From 34d7be5dc28041b95254da517fd0f0f740544ece Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Thu, 24 Aug 2017 12:32:55 +0200
Subject: drm/amdgpu: fix and cleanup VM ready check
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Stop checking the mapped BO itself, cause that one is
certainly not a page table.

Additional to that move the code into amdgpu_vm.c

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 33 ++-------------------------------
 1 file changed, 2 insertions(+), 31 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 7171968f261e..9b1b6bdd4841 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -127,35 +127,6 @@ int amdgpu_gem_object_open(struct drm_gem_object *obj,
 	return 0;
 }
 
-static int amdgpu_gem_vm_check(void *param, struct amdgpu_bo *bo)
-{
-	/* if anything is swapped out don't swap it in here,
-	   just abort and wait for the next CS */
-	if (!amdgpu_bo_gpu_accessible(bo))
-		return -ERESTARTSYS;
-
-	if (bo->shadow && !amdgpu_bo_gpu_accessible(bo->shadow))
-		return -ERESTARTSYS;
-
-	return 0;
-}
-
-static bool amdgpu_gem_vm_ready(struct amdgpu_device *adev,
-				struct amdgpu_vm *vm,
-				struct list_head *list)
-{
-	struct ttm_validate_buffer *entry;
-
-	list_for_each_entry(entry, list, head) {
-		struct amdgpu_bo *bo =
-			container_of(entry->bo, struct amdgpu_bo, tbo);
-		if (amdgpu_gem_vm_check(NULL, bo))
-			return false;
-	}
-
-	return !amdgpu_vm_validate_pt_bos(adev, vm, amdgpu_gem_vm_check, NULL);
-}
-
 void amdgpu_gem_object_close(struct drm_gem_object *obj,
 			     struct drm_file *file_priv)
 {
@@ -189,7 +160,7 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj,
 	if (bo_va && --bo_va->ref_count == 0) {
 		amdgpu_vm_bo_rmv(adev, bo_va);
 
-		if (amdgpu_gem_vm_ready(adev, vm, &list)) {
+		if (amdgpu_vm_ready(adev, vm)) {
 			struct dma_fence *fence = NULL;
 
 			r = amdgpu_vm_clear_freed(adev, vm, &fence);
@@ -513,7 +484,7 @@ static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev,
 {
 	int r = -ERESTARTSYS;
 
-	if (!amdgpu_gem_vm_ready(adev, vm, list))
+	if (!amdgpu_vm_ready(adev, vm))
 		goto error;
 
 	r = amdgpu_vm_update_directories(adev, vm);
-- 
cgit v1.2.3


From 6ac7defb5ccf2c5e7b3fc9eb648535f3b3d17bc0 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Wed, 23 Aug 2017 20:11:25 +0200
Subject: drm/amdgpu: cleanup GWS, GDS and OA allocation
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Those are certainly not kernel allocations, instead set the NO_CPU_ACCESS flag.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 9b1b6bdd4841..ba012933e6aa 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -186,17 +186,17 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 {
 	struct amdgpu_device *adev = dev->dev_private;
 	union drm_amdgpu_gem_create *args = data;
+	uint64_t flags = args->in.domain_flags;
 	uint64_t size = args->in.bo_size;
 	struct drm_gem_object *gobj;
 	uint32_t handle;
-	bool kernel = false;
 	int r;
 
 	/* reject invalid gem flags */
-	if (args->in.domain_flags & ~(AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
-				      AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
-				      AMDGPU_GEM_CREATE_CPU_GTT_USWC |
-				      AMDGPU_GEM_CREATE_VRAM_CLEARED))
+	if (flags & ~(AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
+		      AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
+		      AMDGPU_GEM_CREATE_CPU_GTT_USWC |
+		      AMDGPU_GEM_CREATE_VRAM_CLEARED))
 		return -EINVAL;
 
 	/* reject invalid gem domains */
@@ -211,7 +211,7 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 	/* create a gem object to contain this object in */
 	if (args->in.domains & (AMDGPU_GEM_DOMAIN_GDS |
 	    AMDGPU_GEM_DOMAIN_GWS | AMDGPU_GEM_DOMAIN_OA)) {
-		kernel = true;
+		flags |= AMDGPU_GEM_CREATE_NO_CPU_ACCESS;
 		if (args->in.domains == AMDGPU_GEM_DOMAIN_GDS)
 			size = size << AMDGPU_GDS_SHIFT;
 		else if (args->in.domains == AMDGPU_GEM_DOMAIN_GWS)
@@ -225,8 +225,7 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 
 	r = amdgpu_gem_object_create(adev, size, args->in.alignment,
 				     (u32)(0xffffffff & args->in.domains),
-				     args->in.domain_flags,
-				     kernel, &gobj);
+				     flags, false, &gobj);
 	if (r)
 		return r;
 
-- 
cgit v1.2.3


From 3f3333f8a0e90ac26f84ed7b0aa344efce695c08 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Thu, 3 Aug 2017 14:02:13 +0200
Subject: drm/amdgpu: track evicted page tables v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Instead of validating all page tables when one was evicted,
track which one needs a validation.

v2: simplify amdgpu_vm_ready as well

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1)
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c     |   7 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c    |   8 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c |   2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c     | 227 +++++++++++++----------------
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h     |  16 +-
 5 files changed, 119 insertions(+), 141 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index d6ddd5562c16..8bf178a912f2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -636,9 +636,6 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 
 	amdgpu_cs_report_moved_bytes(p->adev, p->bytes_moved,
 				     p->bytes_moved_vis);
-	fpriv->vm.last_eviction_counter =
-		atomic64_read(&p->adev->num_evictions);
-
 	if (p->bo_list) {
 		struct amdgpu_bo *gds = p->bo_list->gds_obj;
 		struct amdgpu_bo *gws = p->bo_list->gws_obj;
@@ -835,7 +832,7 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p)
 			if (!bo)
 				continue;
 
-			amdgpu_vm_bo_invalidate(adev, bo);
+			amdgpu_vm_bo_invalidate(adev, bo, false);
 		}
 	}
 
@@ -860,7 +857,7 @@ static int amdgpu_cs_ib_vm_chunk(struct amdgpu_device *adev,
 	}
 
 	if (p->job->vm) {
-		p->job->vm_pd_addr = amdgpu_bo_gpu_offset(vm->root.bo);
+		p->job->vm_pd_addr = amdgpu_bo_gpu_offset(vm->root.base.bo);
 
 		r = amdgpu_bo_vm_update_pte(p);
 		if (r)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index ba012933e6aa..d02880640ee7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -160,7 +160,7 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj,
 	if (bo_va && --bo_va->ref_count == 0) {
 		amdgpu_vm_bo_rmv(adev, bo_va);
 
-		if (amdgpu_vm_ready(adev, vm)) {
+		if (amdgpu_vm_ready(vm)) {
 			struct dma_fence *fence = NULL;
 
 			r = amdgpu_vm_clear_freed(adev, vm, &fence);
@@ -481,10 +481,10 @@ static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev,
 				    struct list_head *list,
 				    uint32_t operation)
 {
-	int r = -ERESTARTSYS;
+	int r;
 
-	if (!amdgpu_vm_ready(adev, vm))
-		goto error;
+	if (!amdgpu_vm_ready(vm))
+		return;
 
 	r = amdgpu_vm_update_directories(adev, vm);
 	if (r)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 9e495da0bb03..52d0109c0d9c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -929,7 +929,7 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
 		return;
 
 	abo = container_of(bo, struct amdgpu_bo, tbo);
-	amdgpu_vm_bo_invalidate(adev, abo);
+	amdgpu_vm_bo_invalidate(adev, abo, evict);
 
 	amdgpu_bo_kunmap(abo);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 1334bbb82634..6ff3c1bf035e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -140,7 +140,7 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm,
 			 struct list_head *validated,
 			 struct amdgpu_bo_list_entry *entry)
 {
-	entry->robj = vm->root.bo;
+	entry->robj = vm->root.base.bo;
 	entry->priority = 0;
 	entry->tv.bo = &entry->robj->tbo;
 	entry->tv.shared = true;
@@ -148,61 +148,6 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm,
 	list_add(&entry->tv.head, validated);
 }
 
