On Thu, 7 May 2026 08:35:51 GMT, Tobias Hartmann <[email protected]> wrote:
> [JDK-8381472](https://bugs.openjdk.org/browse/JDK-8381472) added a new fast > path for acmp that emits "unsafe" loads: > https://github.com/openjdk/valhalla/blob/46dafa35ea274abed90c4c1475fbbff69bb14d6b/src/hotspot/share/opto/parse2.cpp#L2331-L2334 > > Although those loads are marked as unsafe themselves, they are still emitted > even if `InlineUnsafeOps` is false. The assert in > `Compile::flatten_alias_type` needs to be adjusted to account for this. Not > adding a regression test because this easily reproduces with an existing test > (in higher tiers). > > Thanks, > Tobias > > --------- > - [x] I confirm that I make this contribution in accordance with the [OpenJDK > Interim AI Policy](https://openjdk.org/legal/ai). This pull request has now been integrated. Changeset: 31ee5a82 Author: Tobias Hartmann <[email protected]> URL: https://git.openjdk.org/valhalla/commit/31ee5a8259aaa0e009f0231c2763cdf614b89316 Stats: 2 lines in 2 files changed: 1 ins; 0 del; 1 mod 8383991: [lworld] C2: assert(InlineUnsafeOps || StressReflectiveCode) failed: indeterminate pointers come only from unsafe ops Reviewed-by: mchevalier ------------- PR: https://git.openjdk.org/valhalla/pull/2407
