23 files changed, 284 insertions, 211 deletions

diff --git a/drivers/gpu/drm/qxl/qxl_release.c b/drivers/gpu/drm/qxl/qxl_release.c
index a6e19c83143e..446e71ca36cb 100644
--- a/drivers/gpu/drm/qxl/qxl_release.c
+++ b/drivers/gpu/drm/qxl/qxl_release.c
@@ -226,6 +226,7 @@ int qxl_release_list_add(struct qxl_release *release, struct qxl_bo *bo)
 
         qxl_bo_ref(bo);
         entry->tv.bo = &bo->tbo;
+        entry->tv.shared = false;
         list_add_tail(&entry->tv.head, &release->bos);
         return 0;
 }
diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
index 1f598ab3b9a7..0b5a230d8b96 100644
--- a/drivers/gpu/drm/radeon/cik.c
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -3959,18 +3959,19 @@ bool cik_semaphore_ring_emit(struct radeon_device *rdev,
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the CP DMA engine (CIK+).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int cik_copy_cpdma(struct radeon_device *rdev,
-                   uint64_t src_offset, uint64_t dst_offset,
-                   unsigned num_gpu_pages,
-                   struct radeon_fence **fence)
+struct radeon_fence *cik_copy_cpdma(struct radeon_device *rdev,
+                                    uint64_t src_offset, uint64_t dst_offset,
+                                    unsigned num_gpu_pages,
+                                    struct reservation_object *resv)
 {
         struct radeon_semaphore *sem = NULL;
+        struct radeon_fence *fence;
         int ring_index = rdev->asic->copy.blit_ring_index;
         struct radeon_ring *ring = &rdev->ring[ring_index];
         u32 size_in_bytes, cur_size_in_bytes, control;
@@ -3980,7 +3981,7 @@ int cik_copy_cpdma(struct radeon_device *rdev,
         r = radeon_semaphore_create(rdev, &sem);
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
-                return r;
+                return ERR_PTR(r);
         }
 
         size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
@@ -3989,10 +3990,10 @@ int cik_copy_cpdma(struct radeon_device *rdev,
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
-        radeon_semaphore_sync_to(sem, *fence);
+        radeon_semaphore_sync_resv(sem, resv, false);
         radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
         for (i = 0; i < num_loops; i++) {
@@ -4014,17 +4015,17 @@ int cik_copy_cpdma(struct radeon_device *rdev,
                 dst_offset += cur_size_in_bytes;
         }
 
-        r = radeon_fence_emit(rdev, fence, ring->idx);
+        r = radeon_fence_emit(rdev, &fence, ring->idx);
         if (r) {
                 radeon_ring_unlock_undo(rdev, ring);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
         radeon_ring_unlock_commit(rdev, ring, false);
-        radeon_semaphore_free(rdev, &sem, *fence);
+        radeon_semaphore_free(rdev, &sem, fence);
 
-        return r;
+        return fence;
 }
 
 /*
diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c
index 192278bc993c..c01a6100c318 100644
--- a/drivers/gpu/drm/radeon/cik_sdma.c
+++ b/drivers/gpu/drm/radeon/cik_sdma.c
@@ -537,18 +537,19 @@ void cik_sdma_fini(struct radeon_device *rdev)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (CIK).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int cik_copy_dma(struct radeon_device *rdev,
-                 uint64_t src_offset, uint64_t dst_offset,
-                 unsigned num_gpu_pages,
-                 struct radeon_fence **fence)
+struct radeon_fence *cik_copy_dma(struct radeon_device *rdev,
+                                  uint64_t src_offset, uint64_t dst_offset,
+                                  unsigned num_gpu_pages,
+                                  struct reservation_object *resv)
 {
         struct radeon_semaphore *sem = NULL;
+        struct radeon_fence *fence;
         int ring_index = rdev->asic->copy.dma_ring_index;
         struct radeon_ring *ring = &rdev->ring[ring_index];
         u32 size_in_bytes, cur_size_in_bytes;
@@ -558,7 +559,7 @@ int cik_copy_dma(struct radeon_device *rdev,
         r = radeon_semaphore_create(rdev, &sem);
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
-                return r;
+                return ERR_PTR(r);
         }
 
         size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
@@ -567,10 +568,10 @@ int cik_copy_dma(struct radeon_device *rdev,
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
-        radeon_semaphore_sync_to(sem, *fence);
+        radeon_semaphore_sync_resv(sem, resv, false);
         radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
         for (i = 0; i < num_loops; i++) {
@@ -589,17 +590,17 @@ int cik_copy_dma(struct radeon_device *rdev,
                 dst_offset += cur_size_in_bytes;
         }
 
-        r = radeon_fence_emit(rdev, fence, ring->idx);
+        r = radeon_fence_emit(rdev, &fence, ring->idx);
         if (r) {
                 radeon_ring_unlock_undo(rdev, ring);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
         radeon_ring_unlock_commit(rdev, ring, false);
-        radeon_semaphore_free(rdev, &sem, *fence);
+        radeon_semaphore_free(rdev, &sem, fence);
 
-        return r;
+        return fence;
 }
 
 /**
diff --git a/drivers/gpu/drm/radeon/evergreen_dma.c b/drivers/gpu/drm/radeon/evergreen_dma.c
index afaba388c36d..946f37d0b469 100644
--- a/drivers/gpu/drm/radeon/evergreen_dma.c
+++ b/drivers/gpu/drm/radeon/evergreen_dma.c
@@ -104,12 +104,14 @@ void evergreen_dma_ring_ib_execute(struct radeon_device *rdev,
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int evergreen_copy_dma(struct radeon_device *rdev,
-                       uint64_t src_offset, uint64_t dst_offset,
-                       unsigned num_gpu_pages,
-                       struct radeon_fence **fence)
+struct radeon_fence *evergreen_copy_dma(struct radeon_device *rdev,
+                                        uint64_t src_offset,
+                                        uint64_t dst_offset,
+                                        unsigned num_gpu_pages,
+                                        struct reservation_object *resv)
 {
         struct radeon_semaphore *sem = NULL;
+        struct radeon_fence *fence;
         int ring_index = rdev->asic->copy.dma_ring_index;
         struct radeon_ring *ring = &rdev->ring[ring_index];
         u32 size_in_dw, cur_size_in_dw;
@@ -119,7 +121,7 @@ int evergreen_copy_dma(struct radeon_device *rdev,
         r = radeon_semaphore_create(rdev, &sem);
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
-                return r;
+                return ERR_PTR(r);
         }
 
         size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4;
@@ -128,10 +130,10 @@ int evergreen_copy_dma(struct radeon_device *rdev,
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
-        radeon_semaphore_sync_to(sem, *fence);
+        radeon_semaphore_sync_resv(sem, resv, false);
         radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
         for (i = 0; i < num_loops; i++) {
@@ -148,17 +150,17 @@ int evergreen_copy_dma(struct radeon_device *rdev,
                 dst_offset += cur_size_in_dw * 4;
         }
 
