摘要:
A method, apparatus, and system are provided for monitoring locks using monitor-memory wait. In one embodiment, a memory to store instructions to perform functions of a monitoring mechanism is provided. The monitoring mechanism having a first logic to cause a processor to exit a sleep state in response to an event, wherein exiting the sleep state comprises resuming control of processing resources that were relinquished by the processor during the sleep state. The monitoring mechanism having a second logic to disable monitoring of a node associated with a contended lock after the processor exits the sleep state.
摘要:
A method for a mechanism for processor power state aware distribution of lowest priority interrupts. The method of one embodiment comprises receiving first power state information from a first component and second power state information from a second component. First task priority information from the first component and second task priority from the second component are also received. An interrupt request from a first device for servicing is received. Power state and task priority information for the first and second components are evaluated to determine which component should service the interrupt request. Either the first component or the second component is selected to be a destination component to service the interrupt request based on the power state and task priority information. The interrupt request is communicated to the destination component.
摘要:
A method for a mechanism for processor power state aware distribution of lowest priority interrupts. The method of one embodiment comprises receiving first power state information from a first component and second power state information from a second component. First task priority information from the first component and second task priority from the second component are also received. An interrupt request from a first device for servicing is received. Power state and task priority information for the first and second components are evaluated to determine which component should service the interrupt request. Either the first component or the second component is selected to be a destination component to service the interrupt request based on the power state and task priority information. The interrupt request is communicated to the destination component.
摘要:
In some embodiments, an apparatus includes processors, signal storage circuitry, and processor selection logic. The signal storage circuitry is to hold willingness indication signals each indicative of a willingness level of an associated one of the processors to receive an interrupt and to hold priority indication signals each indicative of a processor priority level of an associated one of the processors, wherein there are multiple possible willingness levels and multiple possible processor priority levels. The processor selection logic is to select one of the processors to receive an interrupt based at least on the willingness indication signals. Other embodiments are described.
摘要:
According to an embodiment of the invention, a method and apparatus for inter-processor interrupts in a multi-processor system are described. An embodiment comprises writing an inter-processor interrupt request to a first memory location; monitoring the first memory location; detecting the inter-processor interrupt request in the first memory location; calling a function for the inter-processor interrupt request; and performing the function for the inter-processor interrupt request.
摘要:
In some embodiments, an apparatus includes processor selection logic to receive logical destination identification numbers that are associated with interrupts each having a processor cluster identification number to identify a cluster of processors to which the interrupts are directed. The logical destination identification numbers are each to identify which processors within the identified cluster of processors are available to receive the corresponding one of interrupts. The processor selection logic is to select one of the available processors to receive the interrupt, and the selected one of the available processors is identified through a relative position of a corresponding bit in the logical destination identification numbers. Other embodiments are described.
摘要:
A method, apparatus, and system are provided for monitoring locks using monitor-memory wait. According to one embodiment, a node associated with a contended lock is monitored; and a processor seeking the contended lock is put to sleep until a monitor event occurs.
摘要:
In some embodiments, an apparatus includes processors, signal storage circuitry, and processor selection logic. The signal storage circuitry is to hold willingness indication signals each indicative of a willingness level of an associated one of the processors to receive an interrupt and to hold priority indication signals each indicative of a processor priority level of an associated one of the processors, wherein there are multiple possible willingness levels and multiple possible processor priority levels. The processor selection logic is to select one of the processors to receive an interrupt based at least on the willingness indication signals. Other embodiments are described.
摘要:
A method, apparatus, and system are provided for monitoring locks using monitor-memory wait. According to one embodiment, a node associated with a contended lock is monitored; and a processor seeking the contended lock is put to sleep until a monitor event occurs.
摘要:
A firmware-based mechanism for creating, storing and retrieving variable-length records associated with error events occurring in a computer platform. The mechanism responds to error notifications by invoking a firmware-based error-handling module. The error-handling module retrieves processor-specific error information and may also interrogate the other components of the computer platform to determine their error status. Then, according to the nature of the discovered errors, the error-handling module may assemble the retrieved error information and status information into a variable-length error record, which the error-handling module may then store in a memory. On request from a processing agent, the error-handling module may retrieve a previously-stored error record and present it to the requesting agent. Thus, the invention provides a unified and standardized approach to computer error handling at the firmware level.