← Design Verification Interviews
Interview Question · Debug
A Config Database Get That Found Nothing
The Question
The interviewer asks
The driver stops with this fatal error, yet the top module does set the interface. Config tracing was on. What is wrong?
UVM_INFO uvm/base/uvm_config_db_implementation.svh(485) @ 0: reporter [CFGDB/SET] Configuration scope='uvm_test_top.env.agent.*' name='vif' (type PARAMTYPEDTYPE 'T') set accessor= = (PARAMTYPEDTYPE 'T') 3815468544
UVM_INFO @ 0: reporter [RNTST] Running test smoke_test...
UVM_INFO uvm/base/uvm_config_db_implementation.svh(485) @ 0: reporter [CFGDB/GET] Configuration scope='uvm_test_top.env.agt.drv' name='vif' (type PARAMTYPEDTYPE 'T') read accessor=uvm_test_top.env.agt.drv = null (failed lookup)
UVM_FATAL main.sv(14) @ 0: uvm_test_top.env.agt.drv [NOVIF] no virtual interface for uvm_test_top.env.agt.drvYou have two and a half minutes. Speak or type, then press Get Feedback under your answer. No credits needed.
Speech-to-text runs in your browser. Chrome and Edge may send audio to their own speech service.
You can edit your answer before you send it. Feedback uses AI and checks your answer only against our written answer. No credits are used and your answer is not saved.
Reveal the Model Answer
A get matches only when three things agree: the type parameter, the scope and the field name. It returns 0 and never says which one differed. The trace lines show the cause here. The set used the scope uvm_test_top.env.agent.*, but the driver is uvm_test_top.env.agt.drv. Compare each set with the failing get, string by string.
uvm_config_db, and from the table to the code that reads it. A get travels against the arrow; the value, if found, travels with it.- Before run_test0 ns (illustrative)
The top module stores the interface handle with the scope
uvm_test_top.env.agent, then callsrun_test.Follow the calls, values and waits
- data vif = bus: tb module → uvm_config_db
- The test builds0 ns (illustrative)
The test stores
count=5with the scopeuvm_test_top.env.agnet.sqr.*. The scope is misspelled, and nothing checks a scope when it is stored.Follow the calls, values and waits
- data count = 5: incr_test → uvm_config_db
- The agent builds0 ns (illustrative)
The agent looks up its own path,
uvm_test_top.env.agent. Path, field and type all match, sogetreturns 1 and the agent passes the handle to its driver and monitor.Follow the calls, values and waits
- call get vif: env.agent → uvm_config_db
- data vif: uvm_config_db → env.agent
- data vif: env.agent → drv and mon
- The sequence starts0 ns (illustrative)
In
body()the sequence looks upuvm_test_top.env.agent.sqr.seq. No scope matches that path, sogetreturns 0 andcountkeeps its default, 3.Follow the calls, values and waits
- call get count: seq → uvm_config_db
- return no match: uvm_config_db → seq
- The report phase60 ns (illustrative)
The scoreboard reports 3 checked.
check_config_usage()lists the entry nobody read, and its scope shows the typo.
Read the diagram as text
- tb module: stores vif
- incr_test: stores count
- uvm_config_db: scope, field, type, value
- env.agent: get(this, "", "vif", vif)
- seq: get(null, get_full_name(), "count", count)
- drv and mon: drive and watch the pins
- tb module → uvm_config_db: set vif
- incr_test → uvm_config_db: set count
- uvm_config_db → env.agent: vif
- uvm_config_db → seq: count
- env.agent → drv and mon: vif handed down
A scope is a string. Only a lookup that matches it, or check_config_usage(), shows whether it is right.
Self-Check
Tick each key point you covered in your answer.
What Interviewers Listen For (Junior)
A good junior answer lists the three things a get must match: type, scope and field name. It checks each one of the set against the get.
What Interviewers Listen For (Mid-Level)
A good mid-level answer reads the trace lines to find the mismatch. It then adds check_config_usage() to report sets that nobody read.
Cross-Questions
Where the interviewer goes next. Open a node to practise that follow-up.
What IfProve ItBreak ItCompare
- A Config Database Get That Found Nothing
-
Prove ItTracing the Config DatabaseL1 Junior
Show how you would trace every set and get in the config database.
2:30You have two and a half minutes. Speak or type, then press Get Feedback under your answer. No credits needed.
Speech-to-text runs in your browser. Chrome and Edge may send audio to their own speech service.
You can edit your answer before you send it. Feedback uses AI and checks your answer only against our written answer. No credits are used and your answer is not saved.
Reveal the Model Answer
Run with
+UVM_CONFIG_DB_TRACE, or calluvm_config_db_options::turn_on_tracing()beforerun_test. Each set and each get prints a line with its scope and field. At the end,check_config_usage()lists sets that no component ever read, which is how typos usually show up.Self-Check
Tick each key point you covered in your answer.
Tool as Truth
"Raise the verbosity to UVM_DEBUG to see config lookups."
Verbosity alone does not print them. Config tracing is a separate switch.
Source uvm-core 2020.3.1 (IEEE 1800.2 library) source, base/uvm_cmdline_processor.svh, +UVM_CONFIG_DB_TRACE
-
What IfReading a Set and Get That Do Not MeetL1 JuniorL2 Mid-Level
What if the trace shows the set and the get, but they never meet? What do you compare?
2:30You have two and a half minutes. Speak or type, then press Get Feedback under your answer. No credits needed.
Speech-to-text runs in your browser. Chrome and Edge may send audio to their own speech service.
You can edit your answer before you send it. Feedback uses AI and checks your answer only against our written answer. No credits are used and your answer is not saved.
