drm/amdgpu: Add RLCG interface driver implementation for gfx v9.4.3 (v3)

Add RLCG interface support for gfx v9.4.3 and multiple XCCs.
Do not enable it yet.

v2: Fix amdgpu_rlcg_reg_access_ctrl init, add support for multiple XCCs
    in amdgpu_mm_wreg_mmio_rlc

v3: Use GET_INST() when indexing amdgpu_rlcg_reg_access_ctrl

Signed-off-by: Victor Lu <victorchengchi.lu@amd.com>
Reviewed-by: Zhigang Luo <zhigang.luo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

11 files changed, 81 insertions, 51 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 405f93b639d1..43270613bbcb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1128,7 +1128,7 @@ void amdgpu_device_wreg(struct amdgpu_device *adev,
 void amdgpu_device_indirect_wreg_ext(struct amdgpu_device *adev,
 				     u64 reg_addr, u32 reg_data);
 void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev,
-			     uint32_t reg, uint32_t v);
+			     uint32_t reg, uint32_t v, uint32_t xcc_id);
 void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value);
 uint8_t amdgpu_mm_rreg8(struct amdgpu_device *adev, uint32_t offset);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index 56e89e76ff17..00ab0b3c8277 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -154,7 +154,7 @@ static int  amdgpu_debugfs_process_reg_op(bool read, struct file *f,
 		} else {
 			r = get_user(value, (uint32_t *)buf);
 			if (!r)
-				amdgpu_mm_wreg_mmio_rlc(adev, *pos >> 2, value);
+				amdgpu_mm_wreg_mmio_rlc(adev, *pos >> 2, value, 0);
 		}
 		if (r) {
 			result = r;
@@ -283,7 +283,7 @@ static ssize_t amdgpu_debugfs_regs2_op(struct file *f, char __user *buf, u32 off
 		} else {
 			r = get_user(value, (uint32_t *)buf);
 			if (!r)
-				amdgpu_mm_wreg_mmio_rlc(adev, offset >> 2, value);
+				amdgpu_mm_wreg_mmio_rlc(adev, offset >> 2, value, rd->id.xcc_id);
 		}
 		if (r) {
 			result = r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index b593ebb1d677..28e6419023b7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -571,7 +571,8 @@ void amdgpu_device_wreg(struct amdgpu_device *adev,
  * this function is invoked only for the debugfs register access
  */
 void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev,
-			     uint32_t reg, uint32_t v)
+			     uint32_t reg, uint32_t v,
+			     uint32_t xcc_id)
 {
 	if (amdgpu_device_skip_hw_access(adev))
 		return;
@@ -580,7 +581,7 @@ void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev,
 	    adev->gfx.rlc.funcs &&
 	    adev->gfx.rlc.funcs->is_rlcg_access_range) {
 		if (adev->gfx.rlc.funcs->is_rlcg_access_range(adev, reg))
-			return amdgpu_sriov_wreg(adev, reg, v, 0, 0);
+			return amdgpu_sriov_wreg(adev, reg, v, 0, 0, xcc_id);
 	} else if ((reg * 4) >= adev->rmmio_size) {
 		adev->pcie_wreg(adev, reg * 4, v);
 	} else {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
index 80b263646966..b591d33af264 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
@@ -26,6 +26,8 @@
 
 #include "clearstate_defs.h"
 
+#define AMDGPU_MAX_RLC_INSTANCES	8
+
 /* firmware ID used in rlc toc */
 typedef enum _FIRMWARE_ID_ {
 	FIRMWARE_ID_INVALID					= 0,
@@ -201,7 +203,7 @@ struct amdgpu_rlc {
 	u32                     cp_table_size;
 
 	/* safe mode for updating CG/PG state */
-	bool in_safe_mode[8];
+	bool in_safe_mode[AMDGPU_MAX_RLC_INSTANCES];
 	const struct amdgpu_rlc_funcs *funcs;
 
 	/* for firmware data */
@@ -257,7 +259,7 @@ struct amdgpu_rlc {
 
 	bool rlcg_reg_access_supported;
 	/* registers for rlcg indirect reg access */
-	struct amdgpu_rlcg_reg_access_ctrl reg_access_ctrl;
+	struct amdgpu_rlcg_reg_access_ctrl reg_access_ctrl[AMDGPU_MAX_RLC_INSTANCES];
 };
 
 void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev, int xcc_id);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 41aa853a07d2..770eb9725ec1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -954,7 +954,7 @@ static bool amdgpu_virt_get_rlcg_reg_access_flag(struct amdgpu_device *adev,
 	return ret;
 }
 
-static u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag)
+static u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag, u32 xcc_id)
 {
 	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
 	uint32_t timeout = 50000;
@@ -972,7 +972,12 @@ static u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v
 		return 0;
 	}
 
-	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl;
+	if (adev->gfx.xcc_mask && (((1 << xcc_id) & adev->gfx.xcc_mask) == 0)) {
+		dev_err(adev->dev, "invalid xcc\n");
+		return 0;
+	}
+
+	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[xcc_id];
 	scratch_reg0 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg0;
 	scratch_reg1 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg1;
 	scratch_reg2 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg2;
@@ -1037,13 +1042,13 @@ static u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v
 
 void amdgpu_sriov_wreg(struct amdgpu_device *adev,
 		       u32 offset, u32 value,
-		       u32 acc_flags, u32 hwip)
+		       u32 acc_flags, u32 hwip, u32 xcc_id)
 {
 	u32 rlcg_flag;
 
 	if (!amdgpu_sriov_runtime(adev) &&
 		amdgpu_virt_get_rlcg_reg_access_flag(adev, acc_flags, hwip, true, &rlcg_flag)) {
-		amdgpu_virt_rlcg_reg_rw(adev, offset, value, rlcg_flag);
+		amdgpu_virt_rlcg_reg_rw(adev, offset, value, rlcg_flag, xcc_id);
 		return;
 	}
 
@@ -1054,13 +1059,13 @@ void amdgpu_sriov_wreg(struct amdgpu_device *adev,
 }
 
 u32 amdgpu_sriov_rreg(struct amdgpu_device *adev,
-		      u32 offset, u32 acc_flags, u32 hwip)
+		      u32 offset, u32 acc_flags, u32 hwip, u32 xcc_id)
 {
 	u32 rlcg_flag;
 
 	if (!amdgpu_sriov_runtime(adev) &&
 		amdgpu_virt_get_rlcg_reg_access_flag(adev, acc_flags, hwip, false, &rlcg_flag))
-		return amdgpu_virt_rlcg_reg_rw(adev, offset, 0, rlcg_flag);
+		return amdgpu_virt_rlcg_reg_rw(adev, offset, 0, rlcg_flag, xcc_id);
 
 	if (acc_flags & AMDGPU_REGS_NO_KIQ)
 		return RREG32_NO_KIQ(offset);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
index 4f7bab52282a..d1f7509a44cb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
@@ -355,9 +355,9 @@ void amdgpu_virt_update_sriov_video_codec(struct amdgpu_device *adev,
 			struct amdgpu_video_codec_info *decode, uint32_t decode_array_size);
 void amdgpu_sriov_wreg(struct amdgpu_device *adev,
 		       u32 offset, u32 value,
-		       u32 acc_flags, u32 hwip);
+		       u32 acc_flags, u32 hwip, u32 xcc_id);
 u32 amdgpu_sriov_rreg(struct amdgpu_device *adev,
-		      u32 offset, u32 acc_flags, u32 hwip);
+		      u32 offset, u32 acc_flags, u32 hwip, u32 xcc_id);
 bool amdgpu_virt_fw_load_skip_check(struct amdgpu_device *adev,
 			uint32_t ucode_id);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 1d671c330475..0aee9c8288a2 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -4137,7 +4137,7 @@ static void gfx_v10_0_init_rlcg_reg_access_ctrl(struct amdgpu_device *adev)
 {
 	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
 
-	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl;
+	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[0];
 	reg_access_ctrl->scratch_reg0 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG0);
 	reg_access_ctrl->scratch_reg1 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG1);
 	reg_access_ctrl->scratch_reg2 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG2);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
index 3a7af59e83ca..4d53d6f13c3b 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -663,7 +663,7 @@ static void gfx_v11_0_init_rlcg_reg_access_ctrl(struct amdgpu_device *adev)
 {
 	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
 
-	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl;
+	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[0];
 	reg_access_ctrl->scratch_reg0 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG0);
 	reg_access_ctrl->scratch_reg1 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG1);
 	reg_access_ctrl->scratch_reg2 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG2);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index fc179e5f8dc1..0a74116b3142 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -1634,7 +1634,7 @@ static void gfx_v9_0_init_rlcg_reg_access_ctrl(struct amdgpu_device *adev)
 {
 	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
 
-	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl;
+	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[0];
 	reg_access_ctrl->scratch_reg0 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG0);
 	reg_access_ctrl->scratch_reg1 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG1);
 	reg_access_ctrl->scratch_reg2 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG2);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
index a20b77683924..b594d5c05149 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
@@ -1072,6 +1072,25 @@ static void gfx_v9_4_3_xcc_unset_safe_mode(struct amdgpu_device *adev,
 	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SAFE_MODE, data);
 }
 
+static void gfx_v9_4_3_init_rlcg_reg_access_ctrl(struct amdgpu_device *adev)
+{
+	int xcc_id;
+	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
+
+	for (xcc_id = 0; xcc_id < AMDGPU_MAX_RLC_INSTANCES; xcc_id++) {
+		if (((1 << xcc_id) & adev->gfx.xcc_mask) == 0)
+			continue;
+		reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[GET_INST(GC, xcc_id)];
+		reg_access_ctrl->scratch_reg0 = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regSCRATCH_REG0);
+		reg_access_ctrl->scratch_reg1 = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regSCRATCH_REG1);
+		reg_access_ctrl->scratch_reg2 = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regSCRATCH_REG2);
+		reg_access_ctrl->scratch_reg3 = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regSCRATCH_REG3);
+		reg_access_ctrl->grbm_cntl = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regGRBM_GFX_CNTL);
+		reg_access_ctrl->grbm_idx = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regGRBM_GFX_INDEX);
+		reg_access_ctrl->spare_int = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regRLC_SPARE_INT);
+	}
+}
+
 static int gfx_v9_4_3_rlc_init(struct amdgpu_device *adev)
 {
 	/* init spm vmid with 0xf */
@@ -2149,6 +2168,9 @@ static int gfx_v9_4_3_early_init(void *handle)
 	gfx_v9_4_3_set_gds_init(adev);
 	gfx_v9_4_3_set_rlc_funcs(adev);
 
+	/* init rlcg reg access ctrl */
+	gfx_v9_4_3_init_rlcg_reg_access_ctrl(adev);
+
 	return gfx_v9_4_3_init_microcode(adev);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15_common.h b/drivers/gpu/drm/amd/amdgpu/soc15_common.h
index 96948a59f8dd..da683afa0222 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15_common.h
+++ b/drivers/gpu/drm/amd/amdgpu/soc15_common.h
@@ -37,65 +37,65 @@
 #define SOC15_REG_OFFSET1(ip, inst, reg, offset) \
 	(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + (reg)+(offset))
 
-#define __WREG32_SOC15_RLC__(reg, value, flag, hwip) \
+#define __WREG32_SOC15_RLC__(reg, value, flag, hwip, inst) \
 	((amdgpu_sriov_vf(adev) && adev->gfx.rlc.funcs && adev->gfx.rlc.rlcg_reg_access_supported) ? \
-	 amdgpu_sriov_wreg(adev, reg, value, flag, hwip) : \
+	 amdgpu_sriov_wreg(adev, reg, value, flag, hwip, inst) : \
 	 WREG32(reg, value))
 
-#define __RREG32_SOC15_RLC__(reg, flag, hwip) \
+#define __RREG32_SOC15_RLC__(reg, flag, hwip, inst) \
 	((amdgpu_sriov_vf(adev) && adev->gfx.rlc.funcs && adev->gfx.rlc.rlcg_reg_access_supported) ? \
-	 amdgpu_sriov_rreg(adev, reg, flag, hwip) : \
+	 amdgpu_sriov_rreg(adev, reg, flag, hwip, inst) : \
 	 RREG32(reg))
 
 #define WREG32_FIELD15(ip, idx, reg, field, val)	\
 	 __WREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg,	\
 				(__RREG32_SOC15_RLC__( \
 					adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg, \
-					0, ip##_HWIP) & \
+					0, ip##_HWIP, idx) & \
 				~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field), \
-			      0, ip##_HWIP)
+			      0, ip##_HWIP, idx)
 
 #define WREG32_FIELD15_PREREG(ip, idx, reg_name, field, val)        \
 	__WREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][idx][reg##reg_name##_BASE_IDX] + reg##reg_name,   \
 			(__RREG32_SOC15_RLC__( \
 					adev->reg_offset[ip##_HWIP][idx][reg##reg_name##_BASE_IDX] + reg##reg_name, \
-					0, ip##_HWIP) & \
+					0, ip##_HWIP, idx) & \
 					~REG_FIELD_MASK(reg_name, field)) | (val) << REG_FIELD_SHIFT(reg_name, field), \
-			0, ip##_HWIP)
+			0, ip##_HWIP, idx)
 
 #define RREG32_SOC15(ip, inst, reg) \
 	__RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, \
-			 0, ip##_HWIP)
+			 0, ip##_HWIP, inst)
 
-#define RREG32_SOC15_IP(ip, reg) __RREG32_SOC15_RLC__(reg, 0, ip##_HWIP)
+#define RREG32_SOC15_IP(ip, reg) __RREG32_SOC15_RLC__(reg, 0, ip##_HWIP, 0)
 
-#define RREG32_SOC15_IP_NO_KIQ(ip, reg) __RREG32_SOC15_RLC__(reg, AMDGPU_REGS_NO_KIQ, ip##_HWIP)
+#define RREG32_SOC15_IP_NO_KIQ(ip, reg) __RREG32_SOC15_RLC__(reg, AMDGPU_REGS_NO_KIQ, ip##_HWIP, 0)
 
 #define RREG32_SOC15_NO_KIQ(ip, inst, reg) \
 	__RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, \
-			 AMDGPU_REGS_NO_KIQ, ip##_HWIP)
+			 AMDGPU_REGS_NO_KIQ, ip##_HWIP, inst)
 
 #define RREG32_SOC15_OFFSET(ip, inst, reg, offset) \
 	 __RREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + (reg)) + \
-			 (offset), 0, ip##_HWIP)
+			 (offset), 0, ip##_HWIP, inst)
 
 #define WREG32_SOC15(ip, inst, reg, value) \
 	 __WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg), \
-			  value, 0, ip##_HWIP)
+			  value, 0, ip##_HWIP, inst)
 
 #define WREG32_SOC15_IP(ip, reg, value) \
-	 __WREG32_SOC15_RLC__(reg, value, 0, ip##_HWIP)
+	 __WREG32_SOC15_RLC__(reg, value, 0, ip##_HWIP, 0)
 
 #define WREG32_SOC15_IP_NO_KIQ(ip, reg, value) \
-	 __WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_NO_KIQ, ip##_HWIP)
+	 __WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_NO_KIQ, ip##_HWIP, 0)
 
 #define WREG32_SOC15_NO_KIQ(ip, inst, reg, value) \
 	__WREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, \
-			     value, AMDGPU_REGS_NO_KIQ, ip##_HWIP)
+			     value, AMDGPU_REGS_NO_KIQ, ip##_HWIP, inst)
 
 #define WREG32_SOC15_OFFSET(ip, inst, reg, offset, value) \
 	 __WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, \
-			  value, 0, ip##_HWIP)
+			  value, 0, ip##_HWIP, inst)
 
 #define SOC15_WAIT_ON_RREG(ip, inst, reg, expected_value, mask)      \
 	amdgpu_device_wait_on_rreg(adev, inst,                       \
@@ -108,16 +108,16 @@
 	#reg, expected_value, mask)
 
 #define WREG32_RLC(reg, value) \
-	__WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_RLC, GC_HWIP)
+	__WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_RLC, GC_HWIP, 0)
 
-#define WREG32_RLC_EX(prefix, reg, value) \
+#define WREG32_RLC_EX(prefix, reg, value, inst) \
 	do {							\
 		if (amdgpu_sriov_fullaccess(adev)) {    \
 			uint32_t i = 0;	\
 			uint32_t retries = 50000;	\
-			uint32_t r0 = adev->reg_offset[GC_HWIP][0][prefix##SCRATCH_REG0_BASE_IDX] + prefix##SCRATCH_REG0;	\
-			uint32_t r1 = adev->reg_offset[GC_HWIP][0][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG1;	\
-			uint32_t spare_int = adev->reg_offset[GC_HWIP][0][prefix##RLC_SPARE_INT_BASE_IDX] + prefix##RLC_SPARE_INT;	\
+			uint32_t r0 = adev->reg_offset[GC_HWIP][inst][prefix##SCRATCH_REG0_BASE_IDX] + prefix##SCRATCH_REG0;	\
+			uint32_t r1 = adev->reg_offset[GC_HWIP][inst][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG1;	\
+			uint32_t spare_int = adev->reg_offset[GC_HWIP][inst][prefix##RLC_SPARE_INT_BASE_IDX] + prefix##RLC_SPARE_INT;	\
 			WREG32(r0, value);	\
 			WREG32(r1, (reg | 0x80000000));	\
 			WREG32(spare_int, 0x1);	\
@@ -136,17 +136,17 @@
 
 /* shadow the registers in the callback function */
 #define WREG32_SOC15_RLC_SHADOW(ip, inst, reg, value) \
-	__WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg), value, AMDGPU_REGS_RLC, GC_HWIP)
+	__WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg), value, AMDGPU_REGS_RLC, GC_HWIP, inst)
 
 /* for GC only */
 #define RREG32_RLC(reg) \
 	__RREG32_SOC15_RLC__(reg, AMDGPU_REGS_RLC, GC_HWIP)
 
 #define WREG32_RLC_NO_KIQ(reg, value, hwip) \
-	__WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_NO_KIQ | AMDGPU_REGS_RLC, hwip)
+	__WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_NO_KIQ | AMDGPU_REGS_RLC, hwip, 0)
 
 #define RREG32_RLC_NO_KIQ(reg, hwip) \
-	__RREG32_SOC15_RLC__(reg, AMDGPU_REGS_NO_KIQ | AMDGPU_REGS_RLC, hwip)
+	__RREG32_SOC15_RLC__(reg, AMDGPU_REGS_NO_KIQ | AMDGPU_REGS_RLC, hwip, 0)
 
 #define WREG32_SOC15_RLC_SHADOW_EX(prefix, ip, inst, reg, value) \
 	do {							\
@@ -167,32 +167,32 @@
 	} while (0)
 
 #define RREG32_SOC15_RLC(ip, inst, reg) \
-	__RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, AMDGPU_REGS_RLC, ip##_HWIP)
+	__RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, AMDGPU_REGS_RLC, ip##_HWIP, inst)
 
 #define WREG32_SOC15_RLC(ip, inst, reg, value) \
 	do {							\
 		uint32_t target_reg = adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg;\
-		__WREG32_SOC15_RLC__(target_reg, value, AMDGPU_REGS_RLC, ip##_HWIP); \
+		__WREG32_SOC15_RLC__(target_reg, value, AMDGPU_REGS_RLC, ip##_HWIP, inst); \
 	} while (0)
 
 #define WREG32_SOC15_RLC_EX(prefix, ip, inst, reg, value) \
 	do {							\
 			uint32_t target_reg = adev->reg_offset[GC_HWIP][inst][reg##_BASE_IDX] + reg;\
-			WREG32_RLC_EX(prefix, target_reg, value); \
+			WREG32_RLC_EX(prefix, target_reg, value, inst); \
 	} while (0)
 
 #define WREG32_FIELD15_RLC(ip, idx, reg, field, val)   \
 	__WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg), \
 			     (__RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg, \
-						   AMDGPU_REGS_RLC, ip##_HWIP) & \
+						   AMDGPU_REGS_RLC, ip##_HWIP, idx) & \
 			      ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field), \
-			     AMDGPU_REGS_RLC, ip##_HWIP)
+			     AMDGPU_REGS_RLC, ip##_HWIP, idx)
 
 #define WREG32_SOC15_OFFSET_RLC(ip, inst, reg, offset, value) \
-	__WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, value, AMDGPU_REGS_RLC, ip##_HWIP)
+	__WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, value, AMDGPU_REGS_RLC, ip##_HWIP, inst)
 
 #define RREG32_SOC15_OFFSET_RLC(ip, inst, reg, offset) \
-	__RREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, AMDGPU_REGS_RLC, ip##_HWIP)
+	__RREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, AMDGPU_REGS_RLC, ip##_HWIP, inst)
 
 /* inst equals to ext for some IPs */
 #define RREG32_SOC15_EXT(ip, inst, reg, ext) \