.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201109005432.861936-3-airlied@gmail.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 136 +++-----------------------------
 1 file changed, 12 insertions(+), 124 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index fad7ca959316..676fb520e044 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -512,119 +512,6 @@ error:
 	return r;
 }
 
-/**
- * amdgpu_move_vram_ram - Copy VRAM buffer to RAM buffer
- *
- * Called by amdgpu_bo_move().
- */
-static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, bool evict,
-				struct ttm_operation_ctx *ctx,
-				struct ttm_resource *new_mem)
-{
-	struct ttm_resource *old_mem = &bo->mem;
-	struct ttm_resource tmp_mem;
-	struct ttm_place placements;
-	struct ttm_placement placement;
-	int r;
-
-	/* create space/pages for new_mem in GTT space */
-	tmp_mem = *new_mem;
-	tmp_mem.mm_node = NULL;
-	placement.num_placement = 1;
-	placement.placement = &placements;
-	placement.num_busy_placement = 1;
-	placement.busy_placement = &placements;
-	placements.fpfn = 0;
-	placements.lpfn = 0;
-	placements.mem_type = TTM_PL_TT;
-	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;
-	}
-
-	r = ttm_tt_populate(bo->bdev, bo->ttm, ctx);
-	if (unlikely(r))
-		goto out_cleanup;
-
-	/* Bind the memory to the GTT space */
-	r = amdgpu_ttm_backend_bind(bo->bdev, bo->ttm, &tmp_mem);
-	if (unlikely(r)) {
-		goto out_cleanup;
-	}
-
-	/* blit VRAM to GTT */
-	r = amdgpu_move_blit(bo, evict, &tmp_mem, old_mem);
-	if (unlikely(r)) {
-		goto out_cleanup;
-	}
-
-	r = ttm_bo_wait_ctx(bo, ctx);
-	if (unlikely(r))
-		goto out_cleanup;
-
-	amdgpu_ttm_backend_unbind(bo->bdev, bo->ttm);
-	ttm_resource_free(bo, &bo->mem);
-	ttm_bo_assign_mem(bo, new_mem);
-out_cleanup:
-	ttm_resource_free(bo, &tmp_mem);
-	return r;
-}
-
-/**
- * amdgpu_move_ram_vram - Copy buffer from RAM to VRAM
- *
- * Called by amdgpu_bo_move().
- */
-static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo, bool evict,
-				struct ttm_operation_ctx *ctx,
-				struct ttm_resource *new_mem)
-{
-	struct ttm_resource *old_mem = &bo->mem;
-	struct ttm_resource tmp_mem;
-	struct ttm_placement placement;
-	struct ttm_place placements;
-	int r;
-
-	/* make space in GTT for old_mem buffer */
-	tmp_mem = *new_mem;
-	tmp_mem.mm_node = NULL;
-	placement.num_placement = 1;
-	placement.placement = &placements;
-	placement.num_busy_placement = 1;
-	placement.busy_placement = &placements;
-	placements.fpfn = 0;
-	placements.lpfn = 0;
-	placements.mem_type = TTM_PL_TT;
-	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");
-		return r;
-	}
-
-	/* move/bind old memory to GTT space */
-	r = ttm_tt_populate(bo->bdev, bo->ttm, ctx);
-	if (unlikely(r))
-		return r;
-
-	r = amdgpu_ttm_backend_bind(bo->bdev, bo->ttm, &tmp_mem);
-	if (unlikely(r)) {
-		goto out_cleanup;
-	}
-
-	ttm_bo_assign_mem(bo, &tmp_mem);
-	/* copy to VRAM */
-	r = amdgpu_move_blit(bo, evict, new_mem, old_mem);
-	if (unlikely(r)) {
-		goto out_cleanup;
-	}
-out_cleanup:
-	ttm_resource_free(bo, &tmp_mem);
-	return r;
-}
-
 /**
  * amdgpu_mem_visible - Check that memory can be accessed by ttm_bo_move_memcpy
  *
@@ -664,6 +551,17 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 	struct ttm_resource *old_mem = &bo->mem;
 	int r;
 
