Re: [PATCH 6/9] signal: Always call do_notify_parent_cldstop with siglock held

[Oleg Nesterov]

On 04/28, Peter Zijlstra wrote:
>
> I've not had time to fully appreciate the nested locking here, but if it
> is possible to rework things to always take both locks at the same time,
> then it would be possible to impose an arbitrary lock order on things
> and break the cycle that way.

This is clear, but this is not that simple.

For example (with this series at least), ptrace_stop() already holds
current->sighand->siglock which (in particular) we need to protect
current->parent, but then we need current->parent->sighand->siglock
in do_notify_parent_cldstop().

Oleg.