From owner-freebsd-arch@freebsd.org Fri Feb 17 21:24:27 2017 Return-Path: Delivered-To: freebsd-arch@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id BECE8CE34B8 for ; Fri, 17 Feb 2017 21:24:27 +0000 (UTC) (envelope-from jhb@freebsd.org) Received: from mail.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 8E8DF1506; Fri, 17 Feb 2017 21:24:27 +0000 (UTC) (envelope-from jhb@freebsd.org) Received: from ralph.baldwin.cx (c-73-231-226-104.hsd1.ca.comcast.net [73.231.226.104]) by mail.baldwin.cx (Postfix) with ESMTPSA id 442E410A791; Fri, 17 Feb 2017 16:24:26 -0500 (EST) From: John Baldwin To: Sean Bruno Cc: freebsd-arch@freebsd.org, Warner Losh Subject: Re: Thinning out GENERIC Date: Fri, 17 Feb 2017 13:23:01 -0800 Message-ID: <6526794.FlmyzlfvMp@ralph.baldwin.cx> User-Agent: KMail/4.14.10 (FreeBSD/11.0-STABLE; KDE/4.14.10; amd64; ; ) In-Reply-To: References: <34e4e9bd-c7a8-e3e2-4cef-98a691ea7944@freebsd.org> <1907688.dXvDEfTesD@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-4.4.3 (mail.baldwin.cx); Fri, 17 Feb 2017 16:24:26 -0500 (EST) X-Virus-Scanned: clamav-milter 0.99.2 at mail.baldwin.cx X-Virus-Status: Clean X-BeenThere: freebsd-arch@freebsd.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: Discussion related to FreeBSD architecture List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 17 Feb 2017 21:24:27 -0000 On Friday, February 17, 2017 12:54:29 PM Sean Bruno wrote: > > On 02/17/17 12:49, John Baldwin wrote: > > On Friday, February 17, 2017 11:43:45 AM Sean Bruno wrote: > >> > >> On 02/13/17 11:33, John Baldwin wrote: > >>> On Monday, February 13, 2017 10:51:36 AM Warner Losh wrote: > >>>> [[ It isn't just GENERIC, it's in every single kernel ]] > >>>> > >>>> On Sun, Feb 12, 2017 at 7:59 AM, Sean Bruno wrote: > >>>>> Index: sys/conf/files > >>>>> =================================================================== > >>>>> --- sys/conf/files (revision 313683) > >>>>> +++ sys/conf/files (working copy) > >>>>> @@ -2419,7 +2419,7 @@ > >>>>> dev/pci/pci_user.c optional pci > >>>>> dev/pci/pcib_if.m standard > >>>>> dev/pci/pcib_support.c standard > >>>>> -dev/pci/vga_pci.c optional pci > >>>>> +#dev/pci/vga_pci.c optional pci > >>>>> dev/pcn/if_pcn.c optional pcn pci > >>>>> dev/pdq/if_fea.c optional fea eisa > >>>>> dev/pdq/if_fpa.c optional fpa pci > >>>>> > >>>>> I'm playing around with small x86 builds and wanted to come up with a > >>>>> knob that disables the vga_pci driver. Obviously I still want pci(4) to > >>>>> be enabled, but I definitely do not need the vga interfaces for > >>>>> anything. Any suggestions? > >>>> > >>>> You could make it dependent on vt or sc as well as pci, no? > >>> > >>> This is a fairly small driver: > >>> > >>> 6139 340 8 6487 0x1957 /usr/obj/usr/src/sys/GENERIC > >>> /vga_pci.o > >>> > >>> (From an 11.0-stable build) > >>> > >>> device ppc and friends is a lot more for example: > >>> > >>> text data bss dec hex filename > >>> 6238 0 0 6238 0x185e /usr/obj/usr/src/sys/GENERIC/ppb_1284.o > >>> 1675 28 0 1703 0x6a7 /usr/obj/usr/src/sys/GENERIC/ppb_base.o > >>> 2042 0 0 2042 0x7fa /usr/obj/usr/src/sys/GENERIC/ppb_msq.o > >>> 6025 296 8 6329 0x18b9 /usr/obj/usr/src/sys/GENERIC/ppbconf.o > >>> 0 168 0 168 0xa8 /usr/obj/usr/src/sys/GENERIC/ppbus_if.o > >>> 9099 36 0 9135 0x23af /usr/obj/usr/src/sys/GENERIC/ppc.o > >>> 495 256 0 751 0x2ef /usr/obj/usr/src/sys/GENERIC/ppc_acpi.o > >>> 1671 256 0 1927 0x787 /usr/obj/usr/src/sys/GENERIC/ppc_isa.o > >>> 1124 208 0 1332 0x534 /usr/obj/usr/src/sys/GENERIC/ppc_pci.o > >>> 451 208 0 659 0x293 /usr/obj/usr/src/sys/GENERIC/ppc_puc.o > >>> 2751 428 8 3187 0xc73 /usr/obj/usr/src/sys/GENERIC/ppi.o > >>> > >>> It is needed for drm. > >>> > >> > >> Ok, so ... make it dependant on drm, vt, sc and pci? > > > > You can kldload drm at runtime, so a static dependency isn't sufficient. > > It's not clear why removing this from GENERIC is that important. ppc is on far > > fewer machines than vga_pci and takes up 10x more space. > > > > I don't want to remove it from GENERIC. I want to add the capability to > remove it. Sorry if that wasn't clear from my subject line, because it > doesn't look clear now that I re-read it. Why not make this conditional on 'vga'? GENERIC already contains 'device vga' that the VGA-specific bits of 'vt', etc. are dependent on. -- John Baldwin