Re: Strange interrupt behaviour
Perry Harrington (pedward@sun4.apsoft.com)
Thu, 16 Jul 1998 14:23:50 -0700 (PDT)
>
> On Thu, 16 Jul 1998, Rogier Wolff wrote:
> >
> > The "right" way to handle this, is to make kmalloc/gfp/whatever
> > implement this, and allow just a few places to set the flag that
> > enables getting stuff from the "last resort" pool.
>
> The problem with that is that we really want to have multiple
> "last-resort" pools.
>
> For example, NFS would probably also be happier with a last-resort pool
> for kmallocs (of both 8kB and 16kB allocations - we don' tneed a very deep
> pool, just 2-4 allocations would probably be fine to handle the bad
> cases), but we do _not_ want to allow NFS starvation to starve fork()'s or
> vice versa.
>
> I certainly agree with making this a more generic thing, but it mustn't be
> just a flag that everybody then fights over.
Why don't you simply allocate a spare slab when the NFS code is initialized, then
snarf buffers from that when needed, this removes kmalloc hacks. You can apply
this methodology to just about any kernel buffer pool.
>
> Kubys
>
--Perry
--
Perry Harrington Linux rules all OSes. APSoft ()
email: perry@apsoft.com Think Blue. /\
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html