RFR: 8352112: [ubsan] hotspot/share/code/relocInfo.cpp:130:37: runtime error: applying non-zero offset 18446744073709551614 to null pointer

Tue Mar 18 22:10:06 UTC 2025

On Tue, 18 Mar 2025 20:34:06 GMT, Boris Ulasevich <bulasevich at openjdk.org> wrote:

> If we stick to setting reasonable value to _mutable_data, I have concerns about the chosen value. Why blob_end? We can't even be sure it's within the CodeCache range. Wouldn't it be better to set _mutable_data = header_begin()?

I really think we should use the same value for `_mutable_data` and `_immutable_data`. And I have concern about using `header_begin()` which address of code blob. `blob_end()` is already tested for `_immutable_data` - it is safe choice.

> Also, I think that we should update the _mutable_data initial value in the second CodeBlob constructor as well. 

I update it in both constructors.

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

PR Comment: https://git.openjdk.org/jdk/pull/24102#issuecomment-2734847573
PR Comment: https://git.openjdk.org/jdk/pull/24102#issuecomment-2734848300