欢迎访问ic37.com |
会员登录 免费注册
发布采购

78Q8430-100IGT/F 参数 Datasheet PDF下载

78Q8430-100IGT/F图片预览
型号: 78Q8430-100IGT/F
PDF下载: 下载PDF文件 查看货源
内容描述: 10/100以太网MAC和PHY [10/100 Ethernet MAC and PHY]
分类和应用: 电信集成电路编码器以太网局域网(LAN)标准
文件页数/大小: 88 页 / 1209 K
品牌: TERIDIAN [ TERIDIAN SEMICONDUCTOR CORPORATION ]
 浏览型号78Q8430-100IGT/F的Datasheet PDF文件第25页浏览型号78Q8430-100IGT/F的Datasheet PDF文件第26页浏览型号78Q8430-100IGT/F的Datasheet PDF文件第27页浏览型号78Q8430-100IGT/F的Datasheet PDF文件第28页浏览型号78Q8430-100IGT/F的Datasheet PDF文件第30页浏览型号78Q8430-100IGT/F的Datasheet PDF文件第31页浏览型号78Q8430-100IGT/F的Datasheet PDF文件第32页浏览型号78Q8430-100IGT/F的Datasheet PDF文件第33页  
DS_8430_001  
78Q8430 Data Sheet  
If small packets are a problem, then the Mode setting of 10b can be used. In this mode, the QDR bit for  
the QUE is set only when there is an EOF in the QUE, or in other words, the QUE contains at least one  
entire frame. In this mode, TX FIFO under-runs are not possible since the QUE will not begin to transmit  
until it contains the entire frame. The draw-back to this mode is with very large frames. If a frame is too  
large to fit into the QUE all at one time then it will never begin transmitting and the QUE will be stalled.  
If both small and large packets are to be handled then a Mode setting of 11b should be used. In this  
mode, the QDR bit for the QUE is set anytime there is an EOF in the QUE or the number of BLOCKs in  
the QUE is above the threshold. In this way, large packets can preload a fixed number of BLOCKs while  
small packets are guaranteed to transmit.  
To facilitate the handling of very large packets by a fast host, an interrupt that is tied to the QSR  
Threshold is provided. To make use of this, the host sets the Threshold field based on the interrupt  
latency. The host then preloads the QUE with some number of BLOCKs. As soon as the total number of  
blocks left in the QUE falls below the Threshold, an interrupt is generated. In response to the interrupt  
the host writes more data to the QUE to put the number back above the threshold. The host can then go  
on to other tasks until the next interrupt. This cycle is repeated until the frame is completed.  
6.3.3 DMA Slave Mode Access  
Reading or writing large amounts of data into and out of a single QUE involves accessing the same RDR  
or TDR register repeatedly. A DMA Slave Mode is implemented to facilitate this activity and reduce  
overhead on the host side. While in DMA Slave Mode, the address bus on the host interface is ignored  
and all access is assumed to be to the programmed address until DMA mode is terminated. In this way  
the host can use a DMA engine or block transfer facility to write or read QUE data without regard to the  
addresses generated.  
DMA Slave Mode is controlled by the DMA Register (DMA) at address 0x100.  
To read data from a QUE using DMA Slave Mode, the host writes the address of the RDR for the desired  
QUE into bits nine through zero and sets bit 17, the Read Mode bit, in the DMA register at address  
0x100. The host then starts the DMA transfer and all read access to the host interface will go to the  
programmed RDR address. When the DMA transfer is complete, the DMA Mode is terminated by writing  
a zero to bit 17 of the DMA Register (DMA).  
To write data to a QUE using DMA Slave Mode, the host writes the address of the TDR for the desired  
QUE into bits nine through zero and sets bit 16, the Write Mode bit, in the DMA Register. The host then  
starts the DMA transfer and all write access to the host interface will go to the programmed TDR address.  
When the DMA transfer is complete, reading the cleared Write Mode bit from the DMA Register  
terminates the DMA Mode.  
DMA Slave Mode does not have any effect on other operations of the interface such that, for example,  
the ENDIAN settings, STDR settings, etc. are all in effect during a DMA Mode transfer. During a DMA  
mode transfer, the actual register address of the host bus access is ignored. This means that using DMA  
Slave Mode to transfer data out of order is not supported. Data words must always be written to a  
transmit QUE in the desired transmit order and are always read from a receive QUE in received order.  
6.4 Snoop Mode Access  
The Snoop Interface provides a means by which QUE data can be inspected and modified in situ, leaving  
the state of the QUE unchanged. The Snoop Interface works by presenting the contents of a specific  
BLOCK of QUE memory at the SNOOP address space 0x300-0x3FF. The Snoop Control Register  
(SNCR) is used to set which BLOCK of QUE memory is mapped into the SNOOP address space.  
For example, an application that wishes to inspect or modify the contents of the first frame in the receive  
QUE, QUE0, first reads the value of the FIRST BLOCK in QUE0 from its QFLR Register (QFLR). The  
pointer to the FIRST BLOCK in QUE0 is then written to the SNCR register. Now that the SNCR Register  
has programmed the SNOOP interface to point to the FIRST BLOCK for QUE0, accessing registers in the  
address space from 0x300 to 0x3FF will be directly accessing the data for the first frame contained in  
QUE0.  
Rev. 1.2  
29