Re: [PATCH] Migration of standard timers

[Dimitri Sivanich <sivanich@sgi.com>]

On Fri, Sep 15, 2006 at 06:39:19PM +0200, Thomas Gleixner wrote:
> Are you trying to work around the latencies caused by long running timer
> callbacks ? I'm not convinced that this is not curing the symptoms
> instead of the root cause.

Yes, both latency from long running timer callbacks as well as
potential latency from a temporal grouping of timer callbacks
(those occuring on the same tick).

While I agree that root causes of the former should be addressed,
more latencies of this type can always easily creep in.  Timer
migration works as a long term preventative aid, not just a fix
for the problem of the moment.  And adding this needn't restrict
anyone from looking at the aforementioned root causes.