drm/amdgpu: implement querying ras error count for mmhub9.4

Get mmhub error counter by accessing EDC_CNT registers.

v2: Add mmhub_v9_4_ prefix for local static variable and function
Signed-off-by: NDennis Li <dennis.li@amd.com>
Reviewed-by: NHawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: NTao Zhou <tao.zhou1@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c

@@ -658,6 +658,9 @@ static void gmc_v9_0_set_mmhub_funcs(struct amdgpu_device *adev)
 	case CHIP_VEGA20:
 		adev->mmhub.funcs = &mmhub_v1_0_funcs;
 		break;
+	case CHIP_ARCTURUS:
+		adev->mmhub.funcs = &mmhub_v9_4_funcs;
+		break;
 	default:
 		break;
 	}

drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c

@@ -21,6 +21,7 @@
  *
  */
 #include "amdgpu.h"
+#include "amdgpu_ras.h"
 #include "mmhub_v9_4.h"
 
 #include "mmhub/mmhub_9_4_1_offset.h"
@@ -29,7 +30,7 @@
 #include "athub/athub_1_0_offset.h"
 #include "athub/athub_1_0_sh_mask.h"
 #include "vega10_enum.h"
-
+#include "soc15.h"
 #include "soc15_common.h"
 
 #define MMHUB_NUM_INSTANCES			2
@@ -651,3 +652,253 @@ void mmhub_v9_4_get_clockgating(struct amdgpu_device *adev, u32 *flags)
 	if (data & ATCL2_0_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK)
 		*flags |= AMD_CG_SUPPORT_MC_LS;
 }
+
+static const struct soc15_ras_field_entry mmhub_v9_4_ras_fields[] = {
+	{ "MMEA0_DRAMRD_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, DRAMRD_CMDMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, DRAMRD_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA0_DRAMWR_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, DRAMWR_CMDMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, DRAMWR_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA0_DRAMWR_DATAMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, DRAMWR_DATAMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, DRAMWR_DATAMEM_DED_COUNT),
+	},
+	{ "MMEA0_RRET_TAGMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, RRET_TAGMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, RRET_TAGMEM_DED_COUNT),
+	},
+	{ "MMEA0_WRET_TAGMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, WRET_TAGMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, WRET_TAGMEM_DED_COUNT),
+	},
+	{ "MMEA0_DRAMRD_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, DRAMRD_PAGEMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA0_DRAMWR_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, DRAMWR_PAGEMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA0_IORD_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, IORD_CMDMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA0_IOWR_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, IOWR_CMDMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA0_IOWR_DATAMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT, IOWR_DATAMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA0_GMIRD_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT2),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT2, GMIRD_CMDMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT2, GMIRD_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA0_GMIWR_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT2),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT2, GMIWR_CMDMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT2, GMIWR_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA0_GMIWR_DATAMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT2),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT2, GMIWR_DATAMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT2, GMIWR_DATAMEM_DED_COUNT),
+	},
+	{ "MMEA0_GMIRD_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT2),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT2, GMIRD_PAGEMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA0_GMIWR_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT2),
+	SOC15_REG_FIELD(MMEA0_EDC_CNT2, GMIWR_PAGEMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA0_DRAMRD_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA0_EDC_CNT3, DRAMRD_PAGEMEM_DED_COUNT),
+	},
+	{ "MMEA0_DRAMWR_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA0_EDC_CNT3, DRAMWR_PAGEMEM_DED_COUNT),
+	},
+	{ "MMEA0_IORD_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA0_EDC_CNT3, IORD_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA0_IOWR_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA0_EDC_CNT3, IOWR_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA0_IOWR_DATAMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA0_EDC_CNT3, IOWR_DATAMEM_DED_COUNT),
+	},
+	{ "MMEA0_GMIRD_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA0_EDC_CNT3, GMIRD_PAGEMEM_DED_COUNT),
+	},
+	{ "MMEA0_GMIWR_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA0_EDC_CNT3, GMIWR_PAGEMEM_DED_COUNT),
+	},
+	{ "MMEA1_DRAMRD_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, DRAMRD_CMDMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, DRAMRD_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA1_DRAMWR_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, DRAMWR_CMDMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, DRAMWR_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA1_DRAMWR_DATAMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, DRAMWR_DATAMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, DRAMWR_DATAMEM_DED_COUNT),
+	},
+	{ "MMEA1_RRET_TAGMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, RRET_TAGMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, RRET_TAGMEM_DED_COUNT),
