Interface Protocols · All levels
AMBA Family (APB / AHB / AXI): Tricky Q&A
Senior interview and review questions for AMBA Family (APB / AHB / AXI).
Section Q&A bank
Use these drills after completing all topics in AMBA Family (APB / AHB / AXI). Answer with layer, transaction, metric, artifact, and regression.
Why does AXI split reads and writes into separate channels?
[INT][PROT][AMBA-FAMILY]
Q: Why does AXI split reads and writes into separate channels?
A:
So address, data, and response for reads and writes can progress independently, improving throughput under latency.
FOLLOW-UP TRAP: Describing AXI as a single shared bus.What must match between AW and W for a write burst?
[INT][PROT][AMBA-FAMILY]
Q: What must match between AW and W for a write burst?
A:
Burst length, size, ID, and address alignment; W beats must follow AW rules.
FOLLOW-UP TRAP: Checking only the first W beat.When is AXI-Lite preferred over full AXI?
[INT][PROT][AMBA-FAMILY]
Q: When is AXI-Lite preferred over full AXI?
A:
Simple register access without bursts or reorder — control paths and small peripherals.
FOLLOW-UP TRAP: Using Lite for DMA traffic.Where do bridge bugs usually hide?
[INT][PROT][AMBA-FAMILY]
Q: Where do bridge bugs usually hide?
A:
ID remap, width conversion, burst splitting, cache attribute propagation, and ordering.
FOLLOW-UP TRAP: Blaming the endpoint master only.Q&A drill guide
LAYER -> TRANSACTION -> METRIC -> ARTIFACT -> FIX -> REGRESSIONSketch while answering
AMBA INTEGRATION MAP
CPU --AXI--> NIC --AXI--> SRAM
| |
+--AXI-Lite--> peripherals (GPIO, timers)
+--AXI-Stream-> video pipe
Every bridge is a contract rewrite: width, ID, burst, cache attrs.Common traps
Hung write: verify AW/W ordering and wlast alignment.
Hung read: check arready stall and rlast per ID.
Decode errors: address map vs interconnect route table.
Key takeaways
Connect every protocol claim to a transaction identity and measurable metric.
Store the artifact (waveform, log, counter) next to every signoff decision.
Common pitfalls
Debugging timeouts without finding the first bad transaction.
Quoting peak bus width without payload efficiency and retry overhead.
Treating VIP compliance as a substitute for system integration replay.