From owner-svn-src-all@FreeBSD.ORG Fri Apr 3 18:55:39 2015 Return-Path: Delivered-To: svn-src-all@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 4146AA64; Fri, 3 Apr 2015 18:55:39 +0000 (UTC) Received: from bigwig.baldwin.cx (bigwig.baldwin.cx [IPv6:2001:470:1f11:75::1]) (using TLSv1 with cipher DHE-RSA-CAMELLIA256-SHA (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 1532B2D1; Fri, 3 Apr 2015 18:55:39 +0000 (UTC) Received: from ralph.baldwin.cx (pool-173-54-116-245.nwrknj.fios.verizon.net [173.54.116.245]) by bigwig.baldwin.cx (Postfix) with ESMTPSA id 1B1F0B98C; Fri, 3 Apr 2015 14:55:38 -0400 (EDT) From: John Baldwin To: Adrian Chadd Subject: Re: svn commit: r280799 - head/sys/dev/ath Date: Fri, 03 Apr 2015 13:15:01 -0400 Message-ID: <1754332.VQBJusLPhT@ralph.baldwin.cx> User-Agent: KMail/4.14.2 (FreeBSD/10.1-STABLE; KDE/4.14.2; amd64; ; ) In-Reply-To: References: <201503282341.t2SNfOZc032386@svn.freebsd.org> <2308891.2rLa0R7ktj@ralph.baldwin.cx> MIME-Version: 1.0 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="us-ascii" X-Greylist: Sender succeeded SMTP AUTH, not delayed by milter-greylist-4.2.7 (bigwig.baldwin.cx); Fri, 03 Apr 2015 14:55:38 -0400 (EDT) Cc: "svn-src-head@freebsd.org" , "svn-src-all@freebsd.org" , "src-committers@freebsd.org" X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.18-1 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 03 Apr 2015 18:55:39 -0000 On Monday, March 30, 2015 01:54:30 PM Adrian Chadd wrote: > On 30 March 2015 at 13:32, John Baldwin wrote: > > On Monday, March 30, 2015 10:10:43 AM Adrian Chadd wrote: > >> Yes, because it's before the dynamic hint stuff starts up. (kern.hintmode.) > >> > >> Hints don't show up in kenv on the mips boards until I set the sysctl > >> to the "import them now!" setting. > > > > Everything in if_ath.c happens well after SYSINITs are going. Are you > > explicitly forcing hintmode to 1 in your kernel config? At least on > > x86 it always is zero and always pulls from kenv. > > > > If this is magic that config does when you have static hints, then why > > not have init_dynamic_kenv() call into subr_hints.c to pull all the > > static hints into the kenv right away? SI_SUB_KMEM is way, way before > > any device drivers start doing anything. > > This is on MIPS and not on x86. And yes, I checked - both arge and ath > probe/attach get called with no hints in kenv (ie, static hints) and I > don't recall the hintmode being explicitly set in a config file. > > I'd love to make it work via hints calls; it'd significantly simplify things. Note that all the hints processing is MI, so MIPS vs x86 shouldn't matter in theory. I think hintmode gets set to 1 "magically" by config if your kernel config includes static hints. (Looks like this is true, config will add 'hintmode = 1' in the hints.c it generates.) I think what we should do is change the kenv code to basically do the equivalent of forcing hintsmode to 2 once the dynamic kenv is ready (it has a SYSINIT for this). If you do that, I think it means you can remove the sysctl as it's more or less useless then. It will increase RAM usage a bit as you will have the duplicate copy of hints in kenv, but hints aren't very large to begin with. You could also at that point replace hintsmode with a simple boolean "use_static_hints" that gets cleared by the SYSINIT that fires to convert static hints to kenv once dynamic kenv is ready. -- John Baldwin