[PATCH 09] introduce the empty struct ptrace_context

[Oleg Nesterov]

On 08/17, Roland McGrath wrote:
>
> Does this mean you intend to abandon the ptrace-revamp draft series that
> added ptrace_context et al cleanups in the pre-utrace code?

Well, yes. I don't think this series makes sense before utrace-ptrace.

Firstly, it adds task_struct->ptrace_ctx (and kzalloc to tracehooks +
kfree to free_task) which utrace-ptrace does not need. Giwen that we
are going to forget about the old ptrace code, this just means more
necessary changes.

And. At least ->parent and ->ptrace_entry need more attention with
utrace-ptrace, it is not easy to migrate them right now. This again
means this draft series adds more problems.

> Or are you
> culling bits from there to integrate differently now?

Yes, I hope. At least partly.

> > Also, ->exit_code and ->last_siginfo should live in engine->data too.
>
> Well, task_struct.exit_code has core non-ptrace uses, i.e. wait.  What you
> mean is stop overloading it for ptrace wait status and resumption signal,
> and use appropriate explicit bookkeeping for that ptrace_context.

Yes, exactly.

Oleg.