It seems like quite a few regressions were fixed in the latest minor release of 23.3, so should we maybe rebase to a newer minor version 23.3.x?

For the record, these issues were explicitly mentioned:

• "[ANNOUNCE] mesa 23.3.1": https://lists.freedesktop.org/archives/mesa-dev/2023-December/226085.html
• "[ANNOUNCE] mesa 23.3.2": https://lists.freedesktop.org/archives/mesa-dev/2023-December/226088.html
• "[ANNOUNCE] mesa 23.3.3": https://lists.freedesktop.org/archives/mesa-dev/2024-January/226096.html

If we do this rebase (which depends on the chance of regressions), we coordinate it so the testing aligns with that of the RHEL 9.4 stable backport (RHEL-1351). Ideally, we also know what impact it might have in terms of hardware generations, so we know what to test.