RDKEMW-19406: tr69hostif migrated to thunder - #525

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:372

• This test fetches RBUS output into rstdout and then calls rbus_get_value(param), which re-runs the RBUS command a second time. Pass the already-fetched output into rbus_get_value (after updating the helper to accept an optional output string) to avoid the duplicate subprocess call.

    param = AUDIO_BASE + ".AudioLevel"
    rstdout = rbus_get_data(param)

    assert RBUS_EXCEPTION_STRING not in rstdout, \
        f"rbus exception getting {param}"
    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    # AudioLevel is a numeric volume level (0-100)
    value = rbus_get_value(param)
    assert value.isdigit() or value.lstrip('-').isdigit(), \
        f"Expected numeric AudioLevel, got: {value}"

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:613

• This test fetches RBUS output into rstdout and then calls rbus_get_value(param), which re-runs the RBUS command a second time. Pass the existing rstdout into rbus_get_value (after updating the helper) to avoid redundant subprocess execution.

    rstdout = rbus_get_data(param)

    assert RBUS_EXCEPTION_STRING not in rstdout, \
        f"rbus exception getting {param}"
    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    # Value should be a positive integer
    value = rbus_get_value(param)
    try:
        assert int(value) >= 1
    except ValueError:
        assert False, f"Expected integer, got: {value}"

src/hostif/profiles/STBService/Components_VideoOutput_Thunder.cpp:46

• This file now uses org.rdk.DisplaySettings.getZoomSetting for AspectRatioBehaviour, but the in-repo mapping document still states this parameter is backed by org.rdk.AVOutput.getZoomMode (src/hostif/profiles/STBService/docs/thunder-migration-mapping.md:20). Please update the mapping doc (or add a short note here) so future maintainers don’t follow the outdated contract.

#define THUNDER_DS_GET_SUPPORTED_VIDEO_DISPLAYS "org.rdk.DisplaySettings.getSupportedVideoDisplays"
#define THUNDER_DS_GET_CURRENT_RESOLUTION       "org.rdk.DisplaySettings.getCurrentResolution"
#define THUNDER_DS_GET_DISPLAY_ASPECT_RATIO     "org.rdk.DisplaySettings.getDisplayAspectRatio"
#define THUNDER_DS_GET_ENABLE_VIDEO_PORT        "org.rdk.DisplaySettings.getEnableVideoPort"
#define THUNDER_DS_GET_ZOOM_SETTINGS            "org.rdk.DisplaySettings.getZoomSetting"

test/test-artifacts/native-platform/thunder-mock-server.js:279

• The mock Thunder response for org.rdk.DisplaySettings.getVideoCodecInfo uses profile strings ('Main', 'Main10') that don’t match what the STBService Thunder implementation expects (e.g., Capabilities_Thunder compares against "MAIN 10" to compute bitrate). This can cause functional tests to exercise a non-representative/unsupported path and produce different downstream values.

  'org.rdk.DisplaySettings.getVideoCodecInfo': {
    result: {
      numberOfEntries: 2,
      entries: [
        { profile: 'Main', level: 5.1 },
        { profile: 'Main10', level: 5.1 }
      ]

test/functional-tests/tests/helper_functions.py:95

• rbus_get_value() always calls rbus_get_data() internally, which means callers that already fetched RBUS output end up invoking rbuscli twice. This adds overhead and can introduce inconsistencies if the value changes between calls. Also, matching any line containing "Value" is looser than needed; it’s safer to key off lines that start with "Value" after trimming leading whitespace.

    output = rbus_get_data(param)
    for line in output.split('\n'):
        if 'Value :' in line or 'Value:' in line:
            # Extract everything after 'Value :' or 'Value:'
            return line.split(':', 1)[1].strip()

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:253

• This test fetches RBUS output into rstdout and then calls rbus_get_value(param), which re-runs the RBUS command a second time. Pass the already-fetched output into rbus_get_value (after updating the helper to accept an optional output string) to keep the test deterministic and faster.

