摘要:
A method, apparatus, and system are disclosed. In one embodiment the method detects a temperature event in a processor and then modifies the bus frequency of an I/O bus coupled to an I/O controller hub in response to the temperature event. In another embodiment, the apparatus includes a temperature detection unit that detects a temperature event in a processor and, additionally, a bus frequency modification unit that will modify the bus frequency of an I/O bus in response to the temperature event.
摘要:
In some embodiments, a method and apparatus for maintaining data integrity following parity error detection are described. In one embodiment, the method includes the blockage of bus transactions in response to detection of a parity error. Once bus transactions are suspended, a parity error handler routine is invoked to perform parity error recovery according to stored transaction information regarding an error bus transaction of the detected parity error. In one embodiment, the stored information includes a bus master that caused the error, as well as an address associated with the corrupt data for which the parity error was asserted. In one embodiment, data logging is performed to track the bus masters associated with error bus transactions to enable identification of problematic or old hardware devices. Other embodiments are described and claims.
摘要:
Method and apparatus for supporting multi-function PCI devices in PCI bridges. Respective pre-fetch buffers are allocated in response to respective initial data transfer requests issued by a multi-function PCI device. A programmable buffer fill watermark is set up for each pre-fetch buffer. While a portion of data corresponding to the data transfer requests fill the pre-fetch buffers, the fill level of each buffer is monitored to determine if it meets or exceeds its buffer fill watermark. In response to such a condition, the multi-function PCI device is connected to the PCI bridge and a virtual buffer is mapped to the pre-fetch buffer. The pre-fetch buffer is then emptied. During subsequent data transfers, each of the pre-fetch buffer becomes filled, the PCI device is connected, and the virtual buffer is mapped to the filled buffer. The process is continued until all data corresponding to the original data transfer request is received by the multi-function PCI device.
摘要:
Embodiments of the present invention provide a method and apparatus to allow an I/O controller to alert an external controller using an enhanced SMBus implementation that enables bi-directional capability on SMBALERT#. I/O controller includes an auxiliary control register and alert output enable (AOEN) register. When host sets AOE bit in auxiliary control register, SMBALERT# signal is configured as an output signal with bi-directional functionality. External controller uses an interface command to write to AOEN register and determine events/conditions it wants to be alerted on. SMBALERT# is activated in response to a detected event/condition. In response to SMBALERT#, external controller determines the alert generation condition using byte read commands on the system management bus, and clears SMBALERT#.
摘要翻译:本发明的实施例提供一种允许I / O控制器使用能够在SMBALERT#上实现双向能力的增强型SMBus实现来向外部控制器发出警报的方法和装置。 I / O控制器包括辅助控制寄存器和报警输出使能(AOEN)寄存器。 当主机在辅助控制寄存器中设置AOE位时,SMBALERT#信号被配置为具有双向功能的输出信号。 外部控制器使用接口命令写入AOEN寄存器,并确定要发送的事件/条件。 响应于检测到的事件/条件,SMBALERT#被激活。 响应SMBALERT#,外部控制器使用系统管理总线上的字节读取命令确定警报生成条件,并清除SMBALERT#。
摘要:
Access to internal configuration registers on a computer system's chipset using an external micro-controller is provided. A SMB configuration read command including a register address may be received from an external micro-controller. Access to an internal bus may be requested from a bus arbiter. If internal bus access is granted, the SMB configuration read command may be forwarded to a device including the identified register address using the internal bus. In response to the SMB configuration read command, configuration register values from the device may be received. The configuration register values may be forwarded to the external micro-controller.
摘要:
A system to monitor performance of a computing device includes a first bridge to interface with a first set of devices, and a second bridge to interface with a second set of devices. Configuration registers store configuration data associated with the second set of devices, and are accessible through the second bridge. A hub interface allows data to transfer downstream from the first bridge to the second bridge, and allows data to transfer upstream from the second bridge to the first bridge. A controller, external to the first and second bridges, accesses the configuration registers via the second bridge. A logic device allows the second bridge to send data to, and receive data from, the controller.
摘要:
Embodiments of the present invention provide a system event log for a computer system. The system event log may comprise a RAM coupled to a system bus. The system event log may be configured to record information in the RAM corresponding to bus transactions on the system bus. The information may be used to de-bug system problems.
摘要:
According to one embodiment of the invention, a first signal line is provided for a serial interface unit (SIU) of an I/O controller to report interrupt requests to an interrupt controller. In one embodiment, a transition of the first signal line from a first level (e.g., low level or logic 0) to a second level (e.g., high level or logic 1) indicates a pending interrupt to the interrupt controller. A pull up resistor is provided to pull the first signal line to the second level when the first signal line is not driven by the SIU. In response to detecting an interrupt request initiated by an I/O device, a transition from the first level to the second level is generated on the first signal line for a predetermined duration to report the pending interrupt request to the interrupt controller.
摘要:
Alpha values associated with video mixing operations are sent to a memory on a low pin count bus. The memory is accessible to a video mixer, which retrieves the alpha values to perform a mixing operation. The alpha values for a field are sent to the memory during the field time for a previous field rather than during the vertical blanking interval. The alpha values may be compressed prior to transmission.
摘要:
A processor-based system may be operated in an effectively “always on” condition. The system may transition from a lower power consumption state to a higher power consumption state in response to the first operation of a power button. In response to a second operation of the power button, the system transitions from the higher power consumption state to the lower power consumption state. However, unless the system is unplugged, the system remains in a power consuming state even when the power button is repeatedly operated.