INTRODUCTION

USB HID devices send and receive data laid out in a device dependent way. The usbhid library contains routines to extract the report descriptor which contains the data layout information and then use this information.

The routines can be divided into four parts: extraction of the descriptor, parsing of the descriptor, translating to/from symbolic names, and data manipulation.

DESCRIPTOR FUNCTIONS

A report descriptor can be obtained by calling hid_get_report_desc() with a file descriptor obtained by opening a uhid(4) device. Alternatively a data buffer containing the report descriptor can be passed into hid_use_report_desc(). The data is copied into an internal structure. When the report descriptor is no longer needed it should be freed by calling hid_dispose_report_desc(). The type report_desc_t is opaque and should be used when calling the parsing functions. If hid_dispose_report_desc() fails it will return NULL.

DESCRIPTOR PARSING FUNCTIONS

To parse the report descriptor the hid_start_parse() function should be called with a report descriptor and a set that describes which items that are interesting. The set is obtained by or-ing together values (1 << k) where k is an item of type hid_kind_t. The report ID (if present) is given by id. The function returns NULL if the initialization fails, otherwise an opaque value to be used in subsequent calls. After parsing the hid_end_parse() function should be called to free internal data structures.

To iterate through all the items in the report descriptor, hid_get_item() should be called while it returns a value greater than 0. When the report descriptor ends it will return 0; a syntax error within the report descriptor will cause a return value less than 0. The struct pointed to by h will be filled with the relevant data for the item. The definition of hid_item_t can be found in <usbhid.h> and the meaning of the components in the USB HID documentation.

Data should be read/written to the device in the size of the report. The size of a report (of a certain kind) can be computed by the hid_report_size() function. If the report is prefixed by an ID byte it is given by id.

To locate a single item the hid_locate() function can be used. It should be given the usage code of the item and its kind and it will fill the item and return non-zero if the item was found.

NAME TRANSLATION FUNCTIONS

The function hid_usage_page() will return the symbolic name of a usage page, and the function hid_usage_in_page() will return the symbolic name of the usage within the page. Both these functions may return a pointer to static data.

The functions hid_parse_usage_page() and hid_parse_usage_in_page() are the inverses of hid_usage_page() and hid_usage_in_page(). They take a usage string and return the number of the usage, or -1 if it cannot be found.

Before any of these functions can be called the usage table must be parsed, this is done by calling hid_init() with the name of the table. Passing NULL to this function will cause it to use the default table.

DATA EXTRACTION FUNCTIONS

Given the data obtained from a HID device and an item in the report descriptor the hid_get_data() function extracts the value of the item. Conversely hid_set_data() can be used to put data into a report (which must be zeroed first).