Re: Commit abb0cd93494 breaks -singlestep -d in_asm,cpu with qemu-system-sh4

[BALATON Zoltan]

On Mon, 25 Oct 2021, Philippe Mathieu-Daudé wrote:
> On 10/25/21 23:16, BALATON Zoltan wrote:
> > Hello,
> >
> > Commit abb0cd93494 (accel/tcg: Split out log_cpu_exec) seems to have
> > broken -singlestep -d in_asm,cpu output with qemu-system-sh4 after a
> > delay slot.
> [...]
> > However I still don't understand how the delayed branch ends up at
> > 0x8c800964 instead of 0x8c801528 above. Is this ouput not showing some
> > already translated TBs even with -singlestep -d in_asm,cpu,nochain and
> > that's why I see those cpu dumps without instructions? What's the
> > correct way to get a trace of all executed instructions?
>
> IIUC this commit you now need to use both cpu,exec to get the output?

Nope, I get the same even adding exec. I think it now also prints 
registers for already translated instructions where in_asm produces no 
ouput whereas before it was only dumping state when in_asm also had ouput. 
So now we get a lot more cpu state but we don't really know what they are 
for. Not sure what's the use of that, the previous output looked more 
useful.

Regards,
BALATON Zoltan