Re: [PATCH] flush_cpu_workqueue: don't flush an empty ->worklist

[Ingo Molnar <mingo@elte.hu>]

* Andrew Morton <akpm@osdl.org> wrote:

> > I would be happy to be corrected if the above impression of 
> > freeze_processes() is corrected ..
> 
> It could be that the freezer needs a bit of work for this application. 
> Obviously we're not interested in the handling of disk I/O, so we'd 
> really like to do a simple 
> try_to_freeze_tasks(FREEZER_USER_SPACE|FREEZER_KERNEL_THREADS), but 
> the code isn't set up to do that (it should be).  The other non-swsusp 
> callers probably want this change as well.  But that's all a minor 
> matter.

yes. The freezer does the fundamentally right thing: it stops all tasks 
in user-space or waits for them to return back to user-space to stop 
them there, or if it's a pure kernel-space task it waits until that 
kernel-space task voluntarily stop.

Once the system is in such a state, and all processing has been stopped, 
all of the kernel's data structures are completely 'unused', and we can:

- patch the kernel freely (kprobes)

- save+stop the kernel (sw-suspend) 

- remove a CPU (CPU hotplug and suspend)

- (We could also use this mechanism to live-migrate the kernel to 
   another system btw., possibly useful for containers)

- (We could also use this mechanism to create a live snapshot of a 
   running kernel, together with an LVM snapshot of filesystem state, 
   for possible restore point later on.)

It is a very powerful mechanism that has really nice properties - we 
should work on this one shared infrastructure instead of adding zillions 
of per-subsystem CPU hotplug locks.

	Ingo