Re: [PATCH 2/2] ptrace/x86: dont delay perf_event_disable() till second pass in ptrace_write_dr7()

[Oleg Nesterov <oleg@redhat.com>]

On 04/17, Frederic Weisbecker wrote:
>
> 2013/4/16 Oleg Nesterov <oleg@redhat.com>:
> >
> > Well, I disagree.
> >
> > To clarify, I agree with WARN_ON_ONCE(), but afaics it has nothing to
> > do with "second_pass",
> >
> >> And these are indeed supposed
> >> to.
> >
> > Indeed, but this is because ptrace_modify_breakpoint() should not fail.
> >
> > So, what do you think if I change the main loop above
> >
> >                 rc = ptrace_modify_breakpoint(...)
> >         -       if (rc)
> >         +       if (WARN_ON_ONCE(rc))
> >                         break;
>
> It can fail in the first pass if dr7 is incorrect. For example passing
> a length
    ^^^^^^

Oops. Indeed, I forgot that ptrace_modify_breakpoint() has other arguments,
not only "disabled" ;)

Thanks for correcting me.

Oleg.