摘要:
An arbitration system receives requests to access a destination during an arbitration window that spans multiple processor clock cycles. During each clock cycle, the destination is monitored to determine whether the destination is suffering from backpressure by receiving more requests than the destination is able to accommodate during the clock cycle. In response to detecting backpressure, a masking index value assigned to a requesting source is incremented, which limits an amount of requests from the source that will be granted destination access during a subsequent arbitration window. Alternatively, in response to detecting an absence of backpressure during an arbitration window, the masking index value is decremented, which increases the amount of requests from the source that will be granted destination access during a subsequent arbitration window. This arbitration process continues for successive arbitration windows, oscillating between incrementing and decrementing the masking index value during the successive arbitration windows.
摘要:
A plurality of resource requesters may be configured to consume a resource to perform a task. Each of the plurality of resource requesters can be allocated a resource budget to consume the resource to perform the task. An arbiter can select one of the plurality of resource requesters to consume the resource based on an arbitration scheme. When a resource requester is selected, the amount of resource consumed by the resource requester can be deducted from its resource budget. When the resource requester is idle for a number of cycles when selected, the corresponding resource budget can be further reduced to account for the actual amount of resource consumed and wasted by the resource requester, which can provide fairness in resource consumption over few rounds of arbitration.
摘要:
According to one general aspect, an apparatus may include a network of node circuits and a central arbiter circuit. The network of node circuits is within an integrated circuit, wherein the network includes a plurality of segments. The central arbiter circuit may be configured to schedule a routing of a message between a pair of node circuits in the network, wherein the routing includes a guaranteed latency between the pair of node circuits.
摘要:
An apparatus comprising: a plurality of processors, a data bus, shared by the plurality of processors, and configured to at least receive data processed by each of the plurality of processors when performing predetermined tasks, the plurality of processors and data bus comprising at least part of a hardware based real-time computing system; a controller configured to provide for control of a maximum data rate at which data is provided to the data bus for transmission thereover by at least one of the plurality of processors in performing at least one of the predetermined tasks, wherein the controller is configured to provide for a maximum data rate at least comprising an impeded rate for the at least one predetermined tasks and an unimpeded rate wherein the impeded rate is less than the unimpeded rate.
摘要:
The present disclosure includes apparatuses and methods related to accessing status information. One example apparatus comprises a host and a memory device coupled to the host. The memory device includes a controller configured to provide, to a status arbiter, a status signal indicating whether a status register of the controller contains generated status information. Responsive to the status signal indicating that the status register contains the generated status information, the controller can also provide the status information from the controller to the status arbiter via a status intermediary.
摘要:
Aspects include acquiring measurement data of a synchronous input/output (I/O) link between an operating system and a recipient. The acquiring measurement data can include monitoring operating system usage of synchronous I/O commands on the synchronous I/O link and storing the operating system usage in a measurement block as the measurement data. Further, the measurement block is accessible by the operating system to determine that the measurement data is acquired.
摘要:
Methods and structure for detecting that arbitration is delaying discovery. One embodiment is a Serial Attached Small Computer System Interface (SAS) expander. The SAS expander includes multiple SAS ports, a port monitor, and a controller. The port monitor is able to track physical link events during arbitration for at least one of the ports while discovery is in progress at the expander, and to detect based on the physical link events that arbitration is delaying discovery. The controller is able to prioritize discovery requests at the expander responsive to detecting that arbitration is delaying discovery.
摘要:
A memory controller has a request holding unit holding a write request and a read request; a transmission unit transmitting any one of the write request and the read request to a memory through a transmission bus; a reception unit receiving read data corresponding to the read request through a reception bus; and a request arbitration unit performing: a first processing of transmitting the write request before the read request, when a first reception time is not later than a second reception time, and a second processing of transmitting the read request before the write request, when the first reception time is later than the second reception time. The first reception time is when reception of the read data is started when the write request is transmitted first, and the second reception time is when the reception of the read data is started when the read request is transmitted first.
摘要:
Peripheral Component Interconnect Express (PCIe) tunneling over Converged Enhanced Ethernet (CEE) networks. The CEE networks comprise devices configured to use PCIe. An initiating device initiates a command. The command is associated with initiator control signals, which are associated with the initiating device. The initiating device requests permission from an arbiter, and receives a request grant from the arbiter. Based on the request grant, a mapping device maps the initiator control signals, an target device address, and the command into a CEE control frame. Based on the mapping, the initiating device transmits an inquiry to the devices. Based on transmitting the inquiry, the initiating device receives a response from a corresponding device. The corresponding device is associated with the target device address. The response comprises target control signals associated with the corresponding device. Based on the response, the initiating device initiates a transaction to the corresponding device through CEE control frames.
摘要:
A memory management and protection system that incorporates device security features that support a distributed, shared memory system. The concept of secure regions of memory and secure code execution is supported, and a mechanism is provided to extend a chain of trust from a known, fixed secure boot ROM to the actual secure code execution. Furthermore, the system keeps a secure address threshold that is only programmable by a secure supervisor, and will only allow secure access requests that are above this threshold.