Messages in this thread | | | Date | Fri, 27 Oct 2006 11:49:40 -0400 | From | Shailabh Nagar <> | Subject | Re: [PATCH: 2.6.18.1] delayacct: cpu_count in taskstats updated correctly |
| |
David Rientjes wrote: > On Fri, 27 Oct 2006, Martin Tostrup Setek wrote: > >> from: Martin T. Setek <martitse@ifi.uio.no> >> >> cpu_count in struct taskstats should be the same as the corresponding (third) >> value found in /proc/<pid>/schedstat > > I disagree in favor of Documentation/accounting/taskstats-struct.txt. > cpu_count is the number of delay values recorded, so accumulating them is > appropriate. > > David
David's right. The delay accounting field cpu_count is measuring how many delay values are recorded in the field cpu_delay_total (so that one can divide one by the other to get an average if needed).
For the delays reported for a single task (ie. __delayacct_add_tsk called only once for a given task), the effect will be what Martin wants i.e. cpu_count will be the same as /proc/pid/schedstat's third field (sched_info->pcnt)
But when the delays are reported for a tgid, where *accumalation* of delays for all constituent pids is being done (by calling __delayacct_add_tsk repeatedly), what is desired is to accumalate both cpu_count and cpu_delay_total.
So the patch proposed by Martin is incorrect.
--Shailabh - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |