Incorrect comments in fast-path lightweight unlocking code in the interpreter

[David Holmes]

Hi Kris,

--- a/src/cpu/x86/vm/interp_masm_x86_32.cpp Thu Mar 29 13:02:11 2012 -0700
+++ b/src/cpu/x86/vm/interp_masm_x86_32.cpp Fri Mar 30 17:44:51 2012 +0800
@@ -782,7 +782,7 @@
      if (os::is_MP()) lock();
      cmpxchgptr(header_reg, Address(obj_reg, 0));

- // zero for recursive case
+ // zero for successful CAS
      jcc(Assembler::zero, done);

The comment is obviously a copy'n'paste typo. But I think the comment 
should indicate what a successful cas implies. Should it be "zero for 
successful unlock" ? If the cas fails what does that mean?

David

On 30/03/2012 7:54 PM, Krystal Mok wrote:
> Hi all,
>
> It seems like there's a couple of incorrect comments in the x86 version
> of InterpreterMacroAssembler::unlock_object().
> The SPARC version is not affected.
>
> Could anyone please review this patch: [1]
> The diff is against current tip of jdk8/jdk8/hotspot
>
> Regards,
> Kris Mok
>
> [1]: https://gist.github.com/2250418