C23.3 Per Instance or per Type

MediumFixVerilator coverage build · serverReview the topic

Fix a per-port coverage report that prints the same number for every port.

Hardware context: A 3-port switch with one covergroup instance per port. Report each port's own coverage, not the type number.

This is a Fix problem: the code in the editor has one realistic defect. Find it and fix it.

A switch has three ports, each with its own instance of port_cg. A report function port_pct(cg) should give one port's coverage. Every port shows the same number, even when one port sent nothing.

Signals in the Timeline

SignalWidthRole
samples8input of each check
Full specification
The report (given)
port_cg pc [3];      // one instance per port, fresh per case
// result = 1 * port_pct(pc[0]) + 2 * port_pct(pc[1]) + 3 * port_pct(pc[2])

The weights 1, 2 and 3 only tell the ports apart in one number. The full check samples more lists than the examples show.

Examples

Example 1

One Packet on Port 0
Input
packets (port, len): (0, 1)
Output
coverage_percent_x100 = 3333

Port 0 has 1 of 3 bins: 33.33% (3333). Ports 1 and 2 have 0: the result is 3333.

Example 2

Port 1 Only
Input
packets (port, len): (1, 3) (1, 9)
Output
coverage_percent_x100 = 13334

Port 1 has 2 of 3 bins (6667), counted twice: 13334.

Example 3

Every Port
Input
packets (port, len): (0, 1) (1, 3) (2, 9)
Output
coverage_percent_x100 = 19998

Each port has 1 of 3 bins: 3333 x (1 + 2 + 3) = 19998.

Constraints and Contract

  • Keep the covergroup and the function header int port_pct(port_cg cg).
  • Keep the bins.

Hints

Hint 1

Which method returns one instance's number, and which the number of the whole type?

Hint 2

Does option.per_instance change what either method returns?

Hint 3

With every port at a different number, the weighted sum must change when one port alone gets a new sample.

What the grader checks: Three instances are sampled per case with 18 lists of packets. The weighted sum of your per-port numbers is compared with the values IEEE 1800-2023 19.8 and 19.11 give. Three lists are shown.

18 checks on Verilator coverage build · server. Run the examples first, then check the full regression. Full checks use the server judge. Downloads contain public samples only. A passing result establishes the tested cases, not a proof for all possible inputs.

See all 3 steps and the final capstone

Report a problem with this page

Your code SystemVerilogSaved locally
Ln 1, Col 1Tab indents · Ctrl/⌘ + Enter runs samples
Not run yet

One Packet on Port 0

Input
packets (port, len): (0, 1)
Expected output
coverage_percent_x100 = 3333

Port 0 has 1 of 3 bins: 33.33% (3333). Ports 1 and 2 have 0: the result is 3333.

Run samples executes all 3 public checks. Check all tests runs 18 checks on the server.

Verilator coverage build · server