Fix: Ensure F-packet replies from gdb are reliably detected

[TechnoMancer]

Detailed description

• This fixes an issue.
• The origin of the issue was the change contributed in #1686 which added the loop in semihosting_get_gdb_response.
• The check for an F-packet reply was too late as gdb_main_loop may have already clobbered the packet by the time this is checked.

Collaborative inspection of debug logs from #1845 and provided by others pointed to inconsistent behaviour dealing with remote file i/o completion, which was narrowed down to the F-packet reply check being done after the packet had been processed and the packet buffer possibly clobbered.

This is corrected by checking if the received packet is an F-packet first then processing it and using the result of the check to exit the loop after the processing.

Your checklist for this pull request

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

Closing issues

Fixes #1845.