Currently the only usage of triggers is by the trace capture device. This method is straightforward when using one trace source. When using multiple trace sources there can be a time disparity between the trace sources that generate a trigger and when the trigger packet appears at the output of the trace port. See the CoreSight Architecture Specification for more information on triggers.
A trigger can be interpreted as an event that occurred. This can be:
Directly from an event such as a pin toggle from the CTI.
A delayed event such as a pin toggle that has been delayed coming through the Trigger Counter Register.
The completion of a flush.
Table 8.4 extends the ETMv3 specification on how a trigger is represented
|0||x||x||No||Yes||Normal trace data|
[a] The trigger packet encoding is required for the current ETMv3 protocol that uses a special encoding for triggers that always occur on the lower bits of TRACEDATA.