Re: [PATCH] fix granularity of task_u/stime(), v2

[Peter Zijlstra <peterz@infradead.org>]

On Tue, 2009-11-17 at 14:08 +0100, Stanislaw Gruszka wrote:
> On Mon, Nov 16, 2009 at 12:32:43PM -0700, Spencer Candland wrote:
> > > seems you have more test cases for utime decreasing issues,
> > > could you send links to me ? Somehow I could not find them 
> > > by my own. Particularly test case used in development this commit
> > > is interested:
> > > 
> > >   commit 49048622eae698e5c4ae61f7e71200f265ccc529
> > >   Author: Balbir Singh <balbir@linux.vnet.ibm.com>
> > >   Date:   Fri Sep 5 18:12:23 2008 +0200
> > >     sched: fix process time monotonicity
> > 
> > I had originally noticed that in a production web server, so my test
> > case was designed to mirror what I was seeing there, which was just
> > running apache with worker mpm, and running a simple apache bench while
> > watching the utime/stime of the apache children.  Unfortunately that
> > method was not terribly reliable at reproducing the issue, which is why
> > I felt it necessary to try to come up with a better test case this time
> > around.
> 
> No wonder I could not find anything on google and in mailing list
> archives :)
> 
> Seems issue reported then was exactly the same as reported now by
> you. Looks like commit 49048622eae698e5c4ae61f7e71200f265ccc529 just
> make probability of bug smaller and you did not note it until now.  
> 
> Could you please test this patch, if it solve all utime decrease
> problems for you:
> 
> http://patchwork.kernel.org/patch/59795/
> 
> If you confirm it work, I think we should apply it. Otherwise
> we need to go to propagate task_{u,s}time everywhere, which is not
> (my) preferred solution.

That patch will create another issue, it will allow a process to hide
from top by arranging to never run when the tick hits.