RFR: 8292590: Product JVM crashes with FLAG_SET_XXX on non-product Flag
Ioi Lam
iklam at openjdk.org
Tue Aug 30 21:52:57 UTC 2022
On Tue, 30 Aug 2022 00:27:05 GMT, David Holmes <dholmes at openjdk.org> wrote:
>>> Hardly trivial as these macros are totally obscure to begin with :)
>>
>> OK I removed the word "trivial" from the PR descroption.
>>
>>> Where is the definition of what the arguments to ALL_FLAGS means? [Found it - allFlags.hpp]
>>>
>>> It appears we have been handling develop and notproduct incorrectly. Doesn't JVMFlagsEnum also need adjustment?
>>
>> Yes, `notproduct` is not handled correctly. See [JDK-8293045](https://bugs.openjdk.org/browse/JDK-8293045), due to which JVMFlagsEnum cannot be adjusted in this PR. That would need to be done as a part of JDK-8293045l.
>
>> part of JDK-8293045l
>
> There is an extra character at the end of that bug id which causes JBS to act very strangely.
>
> I'm not clear why these issues are being separated out. We seem to have a general problem (that we have know about for a couple of years - ref the discussion in [JDK-8236736](https://bugs.openjdk.org/browse/JDK-8236736)) that notproduct flags are handled incorrectly throughout, so why not fix everything together?
Thanks @dholmes-ora and @coleenp for the review!
-------------
PR: https://git.openjdk.org/jdk/pull/10072
More information about the hotspot-runtime-dev
mailing list