RFR (S): 8022494: Make compilation IDs sequential

[Christian Thalinger]

Only minor nits:
src/share/vm/compiler/compileBroker.cpp

+      if (!method->code()) {
Can you compare against NULL?  It makes it more obvious that it is a pointer; a ! always looks like a boolean to me.
src/share/vm/runtime/sharedRuntime.cpp:

+      if(!method->is_static()) {
+        if((ss.type() == T_LONG) || (ss.type() == T_DOUBLE)) {
Can you add a space after "if"?
-  return nm;
+  return;
No need for a return here.

Otherwise this looks good.

-- Chris

On Aug 6, 2013, at 10:55 PM, Albert Noll <albert.noll at oracle.com> wrote:

> Hi all,
> 
> please review this small patch.
> 
> jbs: https://jbs.oracle.com/bugs/browse/JDK-8022494
> webrev: http://cr.openjdk.java.net/~anoll/8022494/webrev.00/
> 
> Problem: Compilation IDs can be non-sequential, if
> AdapterHandlerLibrary::create_native_wrapper(methodHandle method, int compile_id) {}
> returns prior to issuing the compilation, i.e., that happens if
> method->code() is not NULL.
> 
> Solution: Move the check if method is already compiled (method->code() != NULL) prior
> to incrementing the compilation ID. The current patch does NOT guarantee sequential compilation
> IDs if AdapterHandlerLibrary::buffer_blob() returns null. Please let me know if this is an issue that should 
> be considered an a revision of the patch.
> The patch furthermore contains minor code clean-ups.
> 
> Many thanks in advance,
> Albert
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20130807/7cf17952/attachment-0001.html