Messages in this thread | | | Date | Thu, 12 Mar 2009 10:05:18 +0530 | From | Balbir Singh <> | Subject | Re: [RFC PATCH] cpuacct: per-cgroup utime/stime statistics - v1 |
| |
* Bharata B Rao <bharata@linux.vnet.ibm.com> [2009-03-12 09:59:51]:
> On Wed, Mar 11, 2009 at 09:04:34PM +0530, Balbir Singh wrote: > > * KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> [2009-03-11 09:38:12]: > > > > > BTW, I'm not sure but don't we need special handling if > > > CONFIG_VIRT_CPU_ACCOUNTING=y ? > > > > Good point. Bharata, with CONFIG_VIRT_CPU_ACCOUNTING, utime and stime > > is accounted for within the architecture. > > True, but as I replied to Kamezawa in the other thread, these > architectures are still dependent on generic implementations of > account_{system,user}_time() which feed to per-process and system-wide > accounting. And this is where we have hooks for per-cgroup accounting.
Fair enough...
> So I don't see why we need to handle these archs specially. Do I miss > something ?
May be we should cc the patch to linux-arch as well to make sure we handled all architectures correctly.
-- Balbir
| |