Merge tag 'v4.10-rc5' into for-linus

Sync up with mainline to apply fixup to a commit that came through
power supply tree.

1 files changed, 28 insertions, 13 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index ec1282af2479..216a9572d946 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -89,7 +89,7 @@ int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
  * Free an IB (all asics).
  */
 void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib,
-		    struct fence *f)
+		    struct dma_fence *f)
 {
 	amdgpu_sa_bo_free(adev, &ib->sa_bo, f);
 }
@@ -116,15 +116,16 @@ void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib,
  * to SI there was just a DE IB.
  */
 int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
-		       struct amdgpu_ib *ibs, struct fence *last_vm_update,
-		       struct amdgpu_job *job, struct fence **f)
+		       struct amdgpu_ib *ibs, struct dma_fence *last_vm_update,
+		       struct amdgpu_job *job, struct dma_fence **f)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_ib *ib = &ibs[0];
 	bool skip_preamble, need_ctx_switch;
 	unsigned patch_offset = ~0;
 	struct amdgpu_vm *vm;
-	uint64_t ctx;
+	uint64_t fence_ctx;
+	uint32_t status = 0, alloc_size;
 
 	unsigned i;
 	int r = 0;
@@ -135,14 +136,14 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	/* ring tests don't use a job */
 	if (job) {
 		vm = job->vm;
-		ctx = job->ctx;
+		fence_ctx = job->fence_ctx;
 	} else {
 		vm = NULL;
-		ctx = 0;
+		fence_ctx = 0;
 	}
 
 	if (!ring->ready) {
-		dev_err(adev->dev, "couldn't schedule ib\n");
+		dev_err(adev->dev, "couldn't schedule ib on ring <%s>\n", ring->name);
 		return -EINVAL;
 	}
 
@@ -151,13 +152,16 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		return -EINVAL;
 	}
 
-	r = amdgpu_ring_alloc(ring, 256 * num_ibs);
+	alloc_size = ring->funcs->emit_frame_size + num_ibs *
+		ring->funcs->emit_ib_size;
+
+	r = amdgpu_ring_alloc(ring, alloc_size);
 	if (r) {
 		dev_err(adev->dev, "scheduling IB failed (%d).\n", r);
 		return r;
 	}
 
-	if (ring->type == AMDGPU_RING_TYPE_SDMA && ring->funcs->init_cond_exec)
+	if (ring->funcs->init_cond_exec)
 		patch_offset = amdgpu_ring_init_cond_exec(ring);
 
 	if (vm) {
@@ -174,13 +178,22 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	/* always set cond_exec_polling to CONTINUE */
 	*ring->cond_exe_cpu_addr = 1;
 
-	skip_preamble = ring->current_ctx == ctx;
-	need_ctx_switch = ring->current_ctx != ctx;
+	skip_preamble = ring->current_ctx == fence_ctx;
+	need_ctx_switch = ring->current_ctx != fence_ctx;
+	if (job && ring->funcs->emit_cntxcntl) {
+		if (need_ctx_switch)
+			status |= AMDGPU_HAVE_CTX_SWITCH;
+		status |= job->preamble_status;
+		amdgpu_ring_emit_cntxcntl(ring, status);
+	}
+
 	for (i = 0; i < num_ibs; ++i) {
 		ib = &ibs[i];
 
 		/* drop preamble IBs if we don't have a context switch */
-		if ((ib->flags & AMDGPU_IB_FLAG_PREAMBLE) && skip_preamble)
+		if ((ib->flags & AMDGPU_IB_FLAG_PREAMBLE) &&
+			skip_preamble &&
+			!(status & AMDGPU_PREAMBLE_IB_PRESENT_FIRST))
 			continue;
 
 		amdgpu_ring_emit_ib(ring, ib, job ? job->vm_id : 0,
@@ -209,7 +222,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	if (patch_offset != ~0 && ring->funcs->patch_cond_exec)
 		amdgpu_ring_patch_cond_exec(ring, patch_offset);
 
-	ring->current_ctx = ctx;
+	ring->current_ctx = fence_ctx;
+	if (ring->funcs->emit_switch_buffer)
+		amdgpu_ring_emit_switch_buffer(ring);
 	amdgpu_ring_commit(ring);
 	return 0;
 }