On 07/28, Roland McGrath wrote: > > > Questions: should I send the change in signal.c to Andrew right now? > > It can be applied separetely, it doesn't break utrace_report_jctl(). > > Yes, send it now.
OK, will do > It might be best just to send a replacement > for signals-tracehook_notify_jctl-change.patch that rolls this in. Oh. I am looking at signals-tracehook_notify_jctl-change.patch. It is not trivial but doesn't really change do_signal_stop() (except it folds finish_stop() into). Roland, I'd prefer to send this change separately. Otherwise it will be really hard to review the unified patch. > I wouldn't mention particular utrace code in its log message. Yes, yes, sure. > Just something like the old log paragraph: > > Tracing needs the siglock here, held from the time TASK_STOPPED was set, > to avoid potential SIGCONT races if it wants to allow any blocking in its > tracing hooks. > > adjusted appropriately, perhaps: > > Tracing wants to be able to drop and reacquire the siglock here, > before TASK_STOPPED is set, to avoid potential SIGCONT races if it > wants to allow any blocking in its tracing hooks. OK, thanks. Oleg.