Re: fork: out of memory

Jan Echternach (jec@DInet.de)
Sat, 29 Nov 1997 11:05:39 +0100

On 28 November 1997, Zlatko Calusic <Zlatko.Calusic@CARNet.hr> wrote:
> Jan Echternach <jec@DInet.de> writes:
> > Compiling a driver statically into the kernel is not always an option.
> > Maybe you're short on memory, or you want to fiddle with some compile-
> > time options or debug the driver without rebooting once in a minute.
> >
>
> So, you're short of 20,30 KB for driver, and thus decide to allocate
> (mostly) unused 128 or even 256 KB (to be sure DMA requests always
> succeed), so that you can load your driver?

Agreed. The better solution is to keep not just a certain number of
free pages around, but to additionally require that these pages are
DMA-able and unfragmented. But wait, this is wrong. You don't need to
have these pages ready for you in advance (unless you're allocating
a 64 KB DMA puffer in an interrupt handler). It should be sufficient
to free up the required memory (DMA-able or not, depending on what
was requested) in __get_free_pages. This can be done by either swapping
out some pages or by moving pages to a different physical location
(i.e. the equivalent of a swap out + immediate swap in). The latter
is even allowed inside an interrupt handler (GFP_ATOMIC).

There should be no actual performance penalties. When the current
__get_free_pages would fail the improved version probably doesn't fail
but is only a bit slower than a normal __get_free_pages call. There won't
be any user-visible changes except this.

What do the kernel gurus think about this proposal? Have I overlooked
something?

-- 
Jan Echternach


   Delta Internet              http://www.DInet.de/
                               Tel. +49 2932 91 6 161
  Zeit UmZuDenken!             Fax. +49 2932 91 6 230