drm/amdgpu: fix convert bad page retiremt

Pmfw read ecc info registers and store values in
eccinfo_table in the following order

umc0 ch_inst 0, 1, 2 ... 7
umc1 ch_inst 0, 1, 2 ... 7
...
umc3 ch_inst 0, 1, 2 ... 7

Driver should convert eccinfo_table_idx to channel_index according
to channel_idx_tbl.

Signed-off-by: Stanley.Yang <Stanley.Yang@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c
index 526de1c..f5a1ba7 100644 (file)
--- a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c
@@ -58,29 +58,33 @@ static inline uint32_t get_umc_v6_7_channel_index(struct amdgpu_device *adev,
 }
 
 static void umc_v6_7_ecc_info_query_correctable_error_count(struct amdgpu_device *adev,
-                                                  uint32_t channel_index,
+                                                  uint32_t umc_inst, uint32_t ch_inst,
                                                   unsigned long *error_count)
 {
        uint64_t mc_umc_status;
+       uint32_t eccinfo_table_idx;
        struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
+       eccinfo_table_idx = umc_inst * adev->umc.channel_inst_num + ch_inst;
        /* check for SRAM correctable error
          MCUMC_STATUS is a 64 bit register */
-       mc_umc_status = ras->umc_ecc.ecc[channel_index].mca_umc_status;
+       mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
        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_ecc_info_querry_uncorrectable_error_count(struct amdgpu_device *adev,
-                                                     uint32_t channel_index,
+                                                         uint32_t umc_inst, uint32_t ch_inst,
                                                      unsigned long *error_count)
 {
        uint64_t mc_umc_status;
+       uint32_t eccinfo_table_idx;
        struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
+       eccinfo_table_idx = umc_inst * adev->umc.channel_inst_num + ch_inst;
        /* check the MCUMC_STATUS */
-       mc_umc_status = ras->umc_ecc.ecc[channel_index].mca_umc_status;
+       mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
        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 ||
@@ -97,19 +101,15 @@ static void umc_v6_7_ecc_info_query_ras_error_count(struct amdgpu_device *adev,
 
        uint32_t umc_inst        = 0;
        uint32_t ch_inst         = 0;
-       uint32_t channel_index   = 0;
 
        /*TODO: driver needs to toggle DF Cstate to ensure
         * safe access of UMC registers. Will add the protection */
        LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
-               channel_index = get_umc_v6_7_channel_index(adev,
-                                                        umc_inst,
-                                                        ch_inst);
                umc_v6_7_ecc_info_query_correctable_error_count(adev,
-                                                     channel_index,
+                                                     umc_inst, ch_inst,
                                                      &(err_data->ce_count));
                umc_v6_7_ecc_info_querry_uncorrectable_error_count(adev,
-                                                         channel_index,
+                                                     umc_inst, ch_inst,
                                                          &(err_data->ue_count));
        }
 }
@@ -122,12 +122,14 @@ static void umc_v6_7_ecc_info_query_error_address(struct amdgpu_device *adev,
        uint64_t mc_umc_status, err_addr, retired_page;
        struct eeprom_table_record *err_rec;
        uint32_t channel_index;
+       uint32_t eccinfo_table_idx;
        struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
+       eccinfo_table_idx = umc_inst * adev->umc.channel_inst_num + ch_inst;
        channel_index =
                adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst];
 
-       mc_umc_status = ras->umc_ecc.ecc[channel_index].mca_umc_status;
+       mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
 
        if (mc_umc_status == 0)
                return;
@@ -142,7 +144,7 @@ static void umc_v6_7_ecc_info_query_error_address(struct amdgpu_device *adev,
            (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
            REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)) {
 
-               err_addr = ras->umc_ecc.ecc[channel_index].mca_umc_addr;
+               err_addr = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_addr;
                err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
 
                /* translate umc channel address to soc pa, 3 parts are included */