+	},
+	{ "MMEA1_WRET_TAGMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, WRET_TAGMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, WRET_TAGMEM_DED_COUNT),
+	},
+	{ "MMEA1_DRAMRD_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, DRAMRD_PAGEMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA1_DRAMWR_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, DRAMWR_PAGEMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA1_IORD_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, IORD_CMDMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA1_IOWR_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, IOWR_CMDMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA1_IOWR_DATAMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT, IOWR_DATAMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA1_GMIRD_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT2),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT2, GMIRD_CMDMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT2, GMIRD_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA1_GMIWR_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT2),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT2, GMIWR_CMDMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT2, GMIWR_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA1_GMIWR_DATAMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT2),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT2, GMIWR_DATAMEM_SEC_COUNT),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT2, GMIWR_DATAMEM_DED_COUNT),
+	},
+	{ "MMEA1_GMIRD_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT2),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT2, GMIRD_PAGEMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA1_GMIWR_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT2),
+	SOC15_REG_FIELD(MMEA1_EDC_CNT2, GMIWR_PAGEMEM_SED_COUNT),
+	0, 0,
+	},
+	{ "MMEA1_DRAMRD_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA1_EDC_CNT3, DRAMRD_PAGEMEM_DED_COUNT),
+	},
+	{ "MMEA1_DRAMWR_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA1_EDC_CNT3, DRAMWR_PAGEMEM_DED_COUNT),
+	},
+	{ "MMEA1_IORD_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA1_EDC_CNT3, IORD_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA1_IOWR_CMDMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA1_EDC_CNT3, IOWR_CMDMEM_DED_COUNT),
+	},
+	{ "MMEA1_IOWR_DATAMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA1_EDC_CNT3, IOWR_DATAMEM_DED_COUNT),
+	},
+	{ "MMEA1_GMIRD_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA1_EDC_CNT3, GMIRD_PAGEMEM_DED_COUNT),
+	},
+	{ "MMEA1_GMIWR_PAGEMEM", SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT3),
+	0, 0,
+	SOC15_REG_FIELD(MMEA1_EDC_CNT3, GMIWR_PAGEMEM_DED_COUNT),
+	}
+};
+
+static const struct soc15_reg_entry mmhub_v9_4_edc_cnt_regs[] = {
+   { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT), 0, 0, 0},
+   { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT2), 0, 0, 0},
+   { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA0_EDC_CNT3), 0, 0, 0},
+   { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT), 0, 0, 0},
+   { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT2), 0, 0, 0},
+   { SOC15_REG_ENTRY(MMHUB, 0, mmMMEA1_EDC_CNT3), 0, 0, 0},
+};
+
+static int mmhub_v9_4_get_ras_error_count(const struct soc15_reg_entry *reg,
+	uint32_t value, uint32_t *sec_count, uint32_t *ded_count)
+{
+	uint32_t i;
+	uint32_t sec_cnt, ded_cnt;
+
+	for (i = 0; i < ARRAY_SIZE(mmhub_v9_4_ras_fields); i++) {
+		if(mmhub_v9_4_ras_fields[i].reg_offset != reg->reg_offset)
+			continue;
+
+		sec_cnt = (value &
+				mmhub_v9_4_ras_fields[i].sec_count_mask) >>
+				mmhub_v9_4_ras_fields[i].sec_count_shift;
+		if (sec_cnt) {
+			DRM_INFO("MMHUB SubBlock %s, SEC %d\n",
+				mmhub_v9_4_ras_fields[i].name,
+				sec_cnt);
+			*sec_count += sec_cnt;
+		}
+
+		ded_cnt = (value &
+				mmhub_v9_4_ras_fields[i].ded_count_mask) >>
+				mmhub_v9_4_ras_fields[i].ded_count_shift;
+		if (ded_cnt) {
+			DRM_INFO("MMHUB SubBlock %s, DED %d\n",
+				mmhub_v9_4_ras_fields[i].name,
+				ded_cnt);
+			*ded_count += ded_cnt;
+		}
+	}
+
+	return 0;
+}
+
+static void mmhub_v9_4_query_ras_error_count(struct amdgpu_device *adev,
+					   void *ras_error_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+	uint32_t sec_count = 0, ded_count = 0;
+	uint32_t i;
+	uint32_t reg_value;
+
+	err_data->ue_count = 0;
+	err_data->ce_count = 0;
+
+	for (i = 0; i < ARRAY_SIZE(mmhub_v9_4_edc_cnt_regs); i++) {
+		reg_value =
+			RREG32(SOC15_REG_ENTRY_OFFSET(mmhub_v9_4_edc_cnt_regs[i]));
+		if (reg_value)
+			mmhub_v9_4_get_ras_error_count(&mmhub_v9_4_edc_cnt_regs[i],
+				reg_value, &sec_count, &ded_count);
+	}
+
+	err_data->ce_count += sec_count;
+	err_data->ue_count += ded_count;
+}
+
+const struct amdgpu_mmhub_funcs mmhub_v9_4_funcs = {
+	.ras_late_init = amdgpu_mmhub_ras_late_init,
+	.query_ras_error_count = mmhub_v9_4_query_ras_error_count,
+};
\ No newline at end of file

drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.h

@@ -23,6 +23,8 @@
 #ifndef __MMHUB_V9_4_H__
 #define __MMHUB_V9_4_H__
 
+extern const struct amdgpu_mmhub_funcs mmhub_v9_4_funcs;
+
 u64 mmhub_v9_4_get_fb_location(struct amdgpu_device *adev);
 int mmhub_v9_4_gart_enable(struct amdgpu_device *adev);
 void mmhub_v9_4_gart_disable(struct amdgpu_device *adev);