-/**
- * amdgpu_vm_validate_layer - validate a single page table level
- *
- * @parent: parent page table level
- * @validate: callback to do the validation
- * @param: parameter for the validation callback
- *
- * Validate the page table BOs on command submission if neccessary.
- */
-static int amdgpu_vm_validate_level(struct amdgpu_vm_pt *parent,
-				    int (*validate)(void *, struct amdgpu_bo *),
-				    void *param, bool use_cpu_for_update,
-				    struct ttm_bo_global *glob)
-{
-	unsigned i;
-	int r;
-
-	if (use_cpu_for_update) {
-		r = amdgpu_bo_kmap(parent->bo, NULL);
-		if (r)
-			return r;
-	}
-
-	if (!parent->entries)
-		return 0;
-
-	for (i = 0; i <= parent->last_entry_used; ++i) {
-		struct amdgpu_vm_pt *entry = &parent->entries[i];
-
-		if (!entry->bo)
-			continue;
-
-		r = validate(param, entry->bo);
-		if (r)
-			return r;
-
-		spin_lock(&glob->lru_lock);
-		ttm_bo_move_to_lru_tail(&entry->bo->tbo);
-		if (entry->bo->shadow)
-			ttm_bo_move_to_lru_tail(&entry->bo->shadow->tbo);
-		spin_unlock(&glob->lru_lock);
-
-		/*
-		 * Recurse into the sub directory. This is harmless because we
-		 * have only a maximum of 5 layers.
-		 */
-		r = amdgpu_vm_validate_level(entry, validate, param,
-					     use_cpu_for_update, glob);
-		if (r)
-			return r;
-	}
-
-	return r;
-}
-
 /**
  * amdgpu_vm_validate_pt_bos - validate the page table BOs
  *
@@ -217,32 +162,43 @@ int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			      int (*validate)(void *p, struct amdgpu_bo *bo),
 			      void *param)
 {
-	uint64_t num_evictions;
+	struct ttm_bo_global *glob = adev->mman.bdev.glob;
+	int r;
 
-	/* We only need to validate the page tables
-	 * if they aren't already valid.
-	 */
-	num_evictions = atomic64_read(&adev->num_evictions);
-	if (num_evictions == vm->last_eviction_counter)
-		return 0;
+	spin_lock(&vm->status_lock);
+	while (!list_empty(&vm->evicted)) {
+		struct amdgpu_vm_bo_base *bo_base;
+		struct amdgpu_bo *bo;
 
-	return amdgpu_vm_validate_level(&vm->root, validate, param,
-					vm->use_cpu_for_update,
-					adev->mman.bdev.glob);
-}
+		bo_base = list_first_entry(&vm->evicted,
+					   struct amdgpu_vm_bo_base,
+					   vm_status);
+		spin_unlock(&vm->status_lock);
 
-/**
- * amdgpu_vm_check - helper for amdgpu_vm_ready
- */
-static int amdgpu_vm_check(void *param, struct amdgpu_bo *bo)
-{
-	/* if anything is swapped out don't swap it in here,
-	   just abort and wait for the next CS */
-	if (!amdgpu_bo_gpu_accessible(bo))
-		return -ERESTARTSYS;
+		bo = bo_base->bo;
+		BUG_ON(!bo);
+		if (bo->parent) {
+			r = validate(param, bo);
+			if (r)
+				return r;
 
-	if (bo->shadow && !amdgpu_bo_gpu_accessible(bo->shadow))
-		return -ERESTARTSYS;
+			spin_lock(&glob->lru_lock);
+			ttm_bo_move_to_lru_tail(&bo->tbo);
+			if (bo->shadow)
+				ttm_bo_move_to_lru_tail(&bo->shadow->tbo);
+			spin_unlock(&glob->lru_lock);
+		}
+
+		if (vm->use_cpu_for_update) {
+			r = amdgpu_bo_kmap(bo, NULL);
+			if (r)
+				return r;
+		}
+
+		spin_lock(&vm->status_lock);
+		list_del_init(&bo_base->vm_status);
+	}
+	spin_unlock(&vm->status_lock);
 
 	return 0;
 }
@@ -250,17 +206,19 @@ static int amdgpu_vm_check(void *param, struct amdgpu_bo *bo)
 /**
  * amdgpu_vm_ready - check VM is ready for updates
  *
- * @adev: amdgpu device
  * @vm: VM to check
  *
  * Check if all VM PDs/PTs are ready for updates
  */
-bool amdgpu_vm_ready(struct amdgpu_device *adev, struct amdgpu_vm *vm)
+bool amdgpu_vm_ready(struct amdgpu_vm *vm)
 {
-	if (amdgpu_vm_check(NULL, vm->root.bo))
-		return false;
+	bool ready;
+
+	spin_lock(&vm->status_lock);
+	ready = list_empty(&vm->evicted);
+	spin_unlock(&vm->status_lock);
 
-	return !amdgpu_vm_validate_pt_bos(adev, vm, amdgpu_vm_check, NULL);
+	return ready;
 }
 
 /**
@@ -326,11 +284,11 @@ static int amdgpu_vm_alloc_levels(struct amdgpu_device *adev,
 
 	/* walk over the address space and allocate the page tables */
 	for (pt_idx = from; pt_idx <= to; ++pt_idx) {
-		struct reservation_object *resv = vm->root.bo->tbo.resv;
+		struct reservation_object *resv = vm->root.base.bo->tbo.resv;
 		struct amdgpu_vm_pt *entry = &parent->entries[pt_idx];
 		struct amdgpu_bo *pt;
 
-		if (!entry->bo) {
+		if (!entry->base.bo) {
 			r = amdgpu_bo_create(adev,
 					     amdgpu_vm_bo_size(adev, level),
 					     AMDGPU_GPU_PAGE_SIZE, true,
@@ -351,9 +309,12 @@ static int amdgpu_vm_alloc_levels(struct amdgpu_device *adev,
 			/* Keep a reference to the root directory to avoid
 			* freeing them up in the wrong order.
 			*/
-			pt->parent = amdgpu_bo_ref(vm->root.bo);
+			pt->parent = amdgpu_bo_ref(vm->root.base.bo);
 
-			entry->bo = pt;
+			entry->base.vm = vm;
+			entry->base.bo = pt;
+			list_add_tail(&entry->base.bo_list, &pt->va);
+			INIT_LIST_HEAD(&entry->base.vm_status);
 			entry->addr = 0;
 		}
 
@@ -1020,7 +981,7 @@ static int amdgpu_vm_wait_pd(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	int r;
 
 	amdgpu_sync_create(&sync);
-	amdgpu_sync_resv(adev, &sync, vm->root.bo->tbo.resv, owner);
+	amdgpu_sync_resv(adev, &sync, vm->root.base.bo->tbo.resv, owner);
 	r = amdgpu_sync_wait(&sync, true);
 	amdgpu_sync_free(&sync);
 
@@ -1059,10 +1020,10 @@ static int amdgpu_vm_update_level(struct amdgpu_device *adev,
 
 	memset(&params, 0, sizeof(params));
 	params.adev = adev;
-	shadow = parent->bo->shadow;
+	shadow = parent->base.bo->shadow;
 
 	if (vm->use_cpu_for_update) {
-		pd_addr = (unsigned long)amdgpu_bo_kptr(parent->bo);
+		pd_addr = (unsigned long)amdgpu_bo_kptr(parent->base.bo);
 		r = amdgpu_vm_wait_pd(adev, vm, AMDGPU_FENCE_OWNER_VM);
 		if (unlikely(r))
 			return r;
@@ -1078,7 +1039,7 @@ static int amdgpu_vm_update_level(struct amdgpu_device *adev,
 		/* assume the worst case */
 		ndw += parent->last_entry_used * 6;
 
-		pd_addr = amdgpu_bo_gpu_offset(parent->bo);
+		pd_addr = amdgpu_bo_gpu_offset(parent->base.bo);
 
 		if (shadow) {
 			shadow_addr = amdgpu_bo_gpu_offset(shadow);
@@ -1098,7 +1059,7 @@ static int amdgpu_vm_update_level(struct amdgpu_device *adev,
 
 	/* walk over the address space and update the directory */
 	for (pt_idx = 0; pt_idx <= parent->last_entry_used; ++pt_idx) {
-		struct amdgpu_bo *bo = parent->entries[pt_idx].bo;
+		struct amdgpu_bo *bo = parent->entries[pt_idx].base.bo;
 		uint64_t pde, pt;
 
 		if (bo == NULL)
@@ -1141,7 +1102,7 @@ static int amdgpu_vm_update_level(struct amdgpu_device *adev,
 	}
 
 	if (count) {
-		if (vm->root.bo->shadow)
+		if (vm->root.base.bo->shadow)
 			params.func(&params, last_shadow, last_pt,
 				    count, incr, AMDGPU_PTE_VALID);
 
@@ -1154,7 +1115,8 @@ static int amdgpu_vm_update_level(struct amdgpu_device *adev,
 			amdgpu_job_free(job);
 		} else {
 			amdgpu_ring_pad_ib(ring, params.ib);
-			amdgpu_sync_resv(adev, &job->sync, parent->bo->tbo.resv,
+			amdgpu_sync_resv(adev, &job->sync,
+					 parent->base.bo->tbo.resv,
 					 AMDGPU_FENCE_OWNER_VM);
 			if (shadow)
 				amdgpu_sync_resv(adev, &job->sync,
@@ -1167,7 +1129,7 @@ static int amdgpu_vm_update_level(struct amdgpu_device *adev,
 			if (r)
 				goto error_free;
 