Reveal the Model Answer
Compare the scope strings first: a missing level, a typo or a wrong wildcard shows at once. Then compare the field names. Last, compare the types, because the database is keyed by type too. A set of
intis not found by a get ofbit [31:0].Self-Check
Tick each key point you covered in your answer.
Missed Corner
"The same field name means the get will find it."
The type parameter must match as well. A different type is a different entry, so the get misses.
Source uvm-core 2020.3.1 (IEEE 1800.2 library) source, base/uvm_config_db.svh
-
What IfSetting the Interface in Another Initial BlockL2 Mid-Level
What if one
initialblock sets the interface and a differentinitialblock callsrun_test?2:30You have two and a half minutes. Speak or type, then press Get Feedback under your answer. No credits needed.
Speech-to-text runs in your browser. Chrome and Edge may send audio to their own speech service.
You can edit your answer before you send it. Feedback uses AI and checks your answer only against our written answer. No credits are used and your answer is not saved.
Reveal the Model Answer
Both blocks start at time 0, and the standard does not fix their order. The bench may work on one simulator and fail on another, or after an unrelated edit. Put the
setbeforerun_testin the same block, so the order is written in the code.Self-Check
Tick each key point you covered in your answer.
Tool as Truth
"Initial blocks run in the order they appear in the file."
Many tools happen to do that, but the language leaves the order of processes at the same time undefined.
Source IEEE 1800-2017 §4.7
-
Break ItA Wildcard That Gave Two Agents One BusL2 Mid-Level
How can
set(null, "*", "vif", bus0)break a bench with two agents on two buses, without any error?2:30You have two and a half minutes. Speak or type, then press Get Feedback under your answer. No credits needed.
Speech-to-text runs in your browser. Chrome and Edge may send audio to their own speech service.
You can edit your answer before you send it. Feedback uses AI and checks your answer only against our written answer. No credits are used and your answer is not saved.
Reveal the Model Answer
The
*scope matches every component, so both agents getbus0. Neither get fails, so nothing reports a problem. The second agent drives and watches the wrong bus, and its checks see the first bus's traffic. Set each interface for its own agent path.Self-Check
Tick each key point you covered in your answer.
Half-Truth
"The second set for the other agent will override it."
Only if it comes later from the same context, or from higher up. A broad set from the top can still win.
Source Verification Academy, UVM Cookbook, Precedence Rules (PDF p140)
-
Prove ItSolve 6.4.1: Give Each Driver Its Own BusSimulator-graded
Show it, don't just say it: this Silicon Lab problem is graded by the simulator.
-
CompareConfig DB vs Resource DBL3 Senior
Compare
uvm_config_dbwithuvm_resource_db. Why do most benches use only the first?2:30You have two and a half minutes. Speak or type, then press Get Feedback under your answer. No credits needed.
Speech-to-text runs in your browser. Chrome and Edge may send audio to their own speech service.
You can edit your answer before you send it. Feedback uses AI and checks your answer only against our written answer. No credits are used and your answer is not saved.
Reveal the Model Answer
Both store values in the same resource pool.
uvm_config_dbadds the component tree: scopes are built from a context component, and build-phase precedence follows the hierarchy.uvm_resource_dbtakes raw scope strings and has no hierarchy rules. Benches use config_db because its rules match how tests override environments.Self-Check
Tick each key point you covered in your answer.
Misconception
"They are two separate databases."
config_db is a layer over the resource pool, so a value set through one can be seen through the other.
Source Accellera, UVM 1.1 User's Guide, §4.5.3 Choosing between uvm_resource_db and uvm_config_db (PDF p72)
What People Get Wrong
Common wrong answers and why they fail. Run It shows the wrong belief failing on a real simulator.
Misconception
"Any integer type matches a value set as int."
The lookup uses the exact type parameter. int and bit [7:0] are different types, so the get fails and the default stays.
Run It
The test sets width as an int; the agent reads it as bit [7:0].
get returned 0, width=32 FAIL: the test set an int but the agent asked for bit [7:0], so get found nothing and the default stayedSee the Fix
`timescale 1ns/1ns
`include "uvm_macros.svh"
import uvm_pkg::*;
class bus_agent extends uvm_agent;
`uvm_component_utils(bus_agent)
bit [7:0] width = 32;
function new(string name, uvm_component parent); super.new(name, parent); endfunction
function void build_phase(uvm_phase phase);
bit found;
int w;
found = uvm_config_db #(int)::get(this, "", "width", w); // the fix: the same type as the set
if (found) width = 8'(w);
$display("get returned %0d, width=%0d", found, width);
if (!found) $display("FAIL: the test set an int but the agent asked for bit [7:0], so get found nothing and the default stayed");
else $display("PASS: the types match and the agent got the test's width");
endfunction
endclass
class env_t extends uvm_env;
`uvm_component_utils(env_t)
bus_agent agt;
function new(string name, uvm_component parent); super.new(name, parent); endfunction
function void build_phase(uvm_phase phase); agt = bus_agent::type_id::create("agt", this); endfunction
endclass
class width_test extends uvm_test;
`uvm_component_utils(width_test)
env_t env;
function new(string name, uvm_component parent); super.new(name, parent); endfunction
function void build_phase(uvm_phase phase);
uvm_config_db #(int)::set(this, "env.agt", "width", 8);
env = env_t::type_id::create("env", this);
endfunction
endclass
module tb;
initial run_test("width_test");
endmodule
Prints: get returned 1, width=8 PASS: the types match and the agent got the test's width
Misconception
"If nothing matches, get reports an error."
get returns a bit and reports nothing. The message in the log is the bench's own fatal call.
Source IEEE 1800.2-2020 §C.4.2.2.2 get (uvm-core base/uvm_config_db.svh)
Buzzword Without Mechanism
"It's a config_db problem."
That names the area, not the cause. The useful answer says which of type, scope or field differed, and how the trace showed it.