Re: The Slab Allocator

David S. Miller (davem@jenolan.rutgers.edu)
Thu, 7 Aug 1997 03:00:21 -0400


Date: Thu, 7 Aug 97 00:55:22 MDT
From: colin@nyx.net (Colin Plumb)

I'm a little unclear on David Miller's design constraint "2) No
cli()'s on any code path whatsoever". Doesn't __get_free_pages do
all kinds of locking (including possibly synchronous disk I/O!),
and doesn't *any* kernel memory allocator need to call that
sometimes? David, can you clarify?

Here what I meant was "no cli()'s in the local processor cache hit
case", nothing more.