Implement GPU hang detection

This change uses DRM_IOCTL_I915_GET_RESET_STATS to detect
GPU hangs. When such situation is encountered, then
zeCommandQueueSynchronize returns ZE_RESULT_ERROR_DEVICE_LOST.

Related-To: NEO-5313
Signed-off-by: Patryk Wrobel <patryk.wrobel@intel.com>
This commit is contained in:
Patryk Wrobel
2022-01-20 16:56:19 +00:00
committed by Compute-Runtime-Automation
parent 543c854a3b
commit 498cf5e871
37 changed files with 556 additions and 101 deletions

View File

@@ -148,6 +148,7 @@ class Drm : public DriverModel {
MOCKABLE_VIRTUAL void getPrelimVersion(std::string &prelimVersion);
PhysicalDevicePciBusInfo getPciBusInfo() const override;
bool isGpuHangDetected(uint32_t contextId) override;
bool areNonPersistentContextsSupported() const { return nonPersistentContextsSupported; }
void checkNonPersistentContextsSupport();