This issue also appears in the following locations of the same file:

• line 363
• line 603

    rstdout = rbus_get_data(param)

    assert RBUS_EXCEPTION_STRING not in rstdout, \
        f"rbus exception getting {param}"
    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    value = rbus_get_value(param)
    assert value in ("Present", "Absent"), \
        f"Unexpected DisplayDevice Status: {value}"

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:455

• any(v in rstdout ...) can match substrings in the RBUS CLI framing rather than the actual parameter value. Since you now have rbus_get_value(), it’s more robust to assert on the parsed value instead.

    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    assert any(v in rstdout for v in ("Enabled", "Disabled")), \
        f"Unexpected VideoOutput Status: {rstdout}"

run_l2.sh:103

• The script uses a fixed sleep 2 after starting the Thunder mock server. This can be flaky (server may start slower) or unnecessarily slow (server starts faster). Since the mock server exposes a /status endpoint, consider polling it with a short timeout before running the STBService tests, and fail fast if the server never becomes ready.

# Start Thunder mock server for STBService tests
node test/test-artifacts/native-platform/thunder-mock-server.js &
THUNDER_MOCK_PID=$!
sleep 2

src/hostif/profiles/STBService/gtest/gtest_stbservice_thunder.cpp:1270

• The test comment still says “Thunder returns zoom mode”, but the implementation/stub now uses DisplaySettings.getZoomSetting. Updating the comment avoids confusion when maintaining the Thunder API mapping.

/* getAspectRatioBehaviour: Thunder returns zoom mode */

verify_thunder_mock_setup.sh:17

• &> /dev/null is a bash-specific redirection. This repo’s scripts are primarily /bin/sh (e.g., run_l2.sh, run_ut.sh), so using POSIX redirection keeps this check portable.

if ! command -v node &> /dev/null; then

verify_thunder_mock_setup.sh:79

• [[ ... ]] and the *pattern* match are bash-specific. If this script needs to run under /bin/sh (consistent with the rest of the repo), use a POSIX-compatible check (e.g., grep -q).

if [[ "$RESPONSE" != *'"result":true'* ]]; then
    echo "  ✗ FAIL: Unexpected response from mock server"
    echo "  Response: $RESPONSE"
    kill $MOCK_PID 2>/dev/null || true
    exit 1

run_l2.sh:112

• If the Thunder mock server fails to start, the script currently logs an error but continues to run the STBService L2 tests anyway, which can lead to misleading failures/timeouts. This should fail fast with a non-zero exit code.

else
    echo "[L2] ERROR: Thunder mock server failed to start"
    cat /tmp/thunder-mock-server.log
fi

verify_thunder_mock_setup.sh:1

• This new script uses #!/bin/bash, but other repo scripts are #!/bin/sh (e.g., run_l2.sh:1, run_ut.sh:1, test/functional-tests/automatics/automatics_test_gap.sh:1). If this is meant to run in the same embedded/CI environments, it should be POSIX-sh compatible (including removing bashisms like [[ ]] and &>).

This issue also appears in the following locations of the same file:

• line 17
• line 75

#!/bin/bash

test/functional-tests/tests/helper_functions.py:95

• rbus_get_value() always calls rbus_get_data() internally. Callers that already ran rbus_get_data() (like the updated STBService tests) end up executing the RBUS CLI twice per assertion. Allow passing the existing RBUS output so tests can parse once and avoid extra subprocess calls.

def rbus_get_value(param: str):
    """Extract just the value from RBUS output.
    
    RBUS output format:
        Parameter  1:
                      Name  : Device.Services.STBService.1.Param
                      Type  : string
                      Value : ActualValue
    
    Returns just 'ActualValue' string.
    """
    output = rbus_get_data(param)

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:253

• This test calls rbus_get_data(param) and then calls rbus_get_value(param), which re-runs the RBUS CLI a second time. Parsing the already-fetched rstdout avoids double calls and keeps log assertions tied to the same RBUS invocation.

    value = rbus_get_value(param)
    assert value in ("Present", "Absent"), \
        f"Unexpected DisplayDevice Status: {value}"

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:372

• Same as above: rbus_get_value(param) re-invokes the RBUS CLI even though rstdout is already available. Prefer parsing the existing output to avoid duplicate subprocess execution.

