diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c | 37 |
1 files changed, 28 insertions, 9 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c index b12808c0c331..aab8077e5098 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c @@ -161,6 +161,7 @@ static bool __is_ras_eeprom_supported(struct amdgpu_device *adev) case IP_VERSION(13, 0, 10): return true; case IP_VERSION(13, 0, 6): + case IP_VERSION(13, 0, 14): return (adev->gmc.is_app_apu) ? false : true; default: return false; @@ -222,6 +223,7 @@ static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev, return true; case IP_VERSION(13, 0, 6): case IP_VERSION(13, 0, 10): + case IP_VERSION(13, 0, 14): control->i2c_address = EEPROM_I2C_MADDR_4; return true; default: @@ -404,6 +406,22 @@ static int amdgpu_ras_eeprom_correct_header_tag( return res; } +static void amdgpu_ras_set_eeprom_table_version(struct amdgpu_ras_eeprom_control *control) +{ + struct amdgpu_device *adev = to_amdgpu_device(control); + struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr; + + switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) { + case IP_VERSION(8, 10, 0): + case IP_VERSION(12, 0, 0): + hdr->version = RAS_TABLE_VER_V2_1; + return; + default: + hdr->version = RAS_TABLE_VER_V1; + return; + } +} + /** * amdgpu_ras_eeprom_reset_table -- Reset the RAS EEPROM table * @control: pointer to control structure @@ -423,11 +441,7 @@ int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control) mutex_lock(&control->ras_tbl_mutex); hdr->header = RAS_TABLE_HDR_VAL; - if (adev->umc.ras && - adev->umc.ras->set_eeprom_table_version) - adev->umc.ras->set_eeprom_table_version(hdr); - else - hdr->version = RAS_TABLE_VER_V1; + amdgpu_ras_set_eeprom_table_version(control); if (hdr->version == RAS_TABLE_VER_V2_1) { hdr->first_rec_offset = RAS_RECORD_START_V2_1; @@ -736,6 +750,9 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control) control->tbl_rai.health_percent = 0; } + if (amdgpu_bad_page_threshold != -1) + ras->is_rma = true; + /* ignore the -ENOTSUPP return value */ amdgpu_dpm_send_rma_reason(adev); } @@ -994,6 +1011,9 @@ Out: uint32_t amdgpu_ras_eeprom_max_record_count(struct amdgpu_ras_eeprom_control *control) { + /* get available eeprom table version first before eeprom table init */ + amdgpu_ras_set_eeprom_table_version(control); + if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1) return RAS_MAX_RECORD_COUNT_V2_1; else @@ -1307,8 +1327,7 @@ Out: return res == RAS_TABLE_V2_1_INFO_SIZE ? 0 : res; } -int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control, - bool *exceed_err_limit) +int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control) { struct amdgpu_device *adev = to_amdgpu_device(control); unsigned char buf[RAS_TABLE_HEADER_SIZE] = { 0 }; @@ -1316,7 +1335,7 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control, struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); int res; - *exceed_err_limit = false; + ras->is_rma = false; if (!__is_ras_eeprom_supported(adev)) return 0; @@ -1408,7 +1427,7 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control, dev_warn(adev->dev, "GPU will be initialized due to bad_page_threshold = -1."); res = 0; } else { - *exceed_err_limit = true; + ras->is_rma = true; dev_err(adev->dev, "RAS records:%d exceed threshold:%d, " "GPU will not be initialized. Replace this GPU or increase the threshold", |