欢迎访问ic37.com |
会员登录 免费注册
发布采购

319973-003 参数 Datasheet PDF下载

319973-003图片预览
型号: 319973-003
PDF下载: 下载PDF文件 查看货源
内容描述: 英特尔I / O控制器中枢10 [Intel I/O Controller Hub 10]
分类和应用: 控制器
文件页数/大小: 840 页 / 5889 K
品牌: INTEL [ INTEL ]
 浏览型号319973-003的Datasheet PDF文件第98页浏览型号319973-003的Datasheet PDF文件第99页浏览型号319973-003的Datasheet PDF文件第100页浏览型号319973-003的Datasheet PDF文件第101页浏览型号319973-003的Datasheet PDF文件第103页浏览型号319973-003的Datasheet PDF文件第104页浏览型号319973-003的Datasheet PDF文件第105页浏览型号319973-003的Datasheet PDF文件第106页  
Functional Description  
The following summarizes the ICH10 integrated Gigabit Ethernet controller features:  
• Configurable LED operation for customization of LED display.  
• IPv4 and IPv6 Checksum Offload support (receive, transmit, and large send).  
• 64-bit address master support for system using more than 4 GB of physical  
memory.  
• Configurable receive and transmit data FIFO, programmable in 1 KB increments.  
• Intelligent interrupt generation to enhance driver performance.  
• Compliance with Advanced Configuration and Power Interface and PCI Power  
Management standards.  
• ACPI register set and power down functionality supporting D0 and D3 states.  
• Full wake-up support (APM and ACPI).  
• Magic Packet wake-up enable with unique MAC address.  
• Fragmented UDP checksum off load for package reassembly.  
• Jumbo frames supported.  
• LinkSec support (802.3ae compliant)  
• TimeSync support (802.1as compliant)  
5.3.1  
Gigabit Ethernet PCI Bus Interface  
The Gigabit Ethernet controller has a PCI interface to the host processor and host  
memory. The following sections detail the transaction on the bus.  
5.3.1.1  
Transaction Layer  
The upper layer of the host architecture is the transaction layer. The transaction layer  
connects to the device core using an implementation specific protocol. Through this  
core-to-transaction-layer protocol, the application-specific parts of the device interact  
with the subsystem and transmit and receive requests to or from the remote agent,  
respectively.  
5.3.1.2  
Data Alignment  
4 K Boundary  
5.3.1.2.1  
PCI requests must never specify an Address/Length combination that causes a Memory  
Space access to cross a 4 K boundary. It is the HW responsibility to break requests into  
4 K-aligned requests (if needed). This does not pose any requirement on SW. However,  
if SW allocates a buffer across a 4 K boundary, HW will issue multiple requests for the  
buffer. SW should consider aligning buffers to 4 KB boundary in cases where it  
improves performance.  
The alignment to the 4 K boundaries is done in the core. The Transaction layer will not  
do any alignment according to these boundaries.  
5.3.1.2.2  
5.3.1.3  
64 Bytes  
PCI requests are multiples of 64 bytes and aligned to make better use of memory  
controller resources. Writes, however, can be on any boundary and can cross a 64 byte  
alignment boundary  
Configuration Request Retry Status  
The LAN Controller might have a delay in initialization due to NVM read. If the NVM  
configuration read operation is not completed and the device receives a Configuration  
Request, the device will respond with a Configuration Request Retry Completion Status  
to terminate the Request, and thus effectively stall the Configuration Request until such  
time that the subsystem has completed local initialization and is ready to communicate  
with the host.  
102  
Datasheet