drm/amdgpu: RLC to program regs for Vega10 SR-IOV

Under Vega10 SR-IOV, with new RLC's new feature, VF should call RLC
to program some registers if supported

Signed-off-by: Trigger Huang <Trigger.Huang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

4 files changed, 85 insertions, 77 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
index ef3d93b995b2..6574814d2ade 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
@@ -225,8 +225,8 @@ static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
 
 	lock_srbm(kgd, 0, 0, 0, vmid);
 
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_CONFIG), sh_mem_config);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_BASES), sh_mem_bases);
+	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_CONFIG), sh_mem_config);
+	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_BASES), sh_mem_bases);
 	/* APE1 no longer exists on GFX9 */
 
 	unlock_srbm(kgd);
@@ -369,7 +369,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 		value = RREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_CP_SCHEDULERS));
 		value = REG_SET_FIELD(value, RLC_CP_SCHEDULERS, scheduler1,
 			((mec << 5) | (pipe << 3) | queue_id | 0x80));
-		WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_CP_SCHEDULERS), value);
+		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmRLC_CP_SCHEDULERS), value);
 	}
 
 	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */
@@ -378,13 +378,13 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 
 	for (reg = hqd_base;
 	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
-		WREG32(reg, mqd_hqd[reg - hqd_base]);
+		WREG32_RLC(reg, mqd_hqd[reg - hqd_base]);
 
 
 	/* Activate doorbell logic before triggering WPTR poll. */
 	data = REG_SET_FIELD(m->cp_hqd_pq_doorbell_control,
 			     CP_HQD_PQ_DOORBELL_CONTROL, DOORBELL_EN, 1);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL), data);
+	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL), data);
 
 	if (wptr) {
 		/* Don't read wptr with get_user because the user
@@ -413,25 +413,25 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 		guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size - 1);
 		guessed_wptr += (uint64_t)m->cp_hqd_pq_wptr_hi << 32;
 
-		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_LO),
+		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_LO),
 		       lower_32_bits(guessed_wptr));
-		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI),
+		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI),
 		       upper_32_bits(guessed_wptr));
-		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR),
+		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR),
 		       lower_32_bits((uintptr_t)wptr));
-		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),
 		       upper_32_bits((uintptr_t)wptr));
 		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_PQ_WPTR_POLL_CNTL1),
 		       get_queue_mask(adev, pipe_id, queue_id));
 	}
 
 	/* Start the EOP fetcher */
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_EOP_RPTR),
+	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_EOP_RPTR),
 	       REG_SET_FIELD(m->cp_hqd_eop_rptr,
 			     CP_HQD_EOP_RPTR, INIT_FETCHER, 1));
 
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE), data);
+	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE), data);
 
 	release_queue(kgd);
 
@@ -633,7 +633,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	acquire_queue(kgd, pipe_id, queue_id);
 
 	if (m->cp_hqd_vmid == 0)
-		WREG32_FIELD15(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
+		WREG32_FIELD15_RLC(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
 
 	switch (reset_type) {
 	case KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN:
@@ -647,7 +647,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 		break;
 	}
 
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_DEQUEUE_REQUEST), type);
+	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_DEQUEUE_REQUEST), type);
 
 	end_jiffies = (utimeout * HZ / 1000) + jiffies;
 	while (true) {
@@ -838,7 +838,7 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd,
 
 	mutex_lock(&adev->grbm_idx_mutex);
 
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_INDEX), gfx_index_val);
+	WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, gfx_index_val);
 	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CMD), sq_cmd);
 
 	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
@@ -848,7 +848,7 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd,
 	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
 		SE_BROADCAST_WRITES, 1);
 
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_INDEX), data);
+	WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, data);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 49ed53812091..fc6d8c1fd32b 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -1837,7 +1837,7 @@ static void gfx_v9_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh
 	else
 		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SH_INDEX, sh_num);
 
-	WREG32_SOC15(GC, 0, mmGRBM_GFX_INDEX, data);
+	WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, data);
 }
 
 static u32 gfx_v9_0_get_rb_active_bitmap(struct amdgpu_device *adev)
@@ -1905,8 +1905,8 @@ static void gfx_v9_0_init_compute_vmid(struct amdgpu_device *adev)
 	for (i = FIRST_COMPUTE_VMID; i < LAST_COMPUTE_VMID; i++) {
 		soc15_grbm_select(adev, 0, 0, 0, i);
 		/* CP and shaders */
-		WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
-		WREG32_SOC15(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
+		WREG32_SOC15_RLC(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
+		WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
 	}
 	soc15_grbm_select(adev, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
@@ -1917,7 +1917,7 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
 	u32 tmp;
 	int i;
 
-	WREG32_FIELD15(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
+	WREG32_FIELD15_RLC(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
 
 	gfx_v9_0_tiling_mode_table_init(adev);
 
@@ -1960,7 +1960,7 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
 	 */
 	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 
-	WREG32_SOC15(GC, 0, mmPA_SC_FIFO_SIZE,
+	WREG32_SOC15_RLC(GC, 0, mmPA_SC_FIFO_SIZE,
 		   (adev->gfx.config.sc_prim_fifo_size_frontend <<
 			PA_SC_FIFO_SIZE__SC_FRONTEND_PRIM_FIFO_SIZE__SHIFT) |
 		   (adev->gfx.config.sc_prim_fifo_size_backend <<
@@ -2027,11 +2027,11 @@ static void gfx_v9_0_enable_gui_idle_interrupt(struct amdgpu_device *adev,
 static void gfx_v9_0_init_csb(struct amdgpu_device *adev)
 {
 	/* csib */
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_HI),
+	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_HI),
 			adev->gfx.rlc.clear_state_gpu_addr >> 32);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_LO),
+	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_LO),
 			adev->gfx.rlc.clear_state_gpu_addr & 0xfffffffc);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_LENGTH),
+	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_LENGTH),
 			adev->gfx.rlc.clear_state_size);
 }
 
@@ -2501,7 +2501,7 @@ static void gfx_v9_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)
 		for (i = 0; i < adev->gfx.num_gfx_rings; i++)
 			adev->gfx.gfx_ring[i].sched.ready = false;
 	}
-	WREG32_SOC15(GC, 0, mmCP_ME_CNTL, tmp);
+	WREG32_SOC15_RLC(GC, 0, mmCP_ME_CNTL, tmp);
 	udelay(50);
 }
 
@@ -2699,9 +2699,9 @@ static void gfx_v9_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
 	int i;
 
 	if (enable) {
-		WREG32_SOC15(GC, 0, mmCP_MEC_CNTL, 0);
+		WREG32_SOC15_RLC(GC, 0, mmCP_MEC_CNTL, 0);
 	} else {
-		WREG32_SOC15(GC, 0, mmCP_MEC_CNTL,
+		WREG32_SOC15_RLC(GC, 0, mmCP_MEC_CNTL,
 			(CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK));
 		for (i = 0; i < adev->gfx.num_compute_rings; i++)
 			adev->gfx.compute_ring[i].sched.ready = false;
@@ -2762,9 +2762,9 @@ static void gfx_v9_0_kiq_setting(struct amdgpu_ring *ring)
 	tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS);
 	tmp &= 0xffffff00;
 	tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
-	WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
+	WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
 	tmp |= 0x80;
-	WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
+	WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
 }
 
 static int gfx_v9_0_kiq_kcq_enable(struct amdgpu_device *adev)
@@ -2982,67 +2982,67 @@ static int gfx_v9_0_kiq_init_register(struct amdgpu_ring *ring)
 	/* disable wptr polling */
 	WREG32_FIELD15(GC, 0, CP_PQ_WPTR_POLL_CNTL, EN, 0);
 
-	WREG32_SOC15(GC, 0, mmCP_HQD_EOP_BASE_ADDR,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_EOP_BASE_ADDR,
 	       mqd->cp_hqd_eop_base_addr_lo);
-	WREG32_SOC15(GC, 0, mmCP_HQD_EOP_BASE_ADDR_HI,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_EOP_BASE_ADDR_HI,
 	       mqd->cp_hqd_eop_base_addr_hi);
 
 	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
-	WREG32_SOC15(GC, 0, mmCP_HQD_EOP_CONTROL,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_EOP_CONTROL,
 	       mqd->cp_hqd_eop_control);
 
 	/* enable doorbell? */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL,
 	       mqd->cp_hqd_pq_doorbell_control);
 
 	/* disable the queue if it's active */
 	if (RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1) {
-		WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 1);
+		WREG32_SOC15_RLC(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 1);
 		for (j = 0; j < adev->usec_timeout; j++) {
 			if (!(RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1))
 				break;
 			udelay(1);
 		}
-		WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST,
+		WREG32_SOC15_RLC(GC, 0, mmCP_HQD_DEQUEUE_REQUEST,
 		       mqd->cp_hqd_dequeue_request);
-		WREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR,
+		WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_RPTR,
 		       mqd->cp_hqd_pq_rptr);
-		WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_LO,
+		WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_LO,
 		       mqd->cp_hqd_pq_wptr_lo);
-		WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_HI,
+		WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_HI,
 		       mqd->cp_hqd_pq_wptr_hi);
 	}
 
 	/* set the pointer to the MQD */
-	WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR,
+	WREG32_SOC15_RLC(GC, 0, mmCP_MQD_BASE_ADDR,
 	       mqd->cp_mqd_base_addr_lo);
-	WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR_HI,
+	WREG32_SOC15_RLC(GC, 0, mmCP_MQD_BASE_ADDR_HI,
 	       mqd->cp_mqd_base_addr_hi);
 
 	/* set MQD vmid to 0 */
-	WREG32_SOC15(GC, 0, mmCP_MQD_CONTROL,
+	WREG32_SOC15_RLC(GC, 0, mmCP_MQD_CONTROL,
 	       mqd->cp_mqd_control);
 
 	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_BASE,
 	       mqd->cp_hqd_pq_base_lo);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_BASE_HI,
 	       mqd->cp_hqd_pq_base_hi);
 
 	/* set up the HQD, this is similar to CP_RB0_CNTL */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_CONTROL,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_CONTROL,
 	       mqd->cp_hqd_pq_control);
 
 	/* set the wb address whether it's enabled or not */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR_REPORT_ADDR,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_RPTR_REPORT_ADDR,
 				mqd->cp_hqd_pq_rptr_report_addr_lo);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
 				mqd->cp_hqd_pq_rptr_report_addr_hi);
 
 	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR,
 	       mqd->cp_hqd_pq_wptr_poll_addr_lo);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI,
 	       mqd->cp_hqd_pq_wptr_poll_addr_hi);
 
 	/* enable the doorbell if requested */
@@ -3057,19 +3057,19 @@ static int gfx_v9_0_kiq_init_register(struct amdgpu_ring *ring)
 	       mqd->cp_hqd_pq_doorbell_control);
 
 	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_LO,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_LO,
 	       mqd->cp_hqd_pq_wptr_lo);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_HI,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_HI,
 	       mqd->cp_hqd_pq_wptr_hi);
 
 	/* set the vmid for the queue */
-	WREG32_SOC15(GC, 0, mmCP_HQD_VMID, mqd->cp_hqd_vmid);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_VMID, mqd->cp_hqd_vmid);
 
-	WREG32_SOC15(GC, 0, mmCP_HQD_PERSISTENT_STATE,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PERSISTENT_STATE,
 	       mqd->cp_hqd_persistent_state);
 
 	/* activate the queue */
-	WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE,
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_ACTIVE,
 	       mqd->cp_hqd_active);
 
 	if (ring->use_doorbell)
@@ -3086,7 +3086,7 @@ static int gfx_v9_0_kiq_fini_register(struct amdgpu_ring *ring)
 	/* disable the queue if it's active */
 	if (RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1) {
 
-		WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 1);
+		WREG32_SOC15_RLC(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 1);
 
 		for (j = 0; j < adev->usec_timeout; j++) {
 			if (!(RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1))
@@ -3098,21 +3098,21 @@ static int gfx_v9_0_kiq_fini_register(struct amdgpu_ring *ring)
 			DRM_DEBUG("KIQ dequeue request failed.\n");
 
 			/* Manual disable if dequeue request times out */
-			WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE, 0);
+			WREG32_SOC15_RLC(GC, 0, mmCP_HQD_ACTIVE, 0);
 		}
 
-		WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST,
+		WREG32_SOC15_RLC(GC, 0, mmCP_HQD_DEQUEUE_REQUEST,
 		      0);
 	}
 
-	WREG32_SOC15(GC, 0, mmCP_HQD_IQ_TIMER, 0);
-	WREG32_SOC15(GC, 0, mmCP_HQD_IB_CONTROL, 0);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PERSISTENT_STATE, 0);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, 0x40000000);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, 0);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR, 0);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_HI, 0);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_LO, 0);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_IQ_TIMER, 0);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_IB_CONTROL, 0);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PERSISTENT_STATE, 0);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, 0x40000000);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, 0);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_RPTR, 0);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_HI, 0);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_LO, 0);
 
 	return 0;
 }
@@ -4572,8 +4572,8 @@ static void gfx_v9_0_hqd_set_priority(struct amdgpu_device *adev,
 	mutex_lock(&adev->srbm_mutex);
 	soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
 
-	WREG32_SOC15(GC, 0, mmCP_HQD_PIPE_PRIORITY, pipe_priority);
-	WREG32_SOC15(GC, 0, mmCP_HQD_QUEUE_PRIORITY, queue_priority);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PIPE_PRIORITY, pipe_priority);
+	WREG32_SOC15_RLC(GC, 0, mmCP_HQD_QUEUE_PRIORITY, queue_priority);
 
 	soc15_grbm_select(adev, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
index 7bb5359d0bbd..0dc8926111e4 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
@@ -71,12 +71,12 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
 	uint64_t value;
 
 	/* Program the AGP BAR */
-	WREG32_SOC15(GC, 0, mmMC_VM_AGP_BASE, 0);
-	WREG32_SOC15(GC, 0, mmMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
-	WREG32_SOC15(GC, 0, mmMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+	WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_BASE, 0);
+	WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+	WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
 
 	/* Program the system aperture low logical page number. */
-	WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+	WREG32_SOC15_RLC(GC, 0, mmMC_VM_SYSTEM_APERTURE_LOW_ADDR,
 		     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
 
 	if (adev->asic_type == CHIP_RAVEN && adev->rev_id >= 0x8)
@@ -86,11 +86,11 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
 		 * workaround that increase system aperture high address (add 1)
 		 * to get rid of the VM fault and hardware hang.
 		 */
-		WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+		WREG32_SOC15_RLC(GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
 			     max((adev->gmc.fb_end >> 18) + 0x1,
 				 adev->gmc.agp_end >> 18));
 	else
-		WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+		WREG32_SOC15_RLC(GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
 			     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
 
 	/* Set default page address. */
@@ -129,7 +129,7 @@ static void gfxhub_v1_0_init_tlb_regs(struct amdgpu_device *adev)
 			    MTYPE, MTYPE_UC);/* XXX for emulation. */
 	tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1);
 
-	WREG32_SOC15(GC, 0, mmMC_VM_MX_L1_TLB_CNTL, tmp);
+	WREG32_SOC15_RLC(GC, 0, mmMC_VM_MX_L1_TLB_CNTL, tmp);
 }
 
 static void gfxhub_v1_0_init_cache_regs(struct amdgpu_device *adev)
@@ -267,9 +267,9 @@ int gfxhub_v1_0_gart_enable(struct amdgpu_device *adev)
 		 * VF copy registers so vbios post doesn't program them, for
 		 * SRIOV driver need to program them
 		 */
-		WREG32_SOC15(GC, 0, mmMC_VM_FB_LOCATION_BASE,
+		WREG32_SOC15_RLC(GC, 0, mmMC_VM_FB_LOCATION_BASE,
 			     adev->gmc.vram_start >> 24);
-		WREG32_SOC15(GC, 0, mmMC_VM_FB_LOCATION_TOP,
+		WREG32_SOC15_RLC(GC, 0, mmMC_VM_FB_LOCATION_TOP,
 			     adev->gmc.vram_end >> 24);
 	}
 
@@ -303,7 +303,7 @@ void gfxhub_v1_0_gart_disable(struct amdgpu_device *adev)
 				MC_VM_MX_L1_TLB_CNTL,
 				ENABLE_ADVANCED_DRIVER_MODEL,
 				0);
-	WREG32_SOC15(GC, 0, mmMC_VM_MX_L1_TLB_CNTL, tmp);
+	WREG32_SOC15_RLC(GC, 0, mmMC_VM_MX_L1_TLB_CNTL, tmp);
 
 	/* Setup L2 cache */
 	WREG32_FIELD15(GC, 0, VM_L2_CNTL, ENABLE_L2_CACHE, 0);
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index a4ebe9a0ac7f..f9c9cacc5c50 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -232,7 +232,7 @@ void soc15_grbm_select(struct amdgpu_device *adev,
 	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, VMID, vmid);
 	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, QUEUEID, queue);
 
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_CNTL), grbm_gfx_cntl);
+	WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_CNTL, grbm_gfx_cntl);
 }
 
 static void soc15_vga_set_state(struct amdgpu_device *adev, bool state)
@@ -387,7 +387,15 @@ void soc15_program_register_sequence(struct amdgpu_device *adev,
 			tmp &= ~(entry->and_mask);
 			tmp |= entry->or_mask;
 		}
-		WREG32(reg, tmp);
+
+		if (reg == SOC15_REG_OFFSET(GC, 0, mmPA_SC_BINNER_EVENT_CNTL_3) ||
+			reg == SOC15_REG_OFFSET(GC, 0, mmPA_SC_ENHANCE) ||
+			reg == SOC15_REG_OFFSET(GC, 0, mmPA_SC_ENHANCE_1) ||
+			reg == SOC15_REG_OFFSET(GC, 0, mmSH_MEM_CONFIG))
+			WREG32_RLC(reg, tmp);
+		else
+			WREG32(reg, tmp);
+
 	}
 
 }