Re: 2.5.69-mm4

[David Mosberger]

>>>>> On Tue, 13 May 2003 17:15:36 -0700, William Lee Irwin III <[e-mail]> said:

  William> On Mon, May 12, 2003 at 10:55:04PM -0700, Andrew Morton
  William> wrote:
  >> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.69/2.5.69-mm4/
  >> Lots of small things.  thread-info-in-task_struct.patch allow
  >> thread_info to be allocated as part of task_struct

  William> AIUI the task_cache is meant to prevent certain task_t
  William> (dear gawd I can't stand those _struct suffixes)
  William> refcounting pathologies because the task_t has its final
  William> put done by the task itself or something on that order, so
  William> it may be better for ia64 to adapt the task_cache to their
  William> purposes instead of wiping it entirely. Also, making the
  William> task_cache treatment uniform apart from its declaration
  William> would allow the #ifdef to be shoved in a header.

  William> Alternatively, one could alter the timing of the final put
  William> on a task_t so as to handle it similarly to the final
  William> mmput() (though here, too it might be more sightly to
  William> #ifdef the necessary bits in headers).

  William> I think there are already outstanding task_t refcounting
  William> bugs, so I'm not entirely sure where we stand wrt. changing
  William> final put mechanics.

All I really care about is that (a) task_struct, thread_info, and
kernel stack remain contiugous and within a single 64MB page and (b)
that it be fast (as usual ;-).  Other than that, it doesn't really
matter where the memory comes from.

	--david