large-arraybuffers/basic.js jit-test fails on s390x
Categories
(Core :: JavaScript Engine, defect, P5)
Tracking
()
People
(Reporter: mail, Unassigned)
References
(Blocks 2 open bugs)
Details
User Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:96.0) Gecko/20100101 Firefox/96.0
Steps to reproduce:
Running the tests on s390x leads to a failure
Actual results:
/<<PKGBUILDDIR>>/js/src/jit-test/tests/large-arraybuffers/basic.js:32:13 Error: Assertion failed: got -25089, expected -99
Stack:
test1@/<<PKGBUILDDIR>>/js/src/jit-test/tests/large-arraybuffers/basic.js:32:13
@/<<PKGBUILDDIR>>/js/src/jit-test/tests/large-arraybuffers/basic.js:38:1
Exit code: 3
FAIL - large-arraybuffers/basic.js
TEST-UNEXPECTED-FAIL | js/src/jit-test/tests/large-arraybuffers/basic.js | /<<PKGBUILDDIR>>/js/src/jit-test/tests/large-arraybuffers/b
asic.js:32:13 Error: Assertion failed: got -25089, expected -99 (code 3, args "") [0.0 s]
INFO exit-status : 3
INFO timed-out : False
INFO stderr 2> /<<PKGBUILDDIR>>/js/src/jit-test/tests/large-arraybuffers/basic.js:32:13 Error: Assertion failed: got -25089, e
xpected -99
INFO stderr 2> Stack:
INFO stderr 2> test1@/<<PKGBUILDDIR>>/js/src/jit-test/tests/large-arraybuffers/basic.js:32:13
INFO stderr 2> @/<<PKGBUILDDIR>>/js/src/jit-test/tests/large-arraybuffers/basic.js:38:1
Expected results:
Test should have passed
Comment 1•2 years ago
|
||
The Bugbug bot thinks this bug should belong to the 'Core::JavaScript Engine: JIT' component, and is moving the bug to that component. Please revert this change in case you think the bot is wrong.
Comment 2•2 years ago
|
||
-99 == 0x9d
-25089 == 0x9dff
This looks like somebody loaded a signed byte and sign-extended it to 16 bits and then byte-swapped it (s390 is big-endian iirc), but other scenarios are possible.
Comment 3•2 years ago
|
||
Severity: S2 because this has a high impact for s390x users.
Priority: P5 as this is not a Tier 1 platform.
Contributions are welcome.
Updated•2 years ago
|
Comment 4•2 years ago
|
||
Do we have a simulator or anything to test big-endian platforms? I know we don't have any tier-1 big endian platforms, but it would be nice for something like this.
It would also be useful to check whether this fails when run with --jitflags=interp
. If so, this should be moved to the JavaScript Engine component instead of the JIT.
Comment 5•2 years ago
|
||
(In reply to Steve Fink [:sfink] [:s:] from comment #4)
It would also be useful to check whether this fails when run with
--jitflags=interp
. If so, this should be moved to the JavaScript Engine component instead of the JIT.
We don't have a JIT backend for s390x, so this is definitely the interpreter.
Comment 6•2 years ago
|
||
After some discussion, we think this is more appropriately as S4 bug because this is just a test case failure.
Description
•