On 8 June 2015 at 20:35, Liviu Ionescu <i...@livius.net> wrote: > >> On 08 Jun 2015, at 22:27, Peter Maydell <peter.mayd...@linaro.org> wrote: >> >> That would also fix this problem, yes. It would still leave one >> use case wrong: >> * start QEMU >> * [cpu reset happens here; we load sp/pc] >> * in the debugger load an image (with a vector table) by writing it to RAM >> * let CPU run > > I already fixed this for some time, after completing the load the > debugger is configured to issue 'monitor system_reset', and this > fetches again msp & pc from the freshly loaded image:
Right, but only for your debugger. If we did pc/sp loading on exit from reset this would work for every debugger without it having to have extra QEMU specific magic. -- PMM