1609402 - jit-test/tests/gc/bug-1505622.js fails on arm64-sim on a sufficiently busy system

Status: RESOLVED FIXED

(Core :: JavaScript: GC, defect, P2)

Description

[Lars T Hansen [:lth]]

Linux, --enable-debug --disable-optimize --without-intl-api --enable-simulator=arm64:

$ ../jit-test/jit_test.py dist/bin/js gc/bug-1
[176|  0|  0|  0]  91% =========================================>     |  79.5s
/home/lhansen/m-c/js/src/jit-test/tests/gc/bug-1505622.js:38:9 Error: Assertion failed: got 1, expected 0
Stack:
  @/home/lhansen/m-c/js/src/jit-test/tests/gc/bug-1505622.js:38:9
Exit code: 3
FAIL - gc/bug-1505622.js
[192|  1|  0|  0] 100% ==============================================>| 199.0s
FAILURES:
    gc/bug-1505622.js
TIMEOUTS:

$ ../jit-test/jit_test.py dist/bin/js gc/bug-1505
[1|0|0|0] 100% ==========================================================>|  15.6s
PASSED ALL

So far as I can tell, it is completely consistent: running the test alone, it never fails. Running it with a bunch of other tests, and it fails reliably.

Comment 1

[Jon Coppeard (:jonco)]

Attachment: Bug 1609402 - Improve flakey test that relied on being able to allocate 1MB of GC things in under 5 seconds r?sfink

Tests that involve timing are often problematic. Here we're checking that hitting the heap limit inside the five second time limit we set for skipping last ditch GCs doesn't trigger a GC, but allocating enough to hit this limit can take longer than five seconds on a busy system.

The patch makes the max heap size much smaller and so reduces the possibility that that this can happen.

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/86c17304b4ad
Improve flakey test that relied on being able to allocate 1MB of GC things in under 5 seconds r=sfink

Comment 3

[Razvan Maries]

https://hg.mozilla.org/mozilla-central/rev/86c17304b4ad