Fix: cortexar hardware breakpoints

[litui]

Detailed description

• Hardware breakpoints are now working by bypassing the MMU at the top of the virt_to_phys function when an address is located in RAM. Without this bypass, RAM addresses were being mangled due to inaccurate phys_addr being supplied by the address translation machinery.
• Overriding soft breakpoints to hard breakpoints because GDB automatically selects soft breakpoints when a) a memory map xml is provided, and b) the address is located in RAM. This is not incorrect behaviour, but an override is necessary unless/until soft breakpoints are implemented for cortexar on BMP.

Your checklist for this pull request

• I've read the Code of Conduct
• I've read the guidelines for contributing to this repository
• It builds for hardware native (see Building the firmware)
• It builds as BMDA (see Building the BMDA)
• I've tested it to the best of my ability
• My commit messages provide a useful short description of what the commits do

Closing issues

[litui]

There, more or less fixed things back to the way I found them with the exception of the RAM address fix. Tested and continuation from a load works, and breakpoints appear to land in the correct places with the check of the address against the RAM. Will look over it all again tomorrow when I'm awake.

[litui]

For testing purposes (for the Deluge crowd), here are a couple launch.json configs for VSCode and here is the meson cross-file I'm using to build blackmagic

[litui]

This PR will need tweaking after #1916 gets merged in.

Edit: done!

[litui]

I bumped a request on cortex-debug's repo ( Marus/cortex-debug#978 ) for ability to set the break command used in gdb and then stumbled upon a comment in Marus/cortex-debug#555 which noted that OpenOCD, as a gdb server backend, offers the command gdb_breakpoint_override. Setting this override up as a BMP monitor command might be something to think about.