A link runs for a while, then the transmitter stops for good. Fix its credit counter so that lost updates no longer leak credits.
Hardware context: The digital blocks of a serializer/deserializer: line encoder and decoder, scrambler, CRC and error-correcting code. Keep the line balanced and busy, and detect or correct the errors a fast serial link makes.
This is a Fix problem. The transmitter in the editor keeps a credit count and adds one for every credit update. After a while it stops sending and never starts again.
The receiver has 8 slots and frees one every 4 cycles. Each update it sends carries a running limit: 8 plus every packet it has freed so far. The field is 8 bits wide, so the limit wraps at 256.
About 2 updates in 100 are lost on the way. Repair the four functions so the transmitter never overflows the receiver and never stalls while it may send.
The fixture runs a second copy of the link with the same lost updates. A transmitter that obeys the running limit drives it. Yours must send exactly as many packets.
Signals in the Timeline
| Signal | Width | Role |
|---|---|---|
test_number | 8 | input of each check |
Full specification
function automatic void tx_reset();
function automatic void tx_update(input [7:0] value); // the running limit, modulo 256
function automatic bit tx_can_send();
function automatic void tx_send(); // called once per packet sent
| Item | Rule (bench values) |
|---|---|
| Receiver | 8 slots. One packet is freed every 4 cycles when it holds any. |
| Link | A packet or an update takes 3 cycles to cross, plus the cycle it is sent in. |
| Update | Sent each time a slot is freed. It carries (8 + slots freed so far) modulo 256. Some are lost. |
| Order in a cycle | Arrivals first, then a slot may be freed, then tx_can_send() is asked once; if it says 1, a packet goes and tx_send() is called. |
| Start | After tx_reset() the receiver is empty: 8 packets may go before any update arrives. |
The tests run up to 20,000 cycles, long enough for the limit to wrap many times. They report overflows and how many packets you sent short of the running-limit transmitter.
Examples
Example 1
No Updates Lost- Input
no loss; 4,000 cycles
- Output
overflows = 0, packets_short = 0
With nothing lost, "+1" and the running limit agree. Your transmitter sends as many packets as the reference and nothing overflows.
Example 2
2 Updates in 100 Lost- Input
20 per thousand lost; 20,000 cycles
- Output
overflows = 0, packets_short = 0
Every lost "+1" update takes a credit with it. After 8 are gone the transmitter stops. A running limit loses nothing: the next update covers the lost one.
Example 3
2 In 100, Another Seed- Input
20 per thousand lost; 20,000 cycles
- Output
overflows = 0, packets_short = 0
Over 20,000 cycles about 5,000 packets are sent, so the 8-bit limit wraps about 20 times. A comparison that ignores the wrap stalls at the first one.
Constraints and Contract
- Keep the four function names and their arguments.
- Keep
SLOTS = 8. - Decide only from what the functions receive; the fixture's state is not yours to read.
Hints
Hint 1
What does the transmitter lose for ever when one "+1" update goes missing?
Hint 2
The update already tells you the whole story: how many packets you may have sent in all. Keep a count of what you have sent, and compare.
Hint 3
The limit wraps at 256 but your count may not. Keep both 8 bits wide and look at their difference, which wraps too.
What the grader checks: The fixture runs your transmitter on six bench links beside a running-limit transmitter with the same lost updates. It counts overflows and the packets sent short. It does not model credit types, initial advertisement or Recovery.
6 checks on Icarus · browser samples / server checks. 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 1 steps and the final capstone