[PATCH 20/20] drm/amdkfd: add RAS ECC event support (v2)
Alex Deucher
alexdeucher at gmail.com
Tue Mar 5 20:41:36 UTC 2019
From: Eric Huang <JinhuiEric.Huang at amd.com>
RAS ECC event will combine with GPU reset event, due to
ECC interrupts are caused by uncorrectable error that triggers
GPU reset.
v2: Fix misleading-indentation warning
Signed-off-by: Eric Huang <JinhuiEric.Huang at amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling at amd.com>
Reviewed-by: Alex Deucher <alexander.deucher at amd.com>
Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 1 +
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 1 +
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 1 +
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 2 ++
drivers/gpu/drm/amd/amdkfd/kfd_device.c | 11 +++++++++++
drivers/gpu/drm/amd/amdkfd/kfd_events.c | 18 +++++++++++++++++-
drivers/gpu/drm/amd/amdkfd/kfd_priv.h | 3 +++
include/uapi/linux/kfd_ioctl.h | 12 +++++++++++-
8 files changed, 47 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index 0e1711a75b68..e6a503760b62 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -229,5 +229,6 @@ int kgd2kfd_quiesce_mm(struct mm_struct *mm);
int kgd2kfd_resume_mm(struct mm_struct *mm);
int kgd2kfd_schedule_evict_and_restore_process(struct mm_struct *mm,
struct dma_fence *fence);
+void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd);
#endif /* AMDGPU_AMDKFD_H_INCLUDED */
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 88c45f990f05..6bb71f6ee18e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -4805,6 +4805,7 @@ static int gfx_v9_0_process_ras_data_cb(struct amdgpu_device *adev,
struct amdgpu_iv_entry *entry)
{
/* TODO ue will trigger an interrupt. */
+ kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
amdgpu_ras_reset_gpu(adev, 0);
return AMDGPU_RAS_UE;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 1a164fe2dd9a..84904bd680df 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -354,6 +354,7 @@ static int gmc_v9_0_ecc_interrupt_state(struct amdgpu_device *adev,
static int gmc_v9_0_process_ras_data_cb(struct amdgpu_device *adev,
struct amdgpu_iv_entry *entry)
{
+ kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
amdgpu_ras_reset_gpu(adev, 0);
return AMDGPU_RAS_UE;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
index 058b9daec514..f7a6fafd70ae 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
@@ -1851,6 +1851,8 @@ static int sdma_v4_0_process_ras_data_cb(struct amdgpu_device *adev,
return 0;
}
+ kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
+
amdgpu_ras_reset_gpu(adev, 0);
return AMDGPU_RAS_UE;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index 8be9677c0c07..b3cdbf79f47b 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -466,6 +466,8 @@ struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd,
memset(&kfd->doorbell_available_index, 0,
sizeof(kfd->doorbell_available_index));
+ atomic_set(&kfd->sram_ecc_flag, 0);
+
return kfd;
}
@@ -661,6 +663,9 @@ int kgd2kfd_post_reset(struct kfd_dev *kfd)
return ret;
count = atomic_dec_return(&kfd_locked);
WARN_ONCE(count != 0, "KFD reset ref. error");
+
+ atomic_set(&kfd->sram_ecc_flag, 0);
+
return 0;
}
@@ -1024,6 +1029,12 @@ int kfd_gtt_sa_free(struct kfd_dev *kfd, struct kfd_mem_obj *mem_obj)
return 0;
}
+void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd)
+{
+ if (kfd)
+ atomic_inc(&kfd->sram_ecc_flag);
+}
+
#if defined(CONFIG_DEBUG_FS)
/* This function will send a package to HIQ to hang the HWS
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_events.c
index e9f0e0a1b41c..6e1d41c5bf86 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_events.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_events.c
@@ -1011,25 +1011,41 @@ void kfd_signal_vm_fault_event(struct kfd_dev *dev, unsigned int pasid,
void kfd_signal_reset_event(struct kfd_dev *dev)
{
struct kfd_hsa_hw_exception_data hw_exception_data;
+ struct kfd_hsa_memory_exception_data memory_exception_data;
struct kfd_process *p;
struct kfd_event *ev;
unsigned int temp;
uint32_t id, idx;
+ int reset_cause = atomic_read(&dev->sram_ecc_flag) ?
+ KFD_HW_EXCEPTION_ECC :
+ KFD_HW_EXCEPTION_GPU_HANG;
/* Whole gpu reset caused by GPU hang and memory is lost */
memset(&hw_exception_data, 0, sizeof(hw_exception_data));
hw_exception_data.gpu_id = dev->id;
hw_exception_data.memory_lost = 1;
+ hw_exception_data.reset_cause = reset_cause;
+
+ memset(&memory_exception_data, 0, sizeof(memory_exception_data));
+ memory_exception_data.ErrorType = KFD_MEM_ERR_SRAM_ECC;
+ memory_exception_data.gpu_id = dev->id;
+ memory_exception_data.failure.imprecise = true;
idx = srcu_read_lock(&kfd_processes_srcu);
hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
mutex_lock(&p->event_mutex);
id = KFD_FIRST_NONSIGNAL_EVENT_ID;
- idr_for_each_entry_continue(&p->event_idr, ev, id)
+ idr_for_each_entry_continue(&p->event_idr, ev, id) {
if (ev->type == KFD_EVENT_TYPE_HW_EXCEPTION) {
ev->hw_exception_data = hw_exception_data;
set_event(ev);
}
+ if (ev->type == KFD_EVENT_TYPE_MEMORY &&
+ reset_cause == KFD_HW_EXCEPTION_ECC) {
+ ev->memory_exception_data = memory_exception_data;
+ set_event(ev);
+ }
+ }
mutex_unlock(&p->event_mutex);
}
srcu_read_unlock(&kfd_processes_srcu, idx);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
index 0eeee3c6d6dc..9e0230965675 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
@@ -276,6 +276,9 @@ struct kfd_dev {
uint64_t hive_id;
bool pci_atomic_requested;
+
+ /* SRAM ECC flag */
+ atomic_t sram_ecc_flag;
};
enum kfd_mempool {
diff --git a/include/uapi/linux/kfd_ioctl.h b/include/uapi/linux/kfd_ioctl.h
index e622fd1fbd46..dc067ed0b72d 100644
--- a/include/uapi/linux/kfd_ioctl.h
+++ b/include/uapi/linux/kfd_ioctl.h
@@ -211,6 +211,11 @@ struct kfd_ioctl_dbg_wave_control_args {
#define KFD_HW_EXCEPTION_GPU_HANG 0
#define KFD_HW_EXCEPTION_ECC 1
+/* For kfd_hsa_memory_exception_data.ErrorType */
+#define KFD_MEM_ERR_NO_RAS 0
+#define KFD_MEM_ERR_SRAM_ECC 1
+#define KFD_MEM_ERR_POISON_CONSUMED 2
+#define KFD_MEM_ERR_GPU_HANG 3
struct kfd_ioctl_create_event_args {
__u64 event_page_offset; /* from KFD */
@@ -250,7 +255,12 @@ struct kfd_hsa_memory_exception_data {
struct kfd_memory_exception_failure failure;
__u64 va;
__u32 gpu_id;
- __u32 pad;
+ __u32 ErrorType; /* 0 = no RAS error,
+ * 1 = ECC_SRAM,
+ * 2 = Link_SYNFLOOD (poison),
+ * 3 = GPU hang (not attributable to a specific cause),
+ * other values reserved
+ */
};
/* hw exception data */
--
2.20.1
More information about the amd-gfx
mailing list