    value = rbus_get_value(param)
    assert value.isdigit() or value.lstrip('-').isdigit(), \
        f"Expected numeric AudioLevel, got: {value}"

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:455

• This assertion checks for substrings in the full RBUS output rather than validating the extracted Value field. That can mask formatting changes or match unintended lines. It would be more robust to extract the RBUS Value and compare it to ("Enabled", "Disabled").

    assert any(v in rstdout for v in ("Enabled", "Disabled")), \
        f"Unexpected VideoOutput Status: {rstdout}"

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:613

• rbus_get_value(param) triggers a second RBUS CLI call even though rstdout already contains the output. Parsing once avoids extra subprocess execution and reduces flakiness from back-to-back calls.

    value = rbus_get_value(param)
    try:
        assert int(value) >= 1
    except ValueError:
        assert False, f"Expected integer, got: {value}"

run_l2.sh:105

• Thunder mock server startup should (1) quote top_srcdir, (2) fail fast if node is missing, and (3) ensure the background server is always cleaned up via a trap (so it doesn’t stay running after failures/interrupts).

echo "[L2] Starting Thunder mock server on port 9998..."
cd $top_srcdir
node test/test-artifacts/native-platform/thunder-mock-server.js > /tmp/thunder-mock-server.log 2>&1 &
THUNDER_MOCK_PID=$!
sleep 3

verify_thunder_mock_setup.sh:23

• &> /dev/null is a bash-specific redirection. Also, this script relies on curl later but doesn’t check it up-front, so missing curl will cause a less-informative failure under set -e. Use POSIX redirection and add a curl presence check.

# 1. Check Node.js
echo "[1/6] Checking Node.js installation..."
if ! command -v node &> /dev/null; then
    echo "  ✗ FAIL: Node.js not found in PATH"
    exit 1
fi
NODE_VERSION=$(node --version)
echo "  ✓ Node.js installed: $NODE_VERSION"
echo

verify_thunder_mock_setup.sh:56

• pkill -f thunder-mock-server is a very broad pattern and can kill unrelated processes. Match the specific script name (and suppress pkill’s stderr like run_l2.sh does).

pkill -f thunder-mock-server || true

verify_thunder_mock_setup.sh:79

• [[ ... ]] is bash-specific; with the shebang switched to /bin/sh this will break. Use a POSIX case pattern match instead.

if [[ "$RESPONSE" != *'"result":true'* ]]; then
    echo "  ✗ FAIL: Unexpected response from mock server"
    echo "  Response: $RESPONSE"
    kill $MOCK_PID 2>/dev/null || true
    exit 1

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:370

• This re-runs rbuscli via rbus_get_value(param). Since the test already captured the RBUS output in rstdout, pass it through to avoid the extra invocation.

    value = rbus_get_value(param)

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:609

• This re-runs rbuscli via rbus_get_value(param). Since the test already captured the RBUS output in rstdout, pass it through to avoid the extra invocation.

    value = rbus_get_value(param)

verify_thunder_mock_setup.sh:1

• This repo’s test scripts are using POSIX sh (e.g., run_l2.sh starts with #!/bin/sh). Using #!/bin/bash here reduces portability on embedded/busybox environments.

This issue also appears in the following locations of the same file:

• line 15
• line 56
• line 75

#!/bin/bash

run_l2.sh:117

• If the Thunder mock server fails to start, the script currently logs the error but continues into the STBService pytest run, which will produce confusing failures/timeouts. Fail fast here so the root cause (mock server startup) is surfaced clearly in CI logs.

if kill -0 $THUNDER_MOCK_PID 2>/dev/null; then
    echo "[L2] Thunder mock server started (PID: $THUNDER_MOCK_PID)"
else
    echo "[L2] ERROR: Thunder mock server failed to start"
    cat /tmp/thunder-mock-server.log

test/functional-tests/tests/helper_functions.py:95

• rbus_get_value() always re-invokes rbus_get_data(). In tests that already captured the RBUS output, this causes an extra rbuscli call per assertion (slower and can introduce flakiness if values change between calls). Allow callers to pass the already-fetched output for parsing.

def rbus_get_value(param: str):
    """Extract just the value from RBUS output.
    
    RBUS output format:
        Parameter  1:
                      Name  : Device.Services.STBService.1.Param
                      Type  : string
                      Value : ActualValue
    
    Returns just 'ActualValue' string.
    """
    output = rbus_get_data(param)

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:251

• This re-runs rbuscli via rbus_get_value(param). Since the test already captured the RBUS output in rstdout, pass it through to avoid the extra invocation.

This issue also appears in the following locations of the same file:

• line 370
• line 609

    value = rbus_get_value(param)

run_l2.sh:117

• If the Thunder mock server fails to start, the script currently continues to run the STBService test suite, which will likely produce misleading failures. Fail fast here so CI clearly indicates infrastructure setup failure.

else
    echo "[L2] ERROR: Thunder mock server failed to start"
    cat /tmp/thunder-mock-server.log
fi

run_l2.sh:106

• Unquoted directory variables can break if the path contains spaces or glob characters. Quoting keeps the script robust across environments.

cd $top_srcdir

test/functional-tests/tests/helper_functions.py:95

• rbus_get_value() re-runs the RBUS command even when the caller already captured output, which doubles subprocess invocations and can return inconsistent values if the param changes between calls. Allow passing pre-fetched output and parse only lines that start with "Value" for more robust extraction.

def rbus_get_value(param: str):
    """Extract just the value from RBUS output.
    
    RBUS output format:
        Parameter  1:

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:251

• This currently calls RBUS twice (rbus_get_data + rbus_get_value). Pass the already captured stdout into rbus_get_value() to avoid an extra subprocess and potential value drift.

    value = rbus_get_value(param)

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:370

• This currently calls RBUS twice (rbus_get_data + rbus_get_value). Pass the already captured stdout into rbus_get_value() to avoid an extra subprocess and potential value drift.

    value = rbus_get_value(param)

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:609

• This currently calls RBUS twice (rbus_get_data + rbus_get_value). Pass the already captured stdout into rbus_get_value() to avoid an extra subprocess and potential value drift.

    value = rbus_get_value(param)

verify_thunder_mock_setup.sh:1

• Repository scripts are expected to be POSIX-sh compatible; using /bin/bash reduces portability on embedded/CI images that only provide /bin/sh.

#!/bin/bash

verify_thunder_mock_setup.sh:17

• &> is bash-specific. Use POSIX redirection so the script works under /bin/sh.

if ! command -v node &> /dev/null; then

verify_thunder_mock_setup.sh:80

• [[ ... ]] pattern matching is bash-specific. Use a POSIX case pattern so the connectivity check works under /bin/sh.

if [[ "$RESPONSE" != *'"result":true'* ]]; then
    echo "  ✗ FAIL: Unexpected response from mock server"
    echo "  Response: $RESPONSE"
    kill $MOCK_PID 2>/dev/null || true
    exit 1
fi

verify_mock_apis.sh:1

• Repository scripts are expected to be POSIX-sh compatible; using /bin/bash reduces portability on embedded/CI images that only provide /bin/sh.

#!/bin/bash

verify_mock_apis.sh:12

• cd $(pwd) is redundant (already in the current directory), spawns an extra process, and is unquoted. It can be removed.

cd $(pwd)

verify_mock_apis.sh:30

• local is not POSIX; if the script is run under /bin/sh it will fail. Use regular assignments here (variables will be function-scoped in many shells, but globally scoped in POSIX sh).

    local api_name=$1
    local expected_field=$2
    local test_name=$3

debug_thunder_ci.sh:30

• cd $(pwd) is redundant (already in the current directory), spawns an extra process, and is unquoted. It can be removed.

cd $(pwd)

debug_thunder_ci.sh:15

• Use POSIX sh and a built-in Node detection to keep this debug script runnable in minimal CI images ("which" may not exist).

#!/bin/bash

################################################################################
# Thunder Mock Server CI Debugging Script
################################################################################

echo "══════════════════════════════════════════════════════════════════════"
echo " Thunder Mock Server CI Environment Debug"
echo "══════════════════════════════════════════════════════════════════════"

echo ""
echo "[1/8] Node.js Installation:"
which node
node --version 2>&1 || echo "❌ Node.js not found!"

test/functional-tests/tests/helper_functions.py:95

• rbus_get_value() always executes a new rbuscli GET internally. Callers in this PR often already have the full RBUS output (e.g., for exception checks), so this doubles subprocess calls and can make reads inconsistent if the value changes between calls. Consider allowing callers to pass pre-fetched output (while keeping current behavior as default).

def rbus_get_value(param: str):
    """Extract just the value from RBUS output.
    
    RBUS output format:
        Parameter  1:
                      Name  : Device.Services.STBService.1.Param
                      Type  : string
                      Value : ActualValue
    
    Returns just 'ActualValue' string.
    """
    output = rbus_get_data(param)
    for line in output.split('\n'):
        if 'Value :' in line or 'Value:' in line:
            # Extract everything after 'Value :' or 'Value:'
            return line.split(':', 1)[1].strip()
    # If no Value line found, return the full output stripped
    return output.strip()

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:455

• This assertion matches substrings anywhere in the full RBUS CLI output, which can mask unexpected values (e.g., if the output contains additional text including “Enabled/Disabled”). Prefer validating the actual RBUS Value : line.

    assert any(v in rstdout for v in ("Enabled", "Disabled")), \
        f"Unexpected VideoOutput Status: {rstdout}"

run_l2.sh:106

• cd $top_srcdir is unquoted and its failure isn’t checked. If the directory change fails (or the path contains spaces), the mock server may be started from the wrong location and the failure can be hard to diagnose.

cd $top_srcdir

src/hostif/profiles/STBService/Components_VideoOutput_Thunder.cpp:311

• getHDCP() silently treats malformed JSON as “not compliant” when invokeThunderPluginMethod() succeeds but cJSON_Parse() fails. Logging that parse failure (similar to the other Thunder helper extractors) would make diagnosing backend/middleware issues much easier.

        cJSON *root = cJSON_Parse(response.c_str());
        if (root)
        {
            cJSON *result  = cJSON_GetObjectItem(root, "result");
            cJSON *hdcpObj = result ? cJSON_GetObjectItem(result, "HDCPStatus") : NULL;
            cJSON *field   = hdcpObj ? cJSON_GetObjectItem(hdcpObj, "isHDCPCompliant") : NULL;

src/hostif/profiles/STBService/Components_VideoOutput_Thunder.cpp:311

• getHDCP() silently treats malformed/empty Thunder responses as “not compliant” without logging parse/shape problems. This makes it hard to distinguish real “false” from backend/serialization issues.

    if (invokeThunderPluginMethod(THUNDER_HDCP_GET_STATUS, "{}", response))
    {
        cJSON *root = cJSON_Parse(response.c_str());
        if (root)
        {
            cJSON *result  = cJSON_GetObjectItem(root, "result");
            cJSON *hdcpObj = result ? cJSON_GetObjectItem(result, "HDCPStatus") : NULL;
            cJSON *field   = hdcpObj ? cJSON_GetObjectItem(hdcpObj, "isHDCPCompliant") : NULL;
            if (cJSON_IsBool(field))
                hdcpEnabled = cJSON_IsTrue(field);
            else
                RDK_LOG(RDK_LOG_WARN, LOG_TR69HOSTIF, "[%s] Missing isHDCPCompliant in HDCPStatus\n", __FUNCTION__);
            cJSON_Delete(root);
        }
    }

run_l2.sh:109

• The Thunder mock server is started in the background but there’s no trap/cleanup handler for early exits (e.g., server readiness failure, ctrl-c, or a later exit 1). Also cd $top_srcdir is unquoted, which can break if the path contains spaces.

