C Specification
Bits which can be set in VkDeviceFaultInfoKHR::flags providing
information of the status of the fault reported and which additional fields
have been populated by the driver, are:
// Provided by VK_KHR_device_fault
typedef enum VkDeviceFaultFlagBitsKHR {
VK_DEVICE_FAULT_FLAG_DEVICE_LOST_KHR = 0x00000001,
VK_DEVICE_FAULT_FLAG_MEMORY_ADDRESS_KHR = 0x00000002,
VK_DEVICE_FAULT_FLAG_INSTRUCTION_ADDRESS_KHR = 0x00000004,
VK_DEVICE_FAULT_FLAG_VENDOR_KHR = 0x00000008,
VK_DEVICE_FAULT_FLAG_WATCHDOG_TIMEOUT_KHR = 0x00000010,
VK_DEVICE_FAULT_FLAG_OVERFLOW_KHR = 0x00000020,
} VkDeviceFaultFlagBitsKHR;Description
-
VK_DEVICE_FAULT_FLAG_DEVICE_LOST_KHR specifies that the fault has resulted in a device lost condition. No subsequent entries will be returned for this device.
-
VK_DEVICE_FAULT_FLAG_MEMORY_ADDRESS_KHR specifies that the fault has associated memory access address information which is stored in the faultAddressInfo field (see VkDeviceFaultAddressInfoKHR).
-
VK_DEVICE_FAULT_FLAG_INSTRUCTION_ADDRESS_KHR specifies that the fault has an associated instruction address which is stored in the instructionAddressInfo field (see VkDeviceFaultAddressInfoKHR).
-
VK_DEVICE_FAULT_FLAG_VENDOR_KHR specifies that the fault has associated vendor information stored in the vendorInfo field (see VkDeviceFaultVendorInfoKHR).
-
VK_DEVICE_FAULT_FLAG_WATCHDOG_TIMEOUT_KHR specifies that the fault was the result of a GPU timeout. Further information may be made available using other platform specific extensions via the pNext chain of the VkDeviceFaultInfoKHR structure.
-
VK_DEVICE_FAULT_FLAG_OVERFLOW_KHR specifies that prior faults have occurred, but information about them is no longer available. This typically indicates that faults are occurring more rapidly than the calling application is able to read them back.
Document Notes
For more information, see the Vulkan Specification.
This page is extracted from the Vulkan Specification. Fixes and changes should be made to the Specification, not directly.