drm/amdgpu: revise RLCG access path

what changed:
1)provide new implementation interface for the rlcg access path
2)put SQ_CMD/SQ_IND_INDEX to GFX9 RLCG path to let debugfs's reg_op
function can access reg that need RLCG path help

now even debugfs's reg_op can used to dump wave.

tested-by: Monk Liu <monk.liu@amd.com>
tested-by: Zhou pengju <pengju.zhou@amd.com>
Signed-off-by: Zhou pengju <pengju.zhou@amd.com>
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Emily Deng <Emily.Deng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/amdgpu.h

@@ -994,6 +994,8 @@ uint32_t amdgpu_mm_rreg(struct amdgpu_device *adev, uint32_t reg,
 			uint32_t acc_flags);
 void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
 		    uint32_t acc_flags);
+void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
+		    uint32_t acc_flags);
 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);
 

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

@@ -179,7 +179,7 @@ static int  amdgpu_debugfs_process_reg_op(bool read, struct file *f,
 		} else {
 			r = get_user(value, (uint32_t *)buf);
 			if (!r)
-				WREG32(*pos >> 2, value);
+				amdgpu_mm_wreg_mmio_rlc(adev, *pos >> 2, value, 0);
 		}
 		if (r) {
 			result = r;

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

@@ -306,28 +306,10 @@ void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value)
 		BUG();
 }
 
-/**
+void static inline amdgpu_mm_wreg_mmio(struct amdgpu_device *adev, uint32_t reg, uint32_t v, uint32_t acc_flags)
- * amdgpu_mm_wreg - write to a memory mapped IO register
- *
- * @adev: amdgpu_device pointer
- * @reg: dword aligned register offset
- * @v: 32 bit value to write to the register
- * @acc_flags: access flags which require special behavior
- *
- * Writes the value specified to the offset specified.
- */
-void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
-		    uint32_t acc_flags)
 {
 	trace_amdgpu_mm_wreg(adev->pdev->device, reg, v);
 
-	if (adev->asic_type >= CHIP_VEGA10 && reg == 0) {
-		adev->last_mm_index = v;
-	}
-
-	if ((acc_flags & AMDGPU_REGS_KIQ) || (!(acc_flags & AMDGPU_REGS_NO_KIQ) && amdgpu_sriov_runtime(adev)))
-		return amdgpu_kiq_wreg(adev, reg, v);
-
 	if ((reg * 4) < adev->rmmio_size && !(acc_flags & AMDGPU_REGS_IDX))
 		writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
 	else {
@@ -344,6 +326,48 @@ void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
 	}
 }
 
+/**
+ * amdgpu_mm_wreg - write to a memory mapped IO register
+ *
+ * @adev: amdgpu_device pointer
+ * @reg: dword aligned register offset
+ * @v: 32 bit value to write to the register
+ * @acc_flags: access flags which require special behavior
+ *
+ * Writes the value specified to the offset specified.
+ */
+void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
+		    uint32_t acc_flags)
+{
+	if (adev->asic_type >= CHIP_VEGA10 && reg == 0) {
+		adev->last_mm_index = v;
+	}
+
+	if ((acc_flags & AMDGPU_REGS_KIQ) || (!(acc_flags & AMDGPU_REGS_NO_KIQ) && amdgpu_sriov_runtime(adev)))
+		return amdgpu_kiq_wreg(adev, reg, v);
+
+	amdgpu_mm_wreg_mmio(adev, reg, v, acc_flags);
+}
+
+/*
+ * amdgpu_mm_wreg_mmio_rlc -  write register either with mmio or with RLC path if in range
+ *
+ * this function is invoked only the debugfs register access
+ * */
+void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
+		    uint32_t acc_flags)
+{
+	if (amdgpu_sriov_fullaccess(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 adev->gfx.rlc.funcs->rlcg_wreg(adev, reg, v);
+	}
+
+	amdgpu_mm_wreg_mmio(adev, reg, v, acc_flags);
+}
+
 /**
  * amdgpu_io_rreg - read an IO register
  *

drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h

@@ -127,6 +127,8 @@ struct amdgpu_rlc_funcs {
 	void (*reset)(struct amdgpu_device *adev);
 	void (*start)(struct amdgpu_device *adev);
 	void (*update_spm_vmid)(struct amdgpu_device *adev, unsigned vmid);
+	void (*rlcg_wreg)(struct amdgpu_device *adev, u32 offset, u32 v);
+	bool (*is_rlcg_access_range)(struct amdgpu_device *adev, uint32_t reg);
 };
 
 struct amdgpu_rlc {

drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h

@@ -270,6 +270,9 @@ struct amdgpu_virt {
 #define amdgpu_sriov_runtime(adev) \
 ((adev)->virt.caps & AMDGPU_SRIOV_CAPS_RUNTIME)
 
+#define amdgpu_sriov_fullaccess(adev) \
+(amdgpu_sriov_vf((adev)) && !amdgpu_sriov_runtime((adev)))
+
 #define amdgpu_passthrough(adev) \
 ((adev)->virt.caps & AMDGPU_PASSTHROUGH_MODE)
 

drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c

@@ -224,6 +224,49 @@ static const struct soc15_reg_golden golden_settings_gc_10_1_2[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffffffff, 0x00800000)
 };
 
+static void gfx_v10_rlcg_wreg(struct amdgpu_device *adev, u32 offset, u32 v)
+{
+	static void *scratch_reg0;
+	static void *scratch_reg1;
+	static void *scratch_reg2;
+	static void *scratch_reg3;
+	static void *spare_int;
+	static uint32_t grbm_cntl;
+	static uint32_t grbm_idx;
+	uint32_t i = 0;
+	uint32_t retries = 50000;
+
+	scratch_reg0 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG0)*4;
+	scratch_reg1 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG1)*4;
+	scratch_reg2 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG2)*4;
+	scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG3)*4;
+	spare_int = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_BASE_IDX] + mmRLC_SPARE_INT)*4;
+
+	grbm_cntl = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_CNTL_BASE_IDX] + mmGRBM_GFX_CNTL;
+	grbm_idx = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_INDEX_BASE_IDX] + mmGRBM_GFX_INDEX;
+
+	if (amdgpu_sriov_runtime(adev)) {
+		pr_err("shoudn't call rlcg write register during runtime\n");
+		return;
+	}
+
+	writel(v, scratch_reg0);
+	writel(offset | 0x80000000, scratch_reg1);
+	writel(1, spare_int);
+	for (i = 0; i < retries; i++) {
+		u32 tmp;
+
+		tmp = readl(scratch_reg1);
+		if (!(tmp & 0x80000000))
+			break;
+
+		udelay(10);
+	}
+
+	if (i >= retries)
+		pr_err("timeout: rlcg program reg:0x%05x failed !\n", offset);
+}
+
 static const struct soc15_reg_golden golden_settings_gc_10_1_nv14[] =
 {
 	/* Pending on emulation bring up */
@@ -4247,6 +4290,33 @@ static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 	WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data);
 }
 
+static bool gfx_v10_0_check_rlcg_range(struct amdgpu_device *adev,
+					uint32_t offset,
+					struct soc15_reg_rlcg *entries, int arr_size)
+{
+	int i;
+	uint32_t reg;
+
+	if (!entries)
+		return false;
+
+	for (i = 0; i < arr_size; i++) {
+		const struct soc15_reg_rlcg *entry;
+
+		entry = &entries[i];
+		reg = adev->reg_offset[entry->hwip][entry->instance][entry->segment] + entry->reg;
+		if (offset == reg)
+			return true;
+	}
+
+	return false;
+}
+
+static bool gfx_v10_0_is_rlcg_access_range(struct amdgpu_device *adev, u32 offset)
+{
+	return gfx_v10_0_check_rlcg_range(adev, offset, NULL, 0);
+}
+
 static const struct amdgpu_rlc_funcs gfx_v10_0_rlc_funcs = {
 	.is_rlc_enabled = gfx_v10_0_is_rlc_enabled,
 	.set_safe_mode = gfx_v10_0_set_safe_mode,
@@ -4258,7 +4328,9 @@ static const struct amdgpu_rlc_funcs gfx_v10_0_rlc_funcs = {
 	.stop = gfx_v10_0_rlc_stop,
 	.reset = gfx_v10_0_rlc_reset,
 	.start = gfx_v10_0_rlc_start,
-	.update_spm_vmid = gfx_v10_0_update_spm_vmid
+	.update_spm_vmid = gfx_v10_0_update_spm_vmid,
+	.rlcg_wreg = gfx_v10_rlcg_wreg,
+	.is_rlcg_access_range = gfx_v10_0_is_rlcg_access_range,
 };
 
 static int gfx_v10_0_set_powergating_state(void *handle,

drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c

@@ -697,6 +697,11 @@ static const struct soc15_reg_golden golden_settings_gc_9_4_1_arct[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_FIFO_SIZES, 0xffffffff, 0x00000f00),
 };
 
+static const struct soc15_reg_rlcg rlcg_access_gc_9_0[] = {
+	{SOC15_REG_ENTRY(GC, 0, mmGRBM_GFX_INDEX)},
+	{SOC15_REG_ENTRY(GC, 0, mmSQ_IND_INDEX)},
+};
+
 static const u32 GFX_RLC_SRM_INDEX_CNTL_ADDR_OFFSETS[] =
 {
 	mmRLC_SRM_INDEX_CNTL_ADDR_0 - mmRLC_SRM_INDEX_CNTL_ADDR_0,
@@ -721,6 +726,63 @@ static const u32 GFX_RLC_SRM_INDEX_CNTL_DATA_OFFSETS[] =
 	mmRLC_SRM_INDEX_CNTL_DATA_7 - mmRLC_SRM_INDEX_CNTL_DATA_0,
 };
 
+void gfx_v9_0_rlcg_wreg(struct amdgpu_device *adev, u32 offset, u32 v)
+{
+	static void *scratch_reg0;
+	static void *scratch_reg1;
+	static void *scratch_reg2;
+	static void *scratch_reg3;
+	static void *spare_int;
+	static uint32_t grbm_cntl;
+	static uint32_t grbm_idx;
+	bool shadow;
+
+	scratch_reg0 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG0)*4;
+	scratch_reg1 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG1)*4;
+	scratch_reg2 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG2)*4;
+	scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG3)*4;
+	spare_int = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_BASE_IDX] + mmRLC_SPARE_INT)*4;
+
+	grbm_cntl = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_CNTL_BASE_IDX] + mmGRBM_GFX_CNTL;
+	grbm_idx = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_INDEX_BASE_IDX] + mmGRBM_GFX_INDEX;
+
+	if (amdgpu_sriov_runtime(adev)) {
+		pr_err("shoudn't call rlcg write register during runtime\n");
+		return;
+	}
+
+	if (offset == grbm_cntl || offset == grbm_idx)
+		shadow = true;
+
+	if (shadow) {
+		if (offset  == grbm_cntl)
+			writel(v, scratch_reg2);
+		else if (offset == grbm_idx)
+			writel(v, scratch_reg3);
+
+		writel(v, ((void __iomem *)adev->rmmio) + (offset * 4));
+	} else {
+		uint32_t i = 0;
+		uint32_t retries = 50000;
+
+		writel(v, scratch_reg0);
+		writel(offset | 0x80000000, scratch_reg1);
+		writel(1, spare_int);
+		for (i = 0; i < retries; i++) {
+			u32 tmp;
+
+			tmp = readl(scratch_reg1);
+			if (!(tmp & 0x80000000))
+				break;
+
+			udelay(10);
+		}
+		if (i >= retries)
+			pr_err("timeout: rlcg program reg:0x%05x failed !\n", offset);
+	}
+
+}
+
 #define VEGA10_GB_ADDR_CONFIG_GOLDEN 0x2a114042
 #define VEGA12_GB_ADDR_CONFIG_GOLDEN 0x24104041
 #define RAVEN_GB_ADDR_CONFIG_GOLDEN 0x24000042
@@ -1921,7 +1983,7 @@ static int gfx_v9_0_mec_init(struct amdgpu_device *adev)
 
