1945696 - Intermittent [TV] layout/svg/crashtests/1941838.html (finished) | application terminated with exit code 127

Status: RESOLVED DUPLICATE of bug 1819763

(Core :: Layout, defect, P5)

Description

[Treeherder Bug Filer]

Filed by: smolnar [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer?job_id=493108915&repo=autoland
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/eIvvW0J5TfeHaLP4eSMM7A/runs/0/artifacts/public/logs/live_backing.log
Reftest URL: https://hg.mozilla.org/mozilla-central/raw-file/tip/layout/tools/reftest/reftest-analyzer.xhtml#logurl=https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/eIvvW0J5TfeHaLP4eSMM7A/runs/0/artifacts/public/logs/live_backing.log&only_show_unexpected=1

---

REFTEST INFO | Application command: /builds/worker/workspace/build/application/firefox/firefox -marionette -profile /tmp/tmpvo7gojq2.mozrunner
[task 2025-02-04T00:26:31.989Z] 00:26:31     INFO - ### XPCOM_MEM_BLOAT_LOG defined -- logging bloat/leaks to /tmp/tmpvo7gojq2.mozrunner/runreftest_leaks.log
[task 2025-02-04T00:26:31.989Z] 00:26:31     INFO - *** You are running in chaos test mode. See ChaosMode.h. ***
[task 2025-02-04T00:26:32.826Z] 00:26:32     INFO - [WARN  rkv::backend::impl_safe::environment] `load_ratio()` is irrelevant for this storage backend.
[task 2025-02-04T00:26:33.335Z] 00:26:33     INFO - Inconsistency detected by ld.so: ../elf/dl-tls.c: 481: _dl_allocate_tls_init: Assertion `listp->slotinfo[cnt].gen <= GL(dl_tls_generation)' failed!
[task 2025-02-04T00:29:32.051Z] 00:29:32     INFO - REFTEST TEST-UNEXPECTED-FAIL | layout/svg/crashtests/1941838.html (finished) | application terminated with exit code 127
[task 2025-02-04T00:29:32.076Z] 00:29:32     INFO -  Traceback (most recent call last):
[task 2025-02-04T00:29:32.077Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/tests/reftest/runreftest.py", line 1199, in <module>
[task 2025-02-04T00:29:32.077Z] 00:29:32     INFO -      sys.exit(run_test_harness(parser, options))
[task 2025-02-04T00:29:32.077Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/tests/reftest/runreftest.py", line 1189, in run_test_harness
[task 2025-02-04T00:29:32.077Z] 00:29:32     INFO -      result = reftest.verifyTests(options.tests, options)
[task 2025-02-04T00:29:32.078Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/tests/reftest/runreftest.py", line 674, in verifyTests
[task 2025-02-04T00:29:32.078Z] 00:29:32     INFO -      result = step()
[task 2025-02-04T00:29:32.078Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/tests/reftest/runreftest.py", line 632, in step4
[task 2025-02-04T00:29:32.079Z] 00:29:32     INFO -      result = self.runTests(tests, options)
[task 2025-02-04T00:29:32.079Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/tests/reftest/runreftest.py", line 710, in runTests
[task 2025-02-04T00:29:32.079Z] 00:29:32     INFO -      return self.runSerialTests(manifests, options, cmdargs)
[task 2025-02-04T00:29:32.079Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/tests/reftest/runreftest.py", line 1107, in runSerialTests
[task 2025-02-04T00:29:32.079Z] 00:29:32     INFO -      return run()
[task 2025-02-04T00:29:32.080Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/tests/reftest/runreftest.py", line 1075, in run
[task 2025-02-04T00:29:32.080Z] 00:29:32     INFO -      status = self.runApp(
[task 2025-02-04T00:29:32.081Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/tests/reftest/runreftest.py", line 1021, in runApp
[task 2025-02-04T00:29:32.081Z] 00:29:32     INFO -      raise reraise(exc, value, tb)
[task 2025-02-04T00:29:32.081Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/six.py", line 719, in reraise
[task 2025-02-04T00:29:32.081Z] 00:29:32     INFO -      raise value
[task 2025-02-04T00:29:32.081Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/tests/reftest/runreftest.py", line 960, in runApp
[task 2025-02-04T00:29:32.081Z] 00:29:32     INFO -      marionette.start_session()
[task 2025-02-04T00:29:32.081Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/marionette_driver/decorators.py", line 34, in _
[task 2025-02-04T00:29:32.081Z] 00:29:32     INFO -      m._handle_socket_failure()
[task 2025-02-04T00:29:32.081Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/marionette_driver/marionette.py", line 820, in _handle_socket_failure
[task 2025-02-04T00:29:32.082Z] 00:29:32     INFO -      reraise(exc_cls, exc, tb)
[task 2025-02-04T00:29:32.082Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/six.py", line 719, in reraise
[task 2025-02-04T00:29:32.082Z] 00:29:32     INFO -      raise value
[task 2025-02-04T00:29:32.082Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/marionette_driver/decorators.py", line 24, in _
[task 2025-02-04T00:29:32.082Z] 00:29:32     INFO -      return func(*args, **kwargs)
[task 2025-02-04T00:29:32.082Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/marionette_driver/marionette.py", line 1368, in start_session
[task 2025-02-04T00:29:32.082Z] 00:29:32     INFO -      self.raise_for_port(timeout=timeout)
[task 2025-02-04T00:29:32.082Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/marionette_driver/marionette.py", line 744, in raise_for_port
[task 2025-02-04T00:29:32.082Z] 00:29:32     INFO -      raise socket.timeout(
[task 2025-02-04T00:29:32.082Z] 00:29:32     INFO -  socket.timeout: Timed out waiting for connection on 127.0.0.1:2828!
[task 2025-02-04T00:29:32.094Z] 00:29:32     INFO -  Exception ignored in: <function BaseProfile.__del__ at 0x7f4e70d11430>
[task 2025-02-04T00:29:32.094Z] 00:29:32     INFO -  Traceback (most recent call last):
[task 2025-02-04T00:29:32.094Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/mozprofile/profile.py", line 84, in __del__
[task 2025-02-04T00:29:32.095Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/mozprofile/profile.py", line 294, in cleanup
[task 2025-02-04T00:29:32.095Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/mozprofile/profile.py", line 92, in cleanup
[task 2025-02-04T00:29:32.095Z] 00:29:32     INFO -    File "/builds/worker/workspace/build/venv/lib/python3.8/site-packages/mozfile/mozfile.py", line 248, in remove
[task 2025-02-04T00:29:32.096Z] 00:29:32     INFO -  ImportError: sys.meta_path is None, Python is likely shutting down
[task 2025-02-04T00:29:32.104Z] 00:29:32     INFO - Return code: 1
[task 2025-02-04T00:29:32.104Z] 00:29:32     INFO - TinderboxPrint: reftest-crashtest<br/>81/0/27
[task 2025-02-04T00:29:32.105Z] 00:29:32  WARNING - setting return code to 2
[task 2025-02-04T00:29:32.105Z] 00:29:32    ERROR - TinderboxPrint: Per-test run of .../layout/svg/crashtests/1941838.html<br/>: FAILURE
[task 2025-02-04T00:29:32.105Z] 00:29:32     INFO - Per-test suites: {}
[task 2025-02-04T00:29:32.105Z] 00:29:32     INFO - Running post-action listener: _package_coverage_data
[task 2025-02-04T00:29:32.105Z] 00:29:32     INFO - Running post-action listener: _resource_record_post_action
[task 2025-02-04T00:29:32.105Z] 00:29:32     INFO - Running post-action listener: process_java_coverage_data
[task 2025-02-04T00:29:32.105Z] 00:29:32     INFO - [mozharness: 2025-02-04 00:29:32.105565Z] Finished run-tests step (success)
[task 2025-02-04T00:29:32.105Z] 00:29:32     INFO - [mozharness: 2025-02-04 00:29:32.105604Z] Running uninstall step.
[task 2025-02-04T00:29:32.105Z] 00:29:32     INFO - Running pre-action listener: _resource_record_pre_action
[task 2025-02-04T00:29:32.105Z] 00:29:32     INFO - Running main action method: uninstall

