From owner-freebsd-hackers Fri Oct 31 22:31:29 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id WAA03787 for hackers-outgoing; Fri, 31 Oct 1997 22:31:29 -0800 (PST) (envelope-from owner-freebsd-hackers) Received: from rf900.physics.usyd.edu.au (rf900.physics.usyd.edu.au [129.78.129.109]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id WAA03779 for ; Fri, 31 Oct 1997 22:30:59 -0800 (PST) (envelope-from dawes@rf900.physics.usyd.edu.au) Received: (from dawes@localhost) by rf900.physics.usyd.edu.au (8.8.5/8.8.2) id RAA03824; Sat, 1 Nov 1997 17:30:31 +1100 (EST) Message-ID: <19971101173031.65071@rf900.physics.usyd.edu.au> Date: Sat, 1 Nov 1997 17:30:31 +1100 From: David Dawes To: freebsd-hackers@FreeBSD.ORG Subject: Re: svgalib? References: <199710290633.RAA21328@holly.rd.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.69 In-Reply-To: <199710290633.RAA21328@holly.rd.net>; from Daniel J. O'Connor on Wed, Oct 29, 1997 at 05:03:47PM +1030 Sender: owner-freebsd-hackers@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk I just have a few comments about some of the things that were mentioned in this thread. spork wrote: >Personally, I think it has alot to do with whether or not the user has the >full array of documentation for all their video hardware. I found it easy >with all the books in front of me, but if you don't know the max refresh >on your monitor, or you have an unknown video card in the system, >questions about your "dot clock", etc. seem rather daunting. How many modern video cards require you know about "dot clocks"? In most cases, you run XF86Setup and pick your card out of the list. Sure, there are plenty of cases where it isn't as simple as that. The majority of these are either old hardware or very new (ie unsupported) hardware. In my experience on the XFree86 support list, a lot of people make things more complicated than necessary either because of past experience with (very) old XFree86 versions, or because they've been reading out of date or simply misleading documentation. If you have a modern card, you shouldn't need to know anything about "dot clocks". Amancio Hasty wrote: >The problem can still be contained by way of knowing the PCI ID of the >graphic card so at the very least XF86setup can limit the query to >a few cards if not just one. The monitor setup is a little easier to >manage by having a database of popular models and modes for the >VESA monitors. PCI cards are the easy ones (relatively speaking). If we could simply forget about all the ISA and VLB cards still out there, life would be a lot easier. However, even with PCI cards, knowing the PCI ID (including chip revision) isn't the end of the story. As you say though, it could be used to at least reduce the number of cards presented as options. As for monitors, XF86Setup provides a simple choice based on the resolution/refresh rate the monitor can do. Picking one of those options should be enough to get things going in most cases. If you have the specs in front of you, sure, you can choose to enter in the exact values. Jordan K. Hubbard wrote: >... and I'm sure that the XFree86 folks would just love your help in >doing that in a fashion which works on a reasonable number of the >operating systems they support. :-) > >Otherwise, the chances of this actually happening are close to nil, I >suspect. We've been waiting for a more automatic setup tool for X for >literally years now, and if any improvements have been done it's been >primarily to make the manual setup process a bit more palatable, not >more automatic. Everyone agrees that it should be done, nobody writes >the code. :-( Yes, very true. The person who did the first cut of XF86Setup changed jobs shortly after that was done, and since then has had very little time to work on it. It has only been tweaked around the edges by others since then. Alex wrote: >Read if it works on Linux, they'll bite. ;-) Really? At least some of us who have some say in the directions of XFree86 don't use Linux (at least not by choice). From my point of view, it has to work on all the free ix86 BSD variants, Linux, and preferably also the major SVR4-based ix86 OSs. Something that is Linux-only won't get my support because it is useless to me. >What about SuperProbe? It appears to already do some PCI probing. Why not >attempt to extend it into some sort of autoconfigurator thingy. If it's PCI probing is easy. One of the original plans for SuperProbe was to use it to assist in Xserver configuration. It was originally written in the pre-PCI days, and it became clear then that it couldn't reliably detect the hardware available at the time without often serious side-effects (take a look at the man page for SuperProbe and read about some of its command line options). >gpl'd one could hack together some source, otherwise I'm sure it would be >farily easy to grep and sed your way through the output of that to >automagically set everything up. You're lucky the original author of SuperProbe isn't reading this :-). It is most definitely not GPL'd (there is no GPL'd source in XFree86, and there likely never will be). The XFree86 copyright is similar in style to the BSD copyrights. Snob Art Genre wrote: >My video card setup was a breeze, but I never did get my mouse working as >a PS/2 . . . probably because I had a spare serial port and a serial >adapter for the mouse (a Logitech MouseMan, and a very nice mouse indeed), >and so I took the easy way out. I have such a mouse, and it works fine as a PS/2 mouse (I need all the serial ports for other uses). All I did was set the appropriate jumper on my motherboard to enable the mouse port, build a kernel with psm included, and set the protocol in the XF86Config to "PS/2". Kazutaka YOKOTA wrote: >PS: I am currently working on mouse drivers and moused so that moused >and XFree86 will be identify appropriate mouse protocol automatically. This is something which should help a lot. >2. XF86Setup isn't good at handling mice. If you are using it to > configure XF86Config, do not touch your mouse until you have specified > everything about the mouse. (device: /dev/psm0, protocol: PS/2) > And do NOT change mouse settings again until you quit XF86Setup. This is particularly true for PS/2 mice. Finally, as for GGI, I'm going to keep an open mind on this until they have something to show. On the subject of XFree86-DGA, it was based very loosely on Sun's DGA, and it came about primarily from prompting by some people at Id, so that they could get reasonable performance from Doom/Quake on Linux/X. Development of XFree86-DGA and related things seems likely to continue, with added impetus from the group working on 3D/XGL support for XFree86. David