Re: [PATCH 01/04] Adding cipher mode context information to crypto_tfm

[Matt Mackall <mpm@selenic.com>]

On Thu, Feb 10, 2005 at 12:17:24PM +0100, Fruhwirth Clemens wrote:
> On Thu, 2005-02-10 at 02:33 -0800, Andrew Morton wrote:
> > Fruhwirth Clemens <clemens@endorphin.org> wrote:
> > >
> > > On Wed, 2005-02-09 at 17:19 -0800, Andrew Morton wrote:
> > > > Fruhwirth Clemens <clemens@endorphin.org> wrote:
> > > > Adding a few more fixmap slots wouldn't hurt anyone.  But if you want an
> > > > arbitrarily large number of them then no, we cannot do that.
> > > 
> > > What magnitude is "few more"? 2, 10, 100?
> > 
> > Not 100.  10 would seem excessive.
> 
> Out of curiosity: Where does this limitation even come from? What
> prevents kmap_atomic from adding slots dynamically?

There's a single page of PTEs for mapping high memory and the atomic
slots are a small subset of that. They're fixed in number for
complexity reasons - we don't want to have an allocator here:

/*
 * kmap_atomic/kunmap_atomic is significantly faster than kmap/kunmap because
 * no global lock is needed and because the kmap code must perform a global TLB
 * invalidation when the kmap pool wraps.
 *

-- 
Mathematics is the supreme nostalgia of our time.