RFR: 8308899: Introduce Classfile.Context and improve Classfile.Option(s) [v10]
Brian Goetz
briangoetz at openjdk.org
Thu Jun 8 14:09:58 UTC 2023
On Thu, 8 Jun 2023 13:37:33 GMT, Adam Sotona <asotona at openjdk.org> wrote:
>> Classfile context object and multi-state options have been discussed at https://mail.openjdk.org/pipermail/classfile-api-dev/2023-May/000321.html
>> This patch implements the proposed changes in Classfile API and fixes all affected code across JDK sources and tests.
>>
>> Please review.
>>
>> Thanks,
>> Adam
>
> Adam Sotona has updated the pull request incrementally with one additional commit since the last revision:
>
> Revert "ClassHierarchyResolver::ofSystem is now thread-unsafe"
>
> This reverts commit f3099cd5b252924392995bf65edc710c27822d2b.
Most entities in the system are immutable and therefore freely sharable
without additional coordination. Some have lazily-inflated single-field
caches that conform to the "benign race" (there is only one possible
non-default value) criteria.
The main hole is the CHA resolution cache. There are two ways to
address this: protect the cache, or don't share the context.
Unfortunately now that ClassModel refers back to the context, there's a
risk of unexpected context sharing. Here are a few options:
- Make the CHA cache thread-safe using, say, CHM;
- Make the CHA cache unshared, by using a ThreadLocal (this brings the
context back to being immutable)
- Detect when implicit use of a context happens on a thread other than
the creating thread (currently the only vector for this is
Classfile::transform), and inflate a new context with empty cache in
that case;
- Always create a fresh context with an empty cache in
Classfile::transform
On 6/8/2023 9:24 AM, Adam Sotona wrote:
>
> Unfortunately thread-unsafe context makes sharing of it in tests
> executed in parallel a nightmare.
> I can fix our Corpus tests and hope the race condition won't raise
> also somewhere else later.
> However how to explain this limitation to users?
>
> I suggest to make it always thread-safe (as the context is primary
> expected to be shared in multi-threaded environment) and users may
> make it faster in specific cases by providing non-synchronized map.
>
> —
> Reply to this email directly, view it on GitHub
> <https://github.com/openjdk/jdk/pull/14180#issuecomment-1582576426>,
> or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/AABJ4RHSMWR5Q3ANYK3RPSLXKHG7RANCNFSM6AAAAAAYQLACYQ>.
> You are receiving this because you commented.Message ID:
> ***@***.***>
>
-------------
PR Comment: https://git.openjdk.org/jdk/pull/14180#issuecomment-1582643016
More information about the core-libs-dev
mailing list