Messages in this thread | | | Subject | Re: [RFC GIT PULL] nohz: Basic cputime accounting for adaptive tickless | From | Peter Zijlstra <> | Date | Thu, 14 Jun 2012 11:10:15 +0200 |
| |
On Thu, 2012-06-14 at 11:07 +0200, Peter Zijlstra wrote: > On Wed, 2012-06-13 at 18:19 +0200, Frederic Weisbecker wrote: > > Ingo, Thomas, > > > > This starts the basic code that allow accounting of cputime spent tickless > > outside idle, which is a first step to prepare for the adaptive nohz > > infrastructure. > > > > I hope we can set a tree in -tip for that. If you're fine with it > > this is pullable from: > > There's a number of architecture that already does fine grained > user/kernel time accounting on syscall boundaries etc.. s390, powerpc > and ia64. > > You're now adding a 3rd way of accounting user/kernel time.. I'm not > much looking fwd to that..
Note there's also the whole IRQ_TIME_ACCOUNTING muck.. all in all its getting quite ridiculous.
| |