-        r = radeon_fence_emit(rdev, fence, ring->idx);
+        r = radeon_fence_emit(rdev, &fence, ring->idx);
         if (r) {
                 radeon_ring_unlock_undo(rdev, ring);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
         radeon_ring_unlock_commit(rdev, ring, false);
-        radeon_semaphore_free(rdev, &sem, *fence);
+        radeon_semaphore_free(rdev, &sem, fence);
 
-        return r;
+        return fence;
 }
 
 /**
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 4c5ec44ff328..c6b486f888d5 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -855,13 +855,14 @@ bool r100_semaphore_ring_emit(struct radeon_device *rdev,
         return false;
 }
 
-int r100_copy_blit(struct radeon_device *rdev,
-                   uint64_t src_offset,
-                   uint64_t dst_offset,
-                   unsigned num_gpu_pages,
-                   struct radeon_fence **fence)
+struct radeon_fence *r100_copy_blit(struct radeon_device *rdev,
+                                    uint64_t src_offset,
+                                    uint64_t dst_offset,
+                                    unsigned num_gpu_pages,
+                                    struct reservation_object *resv)
 {
         struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+        struct radeon_fence *fence;
         uint32_t cur_pages;
         uint32_t stride_bytes = RADEON_GPU_PAGE_SIZE;
         uint32_t pitch;
@@ -882,7 +883,7 @@ int r100_copy_blit(struct radeon_device *rdev,
         r = radeon_ring_lock(rdev, ring, ndw);
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d) asking for %u dw.\n", r, ndw);
-                return -EINVAL;
+                return ERR_PTR(-EINVAL);
         }
         while (num_gpu_pages > 0) {
                 cur_pages = num_gpu_pages;
@@ -922,11 +923,13 @@ int r100_copy_blit(struct radeon_device *rdev,
                           RADEON_WAIT_2D_IDLECLEAN |
                           RADEON_WAIT_HOST_IDLECLEAN |
                           RADEON_WAIT_DMA_GUI_IDLE);
-        if (fence) {
-                r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX);
+        r = radeon_fence_emit(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX);
+        if (r) {
+                radeon_ring_unlock_undo(rdev, ring);
+                return ERR_PTR(r);
         }
         radeon_ring_unlock_commit(rdev, ring, false);
-        return r;
+        return fence;
 }
 
 static int r100_cp_wait_for_idle(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c
index 67780374a652..732d4938aab7 100644
--- a/drivers/gpu/drm/radeon/r200.c
+++ b/drivers/gpu/drm/radeon/r200.c
@@ -80,13 +80,14 @@ static int r200_get_vtx_size_0(uint32_t vtx_fmt_0)
         return vtx_size;
 }
 
-int r200_copy_dma(struct radeon_device *rdev,
-                  uint64_t src_offset,
-                  uint64_t dst_offset,
-                  unsigned num_gpu_pages,
-                  struct radeon_fence **fence)
+struct radeon_fence *r200_copy_dma(struct radeon_device *rdev,
+                                   uint64_t src_offset,
+                                   uint64_t dst_offset,
+                                   unsigned num_gpu_pages,
+                                   struct reservation_object *resv)
 {
         struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+        struct radeon_fence *fence;
         uint32_t size;
         uint32_t cur_size;
         int i, num_loops;
@@ -98,7 +99,7 @@ int r200_copy_dma(struct radeon_device *rdev,
         r = radeon_ring_lock(rdev, ring, num_loops * 4 + 64);
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
-                return r;
+                return ERR_PTR(r);
         }
         /* Must wait for 2D idle & clean before DMA or hangs might happen */
         radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
@@ -118,11 +119,13 @@ int r200_copy_dma(struct radeon_device *rdev,
         }
         radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
         radeon_ring_write(ring, RADEON_WAIT_DMA_GUI_IDLE);
-        if (fence) {
-                r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX);
+        r = radeon_fence_emit(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX);
+        if (r) {
+                radeon_ring_unlock_undo(rdev, ring);
+                return ERR_PTR(r);
         }
         radeon_ring_unlock_commit(rdev, ring, false);
-        return r;
+        return fence;
 }
 
 
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index a95ced569d84..5e9146b968b9 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -2894,12 +2894,13 @@ bool r600_semaphore_ring_emit(struct radeon_device *rdev,
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int r600_copy_cpdma(struct radeon_device *rdev,
-                    uint64_t src_offset, uint64_t dst_offset,
-                    unsigned num_gpu_pages,
-                    struct radeon_fence **fence)
+struct radeon_fence *r600_copy_cpdma(struct radeon_device *rdev,
+                                     uint64_t src_offset, uint64_t dst_offset,
+                                     unsigned num_gpu_pages,
+                                     struct reservation_object *resv)
 {
         struct radeon_semaphore *sem = NULL;
+        struct radeon_fence *fence;
         int ring_index = rdev->asic->copy.blit_ring_index;
         struct radeon_ring *ring = &rdev->ring[ring_index];
         u32 size_in_bytes, cur_size_in_bytes, tmp;
@@ -2909,7 +2910,7 @@ int r600_copy_cpdma(struct radeon_device *rdev,
         r = radeon_semaphore_create(rdev, &sem);
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
-                return r;
+                return ERR_PTR(r);
         }
 
         size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
@@ -2918,10 +2919,10 @@ int r600_copy_cpdma(struct radeon_device *rdev,
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
-        radeon_semaphore_sync_to(sem, *fence);
+        radeon_semaphore_sync_resv(sem, resv, false);
         radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
         radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
@@ -2948,17 +2949,17 @@ int r600_copy_cpdma(struct radeon_device *rdev,
         radeon_ring_write(ring, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
         radeon_ring_write(ring, WAIT_CP_DMA_IDLE_bit);
 