Comment 1

[BugBot [:suhaib / :marco/ :calixte]]

:ahale, since you are the author of the regressor, bug 1941838, could you take a look?

For more information, please visit BugBot documentation.

Comment 2

[Ashley Hale [:ahale]]

I don't think this is caused by bug 1941838, if there is a bug in browser startup revealed by chaos mode it is likely unrelated to this crashtest in particular, but this crashtest seems to be very good at triggering a lot of asserts in layout code that deserve investigation.

Comment 3

[Ashley Hale [:ahale]]

NeedInfo: This crashtest layout/svg/crashtests/1941838.html is causing a lot of interesting asserts in layout code, can someone take a look?

Context: The actual purpose of the crashtest was to provoke a panic in WebRender (resolved by https://phabricator.services.mozilla.com/D236306 ), but since layout is lighting up with alarms, and even crashes in chaos mode some of the time, there may be more fires here.

Comment 4

[Daniel Holbert [:dholbert]]

The assertion-failure in comment 0 is:

Inconsistency detected by ld.so: ../elf/dl-tls.c: 481: _dl_allocate_tls_init: Assertion `listp->slotinfo[cnt].gen <= GL(dl_tls_generation)' failed!

...which is bug 1819763.

There are details in that bug, but tl;dr this is a bug in glibc specifically on old versions of Ubuntu (note that the test-run linked in comment 0 here is from Ubuntu 18.04). My understanding is that we're stuck with this randomly happening until we get rid of these Ubuntu 18.04 test runners.

In the meantime, we can ignore this assertion and dupe bugs to bug 1819763 as I'm doing here. If there's a particular test that reliably triggers it (outside of test-verify mode), it might be worth considering flagging that test as skipped on Ubuntu 18.04 I suppose.

Comment 5

[Daniel Holbert [:dholbert]]

Oh, sorry, I guess I do see more non-fatal assertions further down in the test run -- that's probably what you're talking about with the layout assertions.

Those appear to be already annotated as expected failures (they landed as expected in https://hg.mozilla.org/mozilla-central/rev/519a60897139#l3.12 ), so I don't think they're causing the TEST-UNEXPECTED-FAIL status that was quoted in comment 0. Those assertions in layout are non-fatal and are about finding lengths that happen to equal sentinel values, which is not-unusual for a fuzzer testcase with huge lengths as is the case here. So I'm not too concerned about those and they don't appear to be the-thing-that-was-actively-causing-the-test-failure-here; and it's fine to annotate them as expected assertions for this fuzzer testcase.

(Ideally we should make the associated layout code more robust so that large values don't trip assertions like that, but that's not a priority, and it's not entirely clear how best to do that without removing our usage of sentinel values entirely.)

Comment 6

[Ashley Hale [:ahale]]

Yeah I am not concerned with the inconsistency error that you say is bug 1819763.

I was more surprised by having to tell the reftest infrastructure to expect 20 asserts on this crashtest, so thought someone who works on Layout may want to take a look at those, but if you think it's fine to just treat them as expected asserts then I'm satisfied with that.

Comment 7

[Daniel Holbert [:dholbert]]

Thanks - yeah it's probably really just 1-2 assertions, but the testcase has * { ....} rules which apply the same bogus-sized lengths to every element in the page, so we get ~10 copies of those 1-2 assertions (for 10 or so different boxes that are involved and all independently experience the same problem).

So: I'm not concerned, but I appreciate the heads-up. :)