Re: Nick's core remove PageReserved broke vmware...

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Thu, 2005-11-03 at 15:55 +0200, Gleb Natapov wrote:
> On Thu, Nov 03, 2005 at 01:32:15PM +0000, Hugh Dickins wrote:
> > On Thu, 3 Nov 2005, Gleb Natapov wrote:
> > > On Thu, Nov 03, 2005 at 09:22:10AM +1100, Benjamin Herrenschmidt wrote:
> > > > Also, what do you suggest as a good threshold to use on the max amount
> > > > of memory I can let the X server "pin" that way ? I was thinking it as
> > > > equivalent to mlock, thus I could maybe hijack mm->locked_vm & use
> > > > RLIMIT_MEMLOCK or is that too gross ?
> > > > 
> > > This is what infiniband does, so it should be good for you too.
> > 
> > Yes, I noticed that code a couple of days ago, I think you're setting a
> > good example there, for long-term or large-area uses of get_user_pages():
> > Ben, take a look at drivers/infiniband/core/uverbs_mem.c

Ok, thanks guys, will have a look.

Ben.