-        r = radeon_fence_emit(rdev, fence, ring->idx);
+        r = radeon_fence_emit(rdev, &fence, ring->idx);
         if (r) {
                 radeon_ring_unlock_undo(rdev, ring);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
         radeon_ring_unlock_commit(rdev, ring, false);
-        radeon_semaphore_free(rdev, &sem, *fence);
+        radeon_semaphore_free(rdev, &sem, fence);
 
-        return r;
+        return fence;
 }
 
 int r600_set_surface_reg(struct radeon_device *rdev, int reg,
diff --git a/drivers/gpu/drm/radeon/r600_dma.c b/drivers/gpu/drm/radeon/r600_dma.c
index 51fd98553eaf..fc54224ce87b 100644
--- a/drivers/gpu/drm/radeon/r600_dma.c
+++ b/drivers/gpu/drm/radeon/r600_dma.c
@@ -436,18 +436,19 @@ void r600_dma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (r6xx).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int r600_copy_dma(struct radeon_device *rdev,
-                  uint64_t src_offset, uint64_t dst_offset,
-                  unsigned num_gpu_pages,
-                  struct radeon_fence **fence)
+struct radeon_fence *r600_copy_dma(struct radeon_device *rdev,
+                                   uint64_t src_offset, uint64_t dst_offset,
+                                   unsigned num_gpu_pages,
+                                   struct reservation_object *resv)
 {
         struct radeon_semaphore *sem = NULL;
+        struct radeon_fence *fence;
         int ring_index = rdev->asic->copy.dma_ring_index;
         struct radeon_ring *ring = &rdev->ring[ring_index];
         u32 size_in_dw, cur_size_in_dw;
@@ -457,7 +458,7 @@ int r600_copy_dma(struct radeon_device *rdev,
         r = radeon_semaphore_create(rdev, &sem);
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
-                return r;
+                return ERR_PTR(r);
         }
 
         size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4;
@@ -466,10 +467,10 @@ int r600_copy_dma(struct radeon_device *rdev,
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
-        radeon_semaphore_sync_to(sem, *fence);
+        radeon_semaphore_sync_resv(sem, resv, false);
         radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
         for (i = 0; i < num_loops; i++) {
@@ -486,15 +487,15 @@ int r600_copy_dma(struct radeon_device *rdev,
                 dst_offset += cur_size_in_dw * 4;
         }
 
-        r = radeon_fence_emit(rdev, fence, ring->idx);
+        r = radeon_fence_emit(rdev, &fence, ring->idx);
         if (r) {
                 radeon_ring_unlock_undo(rdev, ring);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
         radeon_ring_unlock_commit(rdev, ring, false);
-        radeon_semaphore_free(rdev, &sem, *fence);
+        radeon_semaphore_free(rdev, &sem, fence);
 
