USB System Architecture

Configuring the Hub

Hubs must be configured like any other device, but this also involves identifying other devices that may be attached to the port. The steps taken by configuration software include:

•Reading the standard device descriptors to obtain a variety of information needed to configure the device.

•Assigning a unique address to the hub.

•Powering the ports.

•Checking the hub status change endpoint to detect port events.

•Reading status information to determine the nature of the event.

•Enabling the port to provide access to the attached device.

As indicated above, a hub must implement a status change port in addition to the default port. Figure 20-1 illustrates the required hub endpoints. The default control port provides access to the descriptors that define the type of device requiring configuration. A hub may also be implemented as part of a compound device; hence, the descriptors may describe additional functions beyond the minimum required of the hub alone.

The Default Pipe

All devices, including hubs, have a default control pipe at endpoint zero. Host software owns the default control pipe that is used to configure hubs and USB devices. This communications pipe is established during initialization so that USB devices can be accessed based on established defaults. The configuration process requires numerous default pipe accesses for configuring and controlling hub features, including: reading the device descriptors, powering hub ports, resetting ports, reading port status, and enabling ports.

The Status Change Pipe

Hubs must implement a status change endpoint that can be polled to detect status changes that have occurred at the hub ports (e.g., device attachment and detachment). Note that the status endpoint provides status information for all hub ports. Configuration software determines the characteristics and the endpoint number of the status change register by reading the endpoint descriptor.

376