Bug 204511
| Summary: | [JSCOnly EWS] 39 JSC test failures on trunk | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Aakash Jain <aakash_jain> |
| Component: | New Bugs | Assignee: | Paulo Matos <pmatos> |
| Status: | RESOLVED FIXED | ||
| Severity: | Normal | CC: | ap, guijemont, jsc32, pmatos, webkit-bot-watchers-bugzilla |
| Priority: | P2 | ||
| Version: | Other | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| See Also: | https://bugs.webkit.org/show_bug.cgi?id=204522 | ||
Aakash Jain
~39 pre-existing JSC test failures are noticed on JSCONLY EWS on trunk, both on MIPSEL and JSCv7.
e.g.:
MIPSEL: https://ews-build.webkit.org/#/builders/25/builds/168/steps/17/logs/stdio
JSCv7: https://ews-build.webkit.org/#/builders/26/builds/147/steps/17/logs/stdio
microbenchmarks/bit-test-constant.js.default
microbenchmarks/bit-test-constant.js.dfg-eager
microbenchmarks/bit-test-constant.js.dfg-eager-no-cjit-validate
microbenchmarks/bit-test-constant.js.eager-jettison-no-cjit
microbenchmarks/bit-test-constant.js.mini-mode
microbenchmarks/bit-test-constant.js.no-cjit-collect-continuously
microbenchmarks/bit-test-constant.js.no-cjit-validate-phases
microbenchmarks/bit-test-constant.js.no-llint
microbenchmarks/bit-test-nonconstant.js.default
microbenchmarks/bit-test-nonconstant.js.dfg-eager
microbenchmarks/bit-test-nonconstant.js.dfg-eager-no-cjit-validate
microbenchmarks/bit-test-nonconstant.js.eager-jettison-no-cjit
microbenchmarks/bit-test-nonconstant.js.mini-mode
microbenchmarks/bit-test-nonconstant.js.no-cjit-collect-continuously
microbenchmarks/bit-test-nonconstant.js.no-cjit-validate-phases
microbenchmarks/bit-test-nonconstant.js.no-llint
microbenchmarks/memcpy-typed-loop-small.js.dfg-eager
microbenchmarks/memcpy-typed-loop-small.js.dfg-eager-no-cjit-validate
microbenchmarks/memcpy-typed-loop-small.js.eager-jettison-no-cjit
microbenchmarks/memcpy-typed-loop-small.js.no-cjit-collect-continuously
microbenchmarks/memcpy-typed-loop-small.js.no-cjit-validate-phases
microbenchmarks/memcpy-typed-loop-small.js.no-llint
microbenchmarks/memcpy-typed-loop.js.default
microbenchmarks/memcpy-typed-loop.js.dfg-eager
microbenchmarks/memcpy-typed-loop.js.dfg-eager-no-cjit-validate
microbenchmarks/memcpy-typed-loop.js.eager-jettison-no-cjit
microbenchmarks/memcpy-typed-loop.js.mini-mode
microbenchmarks/memcpy-typed-loop.js.no-cjit-collect-continuously
microbenchmarks/memcpy-typed-loop.js.no-cjit-validate-phases
microbenchmarks/memcpy-typed-loop.js.no-llint
stress/stack-overflow-in-yarr-byteCompile.js.default
stress/toctou-having-a-bad-time-new-array.js.default
stress/toctou-having-a-bad-time-new-array.js.dfg-eager
stress/toctou-having-a-bad-time-new-array.js.dfg-eager-no-cjit-validate
stress/toctou-having-a-bad-time-new-array.js.eager-jettison-no-cjit
stress/toctou-having-a-bad-time-new-array.js.mini-mode
stress/toctou-having-a-bad-time-new-array.js.no-cjit-collect-continuously
stress/toctou-having-a-bad-time-new-array.js.no-cjit-validate-phases
stress/toctou-having-a-bad-time-new-array.js.no-llint
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Aakash Jain
These failures do not seems to happen on build.webkit.org bots. So it might be a configuration issue.
e.g.:
https://build.webkit.org/builders/JSCOnly%20Linux%20MIPS32el%20Release/builds/4016
https://build.webkit.org/builders/JSCOnly%20Linux%20ARMv7%20Thumb2%20Release/builds/9942
Guillaume Emont
I can reproduce 3 of them locally both on arm and mips, 2 of which have just appeared on the arm buildbot: https://build.webkit.org/builders/JSCOnly%20Linux%20ARMv7%20Thumb2%20Release/builds/10119/steps/jscore-test/logs/stdio
I made a patch to (properly) skip them that's about to land:
https://bugs.webkit.org/show_bug.cgi?id=204522
(2 are timeouts and one is supposed to be skipped but the runDefault was added before the skip in a recent change).
Paulo Matos
This has now been fixed and both ARMv7 and MIPSel are green with JIT enabled since r257466 (bug 208196).
Closing this.