-        return r;
+        return fence;
 }
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 79c988db79ad..2e41dc12cd00 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -585,8 +585,11 @@ bool radeon_semaphore_emit_signal(struct radeon_device *rdev, int ring,
                                   struct radeon_semaphore *semaphore);
 bool radeon_semaphore_emit_wait(struct radeon_device *rdev, int ring,
                                 struct radeon_semaphore *semaphore);
-void radeon_semaphore_sync_to(struct radeon_semaphore *semaphore,
-                              struct radeon_fence *fence);
+void radeon_semaphore_sync_fence(struct radeon_semaphore *semaphore,
+                                 struct radeon_fence *fence);
+void radeon_semaphore_sync_resv(struct radeon_semaphore *semaphore,
+                                struct reservation_object *resv,
+                                bool shared);
 int radeon_semaphore_sync_rings(struct radeon_device *rdev,
                                 struct radeon_semaphore *semaphore,
                                 int waiting_ring);
@@ -1855,24 +1858,24 @@ struct radeon_asic {
         } display;
         /* copy functions for bo handling */
         struct {
-                int (*blit)(struct radeon_device *rdev,
-                            uint64_t src_offset,
-                            uint64_t dst_offset,
-                            unsigned num_gpu_pages,
-                            struct radeon_fence **fence);
+                struct radeon_fence *(*blit)(struct radeon_device *rdev,
+                                             uint64_t src_offset,
+                                             uint64_t dst_offset,
+                                             unsigned num_gpu_pages,
+                                             struct reservation_object *resv);
                 u32 blit_ring_index;
-                int (*dma)(struct radeon_device *rdev,
-                           uint64_t src_offset,
-                           uint64_t dst_offset,
-                           unsigned num_gpu_pages,
-                           struct radeon_fence **fence);
+                struct radeon_fence *(*dma)(struct radeon_device *rdev,
+                                            uint64_t src_offset,
+                                            uint64_t dst_offset,
+                                            unsigned num_gpu_pages,
+                                            struct reservation_object *resv);
                 u32 dma_ring_index;
                 /* method used for bo copy */
-                int (*copy)(struct radeon_device *rdev,
-                            uint64_t src_offset,
-                            uint64_t dst_offset,
-                            unsigned num_gpu_pages,
-                            struct radeon_fence **fence);
+                struct radeon_fence *(*copy)(struct radeon_device *rdev,
+                                             uint64_t src_offset,
+                                             uint64_t dst_offset,
+                                             unsigned num_gpu_pages,
+                                             struct reservation_object *resv);
                 /* ring used for bo copies */
                 u32 copy_ring_index;
         } copy;
@@ -2833,9 +2836,9 @@ static inline void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
 #define radeon_hdmi_setmode(rdev, e, m) (rdev)->asic->display.hdmi_setmode((e), (m))
 #define radeon_fence_ring_emit(rdev, r, fence) (rdev)->asic->ring[(r)]->emit_fence((rdev), (fence))
 #define radeon_semaphore_ring_emit(rdev, r, cp, semaphore, emit_wait) (rdev)->asic->ring[(r)]->emit_semaphore((rdev), (cp), (semaphore), (emit_wait))
-#define radeon_copy_blit(rdev, s, d, np, f) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (f))
-#define radeon_copy_dma(rdev, s, d, np, f) (rdev)->asic->copy.dma((rdev), (s), (d), (np), (f))
-#define radeon_copy(rdev, s, d, np, f) (rdev)->asic->copy.copy((rdev), (s), (d), (np), (f))
+#define radeon_copy_blit(rdev, s, d, np, resv) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (resv))
+#define radeon_copy_dma(rdev, s, d, np, resv) (rdev)->asic->copy.dma((rdev), (s), (d), (np), (resv))
+#define radeon_copy(rdev, s, d, np, resv) (rdev)->asic->copy.copy((rdev), (s), (d), (np), (resv))
 #define radeon_copy_blit_ring_index(rdev) (rdev)->asic->copy.blit_ring_index
 #define radeon_copy_dma_ring_index(rdev) (rdev)->asic->copy.dma_ring_index
 #define radeon_copy_ring_index(rdev) (rdev)->asic->copy.copy_ring_index
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 987a3b713e06..ca01bb8ea217 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -81,11 +81,11 @@ bool r100_semaphore_ring_emit(struct radeon_device *rdev,
 int r100_cs_parse(struct radeon_cs_parser *p);
 void r100_pll_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 uint32_t r100_pll_rreg(struct radeon_device *rdev, uint32_t reg);
-int r100_copy_blit(struct radeon_device *rdev,
-                   uint64_t src_offset,
-                   uint64_t dst_offset,
-                   unsigned num_gpu_pages,
-                   struct radeon_fence **fence);
+struct radeon_fence *r100_copy_blit(struct radeon_device *rdev,
+                                    uint64_t src_offset,
+                                    uint64_t dst_offset,
+                                    unsigned num_gpu_pages,
+                                    struct reservation_object *resv);
 int r100_set_surface_reg(struct radeon_device *rdev, int reg,
                          uint32_t tiling_flags, uint32_t pitch,
                          uint32_t offset, uint32_t obj_size);
@@ -153,11 +153,11 @@ void r100_ring_hdp_flush(struct radeon_device *rdev,
 /*
  * r200,rv250,rs300,rv280
  */
-extern int r200_copy_dma(struct radeon_device *rdev,
-                         uint64_t src_offset,
-                         uint64_t dst_offset,
-                         unsigned num_gpu_pages,
-                         struct radeon_fence **fence);
+struct radeon_fence *r200_copy_dma(struct radeon_device *rdev,
+                                   uint64_t src_offset,
+                                   uint64_t dst_offset,
+                                   unsigned num_gpu_pages,
+                                   struct reservation_object *resv);
 void r200_set_safe_registers(struct radeon_device *rdev);
 
 /*
@@ -341,12 +341,14 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
 int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
 int r600_dma_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
-int r600_copy_cpdma(struct radeon_device *rdev,
-                    uint64_t src_offset, uint64_t dst_offset,
-                    unsigned num_gpu_pages, struct radeon_fence **fence);
-int r600_copy_dma(struct radeon_device *rdev,
-                  uint64_t src_offset, uint64_t dst_offset,
-                  unsigned num_gpu_pages, struct radeon_fence **fence);
+struct radeon_fence *r600_copy_cpdma(struct radeon_device *rdev,
+                                     uint64_t src_offset, uint64_t dst_offset,
+                                     unsigned num_gpu_pages,
+                                     struct reservation_object *resv);
+struct radeon_fence *r600_copy_dma(struct radeon_device *rdev,
+                                   uint64_t src_offset, uint64_t dst_offset,
+                                   unsigned num_gpu_pages,
+                                   struct reservation_object *resv);
 void r600_hpd_init(struct radeon_device *rdev);
 void r600_hpd_fini(struct radeon_device *rdev);
 bool r600_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd);
@@ -462,10 +464,10 @@ bool rv770_page_flip_pending(struct radeon_device *rdev, int crtc);
 void r700_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc);
 void r700_cp_stop(struct radeon_device *rdev);
 void r700_cp_fini(struct radeon_device *rdev);
-int rv770_copy_dma(struct radeon_device *rdev,
-                  uint64_t src_offset, uint64_t dst_offset,
-                  unsigned num_gpu_pages,
-                   struct radeon_fence **fence);
+struct radeon_fence *rv770_copy_dma(struct radeon_device *rdev,
+                                    uint64_t src_offset, uint64_t dst_offset,
+                                    unsigned num_gpu_pages,
+                                    struct reservation_object *resv);
 u32 rv770_get_xclk(struct radeon_device *rdev);
 int rv770_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
 int rv770_get_temp(struct radeon_device *rdev);
@@ -536,10 +538,10 @@ void evergreen_dma_fence_ring_emit(struct radeon_device *rdev,
                                    struct radeon_fence *fence);
 void evergreen_dma_ring_ib_execute(struct radeon_device *rdev,
                                    struct radeon_ib *ib);
-int evergreen_copy_dma(struct radeon_device *rdev,
-                       uint64_t src_offset, uint64_t dst_offset,
-                       unsigned num_gpu_pages,
-                       struct radeon_fence **fence);
+struct radeon_fence *evergreen_copy_dma(struct radeon_device *rdev,
+                                        uint64_t src_offset, uint64_t dst_offset,
+                                        unsigned num_gpu_pages,
+                                        struct reservation_object *resv);
 void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable);
 void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode);
 int evergreen_get_temp(struct radeon_device *rdev);
@@ -701,10 +703,10 @@ int si_vm_init(struct radeon_device *rdev);
 void si_vm_fini(struct radeon_device *rdev);
 void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
-int si_copy_dma(struct radeon_device *rdev,
-                uint64_t src_offset, uint64_t dst_offset,
-                unsigned num_gpu_pages,
-                struct radeon_fence **fence);
+struct radeon_fence *si_copy_dma(struct radeon_device *rdev,
+                                 uint64_t src_offset, uint64_t dst_offset,
+                                 unsigned num_gpu_pages,
+                                 struct reservation_object *resv);
 
 void si_dma_vm_copy_pages(struct radeon_device *rdev,
                           struct radeon_ib *ib,
@@ -760,14 +762,14 @@ bool cik_sdma_semaphore_ring_emit(struct radeon_device *rdev,
                                   struct radeon_semaphore *semaphore,
                                   bool emit_wait);
 void cik_sdma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
-int cik_copy_dma(struct radeon_device *rdev,
-                 uint64_t src_offset, uint64_t dst_offset,
-                 unsigned num_gpu_pages,
-                 struct radeon_fence **fence);
-int cik_copy_cpdma(struct radeon_device *rdev,
-                   uint64_t src_offset, uint64_t dst_offset,
-                   unsigned num_gpu_pages,
-                   struct radeon_fence **fence);
+struct radeon_fence *cik_copy_dma(struct radeon_device *rdev,
+                                  uint64_t src_offset, uint64_t dst_offset,
+                                  unsigned num_gpu_pages,
+                                  struct reservation_object *resv);
+struct radeon_fence *cik_copy_cpdma(struct radeon_device *rdev,
+                                    uint64_t src_offset, uint64_t dst_offset,
+                                    unsigned num_gpu_pages,
+                                    struct reservation_object *resv);
 int cik_sdma_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
 int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
diff --git a/drivers/gpu/drm/radeon/radeon_benchmark.c b/drivers/gpu/drm/radeon/radeon_benchmark.c
index 69f5695bdab9..1e8855060fc7 100644
--- a/drivers/gpu/drm/radeon/radeon_benchmark.c
+++ b/drivers/gpu/drm/radeon/radeon_benchmark.c
@@ -45,33 +45,29 @@ static int radeon_benchmark_do_move(struct radeon_device *rdev, unsigned size,
         for (i = 0; i < n; i++) {
                 switch (flag) {
                 case RADEON_BENCHMARK_COPY_DMA:
-                        r = radeon_copy_dma(rdev, saddr, daddr,
-                                            size / RADEON_GPU_PAGE_SIZE,
-                                            &fence);
+                        fence = radeon_copy_dma(rdev, saddr, daddr,
+                                                size / RADEON_GPU_PAGE_SIZE,
+                                                NULL);
                         break;
                 case RADEON_BENCHMARK_COPY_BLIT:
-                        r = radeon_copy_blit(rdev, saddr, daddr,
-                                             size / RADEON_GPU_PAGE_SIZE,
-                                             &fence);
+                        fence = radeon_copy_blit(rdev, saddr, daddr,
+                                                 size / RADEON_GPU_PAGE_SIZE,
+                                                 NULL);
                         break;
                 default:
                         DRM_ERROR("Unknown copy method\n");
-                        r = -EINVAL;
+                        return -EINVAL;
                 }
-                if (r)
-                        goto exit_do_move;
+                if (IS_ERR(fence))
+                        return PTR_ERR(fence);
+
                 r = radeon_fence_wait(fence, false);
-                if (r)
-                        goto exit_do_move;
                 radeon_fence_unref(&fence);
+                if (r)
+                        return r;
         }
         end_jiffies = jiffies;
-        r = jiffies_to_msecs(end_jiffies - start_jiffies);
-
-exit_do_move:
-        if (fence)
-                radeon_fence_unref(&fence);
-        return r;
+        return jiffies_to_msecs(end_jiffies - start_jiffies);
 }
 
 
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 6e3d1c8f3483..f662de41ba49 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -183,6 +183,7 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
                 }
 
                 p->relocs[i].tv.bo = &p->relocs[i].robj->tbo;
+                p->relocs[i].tv.shared = !r->write_domain;
                 p->relocs[i].handle = r->handle;
 
                 radeon_cs_buckets_add(&buckets, &p->relocs[i].tv.head,
@@ -254,16 +255,13 @@ static void radeon_cs_sync_rings(struct radeon_cs_parser *p)
 
         for (i = 0; i < p->nrelocs; i++) {
                 struct reservation_object *resv;
-                struct fence *fence;
 
                 if (!p->relocs[i].robj)
                         continue;
 
                 resv = p->relocs[i].robj->tbo.resv;
-                fence = reservation_object_get_excl(resv);
-
-                radeon_semaphore_sync_to(p->ib.semaphore,
-                                         (struct radeon_fence *)fence);
+                radeon_semaphore_sync_resv(p->ib.semaphore, resv,
+                                           p->relocs[i].tv.shared);
         }
 }
 
@@ -568,7 +566,7 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
                 goto out;
         }
         radeon_cs_sync_rings(parser);
-        radeon_semaphore_sync_to(parser->ib.semaphore, vm->fence);
+        radeon_semaphore_sync_fence(parser->ib.semaphore, vm->fence);
 
         if ((rdev->family >= CHIP_TAHITI) &&
             (parser->chunk_const_ib_idx != -1)) {
diff --git a/drivers/gpu/drm/radeon/radeon_ib.c b/drivers/gpu/drm/radeon/radeon_ib.c
index 6fc7461d70c4..3f39fcca4d07 100644
--- a/drivers/gpu/drm/radeon/radeon_ib.c
+++ b/drivers/gpu/drm/radeon/radeon_ib.c
@@ -145,7 +145,7 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
         if (ib->vm) {
                 struct radeon_fence *vm_id_fence;
                 vm_id_fence = radeon_vm_grab_id(rdev, ib->vm, ib->ring);
-                radeon_semaphore_sync_to(ib->semaphore, vm_id_fence);
+                radeon_semaphore_sync_fence(ib->semaphore, vm_id_fence);
         }
 
         /* sync with other rings */
diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c
index 56d9fd66d8ae..5bfbd8372a3c 100644
--- a/drivers/gpu/drm/radeon/radeon_semaphore.c
+++ b/drivers/gpu/drm/radeon/radeon_semaphore.c
@@ -96,15 +96,15 @@ bool radeon_semaphore_emit_wait(struct radeon_device *rdev, int ridx,
 }
 
