Re: [PATCH 7/9] ptrace: Simplify the wait_task_inactive call in ptrace_check_attach

[Eric W. Biederman]

"Eric W. Biederman" <ebiederm@xxxxxxxxxxxx> writes:

> Asking wait_task_inactive to verify that tsk->__state == __TASK_TRACED
> was needed to detect the when ptrace_stop would decide not to stop
> after calling "set_special_state(TASK_TRACED)".  With the recent
> cleanups ptrace_stop will always stop after calling set_special_state.
>
> Take advatnage of this by no longer asking wait_task_inactive to
> verify the state.  If a bug is hit and wait_task_inactive does not
> succeed warn and return -ESRCH.

As Oleg noticed upthread there are more reasons than simply
!current->ptrace for wait_task_inactive to fail.  In particular a fatal
signal can be received any time before JOBCTL_DELAY_SIGKILL.

So this change is not safe.  I will respin this one.

Eric


> Signed-off-by: "Eric W. Biederman" <ebiederm@xxxxxxxxxxxx>
> ---
>  kernel/ptrace.c | 14 +++-----------
>  1 file changed, 3 insertions(+), 11 deletions(-)
>
> diff --git a/kernel/ptrace.c b/kernel/ptrace.c
> index 16d1a84a2cae..0634da7ac685 100644
> --- a/kernel/ptrace.c
> +++ b/kernel/ptrace.c
> @@ -265,17 +265,9 @@ static int ptrace_check_attach(struct task_struct *child, bool ignore_state)
>  	}
>  	read_unlock(&tasklist_lock);
>  
> -	if (!ret && !ignore_state) {
> -		if (!wait_task_inactive(child, __TASK_TRACED)) {
> -			/*
> -			 * This can only happen if may_ptrace_stop() fails and
> -			 * ptrace_stop() changes ->state back to TASK_RUNNING,
> -			 * so we should not worry about leaking __TASK_TRACED.
> -			 */
> -			WARN_ON(READ_ONCE(child->__state) == __TASK_TRACED);
> -			ret = -ESRCH;
> -		}
> -	}
> +	if (!ret && !ignore_state &&
> +	    WARN_ON_ONCE(!wait_task_inactive(child, 0)))
> +		ret = -ESRCH;
>  
>  	return ret;
>  }

Eric