-			amdgpu_bo_fence(parent->bo, fence, true);
+			amdgpu_bo_fence(parent->base.bo, fence, true);
 			dma_fence_put(vm->last_dir_update);
 			vm->last_dir_update = dma_fence_get(fence);
 			dma_fence_put(fence);
@@ -1180,7 +1142,7 @@ static int amdgpu_vm_update_level(struct amdgpu_device *adev,
 	for (pt_idx = 0; pt_idx <= parent->last_entry_used; ++pt_idx) {
 		struct amdgpu_vm_pt *entry = &parent->entries[pt_idx];
 
-		if (!entry->bo)
+		if (!entry->base.bo)
 			continue;
 
 		r = amdgpu_vm_update_level(adev, vm, entry, level + 1);
@@ -1213,7 +1175,7 @@ static void amdgpu_vm_invalidate_level(struct amdgpu_vm_pt *parent)
 	for (pt_idx = 0; pt_idx <= parent->last_entry_used; ++pt_idx) {
 		struct amdgpu_vm_pt *entry = &parent->entries[pt_idx];
 
-		if (!entry->bo)
+		if (!entry->base.bo)
 			continue;
 
 		entry->addr = ~0ULL;
@@ -1268,7 +1230,7 @@ void amdgpu_vm_get_entry(struct amdgpu_pte_update_params *p, uint64_t addr,
 	*entry = &p->vm->root;
 	while ((*entry)->entries) {
 		idx = addr >> (p->adev->vm_manager.block_size * level--);
-		idx %= amdgpu_bo_size((*entry)->bo) / 8;
+		idx %= amdgpu_bo_size((*entry)->base.bo) / 8;
 		*parent = *entry;
 		*entry = &(*entry)->entries[idx];
 	}
@@ -1304,7 +1266,7 @@ static void amdgpu_vm_handle_huge_pages(struct amdgpu_pte_update_params *p,
 	    p->src ||
 	    !(flags & AMDGPU_PTE_VALID)) {
 
-		dst = amdgpu_bo_gpu_offset(entry->bo);
+		dst = amdgpu_bo_gpu_offset(entry->base.bo);
 		dst = amdgpu_gart_get_vm_pde(p->adev, dst);
 		flags = AMDGPU_PTE_VALID;
 	} else {
@@ -1330,18 +1292,18 @@ static void amdgpu_vm_handle_huge_pages(struct amdgpu_pte_update_params *p,
 		tmp = p->pages_addr;
 		p->pages_addr = NULL;
 
-		pd_addr = (unsigned long)amdgpu_bo_kptr(parent->bo);
+		pd_addr = (unsigned long)amdgpu_bo_kptr(parent->base.bo);
 		pde = pd_addr + (entry - parent->entries) * 8;
 		amdgpu_vm_cpu_set_ptes(p, pde, dst, 1, 0, flags);
 
 		p->pages_addr = tmp;
 	} else {
-		if (parent->bo->shadow) {
-			pd_addr = amdgpu_bo_gpu_offset(parent->bo->shadow);
+		if (parent->base.bo->shadow) {
+			pd_addr = amdgpu_bo_gpu_offset(parent->base.bo->shadow);
 			pde = pd_addr + (entry - parent->entries) * 8;
 			amdgpu_vm_do_set_ptes(p, pde, dst, 1, 0, flags);
 		}
-		pd_addr = amdgpu_bo_gpu_offset(parent->bo);
+		pd_addr = amdgpu_bo_gpu_offset(parent->base.bo);
 		pde = pd_addr + (entry - parent->entries) * 8;
 		amdgpu_vm_do_set_ptes(p, pde, dst, 1, 0, flags);
 	}
@@ -1392,7 +1354,7 @@ static int amdgpu_vm_update_ptes(struct amdgpu_pte_update_params *params,
 		if (entry->addr & AMDGPU_PDE_PTE)
 			continue;
 
-		pt = entry->bo;
+		pt = entry->base.bo;
 		if (use_cpu_update) {
 			pe_start = (unsigned long)amdgpu_bo_kptr(pt);
 		} else {
@@ -1612,12 +1574,12 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 	if (r)
 		goto error_free;
 
-	r = amdgpu_sync_resv(adev, &job->sync, vm->root.bo->tbo.resv,
+	r = amdgpu_sync_resv(adev, &job->sync, vm->root.base.bo->tbo.resv,
 			     owner);
 	if (r)
 		goto error_free;
 
-	r = reservation_object_reserve_shared(vm->root.bo->tbo.resv);
+	r = reservation_object_reserve_shared(vm->root.base.bo->tbo.resv);
 	if (r)
 		goto error_free;
 
@@ -1632,7 +1594,7 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 	if (r)
 		goto error_free;
 
-	amdgpu_bo_fence(vm->root.bo, f, true);
+	amdgpu_bo_fence(vm->root.base.bo, f, true);
 	dma_fence_put(*fence);
 	*fence = f;
 	return 0;
@@ -1927,7 +1889,7 @@ static void amdgpu_vm_free_mapping(struct amdgpu_device *adev,
  */
 static void amdgpu_vm_prt_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 {
-	struct reservation_object *resv = vm->root.bo->tbo.resv;
+	struct reservation_object *resv = vm->root.base.bo->tbo.resv;
 	struct dma_fence *excl, **shared;
 	unsigned i, shared_count;
 	int r;
@@ -2414,12 +2376,25 @@ void amdgpu_vm_bo_rmv(struct amdgpu_device *adev,
  * Mark @bo as invalid.
  */
 void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
-			     struct amdgpu_bo *bo)
+			     struct amdgpu_bo *bo, bool evicted)
 {
 	struct amdgpu_vm_bo_base *bo_base;
 
 	list_for_each_entry(bo_base, &bo->va, bo_list) {
+		struct amdgpu_vm *vm = bo_base->vm;
+
 		bo_base->moved = true;
+		if (evicted && bo->tbo.resv == vm->root.base.bo->tbo.resv) {
+			spin_lock(&bo_base->vm->status_lock);
+			list_move(&bo_base->vm_status, &vm->evicted);
+			spin_unlock(&bo_base->vm->status_lock);
+			continue;
+		}
+
+		/* Don't add page tables to the moved state */
+		if (bo->tbo.type == ttm_bo_type_kernel)
+			continue;
+
 		spin_lock(&bo_base->vm->status_lock);
 		list_move(&bo_base->vm_status, &bo_base->vm->moved);
 		spin_unlock(&bo_base->vm->status_lock);
@@ -2507,6 +2482,7 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	for (i = 0; i < AMDGPU_MAX_VMHUBS; i++)
 		vm->reserved_vmid[i] = NULL;
 	spin_lock_init(&vm->status_lock);
+	INIT_LIST_HEAD(&vm->evicted);
 	INIT_LIST_HEAD(&vm->moved);
 	INIT_LIST_HEAD(&vm->freed);
 
@@ -2551,30 +2527,31 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	r = amdgpu_bo_create(adev, amdgpu_vm_bo_size(adev, 0), align, true,
 			     AMDGPU_GEM_DOMAIN_VRAM,
 			     flags,
-			     NULL, NULL, init_pde_value, &vm->root.bo);
+			     NULL, NULL, init_pde_value, &vm->root.base.bo);
 	if (r)
 		goto error_free_sched_entity;
 
-	r = amdgpu_bo_reserve(vm->root.bo, false);
-	if (r)
-		goto error_free_root;
-
-	vm->last_eviction_counter = atomic64_read(&adev->num_evictions);
+	vm->root.base.vm = vm;
+	list_add_tail(&vm->root.base.bo_list, &vm->root.base.bo->va);
+	INIT_LIST_HEAD(&vm->root.base.vm_status);
 
 	if (vm->use_cpu_for_update) {
-		r = amdgpu_bo_kmap(vm->root.bo, NULL);
+		r = amdgpu_bo_reserve(vm->root.base.bo, false);
 		if (r)
 			goto error_free_root;
-	}
 
-	amdgpu_bo_unreserve(vm->root.bo);
+		r = amdgpu_bo_kmap(vm->root.base.bo, NULL);
+		if (r)
+			goto error_free_root;
+		amdgpu_bo_unreserve(vm->root.base.bo);
+	}
 
 	return 0;
 
 error_free_root:
-	amdgpu_bo_unref(&vm->root.bo->shadow);
-	amdgpu_bo_unref(&vm->root.bo);
-	vm->root.bo = NULL;
+	amdgpu_bo_unref(&vm->root.base.bo->shadow);
+	amdgpu_bo_unref(&vm->root.base.bo);
+	vm->root.base.bo = NULL;
 
 error_free_sched_entity:
 	amd_sched_entity_fini(&ring->sched, &vm->entity);
@@ -2593,9 +2570,11 @@ static void amdgpu_vm_free_levels(struct amdgpu_vm_pt *level)
 {
 	unsigned i;
 
