RESOLVED FIXED 224414
[JSC] Do not copy StringSwitchJumpTable 
https://bugs.webkit.org/show_bug.cgi?id=224414
Summary [JSC] Do not copy StringSwitchJumpTable
Yusuke Suzuki
Reported 2021-04-10 22:07:29 PDT
[JSC] Do not copy StringSwitchJumpTable
Attachments
Patch (41.30 KB, patch)
2021-04-10 22:12 PDT, Yusuke Suzuki 		keith_miller: review+
DetailsFormatted DiffDiff
View All Add attachment proposed patch, testcase, etc.
Yusuke Suzuki
Comment 1 2021-04-10 22:12:01 PDT
Created attachment 425699 [details] Patch
Yusuke Suzuki
Comment 2 2021-04-11 02:43:29 PDT
My plan is doing the similar thing for SimpleJumpTable, but that was a bit more complicated.
Keith Miller
Comment 3 2021-04-12 12:57:38 PDT
Comment on attachment 425699 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=425699&action=review r=me > Source/JavaScriptCore/ChangeLog:8 > + Previously, we are copying UnlinkedStringJumpTable to CodeBlock's StringJumpTable because we are embedding CodeLocation pointer Nit: Since previously is past tense the verb should be were rather than are. Also: "because we are embedding" => "because we embed" > Source/JavaScriptCore/ChangeLog:9 > + inside CodeBlock's StringJumpTable. This is copying mostly identical hashtable to each CodeBlock even in DFG and FTL. And this Nit: "copying mostly" => "copying a mostly". You can also drop the "and in the third sentence. > Source/JavaScriptCore/ChangeLog:10 > + even prevents us from inlining op_switch_string in DFG and FTL. Maybe explain why it prevents inlining. > Source/JavaScriptCore/ChangeLog:15 > + from copying unnecessary hashtable, and even this paves the way to inlining switch_string in DFG and FTL. Nit: hashtables*
Yusuke Suzuki
Comment 4 2021-04-12 15:01:33 PDT
Committed r275840 (236408@main): <https://commits.webkit.org/236408@main>
Radar WebKit Bug Importer
Comment 5 2021-04-12 15:02:13 PDT
<rdar://problem/76559929>
Note You need to log in before you can comment on or make changes to this bug.