# Clean up any stale Thunder mock server from previous runs
echo "[L2] Cleaning up stale Thunder mock servers..."
pkill -f "thunder-mock-server.js" 2>/dev/null || true
sleep 1

# Start Thunder mock server for STBService tests
echo "[L2] Starting Thunder mock server on port 9998..."
cd $top_srcdir
node test/test-artifacts/native-platform/thunder-mock-server.js > /tmp/thunder-mock-server.log 2>&1 &
THUNDER_MOCK_PID=$!
sleep 3

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:455

• VideoOutput.1.Status test now checks for substrings in the full rbuscli output (any(v in rstdout ...)), which can yield false positives if those tokens appear outside the actual Value line. Prefer extracting the RBUS value and asserting exact membership.

    param = VIDOUT_BASE + ".Status"
    rstdout = rbus_get_data(param)

    assert RBUS_EXCEPTION_STRING not in rstdout, \
        f"rbus exception getting {param}"
    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    assert any(v in rstdout for v in ("Enabled", "Disabled")), \
        f"Unexpected VideoOutput Status: {rstdout}"

src/hostif/profiles/STBService/Components_VideoOutput_Thunder.cpp:66

• On failure to fetch Thunder video displays, this destroys ifHash and sets it to NULL, but getInstance() unconditionally calls g_hash_table_lookup(ifHash, ...) after buildPortNameHash(). If the Thunder call fails, this can dereference a NULL hash table and crash the process. Keeping ifHash as an empty hash table (or guarding in getInstance) avoids the NULL deref while still returning no instances.

    if (!invokeThunderPluginMethodAndExtractDelimitedStringArrayField(
            THUNDER_DS_GET_SUPPORTED_VIDEO_DISPLAYS, "{}", "supportedVideoDisplays", ",", delimitedPorts))
    {
        RDK_LOG(RDK_LOG_WARN, LOG_TR69HOSTIF, "[%s:%d] Failed to get video displays\n", __FUNCTION__, __LINE__);
        g_hash_table_destroy(ifHash);
        ifHash = NULL;
        return;

src/hostif/profiles/STBService/Components_SPDIF_Thunder.cpp:83

• Same NULL-dereference risk as other STBService components: if the Thunder call fails, this sets ifHash = NULL, but getInstance() calls g_hash_table_lookup(ifHash, ...) without re-checking. Leaving ifHash as an empty hash table avoids a crash on lookup.

    if (!invokeThunderPluginMethodAndExtractDelimitedStringArrayField(
            THUNDER_DS_GET_SUPPORTED_AUDIO_PORTS, "{}", "supportedAudioPorts", ",", delimitedPorts))
    {
        RDK_LOG(RDK_LOG_WARN, LOG_TR69HOSTIF,
                "[%s:%d] Failed to get supported audio ports\n", __FUNCTION__, __LINE__);
        g_hash_table_destroy(ifHash);
        ifHash = NULL;
        return;

src/hostif/profiles/STBService/Components_AudioOutput_Thunder.cpp:80

• If the Thunder call fails, ifHash is destroyed and set to NULL, but getInstance() calls g_hash_table_lookup(ifHash, ...) immediately after buildPortNameHash() without re-checking. This can crash on NULL dereference when Thunder is unavailable. Prefer leaving ifHash as an empty hash table (or add a guard in getInstance).

    if (!invokeThunderPluginMethodAndExtractDelimitedStringArrayField(
            THUNDER_DS_GET_SUPPORTED_AUDIO_PORTS, "{}", "supportedAudioPorts", ",", delimitedPorts))
    {
        RDK_LOG(RDK_LOG_WARN, LOG_TR69HOSTIF,
                "[%s:%d] Failed to get supported audio ports\n", __FUNCTION__, __LINE__);
        g_hash_table_destroy(ifHash);
        ifHash = NULL;
        return;

src/hostif/profiles/STBService/Components_HDMI_Thunder.cpp:62

• If the Thunder call fails, this sets ifHash = NULL, but getInstance() calls g_hash_table_lookup(ifHash, ...) right after buildPortNameHash() without re-checking. That can cause a NULL dereference crash when Thunder is down. Leaving ifHash allocated (empty) avoids the crash while still yielding no instances.

    if (!invokeThunderPluginMethodAndExtractDelimitedStringArrayField(
            THUNDER_DS_GET_SUPPORTED_VIDEO_DISPLAYS, "{}", "supportedVideoDisplays", ",", delimitedPorts))
    {
        RDK_LOG(RDK_LOG_WARN, LOG_TR69HOSTIF, "[%s:%d] Failed to get video displays\n", __FUNCTION__, __LINE__);
        g_hash_table_destroy(ifHash);
        ifHash = NULL;
        return;

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:372

• This test runs rbuscli get twice: once via rbus_get_data(param) and again via rbus_get_value(param) (which calls rbus_get_data internally). That adds avoidable overhead and can duplicate Thunder activity. Parse the value from rstdout to keep the test single-shot.

