Re: scalable kmap (was Re: vm lock contention reduction)

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

"Martin J. Bligh" wrote:
> 
> ...
> > I was discussing this with sct a few days back.  iiuc, the proposal
> > was to create a small per-cpu pool (say, 4-8 pages) which is a
> > "front-end" to regular old kmap().
> >
> > Any time you have one of these pages in use, the process gets
> > pinned onto the current CPU.
> 
> Ewww! That's gross ;-)

Hey.  So is highmem ;)

But sys_sched_affinity() allows you to change the affinity of
other tasks (it takes a pid).  So that's torn that idea.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/