+	if ((old_mem->mem_type == TTM_PL_SYSTEM &&
+	     new_mem->mem_type == TTM_PL_VRAM) ||
+	    (old_mem->mem_type == TTM_PL_VRAM &&
+	     new_mem->mem_type == TTM_PL_SYSTEM)) {
+		hop->fpfn = 0;
+		hop->lpfn = 0;
+		hop->mem_type = TTM_PL_TT;
+		hop->flags = 0;
+		return -EMULTIHOP;
+	}
+
 	if (new_mem->mem_type == TTM_PL_TT) {
 		r = amdgpu_ttm_backend_bind(bo->bdev, bo->ttm, new_mem);
 		if (r)
@@ -717,17 +615,7 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 		goto memcpy;
 	}
 
-	if (old_mem->mem_type == TTM_PL_VRAM &&
-	    new_mem->mem_type == TTM_PL_SYSTEM) {
-		r = amdgpu_move_vram_ram(bo, evict, ctx, new_mem);
-	} else if (old_mem->mem_type == TTM_PL_SYSTEM &&
-		   new_mem->mem_type == TTM_PL_VRAM) {
-		r = amdgpu_move_ram_vram(bo, evict, ctx, new_mem);
-	} else {
-		r = amdgpu_move_blit(bo, evict,
-				     new_mem, old_mem);
-	}
-
+	r = amdgpu_move_blit(bo, evict, new_mem, old_mem);
 	if (r) {
 memcpy:
 		/* Check that all memory is CPU accessible */
-- 
cgit 