    """
    param = AUDIO_BASE + ".AudioLevel"
    rstdout = rbus_get_data(param)

    assert RBUS_EXCEPTION_STRING not in rstdout, \
        f"rbus exception getting {param}"
    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    # AudioLevel is a numeric volume level (0-100)
    value = rbus_get_value(param)
    assert value.isdigit() or value.lstrip('-').isdigit(), \
        f"Expected numeric AudioLevel, got: {value}"

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:613

• Like other updated assertions, this calls rbus_get_value(param) which executes rbuscli get a second time. That doubles runtime and can introduce extra Thunder logs. Parse the Value line from rstdout instead to keep the test deterministic and faster.

    param = (CAPS_BASE +
             ".VideoDecoder.X_RDKCENTRAL-COM_MPEGHPart2.ProfileLevelNumberOfEntries")
    rstdout = rbus_get_data(param)

    assert RBUS_EXCEPTION_STRING not in rstdout, \
        f"rbus exception getting {param}"
    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    # Value should be a positive integer
    value = rbus_get_value(param)
    try:
        assert int(value) >= 1
    except ValueError:
        assert False, f"Expected integer, got: {value}"

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:251

• rbus_get_value(param) re-runs rbuscli get, so this test currently executes the GET twice (once for rstdout, again inside rbus_get_value). That adds extra subprocess overhead and can add extra Thunder calls/log lines, making failures harder to interpret. Parse the value from the already-captured rstdout instead.

This issue also appears in the following locations of the same file:

• line 362
• line 601

    param = DISPDEV_BASE + ".Status"
    rstdout = rbus_get_data(param)

    assert RBUS_EXCEPTION_STRING not in rstdout, \
        f"rbus exception getting {param}"
    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    value = rbus_get_value(param)
    assert value in ("Present", "Absent"), \
        f"Unexpected DisplayDevice Status: {value}"

run_l2.sh:110

• The mock server is started in the background, but if any subsequent command (including the pytest run) fails, the script will exit before reaching the explicit stop section, leaving the mock server running and potentially affecting later steps/reruns. Add an EXIT trap cleanup right after capturing THUNDER_MOCK_PID, and quote $top_srcdir when changing directories.

# Start Thunder mock server for STBService tests
echo "[L2] Starting Thunder mock server on port 9998..."
cd $top_srcdir
node test/test-artifacts/native-platform/thunder-mock-server.js > /tmp/thunder-mock-server.log 2>&1 &
THUNDER_MOCK_PID=$!
sleep 3

src/hostif/profiles/STBService/Components_VideoOutput_Thunder.cpp:66

• buildPortNameHash() destroys ifHash and sets it to NULL on Thunder failure, but getInstance() unconditionally calls g_hash_table_lookup(ifHash, ...) afterwards. If the Thunder call fails at runtime, this can dereference NULL and crash the process.

    {
        RDK_LOG(RDK_LOG_WARN, LOG_TR69HOSTIF, "[%s:%d] Failed to get video displays\n", __FUNCTION__, __LINE__);
        g_hash_table_destroy(ifHash);
        ifHash = NULL;
        return;

src/hostif/profiles/STBService/Components_SPDIF_Thunder.cpp:83

• buildPortNameHash() sets ifHash to NULL on failure, but getInstance() will still call g_hash_table_lookup(ifHash, ...) and can crash if the Thunder call fails.

        RDK_LOG(RDK_LOG_WARN, LOG_TR69HOSTIF,
                "[%s:%d] Failed to get supported audio ports\n", __FUNCTION__, __LINE__);
        g_hash_table_destroy(ifHash);
        ifHash = NULL;
        return;

src/hostif/profiles/STBService/Components_HDMI_Thunder.cpp:62

• buildPortNameHash() sets ifHash to NULL on failure, but getInstance() later calls g_hash_table_lookup(ifHash, ...) without a NULL check. If the Thunder query fails, this can lead to a NULL dereference.