-	if (level->bo) {
-		amdgpu_bo_unref(&level->bo->shadow);
-		amdgpu_bo_unref(&level->bo);
+	if (level->base.bo) {
+		list_del(&level->base.bo_list);
+		list_del(&level->base.vm_status);
+		amdgpu_bo_unref(&level->base.bo->shadow);
+		amdgpu_bo_unref(&level->base.bo);
 	}
 
 	if (level->entries)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index ff093d4b5e11..4e465e817fe8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -111,12 +111,12 @@ struct amdgpu_vm_bo_base {
 };
 
 struct amdgpu_vm_pt {
-	struct amdgpu_bo	*bo;
-	uint64_t		addr;
+	struct amdgpu_vm_bo_base	base;
+	uint64_t			addr;
 
 	/* array of page tables, one for each directory entry */
-	struct amdgpu_vm_pt	*entries;
-	unsigned		last_entry_used;
+	struct amdgpu_vm_pt		*entries;
+	unsigned			last_entry_used;
 };
 
 struct amdgpu_vm {
@@ -126,6 +126,9 @@ struct amdgpu_vm {
 	/* protecting invalidated */
 	spinlock_t		status_lock;
 
+	/* BOs who needs a validation */
+	struct list_head	evicted;
+
 	/* BOs moved, but not yet updated in the PT */
 	struct list_head	moved;
 
@@ -135,7 +138,6 @@ struct amdgpu_vm {
 	/* contains the page directory */
 	struct amdgpu_vm_pt     root;
 	struct dma_fence	*last_dir_update;
-	uint64_t		last_eviction_counter;
 
 	/* protecting freed */
 	spinlock_t		freed_lock;
@@ -225,7 +227,7 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm);
 void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm,
 			 struct list_head *validated,
 			 struct amdgpu_bo_list_entry *entry);
-bool amdgpu_vm_ready(struct amdgpu_device *adev, struct amdgpu_vm *vm);
+bool amdgpu_vm_ready(struct amdgpu_vm *vm);
 int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			      int (*callback)(void *p, struct amdgpu_bo *bo),
 			      void *param);
@@ -250,7 +252,7 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev,
 			struct amdgpu_bo_va *bo_va,
 			bool clear);
 void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
-			     struct amdgpu_bo *bo);
+			     struct amdgpu_bo *bo, bool evicted);
 struct amdgpu_bo_va *amdgpu_vm_bo_find(struct amdgpu_vm *vm,
 				       struct amdgpu_bo *bo);
 struct amdgpu_bo_va *amdgpu_vm_bo_add(struct amdgpu_device *adev,
-- 
cgit v1.2.3


From 4f5839c56ec38e2f8fcc59ca0e01defa8702987b Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Tue, 29 Aug 2017 16:07:31 +0200
Subject: drm/amdgpu: restrict userptr even more
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Don't allow them to be GEM imported into another process.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 6 ++++++
 1 file changed, 6 insertions(+)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index d02880640ee7..e32a2b55b54f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -112,7 +112,13 @@ int amdgpu_gem_object_open(struct drm_gem_object *obj,
 	struct amdgpu_fpriv *fpriv = file_priv->driver_priv;
 	struct amdgpu_vm *vm = &fpriv->vm;
 	struct amdgpu_bo_va *bo_va;
+	struct mm_struct *mm;
 	int r;
+
+	mm = amdgpu_ttm_tt_get_usermm(abo->tbo.ttm);
+	if (mm && mm != current->mm)
+		return -EPERM;
+
 	r = amdgpu_bo_reserve(abo, false);
 	if (r)
 		return r;
-- 
cgit v1.2.3


From e1eb899b45781b9bb77e6d7772d6e67bb0bc1a18 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Fri, 25 Aug 2017 09:14:43 +0200
Subject: drm/amdgpu: add IOCTL interface for per VM BOs v3
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Add the IOCTL interface so that applications can allocate per VM BOs.

Still WIP since not all corner cases are tested yet, but this reduces average
CS overhead for 10K BOs from 21ms down to 48us.

v2: add some extra checks, remove the WIP tag
v3: rename new flag to AMDGPU_GEM_CREATE_VM_ALWAYS_VALID

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h       |  7 ++--
 drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c    |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c   | 63 ++++++++++++++++++++++---------
 drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c |  3 +-
 include/uapi/drm/amdgpu_drm.h             |  2 +
 5 files changed, 55 insertions(+), 22 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 103635ab784c..5809f55e0d9d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -455,9 +455,10 @@ struct amdgpu_sa_bo {
  */
 void amdgpu_gem_force_release(struct amdgpu_device *adev);
 int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
-				int alignment, u32 initial_domain,
-				u64 flags, bool kernel,
-				struct drm_gem_object **obj);
+			     int alignment, u32 initial_domain,
+			     u64 flags, bool kernel,
+			     struct reservation_object *resv,
+			     struct drm_gem_object **obj);
 
 int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_device *dev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
index 9afa9c097e1f..b6cb276f0a70 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
@@ -149,7 +149,7 @@ static int amdgpufb_create_pinned_object(struct amdgpu_fbdev *rfbdev,
 				       AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
 				       AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS |
 				       AMDGPU_GEM_CREATE_VRAM_CLEARED,
-				       true, &gobj);
+				       true, NULL, &gobj);
 	if (ret) {
 		pr_err("failed to allocate framebuffer (%d)\n", aligned_size);
 		return -ENOMEM;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index e32a2b55b54f..f1e61b3df640 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -44,11 +44,12 @@ void amdgpu_gem_object_free(struct drm_gem_object *gobj)
 }
 
 int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
-				int alignment, u32 initial_domain,
-				u64 flags, bool kernel,
-				struct drm_gem_object **obj)
+			     int alignment, u32 initial_domain,
+			     u64 flags, bool kernel,
+			     struct reservation_object *resv,
+			     struct drm_gem_object **obj)
 {
-	struct amdgpu_bo *robj;
+	struct amdgpu_bo *bo;
 	int r;
 
 	*obj = NULL;
@@ -59,7 +60,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 
 retry:
 	r = amdgpu_bo_create(adev, size, alignment, kernel, initial_domain,
-			     flags, NULL, NULL, 0, &robj);
+			     flags, NULL, resv, 0, &bo);
 	if (r) {
 		if (r != -ERESTARTSYS) {
 			if (initial_domain == AMDGPU_GEM_DOMAIN_VRAM) {
@@ -71,7 +72,7 @@ retry:
 		}
 		return r;
 	}
-	*obj = &robj->gem_base;
+	*obj = &bo->gem_base;
 
 	return 0;
 }
@@ -119,6 +120,10 @@ int amdgpu_gem_object_open(struct drm_gem_object *obj,
 	if (mm && mm != current->mm)
 		return -EPERM;
 
+	if (abo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID &&
+	    abo->tbo.resv != vm->root.base.bo->tbo.resv)
+		return -EPERM;
+
 	r = amdgpu_bo_reserve(abo, false);
 	if (r)
 		return r;
@@ -142,13 +147,14 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj,
 	struct amdgpu_vm *vm = &fpriv->vm;
 
 	struct amdgpu_bo_list_entry vm_pd;
-	struct list_head list;
+	struct list_head list, duplicates;
 	struct ttm_validate_buffer tv;
 	struct ww_acquire_ctx ticket;
 	struct amdgpu_bo_va *bo_va;
 	int r;
 
 	INIT_LIST_HEAD(&list);
+	INIT_LIST_HEAD(&duplicates);
 
 	tv.bo = &bo->tbo;
 	tv.shared = true;
@@ -156,7 +162,7 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj,
 
 	amdgpu_vm_get_pd_bo(vm, &list, &vm_pd);
 
-	r = ttm_eu_reserve_buffers(&ticket, &list, false, NULL);
+	r = ttm_eu_reserve_buffers(&ticket, &list, false, &duplicates);
 	if (r) {
 		dev_err(adev->dev, "leaking bo va because "
 			"we fail to reserve bo (%d)\n", r);
@@ -191,9 +197,12 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 			    struct drm_file *filp)
 {
 	struct amdgpu_device *adev = dev->dev_private;
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
+	struct amdgpu_vm *vm = &fpriv->vm;
 	union drm_amdgpu_gem_create *args = data;
 	uint64_t flags = args->in.domain_flags;
 	uint64_t size = args->in.bo_size;
+	struct reservation_object *resv = NULL;
 	struct drm_gem_object *gobj;
 	uint32_t handle;
 	int r;
@@ -202,7 +211,8 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 	if (flags & ~(AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
 		      AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
 		      AMDGPU_GEM_CREATE_CPU_GTT_USWC |
-		      AMDGPU_GEM_CREATE_VRAM_CLEARED))
+		      AMDGPU_GEM_CREATE_VRAM_CLEARED |
+		      AMDGPU_GEM_CREATE_VM_ALWAYS_VALID))
 		return -EINVAL;
 
