Remove isAOT check in order to enable relo logging for JITServer

[dchopra001]

Signed-off-by: Dhruv Chopra Dhruv.C.Chopra@ibm.com

[dchopra001]

@mpirvu Can I get a review please?

[mpirvu]

Previoulsy, the tracing was done only for AOT compilations. After your change, with the right options, tracing can be done for any method and I assume this will work because if there is no relocation data, nothing will be printed.
I would like some validation/reassurance that cg->getAheadOfTimeCompile() is always non-null.

[dchopra001]

@mpirvu The constructor in OMR's CodeGenerator class initializes _aheadOfTimeCompile too NULL.

The z, p, aarch64 and x codgens in OpenJ9 unconditionally initialize the _aheadOfTimeCompile field. Here's an example on Power:
https://github.com/eclipse/openj9/blob/14754c255f9b47bb3d7f664196612e48c1a2adb7/runtime/compiler/p/codegen/J9CodeGenerator.cpp#L54

So I guess technically it's possible for cg->getAheadOfTimeCompile() to be null. I didn't see this issue because I've only debugged in OpenJ9 for x, p, and z so far.

Therefore I've added a cg->getAheadOfTimeCompile() guard here. This should allow relocation tracing to be enabled for remote as well as AOT compiles, and only if the cg->getAheadOfTimeCompile() object exists.

Please take a look.

[mpirvu]

Since cg->getAheadOfTimeCompile() always returns non-null value, it follows that dumpRelocationData(); will always be executed when comp->getOption(TR_TraceRelocatableDataCG) || comp->getOption(TR_TraceRelocatableDataDetailsCG) || comp->getOption(TR_TraceReloCG) returns true. If that is safe to do for non-AOT compilations, then I am ok with this change.

Therefore, I am suggesting a test, non-AOT, non-JITServer, with one the options above enabled.

[dchopra001]

@mpirvu
I just ran a test with the following command line options:

jvm/jre/bin/java -Xjit:count=0,{*HashMap*}\(traceFull,traceCG,log=trace.log,traceRelocatableDataCG,traceRelocatableDataDetailsCG,traceRelocatableDataDetailsRT,traceRelocatableDataRT,traceReloCG\),verbose,vlog=vlog,rtlog=rtlog_name,aotrtDebugLevel=9 -version

and I see the following output in the trace log

<relocatableDataCG>
Code start =        0, Method start pc = b41356fc, Method start pc offset = 0xb41356fc
No relocation data allocated
</relocatableDataCG>

And obviously there are no crashes.

So it looks like we should be okay to merge this.

[mpirvu]

LGTM

[fjeremic]

@genie-omr build all