 /**
- * radeon_semaphore_sync_to - use the semaphore to sync to a fence
+ * radeon_semaphore_sync_fence - use the semaphore to sync to a fence
  *
  * @semaphore: semaphore object to add fence to
  * @fence: fence to sync to
  *
  * Sync to the fence using this semaphore object
  */
-void radeon_semaphore_sync_to(struct radeon_semaphore *semaphore,
-                              struct radeon_fence *fence)
+void radeon_semaphore_sync_fence(struct radeon_semaphore *semaphore,
+                                 struct radeon_fence *fence)
 {
         struct radeon_fence *other;
 
@@ -116,6 +116,38 @@ void radeon_semaphore_sync_to(struct radeon_semaphore *semaphore,
 }
 
 /**
+ * radeon_semaphore_sync_to - use the semaphore to sync to a reservation object
+ *
+ * @sema: semaphore object to add fence from reservation object to
+ * @resv: reservation object with embedded fence
+ * @shared: true if we should onyl sync to the exclusive fence
+ *
+ * Sync to the fence using this semaphore object
+ */
+void radeon_semaphore_sync_resv(struct radeon_semaphore *sema,
+                                struct reservation_object *resv,
+                                bool shared)
+{
+        struct reservation_object_list *flist;
+        struct fence *f;
+        unsigned i;
+
+        /* always sync to the exclusive fence */
+        f = reservation_object_get_excl(resv);
+        radeon_semaphore_sync_fence(sema, (struct radeon_fence*)f);
+
+        flist = reservation_object_get_list(resv);
+        if (shared || !flist)
+                return;
+
+        for (i = 0; i < flist->shared_count; ++i) {
+                f = rcu_dereference_protected(flist->shared[i],
+                                              reservation_object_held(resv));
+                radeon_semaphore_sync_fence(sema, (struct radeon_fence*)f);
+        }
+}
+
+/**
  * radeon_semaphore_sync_rings - sync ring to all registered fences
  *
  * @rdev: radeon_device pointer
diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c
index 17bc3dced9f1..ce943e1a5e51 100644
--- a/drivers/gpu/drm/radeon/radeon_test.c
+++ b/drivers/gpu/drm/radeon/radeon_test.c
@@ -116,11 +116,16 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
                 radeon_bo_kunmap(gtt_obj[i]);
 
                 if (ring == R600_RING_TYPE_DMA_INDEX)
-                        r = radeon_copy_dma(rdev, gtt_addr, vram_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
+                        fence = radeon_copy_dma(rdev, gtt_addr, vram_addr,
+                                                size / RADEON_GPU_PAGE_SIZE,
+                                                NULL);
                 else
-                        r = radeon_copy_blit(rdev, gtt_addr, vram_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
-                if (r) {
+                        fence = radeon_copy_blit(rdev, gtt_addr, vram_addr,
+                                                 size / RADEON_GPU_PAGE_SIZE,
+                                                 NULL);
+                if (IS_ERR(fence)) {
                         DRM_ERROR("Failed GTT->VRAM copy %d\n", i);
+                        r = PTR_ERR(fence);
                         goto out_lclean_unpin;
                 }
 
@@ -162,11 +167,16 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
                 radeon_bo_kunmap(vram_obj);
 
                 if (ring == R600_RING_TYPE_DMA_INDEX)
-                        r = radeon_copy_dma(rdev, vram_addr, gtt_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
+                        fence = radeon_copy_dma(rdev, vram_addr, gtt_addr,
+                                                size / RADEON_GPU_PAGE_SIZE,
+                                                NULL);
                 else
-                        r = radeon_copy_blit(rdev, vram_addr, gtt_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
-                if (r) {
+                        fence = radeon_copy_blit(rdev, vram_addr, gtt_addr,
+                                                 size / RADEON_GPU_PAGE_SIZE,
+                                                 NULL);
+                if (IS_ERR(fence)) {
                         DRM_ERROR("Failed VRAM->GTT copy %d\n", i);
+                        r = PTR_ERR(fence);
                         goto out_lclean_unpin;
                 }
 
@@ -222,7 +232,7 @@ out_lclean:
                         radeon_bo_unreserve(gtt_obj[i]);
                         radeon_bo_unref(&gtt_obj[i]);
                 }
-                if (fence)
+                if (fence && !IS_ERR(fence))
                         radeon_fence_unref(&fence);
                 break;
         }
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 62d1f4d730a2..eca2ce60d440 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -233,6 +233,7 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
         struct radeon_device *rdev;
         uint64_t old_start, new_start;
         struct radeon_fence *fence;
+        unsigned num_pages;
         int r, ridx;
 
         rdev = radeon_get_rdev(bo->bdev);
@@ -269,12 +270,11 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 
         BUILD_BUG_ON((PAGE_SIZE % RADEON_GPU_PAGE_SIZE) != 0);
 
-        /* sync other rings */
-        fence = (struct radeon_fence *)reservation_object_get_excl(bo->resv);
-        r = radeon_copy(rdev, old_start, new_start,
-                        new_mem->num_pages * (PAGE_SIZE / RADEON_GPU_PAGE_SIZE), /* GPU pages */
-                        &fence);
-        /* FIXME: handle copy error */
+        num_pages = new_mem->num_pages * (PAGE_SIZE / RADEON_GPU_PAGE_SIZE);
+        fence = radeon_copy(rdev, old_start, new_start, num_pages, bo->resv);
+        if (IS_ERR(fence))
+                return PTR_ERR(fence);
+
         r = ttm_bo_move_accel_cleanup(bo, &fence->base,
                                       evict, no_wait_gpu, new_mem);
         radeon_fence_unref(&fence);
diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c
index 671ee566aa51..ce870959dff8 100644
--- a/drivers/gpu/drm/radeon/radeon_vm.c
+++ b/drivers/gpu/drm/radeon/radeon_vm.c
@@ -143,6 +143,7 @@ struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev,
         list[0].prefered_domains = RADEON_GEM_DOMAIN_VRAM;
         list[0].allowed_domains = RADEON_GEM_DOMAIN_VRAM;
         list[0].tv.bo = &vm->page_directory->tbo;
+        list[0].tv.shared = false;
         list[0].tiling_flags = 0;
         list[0].handle = 0;
         list_add(&list[0].tv.head, head);
@@ -156,6 +157,7 @@ struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev,
                 list[idx].prefered_domains = RADEON_GEM_DOMAIN_VRAM;
                 list[idx].allowed_domains = RADEON_GEM_DOMAIN_VRAM;
                 list[idx].tv.bo = &list[idx].robj->tbo;
+                list[idx].tv.shared = false;
                 list[idx].tiling_flags = 0;
                 list[idx].handle = 0;
                 list_add(&list[idx++].tv.head, head);
@@ -395,6 +397,7 @@ static int radeon_vm_clear_bo(struct radeon_device *rdev,
 
         memset(&tv, 0, sizeof(tv));
         tv.bo = &bo->tbo;
+        tv.shared = false;
 
