From mboxrd@z Thu Jan 1 00:00:00 1970 From: dietmar.eggemann@arm.com (Dietmar Eggemann) Date: Mon, 14 Jul 2014 18:54:11 +0100 Subject: [PATCH v3 09/12] Revert "sched: Put rq's sched_avg under CONFIG_FAIR_GROUP_SCHED" In-Reply-To: References: <1404144343-18720-1-git-send-email-vincent.guittot@linaro.org> <1404144343-18720-10-git-send-email-vincent.guittot@linaro.org> <20140710131646.GB3935@laptop> <20140711151304.GD3935@laptop> Message-ID: <53C41943.2000601@arm.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org [...] >> In that same discussion ISTR a suggestion about adding avg_running time, >> as opposed to the current avg_runnable. The sum of avg_running should be >> much more accurate, and still react correctly to migrations. > > I haven't look in details but I agree that avg_running would be much > more accurate than avg_runnable and should probably fit the > requirement. Does it means that we could re-add the avg_running (or > something similar) that has disappeared during the review of load avg > tracking patchset ? Are you referring to '[RFC PATCH 14/14] sched: implement usage tracking' https://lkml.org/lkml/2012/2/1/769 from Paul Turner? __update_entity_runnable_avg() has an additional parameter 'running' so that it can be called for a) sched_entities in update_entity_load_avg(): __update_entity_runnable_avg(..., se->on_rq, cfs_rq->curr == se)) b) rq's in update_rq_runnable_avg(): __update_entity_runnable_avg(..., runnable, runnable); I can see how it gives us two different signals for a sched_entity but for a rq? Do I miss something here? -- Dietmar [...]