RFR: 8286876: NMT.test_unaliged_block_address_vm_assert fails if using clang toolchain

[Johan Sjölen]

On Fri, 2 Dec 2022 15:33:05 GMT, Thomas Stuefe <stuefe at openjdk.org> wrote:

>>>I opened https://bugs.openjdk.org/browse/JDK-8298003 to track the printout error separately and will file a PR for this shortly. I'd like to fix this quickly before JDK 20 code freeze.
>> 
>> I see that you got your reviews, so that's good. I'll fix my version by merging on top of yours and doing any necessary adjustments.
>> 
>>>Is it allowed to do that? This surprises me. We use is_aligned(this) in other places too, e.g. in oopStorage. So maybe this needs a more general solution. @kimbarrett?
>> 
>> All I can find is that C++ talks about objects having an "alignment requirement", I take that to mean that the compiler can assume that alignment for something like `this`.
>
>> > I opened https://bugs.openjdk.org/browse/JDK-8298003 to track the printout error separately and will file a PR for this shortly. I'd like to fix this quickly before JDK 20 code freeze.
>> 
>> I see that you got your reviews, so that's good. I'll fix my version by merging on top of yours and doing any necessary adjustments.
> 
> I have to wait 24hrs till tomorrow morning according to rules. Unless you give me a "its trivial" confirmation in my PR. I think its trivial.

@tstuefe , welcome back :). I hope you had a good vacation! Would you mind having a look at this again?

-------------

PR: https://git.openjdk.org/jdk/pull/11465