Re: [BUG] loop.c oopses

[Andrew Morton <akpm@zip.com.au>]

William Lee Irwin III wrote:
> 
> On Tue, Jul 16, 2002 at 10:42:52PM +0300, Jari Ruusu wrote:
> > The patch below fixes that remap issue, plus uncounted number of other loop
> > issues. For example, device backed loops use pre-allocated pages for zero VM
> > pressure.
> 
> I'd like to understand this (and possibly even use it) as I tend to use
> the loopback block driver often. Any chance you could break this up into
> a blow-by-blow series of bugfixes? As it is here, it's a bit much for me
> to digest as a newbie to bio.
> 
> My needs for explanation are perhaps greater than others on the cc: list,
> so it's really optional, but I'd be much obliged if you could do so.
> 

Seconded, please.  It's obviously an important patch but it's
rather large and opaque.  A description of what it is doing and
in particularly *why* it is doing it would really help.

A nice way of providing such a description is inside

/*
 * these thingies
 */

so the information is not lost.

I particular:

- Does the file_operations-based file-backed IO work with all
  crypto setups?

- What are those preallocated pages doing?  Are they really needed
  with the 2.5 VM?  What problem are they solving?

Thanks.

(BTW: check out the ARRAY_SIZE macro ;))

-