RFR (S): add CodeComments functionality to assember stubs

Christian Thalinger christian.thalinger at oracle.com
Fri Sep 21 09:23:05 PDT 2012


Looks good.

-- Chris

On Sep 21, 2012, at 12:16 AM, "Lindenmaier, Goetz" <goetz.lindenmaier at sap.com> wrote:

> Hi,
> 
> I fixed the line feeds.  The new webrev is at the same address: 
> http://cr.openjdk.java.net/~goetz/webrevs/webrev-comments_in_stubs/
> 
> Thanks a lot for reviewing and the positive comments!
>  Goetz
> 
> 
> 
> -----Original Message-----
> From: Christian Thalinger [mailto:christian.thalinger at oracle.com] 
> Sent: Freitag, 21. September 2012 02:43
> To: Lindenmaier, Goetz
> Cc: hotspot-compiler-dev at openjdk.java.net
> Subject: Re: RFR (S): add CodeComments functionality to assember stubs
> 
> 
> On Sep 19, 2012, at 5:20 AM, "Lindenmaier, Goetz" <goetz.lindenmaier at sap.com> wrote:
> 
>> Hi,
>> 
>> The Assembler and CodeBuffer classes supply CodeComment / block_comment()
>> functionality, which does not work with stubs. The comments are
>> not printed with +PrintStubCode or +PrintInterpreter because the
>> comments are lost when the code is turned into a Stub, while
>> they are kept if the code is copied to a CodeBlob.
>> 
>> We fixed this in our SAP JVM, and contributed the change to the
>> ppc-aix-port some while ago, see
>> http://hg.openjdk.java.net/ppc-aix-port/jdk7u/hotspot/rev/d65d0876ab43.
>> 
>> I propose to add this fix to the OpenJDK mainline.
>> A webrev can be found here:
>> http://cr.openjdk.java.net/~goetz/webrevs/webrev-comments_in_stubs/
>> 
>> Basically the change passes the codeBuffer to the Stub 
>> constructor, and adapts the disassembler to print the
>> comments.
>> In the debug build the  InterpreterCodelet Stub has a new field
>> holding the code comments.
>> 
>> I also added some ttyLocks and \\ns to beautify the output.
> 
> +      tty->print("\n");
> 
> Can you replace these with:
> 
> +      tty->cr();
> 
> Otherwise this looks good and we should integrate it.  Thanks for contributing!
> 
> -- Chris
> 
>> 
>> Could somebody please create a bug id for this issue and review the changes?
>> 
>> Thank you and best regards,
>>  Goetz
> 


More information about the hotspot-compiler-dev mailing list