From 75501872b4af93e228a23534a3f7a7b53da86e84 Mon Sep 17 00:00:00 2001
From: Lee Jones <lee.jones@linaro.org>
Date: Thu, 12 Nov 2020 19:00:21 +0000
Subject: drm/amd/amdgpu/amdgpu_ttm: Demote non-conformant kernel-doc headers,
 fix slightly lacking ones
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:464: warning: Function parameter or member 'bo' not described in 'amdgpu_move_blit'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:464: warning: Function parameter or member 'evict' not described in 'amdgpu_move_blit'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:464: warning: Function parameter or member 'new_mem' not described in 'amdgpu_move_blit'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:464: warning: Function parameter or member 'old_mem' not described in 'amdgpu_move_blit'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:522: warning: Function parameter or member 'adev' not described in 'amdgpu_mem_visible'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:522: warning: Function parameter or member 'mem' not described in 'amdgpu_mem_visible'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:548: warning: Function parameter or member 'bo' not described in 'amdgpu_bo_move'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:548: warning: Function parameter or member 'evict' not described in 'amdgpu_bo_move'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:548: warning: Function parameter or member 'ctx' not described in 'amdgpu_bo_move'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:548: warning: Function parameter or member 'new_mem' not described in 'amdgpu_bo_move'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:548: warning: Function parameter or member 'hop' not described in 'amdgpu_bo_move'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:658: warning: Function parameter or member 'bdev' not described in 'amdgpu_ttm_io_mem_reserve'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:658: warning: Function parameter or member 'mem' not described in 'amdgpu_ttm_io_mem_reserve'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:751: warning: Function parameter or member 'bo' not described in 'amdgpu_ttm_tt_get_user_pages'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:751: warning: Function parameter or member 'pages' not described in 'amdgpu_ttm_tt_get_user_pages'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:855: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_tt_get_user_pages_done'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:892: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_tt_set_user_pages'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:892: warning: Function parameter or member 'pages' not described in 'amdgpu_ttm_tt_set_user_pages'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:906: warning: Function parameter or member 'bdev' not described in 'amdgpu_ttm_tt_pin_userptr'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:906: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_tt_pin_userptr'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:944: warning: Function parameter or member 'bdev' not described in 'amdgpu_ttm_tt_unpin_userptr'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:944: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_tt_unpin_userptr'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1029: warning: Function parameter or member 'bdev' not described in 'amdgpu_ttm_backend_bind'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1029: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_backend_bind'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1029: warning: Function parameter or member 'bo_mem' not described in 'amdgpu_ttm_backend_bind'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1087: warning: Function parameter or member 'bo' not described in 'amdgpu_ttm_alloc_gart'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1146: warning: Function parameter or member 'tbo' not described in 'amdgpu_ttm_recover_gart'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1168: warning: Function parameter or member 'bdev' not described in 'amdgpu_ttm_backend_unbind'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1168: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_backend_unbind'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1214: warning: Function parameter or member 'page_flags' not described in 'amdgpu_ttm_tt_create'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1247: warning: Function parameter or member 'bdev' not described in 'amdgpu_ttm_tt_populate'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1247: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_tt_populate'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1247: warning: Function parameter or member 'ctx' not described in 'amdgpu_ttm_tt_populate'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1291: warning: Function parameter or member 'bdev' not described in 'amdgpu_ttm_tt_unpopulate'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1291: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_tt_unpopulate'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1357: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_tt_get_usermm'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1376: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_tt_affect_userptr'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1376: warning: Function parameter or member 'start' not described in 'amdgpu_ttm_tt_affect_userptr'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1376: warning: Function parameter or member 'end' not described in 'amdgpu_ttm_tt_affect_userptr'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1397: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_tt_is_userptr'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1410: warning: Function parameter or member 'ttm' not described in 'amdgpu_ttm_tt_is_readonly'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1449: warning: bad line:
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1454: warning: Function parameter or member 'adev' not described in 'amdgpu_ttm_tt_pte_flags'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1477: warning: Function parameter or member 'bo' not described in 'amdgpu_ttm_bo_eviction_valuable'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1477: warning: Function parameter or member 'place' not described in 'amdgpu_ttm_bo_eviction_valuable'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1787: warning: Function parameter or member 'adev' not described in 'amdgpu_ttm_init'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1915: warning: Function parameter or member 'adev' not described in 'amdgpu_ttm_late_init'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1926: warning: Function parameter or member 'adev' not described in 'amdgpu_ttm_fini'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2250: warning: Function parameter or member 'f' not described in 'amdgpu_ttm_vram_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2250: warning: Function parameter or member 'buf' not described in 'amdgpu_ttm_vram_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2250: warning: Function parameter or member 'size' not described in 'amdgpu_ttm_vram_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2250: warning: Function parameter or member 'pos' not described in 'amdgpu_ttm_vram_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2285: warning: Function parameter or member 'f' not described in 'amdgpu_ttm_vram_write'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2285: warning: Function parameter or member 'buf' not described in 'amdgpu_ttm_vram_write'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2285: warning: Function parameter or member 'size' not described in 'amdgpu_ttm_vram_write'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2285: warning: Function parameter or member 'pos' not described in 'amdgpu_ttm_vram_write'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2336: warning: Function parameter or member 'f' not described in 'amdgpu_ttm_gtt_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2336: warning: Function parameter or member 'buf' not described in 'amdgpu_ttm_gtt_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2336: warning: Function parameter or member 'size' not described in 'amdgpu_ttm_gtt_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2336: warning: Function parameter or member 'pos' not described in 'amdgpu_ttm_gtt_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2390: warning: Function parameter or member 'f' not described in 'amdgpu_iomem_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2390: warning: Function parameter or member 'buf' not described in 'amdgpu_iomem_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2390: warning: Function parameter or member 'size' not described in 'amdgpu_iomem_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2390: warning: Function parameter or member 'pos' not described in 'amdgpu_iomem_read'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2446: warning: Function parameter or member 'f' not described in 'amdgpu_iomem_write'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2446: warning: Function parameter or member 'buf' not described in 'amdgpu_iomem_write'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2446: warning: Function parameter or member 'size' not described in 'amdgpu_iomem_write'
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:2446: warning: Function parameter or member 'pos' not described in 'amdgpu_iomem_write'

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: Jerome Glisse <glisse@freedesktop.org>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 59 +++++++++++++++++----------------
 1 file changed, 30 insertions(+), 29 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 12aa35ac0eb4..761dd68d7ced 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -451,7 +451,7 @@ error:
 	return r;
 }
 
