From owner-freebsd-x11@FreeBSD.ORG Tue May 7 20:38:51 2013 Return-Path: Delivered-To: freebsd-x11@freebsd.org Received: from mx1.freebsd.org (mx1.FreeBSD.org [8.8.178.115]) by hub.freebsd.org (Postfix) with ESMTP id 5BAEA6EF for ; Tue, 7 May 2013 20:38:51 +0000 (UTC) (envelope-from rfg@tristatelogic.com) Received: from outgoing.tristatelogic.com (segfault.tristatelogic.com [69.62.255.118]) by mx1.freebsd.org (Postfix) with ESMTP id 1565AC2 for ; Tue, 7 May 2013 20:38:49 +0000 (UTC) Received: from segfault-nmh-helo.tristatelogic.com (localhost [127.0.0.1]) by segfault.tristatelogic.com (Postfix) with ESMTP id 33EC33AF45 for ; Tue, 7 May 2013 13:38:46 -0700 (PDT) From: "Ronald F. Guilmette" To: freebsd-x11@freebsd.org Subject: Re: Status of ATI video acceleration support + freebsd port code Date: Tue, 07 May 2013 13:38:46 -0700 Message-ID: <23361.1367959126@server1.tristatelogic.com> X-BeenThere: freebsd-x11@freebsd.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: X11 on FreeBSD -- maintaining and support List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 07 May 2013 20:38:51 -0000 Sergio de Almeida Lenzi wrote: >> Why from scratch? Let's do fork of ums code (and forexamle call him ums-ng) >> and piecemeal upgrade him to functionality of kms. I am sure that there >> will be supporters of my ideas. People must have choice ums vs kms like as >> linux vs freebsd! > >>From my point of view, it is just a matter of $$$$, Sigh. If only that were true! (Mere money cannot produce skill and, more importantly, experience. Money can buy more developers, but as Fred Brooks noted several decades ago, adding developers to a late project only makes it later. Nine women cannot have a baby in one month.) >I (and some others) are funding a company to support FreeBSD, I'll ask the obvious question: Why? >I do not have an idea, but, do you people can tell how much does it >would costs to write UMS or KMS code for FreeBSD??? I think I have >the money, I just must know how much is needed.. hehe How much have you got? :-) (If you look in the United States, at least, I'm sure that you will have no trouble finding plenty of people who are ready, willing, and eager to separate you from whatever money you have available.) >I really do not care if the code goes in the kernel or in user mode, >As far as I can test, running Linux vs FreeBSD in the notebooks I tested >the FreeBSD is far more reliable, faster, than Linux. Ummm... well... no, actually. It really depends on what you are doing. For *production servers*, hands down I would go with FreeBSD every time. For most anything to do with networking, Big Iron, serving up zillions of web pages per hour, or crunching massive data bases, FreeBSD performs, and it's built like a tank. Several major corporations... big name players on the web... depend on it, day in and day out. For the desktop however (and I'm sure that saying this here will probably result in my public stoning), as I have learned only slowy and painfully, it actually kinda sucks. As we have been discussing, _full_ support for video chips from two of the three biggest suppliers of such things, AMD and Nvidia, doesn't exist, even for some cards already considered to be obsolete, such as the Radeon HD 5450 which, I gather, was introduced way back in 2011. Support for USB-connected mass storage devices... you know... like the popular plug in portable (rotating media) USB drives that many people, myself included, are using these days... is, to put it generously, buggy. (Just copying a file to one such drive from my FreeBSD system yesterday resulted in a hard crash and automatic reboot, after which I had to manually fsck several partitions, just to get the system back online. And I have gotten system crashes/lockups from just simply un- plugging such devices... something that I was going to look into further, with some kernel debugging turned on, but never found the time to do.) There are other problems too. I'm sure that there are many other kind of peripherals that are simply unsupported because nobody in the FreeBSD community... which I gather is mostly composed of "Big Iron" type people, and considerably less desktop users... cares about the specific peripherals in question. So support for such peripherals is essentially do-it-yourself. More depressing than any of that however is the way an ordinary non-kernel- hacker kind of person can be left effectively twisting in the wind when something unexpected and unusual goes wrong. For example, due to what amounted to