[Webkit-unassigned] [Bug 64582] JSC JIT does not inline GC allocation fast paths

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=64582


Oliver Hunt <oliver at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #100926|review?, commit-queue?      |review-, commit-queue-
               Flag|                            |




--- Comment #8 from Oliver Hunt <oliver at apple.com>  2011-07-14 21:58:36 PST ---
(From update of attachment 100926)
View in context: https://bugs.webkit.org/attachment.cgi?id=100926&action=review

r- but basically just because of the code duplication.  Would be good to get 32bit going as well.

> Source/JavaScriptCore/jit/JITOpcodes.cpp:352
> +    // remove the object from the free list
> +    loadPtr(Address(regT0), regT1);
> +    storePtr(regT1, &sizeClass->firstFreeCell);
> +    
> +    // initialize the object's vtable
> +    storePtr(ImmPtr(m_globalData->jsFinalObjectVPtr), Address(regT0));
> +    
> +    // initialize the object's structure
> +    storePtr(ImmPtr(m_codeBlock->globalObject()->emptyObjectStructure()), Address(regT0, JSCell::structureOffset()));
> +    
> +    // initialize the inheritor ID
> +    storePtr(ImmPtr(0), Address(regT0, JSObject::offsetOfInheritorID()));
> +    
> +    // initialize the object's property storage pointer
> +    addPtr(Imm32(sizeof(JSObject)), regT0, regT1);

I think all of this could be hoisted into a separate helper function, rather than duplicating the code in op_create_this, well i guess with the exception of the initial structure

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.