Messages in this thread | | | Date | Thu, 11 Apr 2013 17:11:13 +0200 | From | Frederic Weisbecker <> | Subject | Re: [RFC GIT PULL] nohz: Full dynticks rq clock handling |
| |
On Wed, Apr 10, 2013 at 12:26:43PM +0200, Peter Zijlstra wrote: > > Another thing I just noticed; our update_curr() assumes its called at > least once every 4 odd seconds (2^32 ns like). > > update_curr() is 'required' for things like task runtime stats and > everything that hangs off of that like cputimers etc. > > Now I'm not entirely sure how the nr_running==1 nohz case deals with > this, but I can imagine it might take a while. In this case we might > need to 'fix' update_curr() to not asume the time delta fits in 32 > bits. >
Good catch, this is not handled yet.
| |