Tests with Unsafe not compiling since October 24. Testbug?

[Lois Foltan]

On 11/8/2018 7:26 AM, Lois Foltan wrote:

> On 11/8/2018 6:22 AM, Lindenmaier, Goetz wrote:
>
>> Hi,
>>
>> I see a row of test with Unsafe failing since about October 24 with 
>> compilation errors:
>>    compiler/c2/Test7190310_unsafe.java
>>    compiler/intrinsics/unsafe/TestCAEAntiDep.java
>> compiler/intrinsics/unsafe/TestUnsafeMismatchedArrayFieldAccess.java
>>    compiler/profiling/UnsafeAccess.java
>>    compiler/unsafe/GetUnsafeObjectG1PreBarrier.java
>>    compiler/unsafe/JdkInternalMiscUnsafeAccessTestObject.java
>>    compiler/unsafe/MixedUnsafeStoreObject.java
>>    compiler/unsafe/OpaqueAccesses.java
>>    compiler/unsafe/UnsafeGetStableArrayElement.java
>>    runtime/Unsafe/GetPutObject.java
>>    runtime/Unsafe/RangeCheck.java
>>
>> Errors look like this:
>>
>> /priv/jvmtests/output_sapjvm12_o_jdk_dbgU_linuxx86_64/grmpf/testdata/jtreg/jtreg_test_12/test/hotspot/jtreg/compiler/c2/Test7190310_unsafe.java:139: 
>> error: cannot find symbol
>>          return unsafe.getReference(ref, referent_offset);
>> /priv/jvmtests/output_sapjvm12_o_jdk_dbgU_linuxx86_64/grmpf/testdata/jtreg/jtreg_test_12/test/hotspot/jtreg/runtime/Unsafe/GetPutObject.java:45: 
>> error: cannot find symbol
>>          assertEquals(t.o, unsafe.getReference(t, offset));
>>
>> Is anybody aware of this issue or already fixing it? Should I open a 
>> bug for this?
>
> Hi Goetz,
>
> Recently jdk.internal.misc.Unsafe::xxxObject and xxxReference were 
> renamed.  See https://bugs.openjdk.java.net/browse/JDK-8207146.  I 
> don't see a change set associated with the bug even though its status 
> is resolved.  So it is unclear to me if this change went into the 
> valhalla or the jdk repo.  Please do go forward and open a bug for 
> these tests.
>
> Thanks for pointing this out!
> Lois

I was able to find the relevant RFR review email for this change. See 
http://mail.openjdk.java.net/pipermail/hotspot-dev/2018-October/034695.html. 
So this was pushed to the jdk repo and these tests are passing currently 
on our end.  So it maybe that you have a source skew issue?

Thanks,
Lois

>
>>
>> Best regards,
>>    Goetz.
>