From owner-svn-src-head@FreeBSD.ORG Thu Jan 15 14:08:42 2015 Return-Path: Delivered-To: svn-src-head@freebsd.org Received: by hub.freebsd.org (Postfix, from userid 1033) id 9835C891; Thu, 15 Jan 2015 14:08:42 +0000 (UTC) Date: Thu, 15 Jan 2015 14:08:42 +0000 From: Alexey Dokuchaev To: Slawa Olhovchenkov Subject: Re: svn commit: r277204 - head/sys/amd64/conf Message-ID: <20150115140842.GA10593@FreeBSD.org> References: <201501150042.t0F0g7Um018059@svn.freebsd.org> <20150115132303.GA245@zxy.spb.ru> <20150115134446.GA92636@FreeBSD.org> <20150115135342.GD3698@zxy.spb.ru> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20150115135342.GD3698@zxy.spb.ru> User-Agent: Mutt/1.5.23 (2014-03-12) Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org, Warner Losh X-BeenThere: svn-src-head@freebsd.org X-Mailman-Version: 2.1.18-1 Precedence: list List-Id: SVN commit messages for the src tree for head/-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 15 Jan 2015 14:08:42 -0000 On Thu, Jan 15, 2015 at 04:53:42PM +0300, Slawa Olhovchenkov wrote: > On Thu, Jan 15, 2015 at 01:44:46PM +0000, Alexey Dokuchaev wrote: > > intention. AFAIR last time we had a discussion about why our default > > kernel is not MINIMAL, it boiled down to two main problems: 1) loader's > > caching of disk reads (which makes loading *.ko's from /boot/loader.conf > > reading large monolitic kernel is slow too. But not nearly as slow as loading 50-60 modules at boot time (on my stable/8 I have 59 right now). When you read one big file *once* you don't have to worry about caching reads. With everything moved to modules, loader does a lot of superfluous disk access, and to remedy this we need smart(er) caching implementation. > /boot/loader.conf (with all modules currently present in GENERIC) may > be instaled by bsdinstall (and may be part of base.txz). That could be done, but not before we solve (1), and from this POV it deems more important than (2). > This is not only space saving. > This is allow to [unload and load modules, change configuration in runtime > with minimal downtime and without reboot]. I think we all know what are the benefits of modularized kernels. But before we solve aforementioned problems, it will remain a custom option for advanced users. ./danfe