         INIT_LIST_HEAD(&head);
         list_add(&tv.head, &head);
@@ -693,15 +696,10 @@ int radeon_vm_update_page_directory(struct radeon_device *rdev,
                                     incr, R600_PTE_VALID);
 
         if (ib.length_dw != 0) {
-                struct fence *fence;
-
                 radeon_asic_vm_pad_ib(rdev, &ib);
 
-                fence = reservation_object_get_excl(pd->tbo.resv);
-                radeon_semaphore_sync_to(ib.semaphore,
-                                         (struct radeon_fence *)fence);
-
-                radeon_semaphore_sync_to(ib.semaphore, vm->last_id_use);
+                radeon_semaphore_sync_resv(ib.semaphore, pd->tbo.resv, false);
+                radeon_semaphore_sync_fence(ib.semaphore, vm->last_id_use);
                 WARN_ON(ib.length_dw > ndw);
                 r = radeon_ib_schedule(rdev, &ib, NULL, false);
                 if (r) {
@@ -826,11 +824,8 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
                 struct radeon_bo *pt = vm->page_tables[pt_idx].bo;
                 unsigned nptes;
                 uint64_t pte;
-                struct fence *fence;
 
-                fence = reservation_object_get_excl(pt->tbo.resv);
-                radeon_semaphore_sync_to(ib->semaphore,
-                                         (struct radeon_fence *)fence);
+                radeon_semaphore_sync_resv(ib->semaphore, pt->tbo.resv, false);
 
                 if ((addr & ~mask) == (end & ~mask))
                         nptes = end - addr;
@@ -972,7 +967,7 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
         radeon_asic_vm_pad_ib(rdev, &ib);
         WARN_ON(ib.length_dw > ndw);
 
-        radeon_semaphore_sync_to(ib.semaphore, vm->fence);
+        radeon_semaphore_sync_fence(ib.semaphore, vm->fence);
         r = radeon_ib_schedule(rdev, &ib, NULL, false);
         if (r) {
                 radeon_ib_free(rdev, &ib);
diff --git a/drivers/gpu/drm/radeon/rv770_dma.c b/drivers/gpu/drm/radeon/rv770_dma.c
index 74426ac2bb5c..c112764adfdf 100644
--- a/drivers/gpu/drm/radeon/rv770_dma.c
+++ b/drivers/gpu/drm/radeon/rv770_dma.c
@@ -33,18 +33,19 @@
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (r7xx).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int rv770_copy_dma(struct radeon_device *rdev,
-                  uint64_t src_offset, uint64_t dst_offset,
-                  unsigned num_gpu_pages,
-                  struct radeon_fence **fence)
+struct radeon_fence *rv770_copy_dma(struct radeon_device *rdev,
+                                    uint64_t src_offset, uint64_t dst_offset,
+                                    unsigned num_gpu_pages,
+                                    struct reservation_object *resv)
 {
         struct radeon_semaphore *sem = NULL;
+        struct radeon_fence *fence;
         int ring_index = rdev->asic->copy.dma_ring_index;
         struct radeon_ring *ring = &rdev->ring[ring_index];
         u32 size_in_dw, cur_size_in_dw;
@@ -54,7 +55,7 @@ int rv770_copy_dma(struct radeon_device *rdev,
         r = radeon_semaphore_create(rdev, &sem);
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
-                return r;
+                return ERR_PTR(r);
         }
 
         size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4;
@@ -63,10 +64,10 @@ int rv770_copy_dma(struct radeon_device *rdev,
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
-        radeon_semaphore_sync_to(sem, *fence);
+        radeon_semaphore_sync_resv(sem, resv, false);
         radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
         for (i = 0; i < num_loops; i++) {
@@ -83,15 +84,15 @@ int rv770_copy_dma(struct radeon_device *rdev,
                 dst_offset += cur_size_in_dw * 4;
         }
 
-        r = radeon_fence_emit(rdev, fence, ring->idx);
+        r = radeon_fence_emit(rdev, &fence, ring->idx);
         if (r) {
                 radeon_ring_unlock_undo(rdev, ring);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
         radeon_ring_unlock_commit(rdev, ring, false);
-        radeon_semaphore_free(rdev, &sem, *fence);
+        radeon_semaphore_free(rdev, &sem, fence);
 
-        return r;
+        return fence;
 }
diff --git a/drivers/gpu/drm/radeon/si_dma.c b/drivers/gpu/drm/radeon/si_dma.c
index 7c22baaf94db..9b0dfbc913f3 100644
--- a/drivers/gpu/drm/radeon/si_dma.c
+++ b/drivers/gpu/drm/radeon/si_dma.c
@@ -218,18 +218,19 @@ void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (SI).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int si_copy_dma(struct radeon_device *rdev,
-                uint64_t src_offset, uint64_t dst_offset,
-                unsigned num_gpu_pages,
-                struct radeon_fence **fence)
+struct radeon_fence *si_copy_dma(struct radeon_device *rdev,
+                                 uint64_t src_offset, uint64_t dst_offset,
+                                 unsigned num_gpu_pages,
+                                 struct reservation_object *resv)
 {
         struct radeon_semaphore *sem = NULL;
+        struct radeon_fence *fence;
         int ring_index = rdev->asic->copy.dma_ring_index;
         struct radeon_ring *ring = &rdev->ring[ring_index];
         u32 size_in_bytes, cur_size_in_bytes;
@@ -239,7 +240,7 @@ int si_copy_dma(struct radeon_device *rdev,
         r = radeon_semaphore_create(rdev, &sem);
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
-                return r;
+                return ERR_PTR(r);
         }
 
         size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
@@ -248,10 +249,10 @@ int si_copy_dma(struct radeon_device *rdev,
         if (r) {
                 DRM_ERROR("radeon: moving bo (%d).\n", r);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
-        radeon_semaphore_sync_to(sem, *fence);
+        radeon_semaphore_sync_resv(sem, resv, false);
         radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
         for (i = 0; i < num_loops; i++) {
@@ -268,16 +269,16 @@ int si_copy_dma(struct radeon_device *rdev,
                 dst_offset += cur_size_in_bytes;
         }
 
