On 25 July 2014 07:58, Richard Henderson <r...@twiddle.net> wrote: > On 07/24/2014 06:37 PM, Jincheng Miao wrote: >> '-singlestep' option will make TB contains only one instruction, >> so that the qemu_log could output trace log when CPU_LOG_EXEC sets, >> and it could help developers to debug control flow. >> >> But currently, in cpu_exec(), it doesn't check singlestep when >> tb_add_jump(), so the TB linked is executed siliently. >> Therefore, this patch adds singlestep check before tb_add_jump(). >> >> Signed-off-by: Jincheng Miao <jm...@redhat.com> > > Reasonable. I've been thinking that we simply shoudn't emit goto_tb under > single-step. That does require fixes to all but 2 or 3 of the backends > though, > and this patch attacks the problem all in one place.
Huh? We already don't emit goto_tb if single-stepping, surely? (Well, I guess some of the backends might well be broken, but in that case they probably don't get the other bits of singlestep support right either...) -- PMM
