Re: pmap_remap_pages and protection (save/restore, part 1)

[Jed Davis <jdev%panix.com@localhost>]

YAMAMOTO Takashi <yamt%mwd.biglobe.ne.jp@localhost> writes:
>
> because our current implementation of the ioctl is a hack.

The ioctl's interface is going to be annoying no matter what the
implementation, I feel.  libxc does have the mmap/ioctl nicely wrapped
up in utility functions, which could potentially be changed if necessary.

> it isn't straightforward to handle faults on the mapping correctly
> within our vm structure because the api has very linux-vm centric design.

_The UVM Virtual Memory System_ (the overview paper, that is) mentions
that a pager implementation can do its own page allocation, for
e.g. mapping a ROM into a process's address space.  I also notice
cdevsw's d_mmap, but that's expected to return a "physical" address.

-- 
(let ((C call-with-current-continuation)) (apply (lambda (x y) (x y)) (map
((lambda (r) ((C C) (lambda (s) (r (lambda l (apply (s s) l))))))  (lambda
(f) (lambda (l) (if (null? l) C (lambda (k) (display (car l)) ((f (cdr l))
(C k)))))))    '((#\J #\d #\D #\v #\s) (#\e #\space #\a #\i #\newline)))))