 	/* reject invalid gem domains */
@@ -229,9 +239,25 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 	}
 	size = roundup(size, PAGE_SIZE);
 
+	if (flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID) {
+		r = amdgpu_bo_reserve(vm->root.base.bo, false);
+		if (r)
+			return r;
+
+		resv = vm->root.base.bo->tbo.resv;
+	}
+
 	r = amdgpu_gem_object_create(adev, size, args->in.alignment,
 				     (u32)(0xffffffff & args->in.domains),
-				     flags, false, &gobj);
+				     flags, false, resv, &gobj);
+	if (flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID) {
+		if (!r) {
+			struct amdgpu_bo *abo = gem_to_amdgpu_bo(gobj);
+
+			abo->parent = amdgpu_bo_ref(vm->root.base.bo);
+		}
+		amdgpu_bo_unreserve(vm->root.base.bo);
+	}
 	if (r)
 		return r;
 
@@ -273,9 +299,8 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 	}
 
 	/* create a gem object to contain this object in */
-	r = amdgpu_gem_object_create(adev, args->size, 0,
-				     AMDGPU_GEM_DOMAIN_CPU, 0,
-				     0, &gobj);
+	r = amdgpu_gem_object_create(adev, args->size, 0, AMDGPU_GEM_DOMAIN_CPU,
+				     0, 0, NULL, &gobj);
 	if (r)
 		return r;
 
@@ -527,7 +552,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	struct amdgpu_bo_list_entry vm_pd;
 	struct ttm_validate_buffer tv;
 	struct ww_acquire_ctx ticket;
-	struct list_head list;
+	struct list_head list, duplicates;
 	uint64_t va_flags;
 	int r = 0;
 
@@ -563,6 +588,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	}
 
 	INIT_LIST_HEAD(&list);
+	INIT_LIST_HEAD(&duplicates);
 	if ((args->operation != AMDGPU_VA_OP_CLEAR) &&
 	    !(args->flags & AMDGPU_VM_PAGE_PRT)) {
 		gobj = drm_gem_object_lookup(filp, args->handle);
@@ -579,7 +605,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 
 	amdgpu_vm_get_pd_bo(&fpriv->vm, &list, &vm_pd);
 
-	r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL);
+	r = ttm_eu_reserve_buffers(&ticket, &list, true, &duplicates);
 	if (r)
 		goto error_unref;
 
@@ -645,6 +671,7 @@ error_unref:
 int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *filp)
 {
+	struct amdgpu_device *adev = dev->dev_private;
 	struct drm_amdgpu_gem_op *args = data;
 	struct drm_gem_object *gobj;
 	struct amdgpu_bo *robj;
@@ -692,6 +719,9 @@ int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data,
 		if (robj->allowed_domains == AMDGPU_GEM_DOMAIN_VRAM)
 			robj->allowed_domains |= AMDGPU_GEM_DOMAIN_GTT;
 
+		if (robj->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID)
+			amdgpu_vm_bo_invalidate(adev, robj, true);
+
 		amdgpu_bo_unreserve(robj);
 		break;
 	default:
@@ -721,8 +751,7 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 	r = amdgpu_gem_object_create(adev, args->size, 0,
 				     AMDGPU_GEM_DOMAIN_VRAM,
 				     AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED,
-				     ttm_bo_type_device,
-				     &gobj);
+				     false, NULL, &gobj);
 	if (r)
 		return -ENOMEM;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
index 5b3f92891f89..7e0826469b5e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c
@@ -136,7 +136,8 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_device *dev,
 {
 	struct amdgpu_bo *bo = gem_to_amdgpu_bo(gobj);
 
-	if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm))
+	if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm) ||
+	    bo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID)
 		return ERR_PTR(-EPERM);
 
 	return drm_gem_prime_export(dev, gobj, flags);
diff --git a/include/uapi/drm/amdgpu_drm.h b/include/uapi/drm/amdgpu_drm.h
index 7b8fa11c2285..e055776f2f4c 100644
--- a/include/uapi/drm/amdgpu_drm.h
+++ b/include/uapi/drm/amdgpu_drm.h
@@ -87,6 +87,8 @@ extern "C" {
 #define AMDGPU_GEM_CREATE_SHADOW		(1 << 4)
 /* Flag that allocating the BO should use linear VRAM */
 #define AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS	(1 << 5)
+/* Flag that BO is always valid in this VM */
+#define AMDGPU_GEM_CREATE_VM_ALWAYS_VALID	(1 << 6)
 
 struct drm_amdgpu_gem_create_in  {
 	/** the requested memory size */
-- 
cgit v1.2.3


From b72cf4fca2bb786e20864b5e8755105aa9626fb4 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Sun, 3 Sep 2017 15:22:06 +0200
Subject: drm/amdgpu: move taking mmap_sem into get_user_pages v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This didn't helped as intended, just simplify the code.

v2: unlock mmap_sem in the error path as well

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  | 12 +-----------
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c |  4 ----
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c |  8 +++++++-
 3 files changed, 8 insertions(+), 16 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 3fe816f6beca..283a216ee758 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -500,18 +500,14 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
 	struct amdgpu_bo_list_entry *e;
 	struct list_head duplicates;
-	bool need_mmap_lock = false;
 	unsigned i, tries = 10;
 	int r;
 
 	INIT_LIST_HEAD(&p->validated);
 
 	p->bo_list = amdgpu_bo_list_get(fpriv, cs->in.bo_list_handle);
-	if (p->bo_list) {
-		need_mmap_lock = p->bo_list->first_userptr !=
-			p->bo_list->num_entries;
+	if (p->bo_list)
 		amdgpu_bo_list_get_list(p->bo_list, &p->validated);
-	}
 
 	INIT_LIST_HEAD(&duplicates);
 	amdgpu_vm_get_pd_bo(&fpriv->vm, &p->validated, &p->vm_pd);
@@ -519,9 +515,6 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 	if (p->uf_entry.robj)
 		list_add(&p->uf_entry.tv.head, &p->validated);
 
-	if (need_mmap_lock)
-		down_read(&current->mm->mmap_sem);
-
 	while (1) {
 		struct list_head need_pages;
 		unsigned i;
@@ -674,9 +667,6 @@ error_validate:
 
 error_free_pages:
 
-	if (need_mmap_lock)
-		up_read(&current->mm->mmap_sem);
-
 	if (p->bo_list) {
 		for (i = p->bo_list->first_userptr;
 		     i < p->bo_list->num_entries; ++i) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index f1e61b3df640..b0d45c8e6bb3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -318,8 +318,6 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 	}
 
 	if (args->flags & AMDGPU_GEM_USERPTR_VALIDATE) {
-		down_read(&current->mm->mmap_sem);
-
 		r = amdgpu_ttm_tt_get_user_pages(bo->tbo.ttm,
 						 bo->tbo.ttm->pages);
 		if (r)
@@ -334,8 +332,6 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 		amdgpu_bo_unreserve(bo);
 		if (r)
 			goto free_pages;
-
-		up_read(&current->mm->mmap_sem);
 	}
 
 	r = drm_gem_handle_create(filp, gobj, &handle);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 28e121984332..ea0378c8b049 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -622,6 +622,8 @@ int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages)
 	if (!(gtt->userflags & AMDGPU_GEM_USERPTR_READONLY))
 		flags |= FOLL_WRITE;
 
+	down_read(&current->mm->mmap_sem);
+
 	if (gtt->userflags & AMDGPU_GEM_USERPTR_ANONONLY) {
 		/* check that we only use anonymous memory
 		   to prevent problems with writeback */
@@ -629,8 +631,10 @@ int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages)
 		struct vm_area_struct *vma;
 
 		vma = find_vma(gtt->usermm, gtt->userptr);
-		if (!vma || vma->vm_file || vma->vm_end < end)
+		if (!vma || vma->vm_file || vma->vm_end < end) {
+			up_read(&current->mm->mmap_sem);
 			return -EPERM;
+		}
 	}
 
 	do {
@@ -657,10 +661,12 @@ int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages)
 
 	} while (pinned < ttm->num_pages);
 
+	up_read(&current->mm->mmap_sem);
 	return 0;
 
 release_pages:
 	release_pages(pages, pinned, 0);
+	up_read(&current->mm->mmap_sem);
 	return r;
 }
 
-- 
cgit v1.2.3


From 177ae09b5d699a5ebd1cafcee78889db968abf54 Mon Sep 17 00:00:00 2001
From: Andres Rodriguez <andresx7@gmail.com>
Date: Fri, 15 Sep 2017 20:44:06 -0400
Subject: drm/amdgpu: introduce AMDGPU_GEM_CREATE_EXPLICIT_SYNC v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Introduce a flag to signal that access to a BO will be synchronized
through an external mechanism.

