From mboxrd@z Thu Jan 1 00:00:00 1970 From: Darren Hart Subject: Re: intel_turbo_max_3 non-modularity Date: Fri, 28 Apr 2017 15:51:53 -0700 Message-ID: <20170428225153.GB32336@fury> References: <20170424113103.0de3ea31@endymion> <20170424140218.GR16239@windriver.com> <20170428094930.6ba0ba1f@endymion> <20170428174437.GO23334@windriver.com> <20170428204949.GA3960@fury> <20170428213434.lljyvywigcguupjm@hirez.programming.kicks-ass.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from bombadil.infradead.org ([65.50.211.133]:57314 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1035982AbdD1Wv4 (ORCPT ); Fri, 28 Apr 2017 18:51:56 -0400 Content-Disposition: inline In-Reply-To: <20170428213434.lljyvywigcguupjm@hirez.programming.kicks-ass.net> Sender: platform-driver-x86-owner@vger.kernel.org List-ID: To: Peter Zijlstra Cc: Paul Gortmaker , Jean Delvare , Andy Shevchenko , platform-driver-x86@vger.kernel.org, Srinivas Pandruvada On Fri, Apr 28, 2017 at 11:34:34PM +0200, Peter Zijlstra wrote: > On Fri, Apr 28, 2017 at 01:49:49PM -0700, Darren Hart wrote: > > > > Non-modular drivers do not scale. This is a serious issue considering > > > > the growth rate of the kernel. > > > > > > Again, this is an oversimplification that doesn't reflect reality. > > > There are many valid reasons why some drivers be built-in only. The > > > whole world isn't a distro looking to say "=m" to every config option > > > simply because they don't have the time or resources to actually triage > > > them all and see what is relevant for their use case(es). > > > > > > > The point is valid though, and something we need to be keeping an eye on, > > especially with all the non-architectural SoC feature drivers we're seeing code > > roll in for. I appreciate you raising the concern Jean. > > FWIW: > > There's things like ACPI-Processor-Idle and intel_idle that need to be > built-in for correctness sake. > > They can mark the TSC unstable, and this should be done before we run > userspace, which with unstable TSC could see wobbly time before we kill > the TSC. > > This is a somewhat non-obvious case, and on the face of it those could > be modular drivers. But building them as modules gets them loaded far > too late if ever. > > These are subtle things and easily overlooked. Thanks Peter, I will be sure to do my due diligence in determining why something may be bool before applying incoming patches to change that. This isn't clearly documented in intel_idle for example, but it is at least explicitly described in the git log for the Kconfig. -- Darren Hart VMware Open Source Technology Center