        RDK_LOG(RDK_LOG_WARN, LOG_TR69HOSTIF, "[%s:%d] Failed to get video displays\n", __FUNCTION__, __LINE__);
        g_hash_table_destroy(ifHash);
        ifHash = NULL;
        return;

src/hostif/profiles/STBService/Components_AudioOutput_Thunder.cpp:80

• buildPortNameHash() sets ifHash to NULL on failure, but getInstance() uses g_hash_table_lookup(ifHash, ...) without checking for NULL. If the supported-audio-ports Thunder call fails, this can crash.

        RDK_LOG(RDK_LOG_WARN, LOG_TR69HOSTIF,
                "[%s:%d] Failed to get supported audio ports\n", __FUNCTION__, __LINE__);
        g_hash_table_destroy(ifHash);
        ifHash = NULL;
        return;

test/functional-tests/tests/helper_functions.py:95

• rbus_get_value() always executes a fresh rbus_get_data() call, which causes duplicate RBUS CLI invocations in tests that already captured rstdout. Allow callers to pass the already-fetched output to avoid extra processes and reduce flakiness.

def rbus_get_value(param: str):
    """Extract just the value from RBUS output.
    
    RBUS output format:
        Parameter  1:

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:370

• This test calls RBUS twice (once for rstdout, then again inside rbus_get_value()). Use the already-fetched rstdout when extracting the value to reduce runtime and avoid inconsistent results between calls.

    assert RBUS_EXCEPTION_STRING not in rstdout, \
        f"rbus exception getting {param}"
    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    # AudioLevel is a numeric volume level (0-100)
    value = rbus_get_value(param)
    assert value.isdigit() or value.lstrip('-').isdigit(), \
        f"Expected numeric AudioLevel, got: {value}"

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:455

• The assertion checks for substrings in the full RBUSCLI output. This can pass even if the actual Value is different (e.g., the token appears elsewhere in the output). Extract and assert on the parsed RBUS Value field instead.

    assert RBUS_EXCEPTION_STRING not in rstdout, \
        f"rbus exception getting {param}"
    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    assert any(v in rstdout for v in ("Enabled", "Disabled")), \
        f"Unexpected VideoOutput Status: {rstdout}"

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:609

• Like the other updated tests, this runs RBUS twice (once for rstdout, then again inside rbus_get_value()). Pass the existing RBUS output to rbus_get_value() so the parsing is done on a single command result.

    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    # Value should be a positive integer
    value = rbus_get_value(param)
    try:
        assert int(value) >= 1
    except ValueError:
        assert False, f"Expected integer, got: {value}"

test/functional-tests/tests/tr69hostif_stbservice_thunder.py:251

• This test fetches RBUS output twice (rbus_get_data() into rstdout, then rbus_get_value() which runs rbus_get_data() again). Pass the captured output into rbus_get_value() to avoid a second RBUS CLI invocation.

    assert CURL_OK_MSG in grep_tr69hostiflogs(CURL_OK_MSG)
    value = rbus_get_value(param)
    assert value in ("Present", "Absent"), \
        f"Unexpected DisplayDevice Status: {value}"

run_l2.sh:169

• Several pytest suites that previously ran as part of the L2 script are now commented out. For a high-risk migration, this reduces regression coverage and can let unrelated breakages slip through CI.

#pytest --json-report --json-report-summary --json-report-file $RESULT_DIR/device_info.json test/functional-tests/tests/tr69hostif_device_info.py
#pytest --json-report --json-report-summary --json-report-file $RESULT_DIR/interfacestack.json test/functional-tests/tests/tr69hostif_interfacestack.py
#pytest --json-report --json-report-summary --json-report-file $RESULT_DIR/opsdevicemgmt_logging.json test/functional-tests/tests/tr69hostif_opsdevicemgmt_logging.py
#pytest --json-report --json-report-summary --json-report-file $RESULT_DIR/opsdevicemgmt_rpc.json test/functional-tests/tests/tr69hostif_opsdevicemgmt_rpc.py
#pytest --json-report --json-report-summary --json-report-file $RESULT_DIR/storageservice.json test/functional-tests/tests/tr69hostif_storageservice.py