On 02/04/2019 07:18 AM, Peter Zijlstra wrote: > On Mon, Feb 04, 2019 at 10:02:11AM +0100, Juri Lelli wrote: >> On 18/01/19 17:46, Juri Lelli wrote: >>> On 18/01/19 08:17, Tejun Heo wrote: >>>> On Thu, Jan 17, 2019 at 09:47:34AM +0100, Juri Lelli wrote: >>>>> Hi, >>>>> >>>>> v6 of a series of patches, originally authored by Mathieu, with the intent >>>>> of fixing a long standing issue of SCHED_DEADLINE bandwidth accounting. >>>>> As originally reported by Steve [1], when hotplug and/or (certain) >>>>> cpuset reconfiguration operations take place, DEADLINE bandwidth >>>>> accounting information is lost since root domains are destroyed and >>>>> recreated. >>>>> >>>>> Mathieu's approach is based on restoring bandwidth accounting info on >>>>> the newly created root domains by iterating through the (DEADLINE) tasks >>>>> belonging to the configured cpuset(s). >>>>> >>>>> Apart from some minor refactoring needed to rebase the set on top of >>>>> Waiman Long's cpuset for cgroup series (now mainline), two changes worth >>>>> of notice: >>>> Generally looks good to me but can you please ask Waiman to take a >>>> look? >>> Argh! I should have cc-ed him in the first instance. >>> >>> Thanks for reviewing. >>> >>> Waiman, do you see anything wrong with this series? Thanks! >>> >>> https://lore.kernel.org/lkml/20190117084739.17078-1-juri.lelli@xxxxxxxxxx/ >> Ping? > Basically looks OK to me; wlthough I think I prefer the callback_lock / > rq->lock ordering to be the other way around. > > Waiman, you OK with this one? Sorry for the late reply. I reviewed the patchset and don't see anything wrong with it. However, my knowledge of the internal operation of the deadline scheduler is limited. Cheers, Longman
