Kirill Radkin
Kirill Radkin
Note: I faced with question when I tried to investigate an OpenOCD issue: wp/bp are not cleared after reset command. I tried to investigate is this behaviour correct or not....
@timsifive, can you confirm that my reasoning is correct?
> I haven't yet tried to reproduce it locally. What values of dmstatus are you getting, please? Are they unexpected? I get values that differs only in `all/any resumeack` bits,...
[1.log](https://github.com/riscv/riscv-openocd/files/12523178/1.log) [2.log](https://github.com/riscv/riscv-openocd/files/12523180/2.log) After second `examine` targets (1-3) in `halted` state (not `unknown`, maybe somebody fix it in new versions), but before first connection all of them were in `running` state.
P.S. `unknown` hart state was fixed here 198edca6d0516c979c787b30a99cc93d2342221a
> The difference seems to be simply that in the first log the harts are running while in the second log they are halted. Yes, but in both runs we...
> One solution might be to clear all halt groups before examining the first hart. Halt groups are used to implement OpenOCD smp target, so I think we should clear...