Bug 176868
| Summary: | [JSC] Investigate DFG inlining for six-speed templatestring-es6 | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Yusuke Suzuki <ysuzuki> |
| Component: | JavaScriptCore | Assignee: | Yusuke Suzuki <ysuzuki> |
| Status: | NEW | ||
| Severity: | Normal | CC: | fpizlo, saam |
| Priority: | P2 | ||
| Version: | WebKit Nightly Build | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
Yusuke Suzuki
Basically this test is super slow because the target function is not inlined.
But, this should be somewhat inlined by poly inlining / simple inlining with cleared stub.
It seems that our CallSiteInfo is cleared. And non-poly function is set for that.
But it has badFunction in exit site data, and poly stub is already cleared.
So it always has couldTakeSlowPath = true.
I'm not sure this is the right behavior for this case. I think we should have a good story to handle the above case well.
https://arewefastyet.com/#machine=29&view=single&suite=six-speed&subtest=templatestring-es6
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |