drm/amdgpu: indirect register access for nv12 sriov

1. expand rlcg interface for gc & mmhub indirect access
2. add rlcg interface for no kiq

v2: squash in fix for gfx9 (Changfeng)
Signed-off-by: Peng Ju Zhou <PengJu.Zhou@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_device.c

@@ -490,7 +490,7 @@ 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 adev->gfx.rlc.funcs->rlcg_wreg(adev, reg, v);
+			return adev->gfx.rlc.funcs->rlcg_wreg(adev, reg, v, 0);
 	} else {
 		writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
 	}

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

@@ -127,7 +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);
+	void (*rlcg_wreg)(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag);
+	u32 (*rlcg_rreg)(struct amdgpu_device *adev, u32 offset, u32 flag);
 	bool (*is_rlcg_access_range)(struct amdgpu_device *adev, uint32_t reg);
 };
 

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

@@ -173,6 +173,11 @@
 #define mmGC_THROTTLE_CTRL_Sienna_Cichlid              0x2030
 #define mmGC_THROTTLE_CTRL_Sienna_Cichlid_BASE_IDX     0
 
+#define GFX_RLCG_GC_WRITE_OLD	(0x8 << 28)
+#define GFX_RLCG_GC_WRITE	(0x0 << 28)
+#define GFX_RLCG_GC_READ	(0x1 << 28)
+#define GFX_RLCG_MMHUB_WRITE	(0x2 << 28)
+
 MODULE_FIRMWARE("amdgpu/navi10_ce.bin");
 MODULE_FIRMWARE("amdgpu/navi10_pfp.bin");
 MODULE_FIRMWARE("amdgpu/navi10_me.bin");
@@ -1418,38 +1423,127 @@ 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 bool gfx_v10_is_rlcg_rw(struct amdgpu_device *adev, u32 offset, uint32_t *flag, bool write)
+{
+	/* always programed by rlcg, only for gc */
+	if (offset == SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_HI) ||
+	    offset == SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_LO) ||
+	    offset == SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_LENGTH) ||
+	    offset == SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_CNTL) ||
+	    offset == SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_INDEX) ||
+	    offset == SOC15_REG_OFFSET(GC, 0, mmCP_ME_CNTL)) {
+		if (!amdgpu_sriov_reg_indirect_gc(adev))
+			*flag = GFX_RLCG_GC_WRITE_OLD;
+		else
+			*flag = write ? GFX_RLCG_GC_WRITE : GFX_RLCG_GC_READ;
+
+		return true;
+	}
+
+	/* currently support gc read/write, mmhub write */
+	if (offset >= SOC15_REG_OFFSET(GC, 0, mmSDMA0_DEC_START) &&
+	    offset <= SOC15_REG_OFFSET(GC, 0, mmRLC_GTS_OFFSET_MSB)) {
+		if (amdgpu_sriov_reg_indirect_gc(adev))
+			*flag = write ? GFX_RLCG_GC_WRITE : GFX_RLCG_GC_READ;
+		else
+			return false;
+	} else {
+		if (amdgpu_sriov_reg_indirect_mmhub(adev))
+			*flag = GFX_RLCG_MMHUB_WRITE;
+		else
+			return false;
+	}
+
+	return true;
+}
+
+static u32 gfx_v10_rlcg_rw(struct amdgpu_device *adev, u32 offset, u32 v, uint32_t flag)
 {
 	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;
+	u32 ret = 0;
+
+	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_REG0_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 (offset == grbm_cntl || offset == grbm_idx) {
+		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 {
+		writel(v, scratch_reg0);
+		writel(offset | flag, scratch_reg1);
+		writel(1, spare_int);
+		for (i = 0; i < retries; i++) {
+			u32 tmp;
+
+			tmp = readl(scratch_reg1);
+			if (!(tmp & flag))
+				break;
 
-	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;
-	spare_int = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_BASE_IDX] + mmRLC_SPARE_INT)*4;
+			udelay(10);
+		}
 
-	if (amdgpu_sriov_runtime(adev)) {
-		pr_err("shouldn't call rlcg write register during runtime\n");
-		return;
+		if (i >= retries)
+			pr_err("timeout: rlcg program reg:0x%05x failed !\n", offset);
 	}
 
-	writel(v, scratch_reg0);
-	writel(offset | 0x80000000, scratch_reg1);
-	writel(1, spare_int);
-	for (i = 0; i < retries; i++) {
-		u32 tmp;
+	ret = readl(scratch_reg0);
 
-		tmp = readl(scratch_reg1);
-		if (!(tmp & 0x80000000))
-			break;
+	return ret;
+}
 
-		udelay(10);
+static void gfx_v10_rlcg_wreg(struct amdgpu_device *adev, u32 offset, u32 value, u32 flag)
+{
+	uint32_t rlcg_flag;
+
+	if (amdgpu_sriov_fullaccess(adev) &&
+	    gfx_v10_is_rlcg_rw(adev, offset, &rlcg_flag, 1)) {
+		gfx_v10_rlcg_rw(adev, offset, value, rlcg_flag);
+
+		return;
 	}
+	if (flag & AMDGPU_REGS_NO_KIQ)
+		WREG32_NO_KIQ(offset, value);
+	else
+		WREG32(offset, value);
+}
+
+static u32 gfx_v10_rlcg_rreg(struct amdgpu_device *adev, u32 offset, u32 flag)
+{
+	uint32_t rlcg_flag;
 
-	if (i >= retries)
-		pr_err("timeout: rlcg program reg:0x%05x failed !\n", offset);
+	if (amdgpu_sriov_fullaccess(adev) &&
+	    gfx_v10_is_rlcg_rw(adev, offset, &rlcg_flag, 0))
+		return gfx_v10_rlcg_rw(adev, offset, 0, rlcg_flag);
+
+	if (flag & AMDGPU_REGS_NO_KIQ)
+		return RREG32_NO_KIQ(offset);
+	else
+		return RREG32(offset);
+
+	return 0;
 }
 
 static const struct soc15_reg_golden golden_settings_gc_10_1_nv14[] =
@@ -7884,6 +7978,7 @@ static const struct amdgpu_rlc_funcs gfx_v10_0_rlc_funcs_sriov = {
 	.start = gfx_v10_0_rlc_start,
 	.update_spm_vmid = gfx_v10_0_update_spm_vmid,
 	.rlcg_wreg = gfx_v10_rlcg_wreg,
+	.rlcg_rreg = gfx_v10_rlcg_rreg,
 	.is_rlcg_access_range = gfx_v10_0_is_rlcg_access_range,
 };
 

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

@@ -734,7 +734,7 @@ static const u32 GFX_RLC_SRM_INDEX_CNTL_DATA_OFFSETS[] =
 	mmRLC_SRM_INDEX_CNTL_DATA_7 - mmRLC_SRM_INDEX_CNTL_DATA_0,
 };
 
-static void gfx_v9_0_rlcg_wreg(struct amdgpu_device *adev, u32 offset, u32 v)
+static void gfx_v9_0_rlcg_rw(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag)
 {
 	static void *scratch_reg0;
 	static void *scratch_reg1;
@@ -787,6 +787,20 @@ static void gfx_v9_0_rlcg_wreg(struct amdgpu_device *adev, u32 offset, u32 v)
 
 }
 
+static void gfx_v9_0_rlcg_wreg(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag)
+{
+	if (amdgpu_sriov_fullaccess(adev)) {
+		gfx_v9_0_rlcg_rw(adev, offset, v, flag);
+
+		return;
+	}
+
+	if (flag & AMDGPU_REGS_NO_KIQ)
+		WREG32_NO_KIQ(offset, v);
+	else
+		WREG32(offset, v);
+}
+
 #define VEGA10_GB_ADDR_CONFIG_GOLDEN 0x2a114042
 #define VEGA12_GB_ADDR_CONFIG_GOLDEN 0x24104041
 #define RAVEN_GB_ADDR_CONFIG_GOLDEN 0x24000042

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

@@ -77,27 +77,11 @@
 })
 
 #define WREG32_RLC(reg, value) \
-	do {							\
-		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;	\
-			uint32_t r1 = adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG1;	\
-			uint32_t spare_int = adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_BASE_IDX] + mmRLC_SPARE_INT;	\
-			WREG32(r0, value);	\
-			WREG32(r1, (reg | 0x80000000));	\
-			WREG32(spare_int, 0x1);	\
-			for (i = 0; i < retries; i++) {	\
-				u32 tmp = RREG32(r1);	\
-				if (!(tmp & 0x80000000))	\
-					break;	\
-				udelay(10);	\
-			}	\
-			if (i >= retries)	\
-				pr_err("timeout: rlcg program reg:0x%05x failed !\n", reg);	\
-		} else {	\
-			WREG32(reg, value); \
-		}	\
+	do { \
+		if (adev->gfx.rlc.funcs->rlcg_wreg) \
+			adev->gfx.rlc.funcs->rlcg_wreg(adev, reg, value, 0); \
+		else \
+			WREG32(reg, value);	\
 	} while (0)
 
 #define WREG32_RLC_EX(prefix, reg, value) \
@@ -125,23 +109,24 @@
 	} while (0)
 
 #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_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;   \
-			uint32_t grbm_idx = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_INDEX_BASE_IDX] + mmGRBM_GFX_INDEX;   \
-			if (target_reg == grbm_cntl) \
-				WREG32(r2, value);	\
-			else if (target_reg == grbm_idx) \
-				WREG32(r3, value);	\
-			WREG32(target_reg, value);	\
-		} else {	\
-			WREG32(target_reg, value); \
-		}	\
+	WREG32_RLC((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg), value)
+
+#define RREG32_RLC(reg) \
+	(adev->gfx.rlc.funcs->rlcg_rreg ? \
+		adev->gfx.rlc.funcs->rlcg_rreg(adev, reg, 0) : RREG32(reg))
+
+#define WREG32_RLC_NO_KIQ(reg, value) \
+	do { \
+		if (adev->gfx.rlc.funcs->rlcg_wreg) \
+			adev->gfx.rlc.funcs->rlcg_wreg(adev, reg, value, AMDGPU_REGS_NO_KIQ); \
+		else \
+			WREG32_NO_KIQ(reg, value);	\
 	} while (0)
 
+#define RREG32_RLC_NO_KIQ(reg) \
+	(adev->gfx.rlc.funcs->rlcg_rreg ? \
+		adev->gfx.rlc.funcs->rlcg_rreg(adev, reg, AMDGPU_REGS_NO_KIQ) : RREG32_NO_KIQ(reg))
+
 #define WREG32_SOC15_RLC_SHADOW_EX(prefix, ip, inst, reg, value) \
 	do {							\
 		uint32_t target_reg = adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg;\
@@ -160,10 +145,13 @@
 		}	\
 	} while (0)
 
+#define RREG32_SOC15_RLC(ip, inst, reg) \
+	RREG32_RLC(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg)
+
 #define WREG32_SOC15_RLC(ip, inst, reg, value) \
 	do {							\
-			uint32_t target_reg = adev->reg_offset[GC_HWIP][0][reg##_BASE_IDX] + reg;\
-			WREG32_RLC(target_reg, value); \
+		uint32_t target_reg = adev->reg_offset[ip##_HWIP][0][reg##_BASE_IDX] + reg;\
+		WREG32_RLC(target_reg, value); \
 	} while (0)
 
 #define WREG32_SOC15_RLC_EX(prefix, ip, inst, reg, value) \
@@ -173,11 +161,14 @@
 	} while (0)
 
 #define WREG32_FIELD15_RLC(ip, idx, reg, field, val)   \
-    WREG32_RLC((adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg), \
-    (RREG32(adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg) \
-    & ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
+	WREG32_RLC((adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg), \
+	(RREG32_RLC(adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg) \
+	& ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
 
 #define WREG32_SOC15_OFFSET_RLC(ip, inst, reg, offset, value) \
-    WREG32_RLC(((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset), value)
+	WREG32_RLC(((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset), value)
+
+#define RREG32_SOC15_OFFSET_RLC(ip, inst, reg, offset) \
+	RREG32_RLC(((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset))
 
 #endif