Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S934778AbcKKSql (ORCPT ); Fri, 11 Nov 2016 13:46:41 -0500 Received: from resqmta-ch2-03v.sys.comcast.net ([69.252.207.35]:57233 "EHLO resqmta-ch2-03v.sys.comcast.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755669AbcKKSqj (ORCPT ); Fri, 11 Nov 2016 13:46:39 -0500 Date: Fri, 11 Nov 2016 12:46:37 -0600 (CST) From: Christoph Lameter X-X-Sender: cl@east.gentwo.org To: Daniel Vacek cc: Daniel Bristot de Oliveira , Tommaso Cucinotta , LKML , linux-rt-users , Peter Zijlstra , Steven Rostedt , Ingo Molnar Subject: Re: [PATCH] sched/rt: RT_RUNTIME_GREED sched feature In-Reply-To: Message-ID: References: <20161108115958.GO3142@twins.programming.kicks-ass.net> <20161108090740.4226ffc9@gandalf.local.home> <20161108165133.GI3117@twins.programming.kicks-ass.net> <20161108121710.3e7eb664@gandalf.local.home> <20161108180548.GN3117@twins.programming.kicks-ass.net> <20161108195015.GP3117@twins.programming.kicks-ass.net> <8a80c2c2-3803-5648-67b2-4dee7381d5a6@redhat.com> Content-Type: text/plain; charset=US-ASCII X-CMAE-Envelope: MS4wfM6Jl6JWH3arcWMYfLHDzpO8y66RSSIwax2qjqNgR8soNFzINoYR+3tVal6nKI3GUKHtErVrHD+l01DkAYfkNDIqSdc1qo3WTwq6rFztr/3zoEth2ee2 3EgEEQ+NcMyF1oy/BWlMnEyM0f2QxqqJgQ5aWobBQH+jVn8/rnCmRwFA77ogpGLK7bex8cYcXiipE/44XaNc48/NBXEaUeOm2oza3dWLr8xq6UxKoG02vgiv CRZEBVzqNtxmlG/CQFFYNJU47nhsP3wa7+3cLXY2j7zBFaBdoJ82vt+thLhbvskbVppl6otvvU4oI+hA9WJDiQ3JULjN3tJLXIgfKhfu9/KHrioT+nJAd08M 8YRVQpF8CxJTzULLKkACyZ5Ebk5/IpWeY3np26REv2Eke+Celc0= Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 675 Lines: 14 On Thu, 10 Nov 2016, Daniel Vacek wrote: > I believe Daniel's patches are the best thing we can do in current > situation as the behavior now seems rather buggy and does not provide above > mentioned expectations set when rt throttling was merged with default > budget of 95% of CPU time. Nor if you configure so that it does (by > disabling RT_RUNTIME_SHARE), it also forces CPUs to go idle needlessly when > there still can be rt task running not really starving anyone. At least > till a proper rework of rt scheduling with DL Server is implemented. This looks like a fix for a bug and the company I work for is suffering as a result. Could we please merge that ASAP?