 static uint32_t wave_read_ind(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t address)
 {
-	WREG32_SOC15(GC, 0, mmSQ_IND_INDEX,
+	WREG32_SOC15_RLC(GC, 0, mmSQ_IND_INDEX,
 		(wave << SQ_IND_INDEX__WAVE_ID__SHIFT) |
 		(simd << SQ_IND_INDEX__SIMD_ID__SHIFT) |
 		(address << SQ_IND_INDEX__INDEX__SHIFT) |
@@ -1933,7 +1995,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd,
 			   uint32_t wave, uint32_t thread,
 			   uint32_t regno, uint32_t num, uint32_t *out)
 {
-	WREG32_SOC15(GC, 0, mmSQ_IND_INDEX,
+	WREG32_SOC15_RLC(GC, 0, mmSQ_IND_INDEX,
 		(wave << SQ_IND_INDEX__WAVE_ID__SHIFT) |
 		(simd << SQ_IND_INDEX__SIMD_ID__SHIFT) |
 		(regno << SQ_IND_INDEX__INDEX__SHIFT) |
@@ -4908,6 +4970,35 @@ static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 	WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data);
 }
 
+static bool gfx_v9_0_check_rlcg_range(struct amdgpu_device *adev,
+					uint32_t offset,
+					struct soc15_reg_rlcg *entries, int arr_size)
+{
+	int i;
+	uint32_t reg;
+
+	if (!entries)
+		return false;
+
+	for (i = 0; i < arr_size; i++) {
+		const struct soc15_reg_rlcg *entry;
+
+		entry = &entries[i];
+		reg = adev->reg_offset[entry->hwip][entry->instance][entry->segment] + entry->reg;
+		if (offset == reg)
+			return true;
+	}
+
+	return false;
+}
+
+static bool gfx_v9_0_is_rlcg_access_range(struct amdgpu_device *adev, u32 offset)
+{
+	return gfx_v9_0_check_rlcg_range(adev, offset,
+					(void *)rlcg_access_gc_9_0,
+					ARRAY_SIZE(rlcg_access_gc_9_0));
+}
+
 static const struct amdgpu_rlc_funcs gfx_v9_0_rlc_funcs = {
 	.is_rlc_enabled = gfx_v9_0_is_rlc_enabled,
 	.set_safe_mode = gfx_v9_0_set_safe_mode,
@@ -4920,7 +5011,9 @@ static const struct amdgpu_rlc_funcs gfx_v9_0_rlc_funcs = {
 	.stop = gfx_v9_0_rlc_stop,
 	.reset = gfx_v9_0_rlc_reset,
 	.start = gfx_v9_0_rlc_start,
-	.update_spm_vmid = gfx_v9_0_update_spm_vmid
+	.update_spm_vmid = gfx_v9_0_update_spm_vmid,
+	.rlcg_wreg = gfx_v9_0_rlcg_wreg,
+	.is_rlcg_access_range = gfx_v9_0_is_rlcg_access_range,
 };
 
 static int gfx_v9_0_set_powergating_state(void *handle,

drivers/gpu/drm/amd/amdgpu/soc15.h

@@ -42,6 +42,13 @@ struct soc15_reg_golden {
 	u32	or_mask;
 };
 
+struct soc15_reg_rlcg {
+	u32	hwip;
+	u32	instance;
+	u32	segment;
+	u32	reg;
+};
+
 struct soc15_reg_entry {
 	uint32_t hwip;
 	uint32_t inst;

drivers/gpu/drm/amd/amdgpu/soc15_common.h

@@ -70,10 +70,9 @@
 		}						\
 	} while (0)
 
-#define AMDGPU_VIRT_SUPPORT_RLC_PRG_REG(a) (amdgpu_sriov_vf((a)) && !amdgpu_sriov_runtime((a)))
 #define WREG32_RLC(reg, value) \
 	do {							\
-		if (AMDGPU_VIRT_SUPPORT_RLC_PRG_REG(adev)) {    \
+		if (amdgpu_sriov_fullaccess(adev)) {    \
 			uint32_t i = 0;	\
 			uint32_t retries = 50000;	\
 			uint32_t r0 = adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG0;	\
@@ -98,7 +97,7 @@
 #define WREG32_SOC15_RLC_SHADOW(ip, inst, reg, value) \
 	do {							\
 		uint32_t target_reg = adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg;\
-		if (AMDGPU_VIRT_SUPPORT_RLC_PRG_REG(adev)) {    \
+		if (amdgpu_sriov_fullaccess(adev)) {    \
 			uint32_t r2 = adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG2;	\
 			uint32_t r3 = adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG3;	\
 			uint32_t grbm_cntl = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_CNTL_BASE_IDX] + mmGRBM_GFX_CNTL;   \