[jdk21u-dev] RFR: 8327036: [macosx-aarch64] SIGBUS in MarkActivationClosure::do_code_blob reached from Unsafe_CopySwapMemory0

[Andrew Haley]

On Thu, 29 Feb 2024 14:49:21 GMT, Richard Reingruber <rrich at openjdk.org> wrote:

> This adds a `ThreadWXEnable` to `JVM_ENTRY_FROM_LEAF`.
> Without we can get similar crashes as described in the bug report when writing to the code cache.
> Note that in jdk21 we cannot get the very same crash because the nmethod sweeper was removed in jdk20 with [JDK-8290025](https://bugs.openjdk.org/browse/JDK-8290025).
> 
> The fix passed our CI testing: JTReg tests: tier1-4 of hotspot and jdk. All of Langtools and jaxp. JCK, SPECjvm2008, SPECjbb2015, Renaissance Suite, and SAP specific tests (also with ParallelGC).
> Testing was done with fastdebug builds on the main platforms and also on Linux/PPC64le.

So that adds a WX transition to every JVM_ENTRY_FROM_LEAF? I think we need to think about the cost of that. Isn't it better to do the transition when we actually need it?

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

PR Comment: https://git.openjdk.org/jdk21u-dev/pull/305#issuecomment-1972714048