s390: stepping && PT_PTRACED (Was: utrace failed to compile for s390x)

Oleg Nesterov oleg at redhat.com
Wed Dec 9 19:54:28 UTC 2009


(add cc's)

On 12/03, Roland McGrath wrote:
>
> > Not sure what should we do right now, s390 needs more attention.
> > Also, it uses PT_PTRACED bit, I am afraid this is not what we wan
> > with utrace.
>
> This looks simple.  Its one use can just be
> tracehook_consider_fatal_signal(SIGTRAP) instead.  I expect you can just
> send that change to the s390 maintainers and they will take it.

Done.

But... I tried to understand these check and failed. Why do we need them?
They look unneeded to me, but of course I know nothing about s390.

For example, do_single_step(). If notify_die(DIE_SSTEP) does not return
NOTIFY_STOP, it is not kprobe. This means the caller is sysc_singlestep:
in entry.S, who else except ptrace could set TIF_SINGLESTEP ?

Oleg.




More information about the utrace-devel mailing list