-/**
+/*
  * amdgpu_move_blit - Copy an entire buffer to another buffer
  *
  * This is a helper called by amdgpu_bo_move() and amdgpu_move_vram_ram() to
@@ -625,7 +625,7 @@ out_cleanup:
 	return r;
 }
 
-/**
+/*
  * amdgpu_mem_visible - Check that memory can be accessed by ttm_bo_move_memcpy
  *
  * Called by amdgpu_bo_move()
@@ -649,7 +649,7 @@ static bool amdgpu_mem_visible(struct amdgpu_device *adev,
 		<= adev->gmc.visible_vram_size;
 }
 
-/**
+/*
  * amdgpu_bo_move - Move a buffer object to a new memory location
  *
  * Called by ttm_bo_handle_move_mem()
@@ -760,7 +760,7 @@ fail:
 	return r;
 }
 
-/**
+/*
  * amdgpu_ttm_io_mem_reserve - Reserve a block of memory during a fault
  *
  * Called by ttm_mem_io_reserve() ultimately via ttm_bo_vm_fault()
@@ -851,7 +851,7 @@ struct amdgpu_ttm_tt {
 };
 
 #ifdef CONFIG_DRM_AMDGPU_USERPTR
-/**
+/*
  * amdgpu_ttm_tt_get_user_pages - get device accessible pages that back user
  * memory and start HMM tracking CPU page table update
  *
@@ -956,7 +956,7 @@ out:
 	return r;
 }
 
-/**
+/*
  * amdgpu_ttm_tt_userptr_range_done - stop HMM track the CPU page table change
  * Check if the pages backing this ttm range have been invalidated
  *
@@ -992,7 +992,7 @@ bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
 }
 #endif
 
-/**
+/*
  * amdgpu_ttm_tt_set_user_pages - Copy pages in, putting old pages as necessary.
  *
  * Called by amdgpu_cs_list_validate(). This creates the page list
@@ -1007,7 +1007,7 @@ void amdgpu_ttm_tt_set_user_pages(struct ttm_tt *ttm, struct page **pages)
 		ttm->pages[i] = pages ? pages[i] : NULL;
 }
 
-/**
+/*
  * amdgpu_ttm_tt_pin_userptr - prepare the sg table with the user pages
  *
  * Called by amdgpu_ttm_backend_bind()
@@ -1047,7 +1047,7 @@ release_sg:
 	return r;
 }
 
-/**
+/*
  * amdgpu_ttm_tt_unpin_userptr - Unpin and unmap userptr pages
  */
 static void amdgpu_ttm_tt_unpin_userptr(struct ttm_bo_device *bdev,
@@ -1128,7 +1128,7 @@ gart_bind_fail:
 	return r;
 }
 
-/**
+/*
  * amdgpu_ttm_backend_bind - Bind GTT memory
  *
  * Called by ttm_tt_bind() on behalf of ttm_bo_handle_move_mem().
@@ -1186,7 +1186,7 @@ static int amdgpu_ttm_backend_bind(struct ttm_bo_device *bdev,
 	return r;
 }
 
-/**
+/*
  * amdgpu_ttm_alloc_gart - Make sure buffer object is accessible either
  * through AGP or GART aperture.
  *
@@ -1247,7 +1247,7 @@ int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo)
 	return 0;
 }
 
-/**
+/*
  * amdgpu_ttm_recover_gart - Rebind GTT pages
  *
  * Called by amdgpu_gtt_mgr_recover() from amdgpu_device_reset() to
@@ -1268,7 +1268,7 @@ int amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo)
 	return r;
 }
 
-/**
+/*
  * amdgpu_ttm_backend_unbind - Unbind GTT mapped pages
  *
  * Called by ttm_tt_unbind() on behalf of ttm_bo_move_ttm() and
@@ -1346,7 +1346,7 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
 	return &gtt->ttm;
 }
 
-/**
+/*
  * amdgpu_ttm_tt_populate - Map GTT pages visible to the device
  *
  * Map the pages of a ttm_tt object to an address space visible
@@ -1391,7 +1391,7 @@ static int amdgpu_ttm_tt_populate(struct ttm_bo_device *bdev,
 	return ttm_pool_alloc(&adev->mman.bdev.pool, ttm, ctx);
 }
 
-/**
+/*
  * amdgpu_ttm_tt_unpopulate - unmap GTT pages and unpopulate page arrays
  *
  * Unmaps pages of a ttm_tt object from the device address space and
@@ -1461,7 +1461,7 @@ int amdgpu_ttm_tt_set_userptr(struct ttm_buffer_object *bo,
 	return 0;
 }
 
-/**
+/*
  * amdgpu_ttm_tt_get_usermm - Return memory manager for ttm_tt object
  */
 struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm)
@@ -1477,7 +1477,7 @@ struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm)
 	return gtt->usertask->mm;
 }
 
-/**
+/*
  * amdgpu_ttm_tt_affect_userptr - Determine if a ttm_tt object lays inside an
  * address range for the current task.
  *
@@ -1501,7 +1501,7 @@ bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start,
 	return true;
 }
 
-/**
+/*
  * amdgpu_ttm_tt_is_userptr - Have the pages backing by userptr?
  */
 bool amdgpu_ttm_tt_is_userptr(struct ttm_tt *ttm)
@@ -1514,7 +1514,7 @@ bool amdgpu_ttm_tt_is_userptr(struct ttm_tt *ttm)
 	return true;
 }
 
-/**
+/*
  * amdgpu_ttm_tt_is_readonly - Is the ttm_tt object read only?
  */
 bool amdgpu_ttm_tt_is_readonly(struct ttm_tt *ttm)
@@ -1555,9 +1555,10 @@ uint64_t amdgpu_ttm_tt_pde_flags(struct ttm_tt *ttm, struct ttm_resource *mem)
 /**
  * amdgpu_ttm_tt_pte_flags - Compute PTE flags for ttm_tt object
  *
+ * @adev: amdgpu_device pointer
  * @ttm: The ttm_tt object to compute the flags for
  * @mem: The memory registry backing this ttm_tt object
-
+ *
  * Figure out the flags to use for a VM PTE (Page Table Entry).
  */
 uint64_t amdgpu_ttm_tt_pte_flags(struct amdgpu_device *adev, struct ttm_tt *ttm,
@@ -1574,7 +1575,7 @@ uint64_t amdgpu_ttm_tt_pte_flags(struct amdgpu_device *adev, struct ttm_tt *ttm,
 	return flags;
 }
 
-/**
+/*
  * amdgpu_ttm_bo_eviction_valuable - Check to see if we can evict a buffer
  * object.
  *
@@ -1885,7 +1886,7 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev)
 	return 0;
 }
 
-/**
+/*
  * amdgpu_ttm_init - Init the memory management (ttm) as well as various
  * gtt/vram related fields.
  *
@@ -2019,7 +2020,7 @@ 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)
@@ -2030,7 +2031,7 @@ void amdgpu_ttm_late_init(struct amdgpu_device *adev)
 	amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL);
 }
 
-/**
+/*
  * amdgpu_ttm_fini - De-initialize the TTM memory pools
  */
 void amdgpu_ttm_fini(struct amdgpu_device *adev)
@@ -2351,7 +2352,7 @@ static const struct drm_info_list amdgpu_ttm_debugfs_list[] = {
 	{"ttm_page_pool", amdgpu_ttm_pool_debugfs, 0, NULL},
 };
 
-/**
+/*
  * amdgpu_ttm_vram_read - Linear read access to VRAM
  *
  * Accesses VRAM via MMIO for debugging purposes.
@@ -2386,7 +2387,7 @@ static ssize_t amdgpu_ttm_vram_read(struct file *f, char __user *buf,
 	return result;
 }
 
-/**
+/*
  * amdgpu_ttm_vram_write - Linear write access to VRAM
  *
  * Accesses VRAM via MMIO for debugging purposes.
@@ -2439,7 +2440,7 @@ static const struct file_operations amdgpu_ttm_vram_fops = {
 
 #ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
 
-/**
+/*
  * amdgpu_ttm_gtt_read - Linear read access to GTT memory
  */
 static ssize_t amdgpu_ttm_gtt_read(struct file *f, char __user *buf,
@@ -2489,7 +2490,7 @@ static const struct file_operations amdgpu_ttm_gtt_fops = {
 
 #endif
 
-/**
+/*
  * amdgpu_iomem_read - Virtual read access to GPU mapped memory
  *
  * This function is used to read memory that has been mapped to the
@@ -2545,7 +2546,7 @@ static ssize_t amdgpu_iomem_read(struct file *f, char __user *buf,
 	return result;
 }
 
-/**
+/*
  * amdgpu_iomem_write - Virtual write access to GPU mapped memory
  *
  * This function is used to write memory that has been mapped to the
-- 
cgit 


From 6abc3f97628cd33726969f76e5c383db7cc8f876 Mon Sep 17 00:00:00 2001
From: Lee Jones <lee.jones@linaro.org>
Date: Mon, 23 Nov 2020 11:18:41 +0000
Subject: drm/amd/amdgpu/amdgpu_ttm: Add description for 'page_flags'
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:1214: warning: Function parameter or member 'page_flags' not described in 'amdgpu_ttm_tt_create'

Acked-by: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: Jerome Glisse <glisse@freedesktop.org>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 1 +
 1 file changed, 1 insertion(+)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 761dd68d7ced..e49b519aab80 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1317,6 +1317,7 @@ static void amdgpu_ttm_backend_destroy(struct ttm_bo_device *bdev,
  * amdgpu_ttm_tt_create - Create a ttm_tt object for a given BO
  *
  * @bo: The buffer object to create a GTT ttm_tt object around
+ * @page_flags: Page flags to be added to the ttm_tt object
  *
  * Called by ttm_tt_create().
  */
-- 
cgit 


From 5f6fab247c7f206731df8bbb096b7624facf9d90 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 4 Dec 2020 11:52:00 -0500
Subject: drm/amdgpu: free the pre-OS console framebuffer after the first
 modeset
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Rather than in late_init to avoid race conditions between freeing the
buffers and the initial modeset.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c           | 8 ++------
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 5 +++++
 2 files changed, 7 insertions(+), 6 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index e49b519aab80..6be131df21f5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -2026,10 +2026,6 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
  */
 void amdgpu_ttm_late_init(struct amdgpu_device *adev)
 {
-	/* return the VGA stolen memory (if any) back to VRAM */
-	if (!adev->mman.keep_stolen_vga_memory)
-		amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL);
-	amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL);
 }
 
 /*
@@ -2042,8 +2038,8 @@ void amdgpu_ttm_fini(struct amdgpu_device *adev)
 
 	amdgpu_ttm_training_reserve_vram_fini(adev);
 	/* return the stolen vga memory back to VRAM */
-	if (adev->mman.keep_stolen_vga_memory)
-		amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL);
 	/* return the IP Discovery TMR memory back to VRAM */
 	amdgpu_bo_free_kernel(&adev->mman.discovery_memory, NULL, NULL);
 	amdgpu_ttm_fw_reserve_vram_fini(adev);
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 20e448e8a5a3..c23896207e9d 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -8462,6 +8462,11 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 
 	drm_atomic_helper_cleanup_planes(dev, state);
 
+	/* return the stolen vga memory back to VRAM */
+	if (!adev->mman.keep_stolen_vga_memory)
+		amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL);
+
 	/*
 	 * Finally, drop a runtime PM reference for each newly disabled CRTC,
 	 * so we can put the GPU into runtime suspend if we're not driving any
-- 
cgit 


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_ttm.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 


From 228ddee8ed9e24084274954eb526b5798b006f1f Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Mon, 16 Nov 2020 20:12:03 +0100
Subject: drm/amdgpu: fix check order in amdgpu_bo_move
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Reorder the code to fix checking if blitting is available.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/401019/
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 53 ++++++++++++++-------------------
 1 file changed, 23 insertions(+), 30 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index b848f9e97613..4d8f19ab1014 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -551,25 +551,12 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 	struct ttm_resource *old_mem = &bo->mem;
 	int r;
 
-	if ((old_mem->mem_type == TTM_PL_SYSTEM &&
-	     new_mem->mem_type == TTM_PL_VRAM) ||
-	    (old_mem->mem_type == TTM_PL_VRAM &&
-	     new_mem->mem_type == TTM_PL_SYSTEM)) {
-		hop->fpfn = 0;
-		hop->lpfn = 0;
-		hop->mem_type = TTM_PL_TT;
-		hop->flags = 0;
-		return -EMULTIHOP;
-	}
-
 	if (new_mem->mem_type == TTM_PL_TT) {
 		r = amdgpu_ttm_backend_bind(bo->bdev, bo->ttm, new_mem);
 		if (r)
 			return r;
 	}
 
-	amdgpu_bo_move_notify(bo, evict, new_mem);
-
 	/* Can't move a pinned BO */
 	abo = ttm_to_amdgpu_bo(bo);
 	if (WARN_ON_ONCE(abo->tbo.pin_count > 0))
@@ -579,24 +566,23 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 
 	if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
 		ttm_bo_move_null(bo, new_mem);
-		return 0;
+		goto out;
 	}
 	if (old_mem->mem_type == TTM_PL_SYSTEM &&
 	    new_mem->mem_type == TTM_PL_TT) {
 		ttm_bo_move_null(bo, new_mem);
-		return 0;
+		goto out;
 	}
-
 	if (old_mem->mem_type == TTM_PL_TT &&
 	    new_mem->mem_type == TTM_PL_SYSTEM) {
 		r = ttm_bo_wait_ctx(bo, ctx);
 		if (r)
-			goto fail;
+			return r;
 
 		amdgpu_ttm_backend_unbind(bo->bdev, bo->ttm);
 		ttm_resource_free(bo, &bo->mem);
 		ttm_bo_assign_mem(bo, new_mem);
-		return 0;
+		goto out;
 	}
 
 	if (old_mem->mem_type == AMDGPU_PL_GDS ||
@@ -607,27 +593,37 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 	    new_mem->mem_type == AMDGPU_PL_OA) {
 		/* Nothing to save here */
 		ttm_bo_move_null(bo, new_mem);
-		return 0;
+		goto out;
 	}
 
-	if (!adev->mman.buffer_funcs_enabled) {
+	if (adev->mman.buffer_funcs_enabled) {
+		if (((old_mem->mem_type == TTM_PL_SYSTEM &&
+		      new_mem->mem_type == TTM_PL_VRAM) ||
+		     (old_mem->mem_type == TTM_PL_VRAM &&
+		      new_mem->mem_type == TTM_PL_SYSTEM))) {
+			hop->fpfn = 0;
+			hop->lpfn = 0;
+			hop->mem_type = TTM_PL_TT;
+			hop->flags = 0;
+			return -EMULTIHOP;
+		}
+
+		r = amdgpu_move_blit(bo, evict, new_mem, old_mem);
+	} else {
 		r = -ENODEV;
-		goto memcpy;
 	}
 
-	r = amdgpu_move_blit(bo, evict, new_mem, old_mem);
 	if (r) {
-memcpy:
 		/* Check that all memory is CPU accessible */
 		if (!amdgpu_mem_visible(adev, old_mem) ||
 		    !amdgpu_mem_visible(adev, new_mem)) {
 			pr_err("Move buffer fallback to memcpy unavailable\n");
-			goto fail;
+			return r;
 		}
 
 		r = ttm_bo_move_memcpy(bo, ctx, new_mem);
 		if (r)
-			goto fail;
+			return r;
 	}
 
 	if (bo->type == ttm_bo_type_device &&
@@ -639,14 +635,11 @@ memcpy:
 		abo->flags &= ~AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 	}
 
+out:
 	/* update statistics */
 	atomic64_add((u64)bo->num_pages << PAGE_SHIFT, &adev->num_bytes_moved);
+	amdgpu_bo_move_notify(bo, evict, new_mem);
 	return 0;
-fail:
-	swap(*new_mem, bo->mem);
-	amdgpu_bo_move_notify(bo, false, new_mem);
-	swap(*new_mem, bo->mem);
-	return r;
 }
 
 /*
-- 
cgit