微控制器中的以太网控制器行为描述正确吗?

Ethernet controller behavior in micro-controller described correctly?

我正在阅读 "Embedded Software Development: the Open Source approach" 这本书,2015 年版。第28-29页有如下句子:

As said previously, the bus segments shown as dashed vertical lines somewhat deviate from the regular structure of the crossbar and are used by components (for instance, the Ethernet controller) that can be both bus masters and slaves.This kind of behavior is needed in some cases because, continuing the example concerning the Ethernet controller,

  • It must be able to autonomously read from memory and write into memory the frames to be transmitted and being received, respectively, and also
  • It must allow the processor to read and write its internal registers, to access status information and control its behavior, respectively.

上面这句话的写法是一个避免的例子,因为当"it"出现在第2个要点时,"it"可能指的是Ethernet-controller或处理器。

我想知道这句话是什么意思,我有以下解释,所以我想知道是否正确:

...在某些情况下需要这种行为,因为继续有关以太网控制器的示例,

所以我猜对了吗?以太网控制器是否有内部寄存器?并执行 Eth.-con。允许处理器读取这些内部寄存器? Eth.-con。允许处理器也写入它们?

所描述的"it must be able to read from and write to memory the frames..."是一个典型的具有DMA能力的设备的例子,它可以容纳CPU并接管总线直接读写系统内存。

显然,CPU 必须能够告诉网卡诸如要使用的地址过滤器之类的东西,以及它应该从中传输帧和接收帧的内存范围,并且它必须告诉网卡何时开始传输 - 此信息由 CPU 写入控制器寄存器,这些寄存器通常映射到 CPU 的 I/O 地址范围。这是 CPU 控制外围设备的正常方式 - 并且适用于硬盘和软盘控制器、USB 和 RTC 外围设备。