Closed Bug 1663480 Opened 4 years ago Closed 3 years ago

Intermittent js/src/jit-test/tests/wasm/regress/baseline-stack-height.js | Unknown (code -11, args "--disable-wasm-huge-memory") [0.5 s]

Categories

(Core :: JavaScript: WebAssembly, defect, P5)

ARM64
Unspecified
defect

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: intermittent-bug-filer, Unassigned)

References

(Blocks 1 open bug)

Details

(Keywords: intermittent-failure)

Filed by: ncsoregi [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer.html#?job_id=315026105&repo=mozilla-central
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/RJ6uee2hRbu3EU2G-MwMoQ/runs/0/artifacts/public/logs/live_backing.log


[task 2020-09-07T16:09:29.465Z] TEST-PASS | js/src/jit-test/tests/wasm/regress/brtable-conditionblock-folding.js | Success (code 0, args "--test-wasm-await-tier2") [0.3 s]
[task 2020-09-07T16:09:29.519Z] Exit code: -11
[task 2020-09-07T16:09:29.519Z] FAIL - wasm/regress/baseline-stack-height.js
[task 2020-09-07T16:09:29.519Z] TEST-UNEXPECTED-FAIL | js/src/jit-test/tests/wasm/regress/baseline-stack-height.js | Unknown (code -11, args "--disable-wasm-huge-memory") [0.5 s]
[task 2020-09-07T16:09:29.520Z] INFO exit-status : -11
[task 2020-09-07T16:09:29.520Z] INFO timed-out : False
[task 2020-09-07T16:09:29.522Z] TEST-PASS | js/src/jit-test/tests/wasm/regress/bug1300546.js | Success (code 0, args "--wasm-compiler=ion") [0.2 s]

Component: JavaScript Engine → Javascript: WebAssembly

This is on arm64-sim and it's possible that this is cranelift, because despite the name of the test there is not a guard that the test is only run with baseline. Since it's intermittent it's possible it was caused by one of the intermediate states of the code over the last week as we worked out the problems of enabling cranelift, but this is definitely worth watching.

Blocks: 1649932
Severity: normal → S3
Priority: P5 → P1
Hardware: Unspecified → ARM64
Status: NEW → RESOLVED
Closed: 4 years ago
Resolution: --- → FIXED

Lars, there was a recent occurrence of this issue here:
https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=316273243&repo=autoland&lineNumber=12742

Should we reopen this bug?

Flags: needinfo?(lhansen)

Yeah, might as well.

Status: RESOLVED → REOPENED
Flags: needinfo?(lhansen)
Resolution: FIXED → ---

Hm, does cranelift support --disable-wasm-huge-memory, I wonder? I suspect probably not. Have asked Benjamin.

(In reply to Lars T Hansen [:lth] from comment #8)

Hm, does cranelift support --disable-wasm-huge-memory, I wonder? I suspect probably not. Have asked Benjamin.

Benjamin says "it does".

I have tried to provoke this locally for the arm64-sim build using both --test-wasm-await-tier2 and --wasm-compiler=cranelift in combination with --disable-wasm-huge-memory, but so far no luck with both release and debug builds.

Not that those switches should really matter. The test is a compile-only test, it doesn't run any wasm code. With some extra logging I have verified that the code is compiled with tiering, so it could be either compiler. There's a segfault, but no stack in the log. If we had fixed bug 1662406 we might know more.

Could be an unhandled OOM on a constrained test system, I suppose - cranelift has been known to have not-great OOM handling in the past.

Priority: P1 → P3
Priority: P3 → P5

No errors for quite some time, and together with speculation that this may have been cranelift, we move on.

Status: REOPENED → RESOLVED
Closed: 4 years ago3 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.