Re: Access to non-RAM pages

[Benjamin Herrenschmidt]

On Sun, 2018-09-02 at 18:38 -0700, Linus Torvalds wrote:
> On Sun, Sep 2, 2018 at 6:32 PM Benjamin Herrenschmidt
> <benh@xxxxxxxxxxxxxxxxxxx> wrote:
> > 
> > Also, if we cross page boundaries with those guys then we have a bigger
> > problem no ? we could fall off a vmalloc page into the nether or into
> > an ioremap mapping no ?
> 
> It's not used for vmalloc stuff. It's just regular kmalloc().
> 
> So it can cross pages, and it can fall off the end of memory, but it
> can't do random stuff.

Ok, it might be worth adding a DEBUG_VM based (or similar) warning in
case somebody ever thinks of passing a vmalloc pointer to it...

As for falling out of the end of memory, yes it could be a real problem
though I don't see why IO is any different than just hitting a non-
mapped area in that regard. So we should probably keep an unused
(readonly if possible) zero page at the end.
 
Cheers,
Ben.