WM8850
Pre-Production
UNSOLICITED RESPONSES
The High Definition Audio Specification provides support for software interrupts to be generated by
the CODEC independent of verb commands sent to the CODEC: these are known as Unsolicited
Responses. This section defines how the WM8850 creates Unsolicited Responses, explains some of
the novel features in the WM8850 to manage them, and what each Unsolicited Response from each
node represents. This section assumes familiarity with the High Definition Audio Specification.
GENERATING UNSOLICITED RESPONSES
Unsolicited Responses can only be sent wherever a null response would normally be sent.
Unsolicited Responses are queued on a frame-basis. That is, if an Unsolicited Response is issued in
the current frame, it is sent on the next frame containing a null response.
PRIORITY QUEUING
Where more than one Unsolicited Response is awaiting transmission, they are queued such that the
Unsolicited Response with the lowest Priority value (set using the Unsolicited Response Priority
Control Verb) is at the front of the queue. Nodes that have the same Priority setting are queued on a
‘first-to-trigger’ basis.
The Unsolicited Response Priority Control Verb is described in the “Ancillary Functions” section
describing the Audio Function Group node.
IN-QUEUE-UPDATE
The WM8850 stores all flags that trigger an Unsolicited Response. If a node triggers an Unsolicited
Response, the Unsolicited Response is queued as described above. Should the same node trigger a
second Unsolicited Response (from a different flag within that node) while the first is being queued,
the Unsolicited Response is updated to show both flags. This is known as an ‘In-Queue-Update’.
Should the same node trigger a second Unsolicited Response (from the same flag within the node)
while the first is being queued, only one Unsolicited Response is sent for both events.
PP, April 2011, Rev 3.2
76
w
WOLFSON [ WOLFSON MICROELECTRONICS PLC ]