Currently all buffers shared between contexts are subject to implicit
synchronization. However, this is only required for protocols that
currently don't support an explicit synchronization mechanism (DRI2/3).

This patch introduces the AMDGPU_GEM_CREATE_EXPLICIT_SYNC, so that
users can specify when it is safe to disable implicit sync.

v2: only disable explicit sync in amdgpu_cs_ioctl

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Andres Rodriguez <andresx7@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c     | 3 ++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c    | 4 +++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.h | 8 ++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c   | 7 +++++--
 drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h   | 3 ++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c    | 5 +++--
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c     | 8 ++++----
 include/uapi/drm/amdgpu_drm.h              | 2 ++
 8 files changed, 29 insertions(+), 11 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index ab83dfcabb41..38027a00f8ab 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -705,7 +705,8 @@ static int amdgpu_cs_sync_rings(struct amdgpu_cs_parser *p)
 
 	list_for_each_entry(e, &p->validated, tv.head) {
 		struct reservation_object *resv = e->robj->tbo.resv;
-		r = amdgpu_sync_resv(p->adev, &p->job->sync, resv, p->filp);
+		r = amdgpu_sync_resv(p->adev, &p->job->sync, resv, p->filp,
+				     amdgpu_bo_explicit_sync(e->robj));
 
 		if (r)
 			return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index b0d45c8e6bb3..21e99366cab3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -212,7 +212,9 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 		      AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
 		      AMDGPU_GEM_CREATE_CPU_GTT_USWC |
 		      AMDGPU_GEM_CREATE_VRAM_CLEARED |
-		      AMDGPU_GEM_CREATE_VM_ALWAYS_VALID))
+		      AMDGPU_GEM_CREATE_VM_ALWAYS_VALID |
+		      AMDGPU_GEM_CREATE_EXPLICIT_SYNC))
+
 		return -EINVAL;
 
 	/* reject invalid gem domains */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
index c26ef53604af..428aae048f4b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
@@ -193,6 +193,14 @@ static inline bool amdgpu_bo_gpu_accessible(struct amdgpu_bo *bo)
 	}
 }
 
+/**
+ * amdgpu_bo_explicit_sync - return whether the bo is explicitly synced
+ */
+static inline bool amdgpu_bo_explicit_sync(struct amdgpu_bo *bo)
+{
+	return bo->flags & AMDGPU_GEM_CREATE_EXPLICIT_SYNC;
+}
+
 int amdgpu_bo_create(struct amdgpu_device *adev,
 			    unsigned long size, int byte_align,
 			    bool kernel, u32 domain, u64 flags,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
index c586f44312f9..a4bf21f8f1c1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
@@ -169,14 +169,14 @@ int amdgpu_sync_fence(struct amdgpu_device *adev, struct amdgpu_sync *sync,
  *
  * @sync: sync object to add fences from reservation object to
  * @resv: reservation object with embedded fence
- * @shared: true if we should only sync to the exclusive fence
+ * @explicit_sync: true if we should only sync to the exclusive fence
  *
  * Sync to the fence
  */
 int amdgpu_sync_resv(struct amdgpu_device *adev,
 		     struct amdgpu_sync *sync,
 		     struct reservation_object *resv,
-		     void *owner)
+		     void *owner, bool explicit_sync)
 {
 	struct reservation_object_list *flist;
 	struct dma_fence *f;
@@ -191,6 +191,9 @@ int amdgpu_sync_resv(struct amdgpu_device *adev,
 	f = reservation_object_get_excl(resv);
 	r = amdgpu_sync_fence(adev, sync, f);
 
+	if (explicit_sync)
+		return r;
+
 	flist = reservation_object_get_list(resv);
 	if (!flist || r)
 		return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
index dc7687993317..70d7e3a279a0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
@@ -45,7 +45,8 @@ int amdgpu_sync_fence(struct amdgpu_device *adev, struct amdgpu_sync *sync,
 int amdgpu_sync_resv(struct amdgpu_device *adev,
 		     struct amdgpu_sync *sync,
 		     struct reservation_object *resv,
-		     void *owner);
+		     void *owner,
+		     bool explicit_sync);
 struct dma_fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync,
 				     struct amdgpu_ring *ring);
 struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 10952c3e5eb6..a2282bacf960 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1489,7 +1489,8 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring, uint64_t src_offset,
 	job->vm_needs_flush = vm_needs_flush;
 	if (resv) {
 		r = amdgpu_sync_resv(adev, &job->sync, resv,
-				     AMDGPU_FENCE_OWNER_UNDEFINED);
+				     AMDGPU_FENCE_OWNER_UNDEFINED,
+				     false);
 		if (r) {
 			DRM_ERROR("sync failed (%d).\n", r);
 			goto error_free;
@@ -1581,7 +1582,7 @@ int amdgpu_fill_buffer(struct amdgpu_bo *bo,
 
 	if (resv) {
 		r = amdgpu_sync_resv(adev, &job->sync, resv,
-				     AMDGPU_FENCE_OWNER_UNDEFINED);
+				     AMDGPU_FENCE_OWNER_UNDEFINED, false);
 		if (r) {
 			DRM_ERROR("sync failed (%d).\n", r);
 			goto error_free;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index eb4a01c14eee..c559d76ff695 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -1035,7 +1035,7 @@ static int amdgpu_vm_wait_pd(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	int r;
 
 	amdgpu_sync_create(&sync);
-	amdgpu_sync_resv(adev, &sync, vm->root.base.bo->tbo.resv, owner);
+	amdgpu_sync_resv(adev, &sync, vm->root.base.bo->tbo.resv, owner, false);
 	r = amdgpu_sync_wait(&sync, true);
 	amdgpu_sync_free(&sync);
 
@@ -1176,11 +1176,11 @@ static int amdgpu_vm_update_level(struct amdgpu_device *adev,
 			amdgpu_ring_pad_ib(ring, params.ib);
 			amdgpu_sync_resv(adev, &job->sync,
 					 parent->base.bo->tbo.resv,
-					 AMDGPU_FENCE_OWNER_VM);
+					 AMDGPU_FENCE_OWNER_VM, false);
 			if (shadow)
 				amdgpu_sync_resv(adev, &job->sync,
 						 shadow->tbo.resv,
-						 AMDGPU_FENCE_OWNER_VM);
+						 AMDGPU_FENCE_OWNER_VM, false);
 
 			WARN_ON(params.ib->length_dw > ndw);
 			r = amdgpu_job_submit(job, ring, &vm->entity,
@@ -1644,7 +1644,7 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 		goto error_free;
 
 	r = amdgpu_sync_resv(adev, &job->sync, vm->root.base.bo->tbo.resv,
-			     owner);
+			     owner, false);
 	if (r)
 		goto error_free;
 
diff --git a/include/uapi/drm/amdgpu_drm.h b/include/uapi/drm/amdgpu_drm.h
index 4c6e8c482ee4..b62484af8ccb 100644
--- a/include/uapi/drm/amdgpu_drm.h
+++ b/include/uapi/drm/amdgpu_drm.h
@@ -91,6 +91,8 @@ extern "C" {
 #define AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS	(1 << 5)
 /* Flag that BO is always valid in this VM */
 #define AMDGPU_GEM_CREATE_VM_ALWAYS_VALID	(1 << 6)
+/* Flag that BO sharing will be explicitly synchronized */
+#define AMDGPU_GEM_CREATE_EXPLICIT_SYNC		(1 << 7)
 
 struct drm_amdgpu_gem_create_in  {
 	/** the requested memory size */
-- 
cgit v1.2.3


From 396bcb41e035df7b98fb150ca950bf213e70ae7b Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Mon, 9 Oct 2017 14:45:09 +0200
Subject: drm/amdgpu: partial revert VRAM lost handling v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Keep blocking the CS, but revert everything else. Mapping BOs and info IOCTL
are harmless and can still happen even when VRAM content ist lost.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  | 11 -----------
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c |  5 -----
 drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 14 ++++++++++----
 3 files changed, 10 insertions(+), 20 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 5de092eab0fa..0c07df72743c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -1272,16 +1272,12 @@ int amdgpu_cs_wait_ioctl(struct drm_device *dev, void *data,
 {
 	union drm_amdgpu_wait_cs *wait = data;
 	struct amdgpu_device *adev = dev->dev_private;
-	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	unsigned long timeout = amdgpu_gem_timeout(wait->in.timeout);
 	struct amdgpu_ring *ring = NULL;
 	struct amdgpu_ctx *ctx;
 	struct dma_fence *fence;
 	long r;
 
