Hacker News new | past | comments | ask | show | jobs | submit login

Makes perfect sense. I would assume you also have to re-implement the routines writing the registers/memory to the non-volatile device, as you can't rely on peripheral registers being consistent. Ideally the host should have independent access to the said volatile memory, but that's getting close to implementing a debugger on host which uses JTAG/SWD to inspect the state of MCU after a crash.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact