drm/amdgpu: add helper funtion to query umc ras error

Add helper functions to query correctable and
uncorrectable umc ras error.
Signed-off-by: NHawking Zhang <Hawking.Zhang@amd.com>
Acked-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NJohn Clements <John.Clements@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/umc_v6_7.c

@@ -24,6 +24,76 @@
 #include "amdgpu_ras.h"
 #include "amdgpu.h"
 
+#include "umc/umc_6_7_0_offset.h"
+#include "umc/umc_6_7_0_sh_mask.h"
+
+static void umc_v6_7_query_correctable_error_count(struct amdgpu_device *adev,
+						   uint32_t umc_reg_offset,
+						   unsigned long *error_count)
+{
+	uint32_t ecc_err_cnt_sel, ecc_err_cnt_sel_addr;
+	uint32_t ecc_err_cnt, ecc_err_cnt_addr;
+	uint64_t mc_umc_status;
+	uint32_t mc_umc_status_addr;
+
+	/* UMC 6_1_1 registers */
+	ecc_err_cnt_sel_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_0_EccErrCntSel);
+	ecc_err_cnt_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_0_EccErrCnt);
+	mc_umc_status_addr =
+		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0);
+
+	/* select the lower chip and check the error count */
+	ecc_err_cnt_sel = RREG32_PCIE((ecc_err_cnt_sel_addr + umc_reg_offset) * 4);
+	ecc_err_cnt_sel = REG_SET_FIELD(ecc_err_cnt_sel, UMCCH0_0_EccErrCntSel,
+					EccErrCntCsSel, 0);
+	WREG32_PCIE((ecc_err_cnt_sel_addr + umc_reg_offset) * 4, ecc_err_cnt_sel);
+
+	ecc_err_cnt = RREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4);
+	*error_count +=
+		(REG_GET_FIELD(ecc_err_cnt, UMCCH0_0_EccErrCnt, EccErrCnt) -
+		 UMC_V6_7_CE_CNT_INIT);
+
+	/* select the higher chip and check the err counter */
+	ecc_err_cnt_sel = REG_SET_FIELD(ecc_err_cnt_sel, UMCCH0_0_EccErrCntSel,
+					EccErrCntCsSel, 1);
+	WREG32_PCIE((ecc_err_cnt_sel_addr + umc_reg_offset) * 4, ecc_err_cnt_sel);
+
+	ecc_err_cnt = RREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4);
+	*error_count +=
+		(REG_GET_FIELD(ecc_err_cnt, UMCCH0_0_EccErrCnt, EccErrCnt) -
+		 UMC_V6_7_CE_CNT_INIT);
+
+	/* check for SRAM correctable error
+	  MCUMC_STATUS is a 64 bit register */
+	mc_umc_status = RREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4);
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)
+		*error_count += 1;
+}
+
+static void umc_v6_7_querry_uncorrectable_error_count(struct amdgpu_device *adev,
+						      uint32_t umc_reg_offset,
+						      unsigned long *error_count)
+{
+	uint64_t mc_umc_status;
+	uint32_t mc_umc_status_addr;
+
+	mc_umc_status_addr =
+		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0);
+
+	/* check the MCUMC_STATUS */
+	mc_umc_status = RREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4);
+	if ((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
+	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1))
+		*error_count += 1;
+}
+
 const struct amdgpu_umc_funcs umc_v6_7_funcs = {
 	.ras_late_init = amdgpu_umc_ras_late_init,
 };

drivers/gpu/drm/amd/amdgpu/umc_v6_7.h

@@ -23,6 +23,13 @@
 #ifndef __UMC_V6_7_H__
 #define __UMC_V6_7_H__
 
+/* EccErrCnt max value */
+#define UMC_V6_7_CE_CNT_MAX		0xffff
+/* umc ce interrupt threshold */
+#define UMC_V6_7_CE_INT_THRESHOLD	0xffff
+/* umc ce count initial value */
+#define UMC_V6_7_CE_CNT_INIT	(UMC_V6_7_CE_CNT_MAX - UMC_V6_7_CE_INT_THRESHOLD)
+
 extern const struct amdgpu_umc_funcs umc_v6_7_funcs;
 
 #endif