From 322a7e005db78b8a46ead91b7e3df3514cb658f0 Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Thu, 2 Feb 2023 20:54:08 +0800 Subject: drm/amdgpu: Add common helper to query ras error (v2) Add common helper to query ras error status and log error information, including memory block id and erorr count. The helpers are applicable to IP blocks that follow the new ras error logging design. For IP blocks that don't support the new design, please still implement ip specific helper to query ras error. v2: optimize struct amdgpu_ras_err_status_reg_entry and the implementaion in helper (Lijo/Tao) Signed-off-by: Hawking Zhang Reviewed-by: Tao Zhou Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 54 +++++++++++++++++++++++++++++++++ 1 file changed, 54 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h index 17b3d1992e80..c820af7f1a4b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h @@ -314,6 +314,43 @@ enum amdgpu_ras_ret { AMDGPU_RAS_PT, }; +/* ras error status reisger fields */ +#define ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT 0x0 +#define ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK 0x00000001L +#define ERR_STATUS_LO__MEMORY_ID__SHIFT 0x18 +#define ERR_STATUS_LO__MEMORY_ID_MASK 0xFF000000L +#define ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT 0x2 +#define ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK 0x00000004L +#define ERR_STATUS__ERR_CNT__SHIFT 0x17 +#define ERR_STATUS__ERR_CNT_MASK 0x03800000L + +#define AMDGPU_RAS_REG_ENTRY(ip, inst, reg_lo, reg_hi) \ + ip##_HWIP, inst, reg_lo##_BASE_IDX, reg_lo, reg_hi##_BASE_IDX, reg_hi + +#define AMDGPU_RAS_REG_ENTRY_OFFSET(hwip, ip_inst, segment, reg) \ + (adev->reg_offset[hwip][ip_inst][segment] + (reg)) + +#define AMDGPU_RAS_ERR_INFO_VALID (1 << 0) +#define AMDGPU_RAS_ERR_STATUS_VALID (1 << 1) +#define AMDGPU_RAS_ERR_ADDRESS_VALID (1 << 2) + +struct amdgpu_ras_err_status_reg_entry { + uint32_t hwip; + uint32_t ip_inst; + uint32_t seg_lo; + uint32_t reg_lo; + uint32_t seg_hi; + uint32_t reg_hi; + uint32_t reg_inst; + uint32_t flags; + const char *block_name; +}; + +struct amdgpu_ras_memory_id_entry { + uint32_t memory_id; + const char *name; +}; + struct ras_common_if { enum amdgpu_ras_block block; enum amdgpu_ras_error_type type; @@ -696,4 +733,21 @@ int amdgpu_ras_set_context(struct amdgpu_device *adev, struct amdgpu_ras *ras_co int amdgpu_ras_register_ras_block(struct amdgpu_device *adev, struct amdgpu_ras_block_object *ras_block_obj); void amdgpu_ras_interrupt_fatal_error_handler(struct amdgpu_device *adev); +void amdgpu_ras_get_error_type_name(uint32_t err_type, char *err_type_name); +bool amdgpu_ras_inst_get_memory_id_field(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_entry, + uint32_t instance, + uint32_t *memory_id); +bool amdgpu_ras_inst_get_err_cnt_field(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_entry, + uint32_t instance, + unsigned long *err_cnt); +void amdgpu_ras_inst_query_ras_error_count(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_list, + uint32_t reg_list_size, + const struct amdgpu_ras_memory_id_entry *mem_list, + uint32_t mem_list_size, + uint32_t instance, + uint32_t err_type, + unsigned long *err_count); #endif -- cgit From e53a3250f76b8a0dd5b533bd0ce0dc821055e77d Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Fri, 3 Feb 2023 16:10:37 +0800 Subject: drm/amdgpu: Add common helper to reset ras error Add common helper to reset ras error status. It applies to IP blocks that follow the new ras error logging register design, and need to write 0 to reset the error status. For IP blocks that don't support the new design, please still implement ip specific helper. Signed-off-by: Hawking Zhang Reviewed-by: Tao Zhou Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h index c820af7f1a4b..e96333d0c269 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h @@ -750,4 +750,8 @@ void amdgpu_ras_inst_query_ras_error_count(struct amdgpu_device *adev, uint32_t instance, uint32_t err_type, unsigned long *err_count); +void amdgpu_ras_inst_reset_ras_error_count(struct amdgpu_device *adev, + const struct amdgpu_ras_err_status_reg_entry *reg_list, + uint32_t reg_list_size, + uint32_t instance); #endif -- cgit From 2c22ed0bdb0cb6da9408593eafa6137325576017 Mon Sep 17 00:00:00 2001 From: Tao Zhou Date: Mon, 27 Feb 2023 18:25:23 +0800 Subject: drm/amdgpu: add instance mask for RAS inject User can specify injected instances by the mask. For backward compatibility, the mask value is incorporated into sub block index without interface change of RAS TA. User uses logical mask and driver should convert it to physical value before sending it to RAS TA. v2: update parameter name. Signed-off-by: Tao Zhou Reviewed-by: Hawking Zhang Reviewed-by: Stanley.Yang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h index e96333d0c269..bc43f7db17cc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h @@ -32,6 +32,11 @@ struct amdgpu_iv_entry; #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS (0x1 << 0) +/* position of instance value in sub_block_index of + * ta_ras_trigger_error_input, the sub block uses lower 12 bits + */ +#define AMDGPU_RAS_INST_MASK 0xfffff000 +#define AMDGPU_RAS_INST_SHIFT 0xc enum amdgpu_ras_block { AMDGPU_RAS_BLOCK__UMC = 0, @@ -508,6 +513,7 @@ struct ras_inject_if { struct ras_common_if head; uint64_t address; uint64_t value; + uint32_t instance_mask; }; struct ras_cure_if { @@ -545,7 +551,8 @@ struct amdgpu_ras_block_object { }; struct amdgpu_ras_block_hw_ops { - int (*ras_error_inject)(struct amdgpu_device *adev, void *inject_if); + int (*ras_error_inject)(struct amdgpu_device *adev, + void *inject_if, uint32_t instance_mask); void (*query_ras_error_count)(struct amdgpu_device *adev, void *ras_error_status); void (*query_ras_error_status)(struct amdgpu_device *adev); void (*query_ras_error_address)(struct amdgpu_device *adev, void *ras_error_status); -- cgit From 6c47a79b3b8ba91faf89f9866da2ec16aac979e7 Mon Sep 17 00:00:00 2001 From: YiPeng Chai Date: Tue, 16 May 2023 17:34:17 +0800 Subject: drm/amdgpu: perform mode2 reset for sdma fed error on gfx v11_0_3 perform mode2 reset for sdma fed error on gfx v11_0_3. Signed-off-by: YiPeng Chai Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h index bc43f7db17cc..46bf1889a9d7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h @@ -339,6 +339,8 @@ enum amdgpu_ras_ret { #define AMDGPU_RAS_ERR_STATUS_VALID (1 << 1) #define AMDGPU_RAS_ERR_ADDRESS_VALID (1 << 2) +#define AMDGPU_RAS_GPU_RESET_MODE2_RESET (0x1 << 0) + struct amdgpu_ras_err_status_reg_entry { uint32_t hwip; uint32_t ip_inst; @@ -427,6 +429,9 @@ struct amdgpu_ras { /* Indicates smu whether need update bad channel info */ bool update_channel_flag; + + /* Record special requirements of gpu reset caller */ + uint32_t gpu_reset_flags; }; struct ras_fs_data { -- cgit