Re: [patch][rfc] mm: have expand_stack honour VM_LOCKED

[KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>]

> > Hi Nick,
> > 
> > > Is this valid?
> > > 
> > > 
> > > It appears that direct callers of expand_stack may not properly lock the newly
> > > expanded stack if they don't call make_pages_present (page fault handlers do
> > > this).
> > 
> > When happend this issue?
> > 
> > I think...
> > 
> > case 1. explit mlock to stack 
> > 
> >    1. mlock to stack
> >         -> make_pages_present is called via mlock(2).
> >    2. stack increased
> >         -> no page fault happened.
> > 
> > case 2. swapout and mlock stack
> > 
> >    1. stack swap out
> >    2. mlock to stack
> >         -> the page doesn't swap in at the time.
> >    3. page fault in the stack
> >         -> the page swap in
> >            (no need make_present_page())
> > 
> > 
> > So, it seems this patch isn't necessary.
> 
> What if you you page fault the stack further than a single page down?
> 

I see. thanks.

But unfortunately, this patch conflicted against unevictable patch series.
I'll make for -mm version patch few days after if you don't like do that.



--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>