[HW] A series of fixes to improve timing

[mp-17]

A series of fixes and modifications to ease timing closure.

Changelog

Fixed

• Bump upload and delete artifact actions
• Fix synthesis-unfriendly constructs

Changed

• Cut CVA6 - Ara combinatorial path on instruction interface
• Parametrize OS support (disabled by default)
• Cut ready-path between VLSU and MASKU
• Relax tight ADDRGEN timing path
• Time multiplex VCPOP and VFIRST ops in MASKU
• Refactor MASKU
• Remove bit-support for tail elements
• Adapt mask tests to this behavior
• Use only one commit port in CVA6

Checklist

• Automated tests pass
• Changelog updated
• Code style guideline is observed