quoted

quoted

quoted

quoted

On Wed, 29 Jan 2003 09:59:49 +0000, Russell King [off-list ref] said:

  Russell> On Tue, Jan 28, 2003 at 10:07:29PM -0800, Andrew Morton
  Russell> wrote:

  >> The frlock code is showing nice speedups, but I think the main
  >> reason we want this is to fix the problem wherein an application
  >> spinning on gettimeofday() can make time stop.

  Russell> I'm slightly concerned about this.  With this patch, we
  Russell> generally seem to do:

  Russell> [snip...]

  Russell> The same is true for other architectures; their
  Russell> gettimeoffset implementations need to be audited by the
  Russell> architecture maintainers to ensure that they are safe to
  Russell> run with (local) interrupts enabled.

Should be fine as far as ia64 is concerned, since gettimeoffset()
currently simply reads the cycle-counter (and I think even HPET-based
interpolation would be lock-free).

	--david
--
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/