[PATCH 0/1] poweroff: change orderly_poweroff() to use schedule_work()

[Oleg Nesterov <oleg@redhat.com>]

On 03/12, Oleg Nesterov wrote:
>
> On 03/12, Linus Torvalds wrote:
> >
> > So yeah, I guess
> > everything could just go into a workqueue.
>
> OK, I'll try to make the patch tomorrow. Should be trivial but it is
> not clear how we should pass "bool force" without allocating the
> work_struct which would be nice to avoid.

Yes, it would be nice to keep it simple and use a single work/arg.

Could you review? The change is trivial but

	- orderly_poweroff() always return 0.

	- the patch assumes that orderly_poweroff(false) after
	  orderly_poweroff(true) acts as "force = true". Only xen
	  uses "false", I hope this is fine.

	  In fact I think we can change poweroff_force argument
	  unconditionally, this "if (force)" check is mostly
	  documentation.

	  But we can add the locking or even allocate work_struct
	  every time if this is wrong (or just looks wrong).

	- The patch assumes that orderly_poweroff() doesn't need
	  the keventd_up() check, I hope this is correct...


Lucas, Andrew, sorry. If this patch will be applied, then

	kernel-sysc-use-the-simpler-call_usermodehelper.patch

should be dropped. Or I can redo this fix on top of -mm cleanup.

> And. It seems there is another problem. argv_split(poweroff_cmd) can
> obviously race with proc_dostring() ?

I'll send another patch...

Oleg.