just an ordinay power failure, my ports tree became "corrupted" in some vague, mysterious, and unspecified way back in March, and not a single person from the "community" ever stepped forward to even explain to me what the actual problem was/is: http://lists.freebsd.org/pipermail/freebsd-questions/2013-March/250191.html As of today, I *still* have not attended to this serious and major headache. Since nobody can or will even tell me what the real problem is, I'm going to have to dredge into the source code for pkg_version myself to try to even just figure how what the hell it is complaining about, exactly, so that I can have some hope of correcting the problem... whatver the hell it is. Untill I do that, I don't dare install any new ports an I don't dare even update any of my existing installed ports for fear that will just "corrupt" things further. Oh yea, one more gripe... Linux users generally have access to the latest and greatest releases of _everything_... at least everything that a desktop user might ever want or need. In the case of FreeBSD, this is *rarely* the case. Port maintainers apparently all have day jobs that take priority. Nobody can blame them for that, but it is a *constant* annoyance to be using this port or that port, to find a bug in it, to report the bug to the original developers, only to have them ask the standard question: "Are you seeing this bug in the latest release?" and then to look at either the output of pkg_info or the corresponding makefile in the ports tree only to find out that no, you are running the latest official FreeBSD port of the thing in question, which is one or two or three releases back from what everybody in the Linux and windows communities are running. (Just one example: XBMC in the ports tree is stuck at 12.0. Th XBMC team just released 12.2 yesterday. I could eaily give you hundreds more examples like this.) For all of the above reasons, after well more than a dozen years of being an ardent user and promoter of FreeBSD, as of this week I am getting ready to finally accept the handwriting that has been on the wall for years already. I'm going to dump FreeBSD and switch myself entirely over to Debian Linux. I'm just getting too old for all this aggravation. (Sorry to dump all of my gripes on you all, but this head of steam has been building up for some time now.) >The only problem with >FreeBSD is the drivers for the radeon and intel video chips... No, that is definitely *not* the only problem. See above. If you need a machine to network (vie ethernet) to anything else, and pretty much no matter what kind of wireless or wired ethernet card you have, or how many of them you need to stick into one single system, FreeBSD is for you, and solid as a rock. If you plan to use and/or plug in and/or unplug any kind of USB peripheral however... well.. my advice is that you should make sure that you have current backups for everything. Likewise, if you need to crunch a big database, use FreeBSD. If you want to use (the latest official releases of) XBMC, VLC, Avidemux, Handbrake, or any other "multimedia" type application with any non-Intel graphics card that has been manufactured this century, then buy a Windoze license. At least that way when it doesn't work you have at least a prayer of getting some support from the vendor. >We will need also code for the wifi, camera, and sleep/wakeup... Yes, and etc., etc., etc... (As noted above, WiFi should be no problem. Cameras however are probably going to be USB devices, so good luck with that. Good luck making it work reliably. Proper support for sleep/suspend/wakeup may or may not be there. Even without knowing anything about it, if I were a betting man I would bet that support for that is _alleged_ to be present in current releases of FreeBSD, but that some aspect of it will be broken or will malfunction, and then you will need to report that as a PR, and maybe sometime in 2015 the relevant maintainer/developer may have time to look at that.) >The idea is to sell notebooks to a closed market, about 100.000 units... >each notebooks can sell for US$400 (4gb of ram, 320Gb of disk) 14inch >screen. >no brand name... > >Operating system FreeBSD9 or 10, gnome 2.32. Good luck with that. I really don't know why you would want to swim against the tide. Why not just use Linux? >OPPINIONS please???? I don't have any of those. As you can see however, I have plenty of opinions, most of which I should probably keep to myself, but that has never stopped me before, so... Regards, rfg