Re: can device drivers return non-ram via vm_ops->nopage?

[Russell King <rmk+lkml@arm.linux.org.uk>]

On Sat, Mar 20, 2004 at 12:50:53PM -0800, William Lee Irwin III wrote:
> On Sat, Mar 20, 2004 at 12:13:45PM -0800, Andrew Morton wrote:
> > I agree that ->nopage implementations should not be doing what that driver
> > is doing.  ->nopage is defined to return a page*: it's crazy to be
> > returning someting from there which isn't covered by mem_map[].
> > I just don't think it's important enough to be able to cope with
> > non-mem_map[] "memory" in do_no_page(), so I agree that requiring ->mmap()
> > to synchronously instantiate the pte's and retaining the debug check in
> > do_no_page() is a good idea.
> 
> There are other reasons for doing it, e.g. unusual TLB attributes
> and/or unusual pagetable structures backing the virtual region. I don't
> see anyone standing up and screaming for more functionality than cache
> coherency and/or disablement now, so as far as I'm concerned,
> remap_area_pages() (or rmk's stuff) kills the issue.

I'm no longer planning on this.  In fact, I see a future where I tell
people who want to use sound on ARM to go screw themselves because
there doesn't seem to be an acceptable solution to this problem.

Of course, this will lead to dirty hacks by many people who *REQUIRE*
sound to work, but I guess we just don't care about that.

(Yes, I'm pissed off over this issue.)

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 PCMCIA      - http://pcmcia.arm.linux.org.uk/
                 2.6 Serial core