[Webkit-unassigned] [Bug 231321] New: Differential testing: Weird behavior in FTL
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Wed Oct 6 12:40:48 PDT 2021
https://bugs.webkit.org/show_bug.cgi?id=231321
Bug ID: 231321
Summary: Differential testing: Weird behavior in FTL
Product: WebKit
Version: WebKit Local Build
Hardware: PC
OS: Linux
Status: NEW
Severity: Normal
Priority: P2
Component: JavaScriptCore
Assignee: webkit-unassigned at lists.webkit.org
Reporter: lukas.bernhard at rub.de
Differential testing identifies the following samples to trigger a miscomputation in FTL.
Tested on e467a9710432ebb3dae9880f897cf93929adc0e6 (Wed Oct 6 16:30:57 2021 +0000)
Sorry I couldn't minimize the testcase further, everything I try to simplify breaks the differential behavior.
Also, the bug description is obviously meaningless due to not having a hunch regarding the root cause.
Release/bin/jsc --validateOptions=true --useConcurrentJIT=false --useConcurrentGC=false --thresholdForJITSoon=10 --thresholdForJITAfterWarmUp=10 --thresholdForOptimizeAfterWarmUp=100 --thresholdForOptimizeAfterLongWarmUp=100 --thresholdForOptimizeSoon=100 --thresholdForFTLOptimizeAfterWarmUp=1000 --thresholdForFTLOptimizeSoon=1000 --validateBCE=true --useFTLJIT=true diff.js
function main() {
let v224;
const v35 = [0, 0, {b:"AAAAA"}];
async function v36(arr) {
edenGC(); // removing this break differential
for (let v198 = 0; v198 < 2; v198++) {
const v200 = [0, 0];
const v201 = `
for (let v205 = 0; v205 < 60000; v205++) { }
async function v215() { } // never called but removing breaks differential
const v222 = {"__proto__":[[]], "a":0, "b":0};
for (const v223 in v222) {
v224 = arr[v223];
v222.__proto__ = {};
}
v200;
`;
eval(v201); // moving code out of eval breaks differential
}
}
v35.filter(v36);
print(v224) // prints undefined in FTL, AAAAA without FTL (also AAAAA in v8)
}
main();
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20211006/2904334a/attachment.htm>
More information about the webkit-unassigned
mailing list