-	if (amdgpu_kms_vram_lost(adev, fpriv))
-		return -ENODEV;
-
 	ctx = amdgpu_ctx_get(filp->driver_priv, wait->in.ctx_id);
 	if (ctx == NULL)
 		return -EINVAL;
@@ -1350,16 +1346,12 @@ int amdgpu_cs_fence_to_handle_ioctl(struct drm_device *dev, void *data,
 				    struct drm_file *filp)
 {
 	struct amdgpu_device *adev = dev->dev_private;
-	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	union drm_amdgpu_fence_to_handle *info = data;
 	struct dma_fence *fence;
 	struct drm_syncobj *syncobj;
 	struct sync_file *sync_file;
 	int fd, r;
 
-	if (amdgpu_kms_vram_lost(adev, fpriv))
-		return -ENODEV;
-
 	fence = amdgpu_cs_get_fence(adev, filp, &info->in.fence);
 	if (IS_ERR(fence))
 		return PTR_ERR(fence);
@@ -1521,15 +1513,12 @@ int amdgpu_cs_wait_fences_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *filp)
 {
 	struct amdgpu_device *adev = dev->dev_private;
-	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	union drm_amdgpu_wait_fences *wait = data;
 	uint32_t fence_count = wait->in.fence_count;
 	struct drm_amdgpu_fence *fences_user;
 	struct drm_amdgpu_fence *fences;
 	int r;
 
-	if (amdgpu_kms_vram_lost(adev, fpriv))
-		return -ENODEV;
 	/* Get the fences from userspace */
 	fences = kmalloc_array(fence_count, sizeof(struct drm_amdgpu_fence),
 			GFP_KERNEL);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 21e99366cab3..fb72edc4c026 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -579,11 +579,6 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 			args->operation);
 		return -EINVAL;
 	}
-	if ((args->operation == AMDGPU_VA_OP_MAP) ||
-	    (args->operation == AMDGPU_VA_OP_REPLACE)) {
-		if (amdgpu_kms_vram_lost(adev, fpriv))
-			return -ENODEV;
-	}
 
 	INIT_LIST_HEAD(&list);
 	INIT_LIST_HEAD(&duplicates);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index 82e8d43b235a..f759836d10ef 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -270,7 +270,6 @@ static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info,
 static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 {
 	struct amdgpu_device *adev = dev->dev_private;
-	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	struct drm_amdgpu_info *info = data;
 	struct amdgpu_mode_info *minfo = &adev->mode_info;
 	void __user *out = (void __user *)(uintptr_t)info->return_pointer;
@@ -283,8 +282,6 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 
 	if (!info->return_size || !info->return_pointer)
 		return -EINVAL;
-	if (amdgpu_kms_vram_lost(adev, fpriv))
-		return -ENODEV;
 
 	switch (info->query) {
 	case AMDGPU_INFO_ACCEL_WORKING:
@@ -792,10 +789,19 @@ void amdgpu_driver_lastclose_kms(struct drm_device *dev)
 	vga_switcheroo_process_delayed_switch();
 }
 
+/**
+ * amdgpu_kms_vram_lost - check if VRAM was lost for this client
+ *
+ * @adev: amdgpu device
+ * @fpriv: client private
+ *
+ * Check if all CS is blocked for the client because of lost VRAM
+ */
 bool amdgpu_kms_vram_lost(struct amdgpu_device *adev,
 			  struct amdgpu_fpriv *fpriv)
 {
-	return fpriv->vram_lost_counter != atomic_read(&adev->vram_lost_counter);
+	return fpriv->vram_lost_counter !=
+		atomic_read(&adev->vram_lost_counter);
 }
 
 /**
-- 
cgit v1.2.3


From ff4cd38943199348a53f980709e16f0bc5c0b8c9 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Mon, 6 Nov 2017 15:25:37 +0100
Subject: drm/amdgpu: make AMDGPU_VA_RESERVED_SIZE 64bit
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Even when it's a small handle it as 64bit value as well.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 5 ++---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h  | 3 ++-
 2 files changed, 4 insertions(+), 4 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index fb72edc4c026..4710e51099c2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -556,9 +556,8 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 
 	if (args->va_address < AMDGPU_VA_RESERVED_SIZE) {
 		dev_err(&dev->pdev->dev,
-			"va_address 0x%lX is in reserved area 0x%X\n",
-			(unsigned long)args->va_address,
-			AMDGPU_VA_RESERVED_SIZE);
+			"va_address 0x%LX is in reserved area 0x%LX\n",
+			args->va_address, AMDGPU_VA_RESERVED_SIZE);
 		return -EINVAL;
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index aa914256b4bc..bae77353447b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -94,7 +94,8 @@ struct amdgpu_bo_list_entry;
 #define AMDGPU_MMHUB				1
 
 /* hardcode that limit for now */
-#define AMDGPU_VA_RESERVED_SIZE			(8 << 20)
+#define AMDGPU_VA_RESERVED_SIZE			(8ULL << 20)
+
 /* max vmids dedicated for process */
 #define AMDGPU_VM_MAX_RESERVED_VMID	1
 
-- 
cgit v1.2.3


From 8e96e372bceceb76b92a81f8e28bc8163287d25c Mon Sep 17 00:00:00 2001
From: Roger He <Hongbo.He@amd.com>
Date: Fri, 10 Nov 2017 20:00:30 +0800
Subject: drm/amd/amdgpu: if visible VRAM allocation fail, fall back to
 invisible try again
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Roger He <Hongbo.He@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 5 +++++
 1 file changed, 5 insertions(+)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 4710e51099c2..951d625bbdd7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -63,6 +63,11 @@ retry:
 			     flags, NULL, resv, 0, &bo);
 	if (r) {
 		if (r != -ERESTARTSYS) {
+			if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) {
+				flags &= ~AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
+				goto retry;
+			}
+
 			if (initial_domain == AMDGPU_GEM_DOMAIN_VRAM) {
 				initial_domain |= AMDGPU_GEM_DOMAIN_GTT;
 				goto retry;
-- 
cgit v1.2.3


From c6f92f9fbe7dbcc8903a67229aa88b4077ae4422 Mon Sep 17 00:00:00 2001
From: Mel Gorman <mgorman@techsingularity.net>
Date: Wed, 15 Nov 2017 17:37:55 -0800
Subject: mm: remove cold parameter for release_pages

All callers of release_pages claim the pages being released are cache
hot.  As no one cares about the hotness of pages being released to the
allocator, just ditch the parameter.

No performance impact is expected as the overhead is marginal.  The
parameter is removed simply because it is a bit stupid to have a useless
parameter copied everywhere.

Link: http://lkml.kernel.org/r/20171018075952.10627-7-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  | 6 ++----
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 2 +-
 drivers/gpu/drm/etnaviv/etnaviv_gem.c   | 6 +++---
 drivers/gpu/drm/i915/i915_gem_userptr.c | 4 ++--
 drivers/gpu/drm/radeon/radeon_ttm.c     | 2 +-
 fs/fuse/dev.c                           | 2 +-
 include/linux/pagemap.h                 | 2 +-
 include/linux/swap.h                    | 2 +-
 mm/swap.c                               | 8 ++++----
 mm/swap_state.c                         | 2 +-
 11 files changed, 18 insertions(+), 20 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 60d8bedb694d..cd664832f9e8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -553,8 +553,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 				 * invalidated it. Free it and try again
 				 */
 				release_pages(e->user_pages,
-					      e->robj->tbo.ttm->num_pages,
-					      false);
+					      e->robj->tbo.ttm->num_pages);
 				kvfree(e->user_pages);
 				e->user_pages = NULL;
 			}
@@ -691,8 +690,7 @@ error_free_pages:
 				continue;
 
 			release_pages(e->user_pages,
-				      e->robj->tbo.ttm->num_pages,
-				      false);
+				      e->robj->tbo.ttm->num_pages);
 			kvfree(e->user_pages);
 		}
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 6149a47fe63d..0bda8f2a188a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -347,7 +347,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 	return 0;
 
 free_pages:
-	release_pages(bo->tbo.ttm->pages, bo->tbo.ttm->num_pages, false);
+	release_pages(bo->tbo.ttm->pages, bo->tbo.ttm->num_pages);
 
 unlock_mmap_sem:
 	up_read(&current->mm->mmap_sem);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index bc746131987f..d792959fac43 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -659,7 +659,7 @@ int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages)
 	return 0;
 
 release_pages:
-	release_pages(pages, pinned, 0);
+	release_pages(pages, pinned);
 	return r;
 }
 
