[PATCHv12 2/4] zbud: add to mm/

[Andrew Morton]

On Wed, 29 May 2013 14:09:02 -0700 (PDT) Dan Magenheimer <dan.magenheimer at oracle.com> wrote:

> > memory_failure() is merely an example of a general problem: code which
> > reads from the memmap[] array and expects its elements to be of type
> > `struct page'.  Other examples might be memory hotplugging, memory leak
> > checkers etc.  I have vague memories of out-of-tree patches
> > (bigphysarea?) doing this as well.
> > 
> > It's a general problem to which we need a general solution.
> 
> <Obi-tmem Kenobe slowly materializes... "use the force, Luke!">
> 
> One could reasonably argue that any code that makes incorrect
> assumptions about the contents of a struct page structure is buggy
> and should be fixed.

Well it has type "struct page" and all code has a right to expect the
contents to match that type.

>  Isn't the "general solution" already described
> in the following comment, excerpted from include/linux/mm.h, which
> implies that "scribbling on existing pageframes" [carefully], is fine?
> (And, if not, shouldn't that comment be fixed, or am I misreading
> it?)
> 
> <start excerpt>
>  * For the non-reserved pages, page_count(page) denotes a reference count.
>  *   page_count() == 0 means the page is free. page->lru is then used for
>  *   freelist management in the buddy allocator.
>  *   page_count() > 0  means the page has been allocated.

Well kinda maybe.  How all the random memmap-peekers handle this I do
not know.  Setting PageReserved is a big hammer which should keep other
little paws out of there, although I guess it's abusive of whatever
PageReserved is supposed to mean.

It's what we used to call a variant record.  The tag is page.flags and
the protocol is, umm,

PageReserved: doesn't refer to a page at all - don't touch
PageSlab: belongs to slab or slub
!PageSlab: regular kernel/user/pagecache page

Are there any more?

So what to do here?  How about

- Position the zbud fields within struct page via the preferred
  means: editing its definition.

- Decide upon and document the means by which the zbud variant is tagged

- Demonstrate how this is safe against existing memmap-peekers

- Do all this without consuming another page flag :)