drm/amdgpu: skip access sdma_v5_0 registers under SRIOV (v2)

Due to the new L1.0b0c011b policy, many SDMA registers are blocked which raise
the violation warning. There are total 6 pair register needed to be skipped
when driver init and de-init.
mmSDMA0/1_CNTL
mmSDMA0/1_F32_CNTL
mmSDMA0/1_UTCL1_PAGE
mmSDMA0/1_UTCL1_CNTL
mmSDMA0/1_CHICKEN_BITS,
mmSDMA0/1_SEM_WAIT_FAIL_TIMER_CNTL

v2: squash in warning fix
Signed-off-by: Yintian Tao <yttao@amd.com>
Reviewed-by: Emily Deng <Emily.Deng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c

@@ -88,6 +88,29 @@ static const struct soc15_reg_golden golden_settings_sdma_5[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_UTCL1_PAGE, 0x00ffffff, 0x000c5c00)
 };
 
+static const struct soc15_reg_golden golden_settings_sdma_5_sriov[] = {
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_GFX_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_PAGE_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_RLC0_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_RLC1_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_RLC2_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_RLC3_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_RLC4_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_RLC5_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_RLC6_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_RLC7_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_GFX_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_PAGE_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_RLC0_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_RLC1_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_RLC2_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_RLC3_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_RLC4_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_RLC5_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_RLC6_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_RLC7_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
+};
+
 static const struct soc15_reg_golden golden_settings_sdma_nv10[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_RLC3_RB_WPTR_POLL_CNTL, 0x0000fff0, 0x00403000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA1_RLC3_RB_WPTR_POLL_CNTL, 0x0000fff0, 0x00403000),
@@ -141,9 +164,14 @@ static void sdma_v5_0_init_golden_registers(struct amdgpu_device *adev)
 						(const u32)ARRAY_SIZE(golden_settings_sdma_nv14));
 		break;
 	case CHIP_NAVI12:
-		soc15_program_register_sequence(adev,
-						golden_settings_sdma_5,
-						(const u32)ARRAY_SIZE(golden_settings_sdma_5));
+		if (amdgpu_sriov_vf(adev))
+			soc15_program_register_sequence(adev,
+							golden_settings_sdma_5_sriov,
+							(const u32)ARRAY_SIZE(golden_settings_sdma_5_sriov));
+		else
+			soc15_program_register_sequence(adev,
+							golden_settings_sdma_5,
+							(const u32)ARRAY_SIZE(golden_settings_sdma_5));
 		soc15_program_register_sequence(adev,
 						golden_settings_sdma_nv12,
 						(const u32)ARRAY_SIZE(golden_settings_sdma_nv12));
@@ -526,7 +554,7 @@ static void sdma_v5_0_rlc_stop(struct amdgpu_device *adev)
  */
 static void sdma_v5_0_ctx_switch_enable(struct amdgpu_device *adev, bool enable)
 {
-	u32 f32_cntl, phase_quantum = 0;
+	u32 f32_cntl = 0, phase_quantum = 0;
 	int i;
 
 	if (amdgpu_sdma_phase_quantum) {
@@ -554,9 +582,12 @@ static void sdma_v5_0_ctx_switch_enable(struct amdgpu_device *adev, bool enable)
 	}
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
-		f32_cntl = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL));
-		f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_CNTL,
-				AUTO_CTXSW_ENABLE, enable ? 1 : 0);
+		if (!amdgpu_sriov_vf(adev)) {
+			f32_cntl = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL));
+			f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_CNTL,
+						 AUTO_CTXSW_ENABLE, enable ? 1 : 0);
+		}
+
 		if (enable && amdgpu_sdma_phase_quantum) {
 			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_PHASE0_QUANTUM),
 			       phase_quantum);
@@ -565,7 +596,8 @@ static void sdma_v5_0_ctx_switch_enable(struct amdgpu_device *adev, bool enable)
 			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_PHASE2_QUANTUM),
 			       phase_quantum);
 		}
-		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL), f32_cntl);
+		if (!amdgpu_sriov_vf(adev))
+			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL), f32_cntl);
 	}
 
 }
@@ -588,6 +620,9 @@ static void sdma_v5_0_enable(struct amdgpu_device *adev, bool enable)
 		sdma_v5_0_rlc_stop(adev);
 	}
 
+	if (amdgpu_sriov_vf(adev))
+		return;
+
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		f32_cntl = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_F32_CNTL));
 		f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, enable ? 0 : 1);
@@ -620,7 +655,8 @@ static int sdma_v5_0_gfx_resume(struct amdgpu_device *adev)
 		ring = &adev->sdma.instance[i].ring;
 		wb_offset = (ring->rptr_offs * 4);
 
-		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0);
+		if (!amdgpu_sriov_vf(adev))
+			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0);
 
 		/* Set ring buffer size in dwords */
 		rb_bufsz = order_base_2(ring->ring_size / 4);
@@ -696,26 +732,28 @@ static int sdma_v5_0_gfx_resume(struct amdgpu_device *adev)
 		/* set minor_ptr_update to 0 after wptr programed */
 		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 0);
 
-		/* set utc l1 enable flag always to 1 */
-		temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL));
-		temp = REG_SET_FIELD(temp, SDMA0_CNTL, UTC_L1_ENABLE, 1);
-
-		/* enable MCBP */
-		temp = REG_SET_FIELD(temp, SDMA0_CNTL, MIDCMD_PREEMPT_ENABLE, 1);
-		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL), temp);
-
-		/* Set up RESP_MODE to non-copy addresses */
-		temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL));
-		temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3);
-		temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9);
-		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL), temp);
-
-		/* program default cache read and write policy */
-		temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE));
-		/* clean read policy and write policy bits */
-		temp &= 0xFF0FFF;
-		temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) | (CACHE_WRITE_POLICY_L2__DEFAULT << 14));
-		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE), temp);
+		if (!amdgpu_sriov_vf(adev)) {
+			/* set utc l1 enable flag always to 1 */
+			temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL));
+			temp = REG_SET_FIELD(temp, SDMA0_CNTL, UTC_L1_ENABLE, 1);
+
+			/* enable MCBP */
+			temp = REG_SET_FIELD(temp, SDMA0_CNTL, MIDCMD_PREEMPT_ENABLE, 1);
+			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL), temp);
+
+			/* Set up RESP_MODE to non-copy addresses */
+			temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL));
+			temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3);
+			temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9);
+			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL), temp);
+
+			/* program default cache read and write policy */
+			temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE));
+			/* clean read policy and write policy bits */
+			temp &= 0xFF0FFF;
+			temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) | (CACHE_WRITE_POLICY_L2__DEFAULT << 14));
+			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE), temp);
+		}
 
 		if (!amdgpu_sriov_vf(adev)) {
 			/* unhalt engine */
@@ -1385,14 +1423,16 @@ static int sdma_v5_0_set_trap_irq_state(struct amdgpu_device *adev,
 {
 	u32 sdma_cntl;
 
-	u32 reg_offset = (type == AMDGPU_SDMA_IRQ_INSTANCE0) ?
-		sdma_v5_0_get_reg_offset(adev, 0, mmSDMA0_CNTL) :
-		sdma_v5_0_get_reg_offset(adev, 1, mmSDMA0_CNTL);
+	if (!amdgpu_sriov_vf(adev)) {
+		u32 reg_offset = (type == AMDGPU_SDMA_IRQ_INSTANCE0) ?
+			sdma_v5_0_get_reg_offset(adev, 0, mmSDMA0_CNTL) :
+			sdma_v5_0_get_reg_offset(adev, 1, mmSDMA0_CNTL);
 
-	sdma_cntl = RREG32(reg_offset);
-	sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA0_CNTL, TRAP_ENABLE,
-		       state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
-	WREG32(reg_offset, sdma_cntl);
+		sdma_cntl = RREG32(reg_offset);
+		sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA0_CNTL, TRAP_ENABLE,
+					  state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+		WREG32(reg_offset, sdma_cntl);
+	}
 
 	return 0;
 }