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gem.c b/drivers/gpu/drm/etnaviv/etnaviv_gem.c
index 57881167ccd2..bcc8c2d7c7c9 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_gem.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gem.c
@@ -779,7 +779,7 @@ static struct page **etnaviv_gem_userptr_do_get_pages(
 	up_read(&mm->mmap_sem);
 
 	if (ret < 0) {
-		release_pages(pvec, pinned, 0);
+		release_pages(pvec, pinned);
 		kvfree(pvec);
 		return ERR_PTR(ret);
 	}
@@ -852,7 +852,7 @@ static int etnaviv_gem_userptr_get_pages(struct etnaviv_gem_object *etnaviv_obj)
 		}
 	}
 
-	release_pages(pvec, pinned, 0);
+	release_pages(pvec, pinned);
 	kvfree(pvec);
 
 	work = kmalloc(sizeof(*work), GFP_KERNEL);
@@ -886,7 +886,7 @@ static void etnaviv_gem_userptr_release(struct etnaviv_gem_object *etnaviv_obj)
 	if (etnaviv_obj->pages) {
 		int npages = etnaviv_obj->base.size >> PAGE_SHIFT;
 
-		release_pages(etnaviv_obj->pages, npages, 0);
+		release_pages(etnaviv_obj->pages, npages);
 		kvfree(etnaviv_obj->pages);
 	}
 	put_task_struct(etnaviv_obj->userptr.task);
diff --git a/drivers/gpu/drm/i915/i915_gem_userptr.c b/drivers/gpu/drm/i915/i915_gem_userptr.c
index 709efe2357ea..aa22361bd5a1 100644
--- a/drivers/gpu/drm/i915/i915_gem_userptr.c
+++ b/drivers/gpu/drm/i915/i915_gem_userptr.c
@@ -554,7 +554,7 @@ __i915_gem_userptr_get_pages_worker(struct work_struct *_work)
 	}
 	mutex_unlock(&obj->mm.lock);
 
-	release_pages(pvec, pinned, 0);
+	release_pages(pvec, pinned);
 	kvfree(pvec);
 
 	i915_gem_object_put(obj);
@@ -668,7 +668,7 @@ i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj)
 		__i915_gem_userptr_set_active(obj, true);
 
 	if (IS_ERR(pages))
-		release_pages(pvec, pinned, 0);
+		release_pages(pvec, pinned);
 	kvfree(pvec);
 
 	return pages;
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index bf69bf9086bf..1fdfc7a46072 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -597,7 +597,7 @@ release_sg:
 	kfree(ttm->sg);
 
 release_pages:
-	release_pages(ttm->pages, pinned, 0);
+	release_pages(ttm->pages, pinned);
 	return r;
 }
 
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index a42d89371748..17f0d05bfd4c 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -1636,7 +1636,7 @@ out_finish:
 
 static void fuse_retrieve_end(struct fuse_conn *fc, struct fuse_req *req)
 {
-	release_pages(req->pages, req->num_pages, false);
+	release_pages(req->pages, req->num_pages);
 }
 
 static int fuse_retrieve(struct fuse_conn *fc, struct inode *inode,
diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
index e0f7181118fe..4c6790bb7afb 100644
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -118,7 +118,7 @@ static inline void mapping_set_gfp_mask(struct address_space *m, gfp_t mask)
 	m->gfp_mask = mask;
 }
 
-void release_pages(struct page **pages, int nr, bool cold);
+void release_pages(struct page **pages, int nr);
 
 /*
  * speculatively take a reference to a page.
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 454f042bcdd5..c2b8128799c1 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -510,7 +510,7 @@ static inline struct swap_info_struct *swp_swap_info(swp_entry_t entry)
 #define free_page_and_swap_cache(page) \
 	put_page(page)
 #define free_pages_and_swap_cache(pages, nr) \
-	release_pages((pages), (nr), false);
+	release_pages((pages), (nr));
 
 static inline void show_swap_cache_info(void)
 {
diff --git a/mm/swap.c b/mm/swap.c
index 88a19b6cdf7c..29cf75f1a860 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -210,7 +210,7 @@ static void pagevec_lru_move_fn(struct pagevec *pvec,
 	}
 	if (pgdat)
 		spin_unlock_irqrestore(&pgdat->lru_lock, flags);
-	release_pages(pvec->pages, pvec->nr, 0);
+	release_pages(pvec->pages, pvec->nr);
 	pagevec_reinit(pvec);
 }
 
@@ -740,7 +740,7 @@ void lru_add_drain_all(void)
  * Decrement the reference count on all the pages in @pages.  If it
  * fell to zero, remove the page from the LRU and free it.
  */
-void release_pages(struct page **pages, int nr, bool cold)
+void release_pages(struct page **pages, int nr)
 {
 	int i;
 	LIST_HEAD(pages_to_free);
@@ -817,7 +817,7 @@ void release_pages(struct page **pages, int nr, bool cold)
 		spin_unlock_irqrestore(&locked_pgdat->lru_lock, flags);
 
 	mem_cgroup_uncharge_list(&pages_to_free);
-	free_hot_cold_page_list(&pages_to_free, cold);
+	free_hot_cold_page_list(&pages_to_free, 0);
 }
 EXPORT_SYMBOL(release_pages);
 
@@ -837,7 +837,7 @@ void __pagevec_release(struct pagevec *pvec)
 		lru_add_drain();
 		pvec->drained = true;
 	}
-	release_pages(pvec->pages, pagevec_count(pvec), 0);
+	release_pages(pvec->pages, pagevec_count(pvec));
 	pagevec_reinit(pvec);
 }
 EXPORT_SYMBOL(__pagevec_release);
diff --git a/mm/swap_state.c b/mm/swap_state.c
index 374d446f7a0a..39ae7cfad90f 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -319,7 +319,7 @@ void free_pages_and_swap_cache(struct page **pages, int nr)
 	lru_add_drain();
 	for (i = 0; i < nr; i++)
 		free_swap_cache(pagep[i]);
-	release_pages(pagep, nr, false);
+	release_pages(pagep, nr);
 }
 
 /*
-- 
cgit v1.2.3


From d5a480b44b189bad1d67df5d4c5ed9e6c750e8f2 Mon Sep 17 00:00:00 2001
From: "Xiangliang.Yu" <Xiangliang.Yu@amd.com>
Date: Fri, 20 Oct 2017 17:21:40 +0800
Subject: drm/amdgpu: fix kernel hang when starting VNC server
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

After starting VNC server or running CTS test, kernel will hang and
can see below call trace:

[961816] INFO: task khugepaged:42 blocked for more than 120 seconds.
[968581]       Tainted: G           OE   4.13.0 #1
[973495] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables
          this message.
[980962] khugepaged      D    0    42      2 0x00000000
[980967] Call Trace:
[980977]  __schedule+0x28d/0x890
[980982]  schedule+0x36/0x80
[980986]  rwsem_down_read_failed+0x139/0x1c0
[980991]  ? update_curr+0x100/0x1c0
[981004]  call_rwsem_down_read_failed+0x18/0x30
[981007]  down_read+0x20/0x40
[981012]  khugepaged_scan_mm_slot+0x78/0x1ac0
[981018]  ? __switch_to+0x23e/0x4a0
[981022]  ? finish_task_switch+0x79/0x240
[981026]  khugepaged+0x146/0x480
[981031]  ? remove_wait_queue+0x60/0x60
[981035]  kthread+0x109/0x140
[981037]  ? khugepaged_scan_mm_slot+0x1ac0/0x1ac0
[981039]  ? kthread_park+0x60/0x60
[981044]  ret_from_fork+0x25/0x30

After checking code and found 'commit b72cf4fca2bb7 ("drm/amdgpu: move
taking mmap_sem into get_user_pages v2")' forget to drop one case of
up_read.

Signed-off-by: Xiangliang.Yu <Xiangliang.Yu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

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

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 951d625bbdd7..14aff2f15a94 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -328,7 +328,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 		r = amdgpu_ttm_tt_get_user_pages(bo->tbo.ttm,
 						 bo->tbo.ttm->pages);
 		if (r)
-			goto unlock_mmap_sem;
+			goto release_object;
 
 		r = amdgpu_bo_reserve(bo, true);
 		if (r)
@@ -353,9 +353,6 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 free_pages:
 	release_pages(bo->tbo.ttm->pages, bo->tbo.ttm->num_pages, false);
 
-unlock_mmap_sem:
-	up_read(&current->mm->mmap_sem);
-
 release_object:
 	drm_gem_object_put_unlocked(gobj);
 
-- 
cgit v1.2.3