[Webkit-unassigned] [Bug 183130] [MIPS] Optimize generated JIT code for branches

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=183130

--- Comment #9 from Yusuke Suzuki <utatane.tea at gmail.com> ---
Comment on attachment 335174
  --> https://bugs.webkit.org/attachment.cgi?id=335174
MIPS branches

View in context: https://bugs.webkit.org/attachment.cgi?id=335174&action=review

>> Source/JavaScriptCore/ChangeLog:15
>> +        Secondly, in branches with immediate operand, the immediate value is 0 in many cases, and in
> 
> What happens if we just use patchableBranch64 and repatch it?
> I don't think this is safe if it becomes broken.
> How about separating patchableBranchPtr, patchableBranchPtrWithPatch etc. from branchPtr implementation and emitting nops in patchable versions?

I mean, using replacing function with PatchableJump which does not come after moveWithPatch.

-- 
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/20180320/e59c5e0c/attachment.html>