Re: [patch 20/24] perfmon: system calls interface

[Oleg Nesterov <oleg@redhat.com>]

On 11/26, Ingo Molnar wrote:
>
> * eranian@googlemail.com <eranian@googlemail.com> wrote:
>
> > +static int pfm_task_incompatible(struct pfm_context *ctx,
> > +				 struct task_struct *task)
> > +{
> > +	/*
> > +	 * cannot attach to a kernel thread
> > +	 */
> > +	if (!task->mm) {
> > +		PFM_DBG("cannot attach to kernel thread [%d]", task->pid);
> > +		return -EPERM;
> > +	}
> > +
> > +	/*
> > +	 * cannot attach to a zombie task
> > +	 */
> > +	if (task->exit_state == EXIT_ZOMBIE || task->exit_state == EXIT_DEAD) {
> > +		PFM_DBG("cannot attach to zombie/dead task [%d]", task->pid);
> > +		return -EBUSY;
> > +	}
> > +	return 0;
> > +}

I agree with Ingo these checks are pointless. Without the locks
the ->mm or ->exit_state can be changed right after the check.

And, as Ingo pointed out, you don't need this function at all,
if ptrace_check_attach() succeeds the task must have ->mm and
its ->exit_state == 0.

But, please note that the task can be SIGKILL'ed right after
ptrace_check_attach(), it can drop ->mm, it can be released.
(i don't understand the patch, perhaps this doesn't matter for
 you, just in case).

Oleg.