Coverage Writer¶

To emit coverage records and create the coverage files, noRTL requires a coverage writer to be present in the testbench.

noRTL comes with a default implementation inside the simulation modules. It's usage and requirements are described in the coverage setup guide.

You may replace it by your own implementation or customize parts of it, e.g. to customize the output directory. If you customize or replace the coverage writer, please implement the Verilog functions listed below.

Write Raw Record¶

The function write_record must write a raw JSON record (already serialized to a Verilog string) to the output.

This low-level function is used, if the CoverageWriter is set to raw mode.

function automatic void write_record(string target, string record, logic clear = 1'b0);

Arguments:

target: Coverage target identifier. Typically, this is the name of the noRTL engine. The function must separate data sent to different targets.
record: The record to write as a JSON string.
clear: If this bit is high, the write shall clear all previous data. E.g., when writing to a file, open it with $fopen(..., "w"), if clear is high, and $fopen(..., "a") if it is low.

Emit Info Record¶

The function emit_info emits an Info record. The test information property is not included in the call signature. It may get added by the coverage writer.

function automatic void emit_info(
    string coverage_target,
    string engine [string]
);

Arguments:

coverage_target: Coverage target identifier. See description for target of the write_record function.
engine: Associative array of data for the engine property of the Info record. For example, the function may get called like this:
```
coverage_writer.emit_info("my_engine", '{
    "name": "my_engine",
    "sha256sum": "1234..."
});
```

Emit State Coverage Record¶

The function emit_state_coverage emits an State coverage record.

function automatic void emit_state_coverage(
    string coverage_target,
    string state,
    logic covered = 1'b1
);

Arguments:

coverage_target: Coverage target identifier. See description for target of the write_record function.
state: Name of the state, e.g. "WORKER_1_STATE_123".
covered: Coverage value, indicates if the state was visited.

Emit Transition Coverage Record¶

The function emit_transition_coverage emits an Transition coverage record.

function automatic void emit_transition_coverage(
    string coverage_target,
    string source,
    string target,
    logic covered = 1'b1
);

Arguments:

coverage_target: Coverage target identifier. See description for target of the write_record function.
source: Name of the source state.
target: Name of the target state.
covered: Coverage value, indicates if the transition has occured.

Emit Cover Point Record¶

The function emit_cover_point emits an cover point record.

function automatic void emit_cover_point(
    string coverage_target,
    string state,
    string condition,
    string point_name,
    string group_name = "",
    int group_coords [] = {},
    logic covered = 1'b1
);

Arguments:

coverage_target: Coverage target identifier. See description for target of the write_record function.
state: Name of the state of the coverpoint.
condition: Condition of the coverpoint, e.g. "adr > 0". The condition is rendered by noRTL in Python.
point_name: Name of the coverpoint.
group_name: Optional name of the covergroup, to which the coverpoint belongs. If the string is empty, it must be set to null or omitted in the JSON record.
group_coords: Optional array of coordinates inside the covergroup. If the array is empty, it must be set to null or ommitted in the JSON record.
covered: Coverage value, indicates if the transition has occured.

Emit Cover Group Record¶

The function emit_cover_group emits an cover group record.

function automatic void emit_cover_group(
    string coverage_target,
    string name,
    string axis [],
    int limits [][]
);

Arguments:

coverage_target: Coverage target identifier. See description for target of the write_record function.
name: Name of the covergroup.
axis: Array of names for the axis of the covergroup.
limits: Array of lower and upper limits for each axis of the covergroup. Each entry in the array is an array of length 2. Due to technical reasons of the SystemVerilog implementation, the inner array length must be set dynamic.