338
By generating read requests to memory addresses, the patrol scrubber allows the
hardware an opportunity to run ECC on a memory address and correct any correctable
errors before they can become uncorrectable errors. Optionally, if an uncorrectable
error is discovered, the patrol scrubber can trigger a hardware interrupt and notify the
software layer of its memory address.
Unconsumed Uncorrectable Memory-Error Persistent Memory
Root-Device Notification
The ACPI specification describes a method for hardware to notify software of
unconsumed uncorrectable errors called the Unconsumed Uncorrectable MemoryError Persistent Memory Root-Device Notification. Using the ACPI-defined framework,
the operating system can subscribe to be notified by the platform whenever an
uncorrectable memory error is detected. It is the platform’s responsibility to receive
notification from persistent memory devices that an uncorrectable error has been
detected and take appropriate action to generate a persistent memory root-device
notification. Upon receipt of root-device notification, the operating system can then use
existing ACPI methods, such as Address Range Scrub (ARS), to discover the address of
the newly created uncorrectable memory error and take appropriate actions.
Address Range Scrub
ARS is a device-specific method (_DSM) defined in the ACPI specification. Privileged
software can call an ACPI _DSM such as ARS at runtime to retrieve or scan for the
locations of uncorrectable memory errors for all persistent memory in the platform.
Because ARS is implemented by the platform, each vendor may implement some of the
functionality differently.
An ARS accepts a given system address range from the caller and, like patrol scrub,
inspects each memory address in that range for memory errors. When ARS completes,
the caller is given a list of memory addresses in the given range that contains memory
errors. Inspection of each memory address may be handled by persistent memory
hardware or by the platform itself. Unlike a patrol scrub, ARS inspects each memory
address at a very high frequency. This increased frequency of the scrub may impact
the quality of service for the persistent memory hardware. Thus, ARS can optionally be
invoked by the caller to return the results of the previous ARS, sometimes referred to as a
short ARS.
Chapter 17 reliability, availability, and ServiCeability (raS)
By generating read requests to memory addresses, the patrol scrubber allows the
hardware an opportunity to run ECC on a memory address and correct any correctable
errors before they can become uncorrectable errors. Optionally, if an uncorrectable
error is discovered, the patrol scrubber can trigger a hardware interrupt and notify the
software layer of its memory address.
Unconsumed Uncorrectable Memory-Error Persistent Memory
Root-Device Notification
The ACPI specification describes a method for hardware to notify software of
unconsumed uncorrectable errors called the Unconsumed Uncorrectable MemoryError Persistent Memory Root-Device Notification. Using the ACPI-defined framework,
the operating system can subscribe to be notified by the platform whenever an
uncorrectable memory error is detected. It is the platform’s responsibility to receive
notification from persistent memory devices that an uncorrectable error has been
detected and take appropriate action to generate a persistent memory root-device
notification. Upon receipt of root-device notification, the operating system can then use
existing ACPI methods, such as Address Range Scrub (ARS), to discover the address of
the newly created uncorrectable memory error and take appropriate actions.
Address Range Scrub
ARS is a device-specific method (_DSM) defined in the ACPI specification. Privileged
software can call an ACPI _DSM such as ARS at runtime to retrieve or scan for the
locations of uncorrectable memory errors for all persistent memory in the platform.
Because ARS is implemented by the platform, each vendor may implement some of the
functionality differently.
An ARS accepts a given system address range from the caller and, like patrol scrub,
inspects each memory address in that range for memory errors. When ARS completes,
the caller is given a list of memory addresses in the given range that contains memory
errors. Inspection of each memory address may be handled by persistent memory
hardware or by the platform itself. Unlike a patrol scrub, ARS inspects each memory
address at a very high frequency. This increased frequency of the scrub may impact
the quality of service for the persistent memory hardware. Thus, ARS can optionally be
invoked by the caller to return the results of the previous ARS, sometimes referred to as a
short ARS.
Chapter 17 reliability, availability, and ServiCeability (raS)