Closed Bug 909826 Opened 8 years ago Closed 8 years ago

OdinMonkey: unify GenerateEntry


(Core :: JavaScript Engine, defect)

Not set





(Reporter: luke, Assigned: luke)



(3 files)

The ARM and x86/x64 GenerateEntry functions are basically the same.  (Along the same lines it'd be nice to unify the FFI paths too but that is for a future patch.)
Attached patch step 1Splinter Review
This patch changes the x86/x64 entry to take a 'globalData' argument, like ARM does.
Attachment #796115 - Flags: review?(bbouvier)
Attached patch step 2Splinter Review
Unify them.  I'm pretty sure StackDecrementForCall is doing the same calculation but I'd certainly appreciate a double check.
Attachment #796116 - Flags: review?(dtc-moz)
While I was in the area...
Attachment #796117 - Flags: review?(dtc-moz)
Comment on attachment 796115 [details] [diff] [review]
step 1

Review of attachment 796115 [details] [diff] [review]:

Cool to finally see this getting factored out!
Attachment #796115 - Flags: review?(bbouvier) → review+
Comment on attachment 796116 [details] [diff] [review]
step 2

Actually, Marty wrote these, so I should give him the reviews.
Attachment #796116 - Flags: review?(dtc-moz) → review?(mrosenberg)
Attachment #796117 - Flags: review?(dtc-moz) → review?(mrosenberg)
Comment on attachment 796116 [details] [diff] [review]
step 2

Review of attachment 796116 [details] [diff] [review]:

::: js/src/jit/AsmJS.cpp
@@ +5256,5 @@
>  #endif
> +    // ARM and x64 have a globally-pinned HeapReg (x86 uses immediates in
> +    // effective addresses).
> +#if defined(JS_CPU_X64) or defined(JS_CPU_ARM)

I am not a huge fan of c++'s English replacements for C operators.
Attachment #796116 - Flags: review?(mrosenberg) → review+
Attachment #796117 - Flags: review?(mrosenberg) → review+
You need to log in before you can comment on or make changes to this bug.