lkml.org 
[lkml]   [2000]   [Sep]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: the new VMt
On Mon, 25 Sep 2000, Linus Torvalds wrote:
> On Mon, 25 Sep 2000, Andrea Arcangeli wrote:
> >
> > But I'd much prefer to pass not only the classzone from allocator
> > to memory balancing, but _also_ the order of the allocation,
> > and then shrink_mmap will know it doesn't worth to free anything
> > that isn't contigous on the order of the allocation that we need.
>
> I suspect that the proper way to do this is to just make another gfp_flag,
> which is basically another hint to the mm layer that we're doing a multi-
> page allocation and that the MM layer should not try forever to handle it.
>
> In fact, that's independent of whether it is a multi-page
> allocation or not. It might be something like __GFP_SOFT - you
> could use it with single pages too.
>
> Thinking about it, we do have it already. It's called
> !__GFP_HIGH, and it used by all the GFP_USER allocations.

Hmm, I think these two are orthagonal.

__GFP_HIGH means that we are allowed to eat deeper into
the free list (maybe needed to avoid a deadlock freeing
pages)

__GFP_SOFT would mean "don't bother waiting for free pages",
which is something very different...

(I wouldn't want a user process to get killed simply because
kswapd is waiting for IO to finish on a swapout, in that case
we really do want to sleep for a while)

regards,

Rik
--
"What you're running that piece of shit Gnome?!?!"
-- Miguel de Icaza, UKUUG 2000

http://www.conectiva.com/ http://www.surriel.com/

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 12:38    [W:0.221 / U:0.196 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site