windows/aarch64 - jittest: Unknown (code -1073741819) failures across various files
Categories
(Core :: JavaScript Engine, defect, P5)
Tracking
()
People
(Reporter: intermittent-bug-filer, Unassigned)
References
(Blocks 1 open bug)
Details
#[markdown(off)]
Filed by: egao [at] mozilla.com
https://treeherder.mozilla.org/logviewer.html#?job_id=226341303&repo=try
Observed this common failure across various files:
- tests\jit-test\jit-test\tests\basic\test-apply-many-args.js
- tests\jit-test\jit-test\tests\ion\bug1408412.js
- tests\jit-test\jit-test\tests\self-test\assertRecoveredOnBailout-1.js
and various other files.
Sample:
19:22:50 INFO - Exit code: -1073741819
19:22:50 INFO - FAIL - bug1490638.js
19:22:50 WARNING - TEST-UNEXPECTED-FAIL | tests\jit-test\jit-test\tests\bug1490638.js | Unknown (code -1073741819, args "") [0.2 s]
19:22:50 INFO - {"action": "test_start", "jitflags": "", "pid": 8108, "source": "jittests", "test": "bug1490638.js", "thread": "main", "time": 1549480970.7489998}
19:22:50 INFO - {"action": "test_end", "extra": {"jitflags": ""}, "jitflags": "", "message": "Unknown", "pid": 8108, "source": "jittests", "status": "FAIL", "test": "bug1490638.js", "thread": "main", "time": 1549480970.988}
19:22:50 INFO - INFO exit-status : -1073741819
19:22:50 INFO - INFO timed-out : False
Updated•6 years ago
|
Comment 1•6 years ago
|
||
:sdetar, here are some jittest failures on windows/arm64, is there a tracking bug I can associate with this set of failures?
Comment hidden (Intermittent Failures Robot) |
Comment 3•6 years ago
|
||
:jmaher, added to Meta bug 1526995 which I am using to track these bugs related to the JS engine. Once we have ARM64 Ion support completed, we can re-validate if these test still fail and need to be fixed.
Error code -1073741819 means a crash. We probably shouldn't group together all the test cases that crash; they are likely to be separate root causes.
Comment 5•6 years ago
|
||
(In reply to David Major [:dmajor] from comment #4)
Error code -1073741819 means a crash. We probably shouldn't group together all the test cases that crash; they are likely to be separate root causes.
Understood, what I can do is re-run the jittest suite with updated m-c code (revision from when bug was filed is probably about 1.5 weeks old now). Once re-run, I will log separate bugs for each crash.
Comment 6•6 years ago
|
||
:dmajor, I realize it has been about two weeks since Jittest was run and analyzed.
The following results are from mozilla-central revision d326a9d
.
Try: https://treeherder.mozilla.org/#/jobs?repo=try&revision=6c6bc5b8f73a27b345f96045a3a5f9762be559f7
Logs: https://taskcluster-artifacts.net/dryYij1hRVC-m5L8i3yZDw/0/public/logs/live_backing.log
As visible in the logs error codes of -1073741819
are no longer reproducible in the tests. However, these have been replaced by a new code, -1073741795
.
I will go ahead and make individual bugs for the crashes as per comment 4.
Sorry to make conflicting requests, but...
0:000> !error 0n-1073741795
Error code: (NTSTATUS) 0xc000001d (3221225501) - {EXCEPTION} Illegal Instruction An attempt was made to execute an illegal instruction.
An illegal instruction error is somewhat different from a crash. The former probably has a single root cause and the latter could be anything. For now we should probably just start with one bug for this error code.
Comment 8•6 years ago
•
|
||
(In reply to David Major [:dmajor] from comment #7)
Sorry to make conflicting requests, but...
0:000> !error 0n-1073741795
Error code: (NTSTATUS) 0xc000001d (3221225501) - {EXCEPTION} Illegal Instruction An attempt was made to execute an illegal instruction.An illegal instruction error is somewhat different from a crash. The former probably has a single root cause and the latter could be anything. For now we should probably just start with one bug for this error code.
No worries :dmajor, filed bug 1531175 to catch all of error code -1073741795
. It has been related to the parent of this bug, 1526015.
Comment hidden (Intermittent Failures Robot) |
Updated•2 years ago
|
Description
•