----- Original Message ----- > > > ----- Original Message ----- > > Hello Dave, > > > > Please refer to the attachment. I think the string of state is not > > correct. > > > > -- > > -- > > Regards > > Qiao Nuohan > > > > > > --- a/task.c > > +++ b/task.c > > @@ -4633,7 +4633,7 @@ task_state_string(ulong task, char *buf, > > int verbose) > > > > if (valid_task_state(_TRACING_STOPPED_) && > > (state & _TRACING_STOPPED_)) { > > - sprintf(buf, "ST"); > > + sprintf(buf, "TS"); > > valid++; > > set++; > > } > ~ > > Actually it is shown as "ST" on purpose, because it goes back to > older kernels where there was no separate "traced" state, and both > "stopped" and "traced" states were essentially the same. The flag > bits keep changing over time, but the original list of states shown > by "ps" has always been: > > crash> help ps > ... > 5. the task state (RU, IN, UN, ZO, ST, DE, SW). > ... > > However I agree with you that for clarity's sake the traced state should > be differentiated. But I think it would be better to change it to "TR" > instead of your "TS" suggestion -- note that the verbose translation > used by the "set" command shows "TASK_TRACED": > > if (state & _TRACING_STOPPED_) > sprintf(&buf[strlen(buf)], "%sTASK_TRACED", > count++ ? "|" : ""); > > I'll update the "ps" help page with the new "TR" identifier. > > Thanks, > Dave And related to the above, I'll also update the "foreach" command's state qualifier to accept a "TR" option. As it is now, there's no way to filter out traced tasks. Dave -- Crash-utility mailing list Crash-utility@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/crash-utility