On 9 October 2015 at 16:55, Sergey Fedorov <serge.f...@gmail.com> wrote: > Thank you for the explanation, Peter. I see, if we do insn translation > then we take the page fault instead of the CPU breakpoint. As of user > mode, can we actually set any CPU breakpoint? If not, as I guess, then > (b) is not applicable to the possible solution that you have described.
You're right, we can only have GDB breakpoints in user-mode. -- PMM