Re: [PATCH 1/7] Freezer: Read PF_BORROWED_MM in a nonracy way

[Oleg Nesterov <oleg@tv-sign.ru>]

On 05/11, Linus Torvalds wrote:
> 
> On Sat, 12 May 2007, Oleg Nesterov wrote:
> > 
> > things change, ->mm is not stable if the kernel thread does use_mm/unuse_mm.
> 
> ->mm is not stable *regardless*!
> 
> Trivial examples:
>  - kernel thread does execve()
>  - user thread does exit().

Yes sure. Quoting myself,
	>
	>  true->false means daemonize() or do_exit(), seems harmless.
	>
	>  false->true means exec from kernel space. That is why FREEZER_KERNEL_THREADS
	>  in fact means all tasks, not only kernel threads.
	>

> The use "use_mm()" and "unuse_mm()" things are total red herrings.
> 
> If the freezer depends on the difference between user and kernel threads, 
> then THAT PATCH IS BUGGY. It's that simple.

This is another story, I can't comment because I am not educated enough.

However, in my opininon THAT PATCH has nothing to do with this problem.
It just improves the code that we already have.

> > However, the return value == 0 does not change in that particular case,
> > exactly because is_user_space() takes task_lock().
> 
> As does exit_mm() etc.

Note the "in that particular case".

> See? The locking was pointless. Exactly because you release the lock 
> before the user can actually do anything about the return value!

Yes. See the "Quoting myself" above.

> Anyway, I think the whole freezer thing is broken. There's no reason to 
> freeze kernel threads. 

It is not perfect. Rafael tries to improve it.

Do we need freezer? Should we freeze kernel threads? I can't judge. I tried
to read a long thread about suspend, and failed to understand it.

I personally think we can simplify things if CPU-hotplug use freezer, at least.

Oleg.