aboutsummaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h33
1 files changed, 10 insertions, 23 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
index dcf1f3dbb5c4..669720a9c60a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
@@ -28,7 +28,6 @@
#include <linux/list.h>
#include <linux/kfifo.h>
#include <linux/radix-tree.h>
-#include <linux/siphash.h>
#include "ta_ras_if.h"
#include "amdgpu_ras_eeprom.h"
#include "amdgpu_smuio.h"
@@ -47,6 +46,8 @@ struct amdgpu_iv_entry;
#define AMDGPU_RAS_GPU_ERR_SOCKET_ID(x) AMDGPU_GET_REG_FIELD(x, 10, 8)
#define AMDGPU_RAS_GPU_ERR_AID_ID(x) AMDGPU_GET_REG_FIELD(x, 12, 11)
#define AMDGPU_RAS_GPU_ERR_HBM_ID(x) AMDGPU_GET_REG_FIELD(x, 14, 13)
+#define AMDGPU_RAS_GPU_ERR_DATA_ABORT(x) AMDGPU_GET_REG_FIELD(x, 29, 29)
+#define AMDGPU_RAS_GPU_ERR_UNKNOWN(x) AMDGPU_GET_REG_FIELD(x, 30, 30)
#define AMDGPU_RAS_BOOT_STATUS_POLLING_LIMIT 100
#define AMDGPU_RAS_BOOT_STEADY_STATUS 0xBA
@@ -476,16 +477,15 @@ struct ras_err_pages {
};
struct ras_ecc_err {
- u64 hash_index;
uint64_t status;
uint64_t ipid;
uint64_t addr;
+ uint64_t pa_pfn;
struct ras_err_pages err_pages;
};
struct ras_ecc_log_info {
struct mutex lock;
- siphash_key_t ecc_key;
struct radix_tree_root de_page_tree;
uint64_t de_queried_count;
uint64_t prev_de_queried_count;
@@ -572,19 +572,11 @@ struct ras_fs_data {
char debugfs_name[32];
};
-struct ras_err_addr {
- struct list_head node;
- uint64_t err_status;
- uint64_t err_ipid;
- uint64_t err_addr;
-};
-
struct ras_err_info {
struct amdgpu_smuio_mcm_config_info mcm_info;
u64 ce_count;
u64 ue_count;
u64 de_count;
- struct list_head err_addr_list;
};
struct ras_err_node {
@@ -941,14 +933,14 @@ void amdgpu_ras_inst_reset_ras_error_count(struct amdgpu_device *adev,
int amdgpu_ras_error_data_init(struct ras_err_data *err_data);
void amdgpu_ras_error_data_fini(struct ras_err_data *err_data);
int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data,
- struct amdgpu_smuio_mcm_config_info *mcm_info,
- struct ras_err_addr *err_addr, u64 count);
+ struct amdgpu_smuio_mcm_config_info *mcm_info,
+ u64 count);
int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
- struct amdgpu_smuio_mcm_config_info *mcm_info,
- struct ras_err_addr *err_addr, u64 count);
+ struct amdgpu_smuio_mcm_config_info *mcm_info,
+ u64 count);
int amdgpu_ras_error_statistic_de_count(struct ras_err_data *err_data,
- struct amdgpu_smuio_mcm_config_info *mcm_info,
- struct ras_err_addr *err_addr, u64 count);
+ struct amdgpu_smuio_mcm_config_info *mcm_info,
+ u64 count);
void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32 num_instances);
int amdgpu_ras_bind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk,
const struct aca_info *aca_info, void *data);
@@ -957,12 +949,6 @@ int amdgpu_ras_unbind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk)
ssize_t amdgpu_ras_aca_sysfs_read(struct device *dev, struct device_attribute *attr,
struct aca_handle *handle, char *buf, void *data);
-void amdgpu_ras_add_mca_err_addr(struct ras_err_info *err_info,
- struct ras_err_addr *err_addr);
-
-void amdgpu_ras_del_mca_err_addr(struct ras_err_info *err_info,
- struct ras_err_addr *mca_err_addr);
-
void amdgpu_ras_set_fed(struct amdgpu_device *adev, bool status);
bool amdgpu_ras_get_fed_status(struct amdgpu_device *adev);
@@ -982,4 +968,5 @@ __printf(3, 4)
void amdgpu_ras_event_log_print(struct amdgpu_device *adev, u64 event_id,
const char *fmt, ...);
+bool amdgpu_ras_is_rma(struct amdgpu_device *adev);
#endif