Perma tests/jit-test/jit-test/tests/gc/weak-marking-01.js | /data/local/tests/jit-tests/jit-tests/tests/gc/weak-marking-01.js:39:3 Error: Incremental GC already in progress (code 3, args "--ion-eager --ion-offthread-compile=off") [0.4 s]
Categories
(Core :: JavaScript Engine, defect, P5)
Tracking
()
Tracking | Status | |
---|---|---|
firefox66 | --- | affected |
People
(Reporter: intermittent-bug-filer, Unassigned)
References
(Blocks 1 open bug)
Details
(Whiteboard: [arm64:m3])
Filed by: bclary [at] mozilla.com
https://treeherder.mozilla.org/logviewer.html#?job_id=221073562&repo=try
While temporarily enabling debug jittests on Android hardware, I ran into this. It requires the additional patch to enable debug. It is reproducible in that it occurred in a retrigger. So far, I have only seen it on pixel2 arm64 debug.
Comment 1•6 years ago
|
||
cpeterson: You may find this interesting.
Comment hidden (Intermittent Failures Robot) |
Comment 3•6 years ago
|
||
Bob says these Jit7 test failures are perma-fail in ARM64 debug builds:
TEST-UNEXPECTED-FAIL | tests/jit-test/jit-test/tests/gc/weak-marking-01.js | /data/local/tests/jit-tests/jit-tests/tests/gc/weak-marking-01.js:39:3 Error: Incremental GC already in progress (code 3, args "--ion-eager --ion-offthread-compile=off") [0.4 s]
TEST-UNEXPECTED-FAIL | tests/jit-test/jit-test/tests/gc/weak-marking-01.js | /data/local/tests/jit-tests/jit-tests/tests/gc/weak-marking-01.js:39:3 Error: Incremental GC already in progress (code 3, args "--baseline-eager") [0.4 s]
TEST-UNEXPECTED-FAIL | tests/jit-test/jit-test/tests/gc/weak-marking-02.js | /data/local/tests/jit-tests/jit-tests/tests/gc/weak-marking-02.js:54:3 Error: Incremental GC already in progress (code 3, args "--ion-eager --ion-offthread-compile=off") [0.4 s]
TEST-UNEXPECTED-FAIL | tests/jit-test/jit-test/tests/gc/weak-marking-02.js | /data/local/tests/jit-tests/jit-tests/tests/gc/weak-marking-02.js:54:3 Error: Incremental GC already in progress (code 3, args "--baseline-eager") [0.3 s]
Adding the [arm64:m3] whiteboard tag because we should fix reproducible test crashes before letting ARM64 Fennec Nightly ride the trains to Beta.
Comment 4•6 years ago
|
||
Steve, can you take a look at this?
(I had a quick look at the test and couldn't see anything. We disable gczeal at the start and gcslice() looks like it runs with unlimited slice if no argument is supplied).
Comment 5•6 years ago
|
||
I guess this was already happening. Fortunately, pbone just fixed this.
Updated•6 years ago
|
Description
•