-        r = radeon_fence_emit(rdev, fence, ring->idx);
+        r = radeon_fence_emit(rdev, &fence, ring->idx);
         if (r) {
                 radeon_ring_unlock_undo(rdev, ring);
                 radeon_semaphore_free(rdev, &sem, NULL);
-                return r;
+                return ERR_PTR(r);
         }
 
         radeon_ring_unlock_commit(rdev, ring, false);
-        radeon_semaphore_free(rdev, &sem, *fence);
+        radeon_semaphore_free(rdev, &sem, fence);
 
-        return r;
+        return fence;
 }
 
diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
index adafc0f8ec06..8ce508e76208 100644
--- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c
+++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
@@ -119,8 +119,14 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
                         ret = -EBUSY;
                 }
 
-                if (!ret)
-                        continue;
+                if (!ret) {
+                        if (!entry->shared)
+                                continue;
+
+                        ret = reservation_object_reserve_shared(bo->resv);
+                        if (!ret)
+                                continue;
+                }
 
                 /* uh oh, we lost out, drop every reservation and try
                  * to only reserve this buffer, then start over if
@@ -136,6 +142,9 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
                         ret = 0;
                 }
 
+                if (!ret && entry->shared)
+                        ret = reservation_object_reserve_shared(bo->resv);
+
                 if (unlikely(ret != 0)) {
                         if (ret == -EINTR)
                                 ret = -ERESTARTSYS;
@@ -183,7 +192,10 @@ void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket,
 
         list_for_each_entry(entry, list, head) {
                 bo = entry->bo;
-                reservation_object_add_excl_fence(bo->resv, fence);
+                if (entry->shared)
+                        reservation_object_add_shared_fence(bo->resv, fence);
+                else
+                        reservation_object_add_excl_fence(bo->resv, fence);
                 ttm_bo_add_to_lru(bo);
                 __ttm_bo_unreserve(bo);
         }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index 0ceaddc8e4f7..b4de3b2a7cc5 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -346,6 +346,7 @@ static int vmw_bo_to_validate_list(struct vmw_sw_context *sw_context,
                 ++sw_context->cur_val_buf;
                 val_buf = &vval_buf->base;
                 val_buf->bo = ttm_bo_reference(bo);
+                val_buf->shared = false;
                 list_add_tail(&val_buf->head, &sw_context->validate_nodes);
                 vval_buf->validate_as_mob = validate_as_mob;
         }
@@ -2670,9 +2671,11 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
         INIT_LIST_HEAD(&validate_list);
 
         pinned_val.bo = ttm_bo_reference(dev_priv->pinned_bo);
+        pinned_val.shared = false;
         list_add_tail(&pinned_val.head, &validate_list);
 
         query_val.bo = ttm_bo_reference(dev_priv->dummy_query_bo);
+        query_val.shared = false;
         list_add_tail(&query_val.head, &validate_list);
 
         ret = ttm_eu_reserve_buffers(&ticket, &validate_list, false);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
index ff0e03b97753..26584316cb78 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
@@ -133,6 +133,7 @@ static void vmw_resource_release(struct kref *kref)
                         struct ttm_validate_buffer val_buf;
 
                         val_buf.bo = bo;
+                        val_buf.shared = false;
                         res->func->unbind(res, false, &val_buf);
                 }
                 res->backup_dirty = false;
@@ -1219,6 +1220,7 @@ vmw_resource_check_buffer(struct vmw_resource *res,
 
         INIT_LIST_HEAD(&val_list);
         val_buf->bo = ttm_bo_reference(&res->backup->base);
+        val_buf->shared = false;
         list_add_tail(&val_buf->head, &val_list);
         ret = ttm_eu_reserve_buffers(NULL, &val_list, interruptible);
         if (unlikely(ret != 0))
@@ -1312,6 +1314,7 @@ int vmw_resource_do_evict(struct vmw_resource *res, bool interruptible)
         BUG_ON(!func->may_evict);
 
         val_buf.bo = NULL;
+        val_buf.shared = false;
         ret = vmw_resource_check_buffer(res, interruptible, &val_buf);
         if (unlikely(ret != 0))
                 return ret;
@@ -1357,6 +1360,7 @@ int vmw_resource_validate(struct vmw_resource *res)
                 return 0;
 
         val_buf.bo = NULL;
+        val_buf.shared = false;
         if (res->backup)
                 val_buf.bo = &res->backup->base;
         do {
@@ -1474,6 +1478,7 @@ void vmw_resource_move_notify(struct ttm_buffer_object *bo,
                 struct ttm_validate_buffer val_buf;
 
                 val_buf.bo = bo;
+                val_buf.shared = false;
 
                 list_for_each_entry_safe(res, n, &dma_buf->res_list, mob_head) {
 
diff --git a/include/drm/ttm/ttm_execbuf_util.h b/include/drm/ttm/ttm_execbuf_util.h
index ff11a424f752..460441714413 100644
--- a/include/drm/ttm/ttm_execbuf_util.h
+++ b/include/drm/ttm/ttm_execbuf_util.h
@@ -39,11 +39,13 @@
  *
  * @head:           list head for thread-private list.
  * @bo:             refcounted buffer object pointer.
+ * @shared:         should the fence be added shared?
  */
 
 struct ttm_validate_buffer {
         struct list_head head;
         struct ttm_buffer_object *bo;
+        bool shared;
 };
 
 /**