RFR: 8361380: ARM32: Atomic stubs should be in pre-universe

[Aleksey Shipilev]

On Fri, 11 Jul 2025 17:02:07 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:

> See bug for more investigation.
> 
> Additional testing:
>  - [x] Linux ARM32 server fastdebug, reproducer does not fail anymore
>  - [ ] Linux x86_64 Zero fastdebug, `bootcycle-images`
>  - [ ] Linux x86_64 server fastdebug, `tier1`
>  - [ ] Linux AArch64 server fastdebug, `tier1`

Hey @adinn, I am probably confused how the stubs machinery is supposed to work. This patch looks simple, but Zero now fails with:


$ CONF=linux-x86_64-zero-fastdebug make images
Building target 'images' in configuration 'linux-x86_64-zero-fastdebug'
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  Internal Error (/home/shade/trunks/jdk/src/hotspot/share/runtime/stubInfo.cpp:169), pid=986601, tid=986604
#  assert(idx2 >= 0 && idx2 >= idx1) failed: bad blob ids first -1 and second -1
#


What's the landmine here?

Must be https://github.com/openjdk/jdk/commit/3d74cbe0ac9b68dfc161a4c079b695a320a0e133, I'll pick it up.

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

PR Comment: https://git.openjdk.org/jdk/pull/26270#issuecomment-3063070110
PR Comment: https://git.openjdk.org/jdk/pull/26270#issuecomment-3063160913