tests: expect pset failures on non-psram devices
What the developer wrote
Authored by Jon Griffiths
tests: expect pset failures on non-psram devices
Because the message stream desyncs when we send too large a message,
skip tests for unsupported config entirely (until a robust workaround is
available for this issue generally).
What changed, and why it matters
This commit only changes test code. It tells the test suite to skip or expect failures for certain large Bitcoin/Liquid transaction tests on devices without extra memory (non-PSRAM). It does not change the actual wallet firmware or fix a security bug in production code.
No security action required for this commit. Treat as a testing-infrastructure change. If investigating the underlying message-stream desync, track it as a reliability/robustness issue rather than a vulnerability unless further evidence shows exploitable behavior.
Security signals we found
Test-only change
References message-stream desync for oversized messages on non-PSRAM devices
No production/firmware code modified
Evidence from the diff
The diff modifies test_jade.py so that test_sign_psbt accepts a has_psram flag. For non-PSRAM devices, it skips PSBT inputs larger than the MAX_INPUT_MSG_SIZE limit and skips/expects failure for PSET (Liquid) cases. This is a test-harness adjustment to avoid false failures caused by message-stream desync on memory-constrained hardware; no runtime code is patched.
Changed components
test_jade.pyInspect captured patch +23 / −6
diff --git a/test_jade.py b/test_jade.py
index 885fda7..381e846 100644
--- a/test_jade.py
+++ b/test_jade.py
@@ -3024,11 +3024,28 @@ def test_liquid_blinded_commitments(jadeapi):
assert rslt == ledger_commitments[1]
-def test_sign_psbt(jadeapi, cases):
+def test_sign_psbt(jadeapi, cases, has_psram):
for txn_data in _get_test_cases(cases):
+ # Expect PSET test cases to fail for non-PSRAM devices
+ psbt_bin = txn_data['input']['psbt']
+
+ expect_pset_failure = False
+ if not has_psram:
+ # Max message size from main/process.h
+ # 69 bytes of overhead for a sign_psbt request
+ MAX_INPUT_MSG_SIZE = 1024 * 17 + 69
+ if len(psbt_bin) + 69 > MAX_INPUT_MSG_SIZE:
+ logger.warning(f'Skipping {txn_data["filename"]} large PSBT on non-psram device')
+ continue
+ if psbt_bin[2] == ord('e'):
+ expect_pset_failure = True
+ continue
+
try:
- rslt = jadeapi.sign_psbt(txn_data['input']['network'], txn_data['input']['psbt'])
+ rslt = jadeapi.sign_psbt(txn_data['input']['network'], psbt_bin)
except JadeError as err:
+ if expect_pset_failure:
+ continue # Trying to parse a PSET on an unsupported device
if 'expected_output' in txn_data:
# We expected this test to pass
assert False, f'FAILED: {err.message}: {txn_data}'
@@ -3737,8 +3754,8 @@ def run_api_tests(jadeapi, isble, qemu, authuser=False):
test_sign_tx(jadeapi, SIGN_LIQUID_TXN_TESTS, has_psram)
# Test sign psbts (app-generated cases)
- test_sign_psbt(jadeapi, SIGN_PSBT_TESTS)
- test_sign_psbt(jadeapi, SIGN_PSET_TESTS)
+ test_sign_psbt(jadeapi, SIGN_PSBT_TESTS, has_psram)
+ test_sign_psbt(jadeapi, SIGN_PSET_TESTS, has_psram)
# Short sanity-test of 12-word mnemonic
test_12word_mnemonic(jadeapi)
@@ -3772,9 +3789,9 @@ def run_api_tests(jadeapi, isble, qemu, authuser=False):
# - Mixed wallet and non-wallet inputs
# - Unusual input and change paths
# - Negative test cases (invalid PSBTs)
- test_sign_psbt(jadeapi, SIGN_PSBT_SS_TESTS)
+ test_sign_psbt(jadeapi, SIGN_PSBT_SS_TESTS, has_psram)
# Singlesig Liquid (PSET) tests
- test_sign_psbt(jadeapi, SIGN_PSET_SS_TESTS)
+ test_sign_psbt(jadeapi, SIGN_PSET_SS_TESTS, has_psram)
# Sign identity (ssh & gpg) tests require a specific mnemonic
rslt = jadeapi.set_mnemonic(TEST_MNEMONIC_12_IDENTITY)
Why this scored 11/100
Community notes
Notes can correct, qualify, or add evidence to the AI analysis. Every note shown here has been validated by a human moderator.
The AI analysis stands alone for now. Submit a note if you can add evidence or important context.