VIA releases Linux Driver source packages: On top of the UniChrome drivers, VIA also offers display driver sources for the VIA ProSavage and ProSavage DDR integrated graphics controllers as well as the integrated Network driver source supporting six VIA Chipset South Bridges.
Nice! Not much else to say.
While thats good, now perhaps Via should also try to get their marketing act together, the only place to buy specifically miniITX boards that use these chip sets is the web, nothing in the stores anywhere even in TW mobo markets, ordinary folks don’t seem to want compact formats. Puts me off from buying one if I can’t see if its acceptable running something.
I downloaded the VIA driver source code but was unable to determine what license (if any) it is under. Is it free software?
I don’t agree with nvidia closed code philosophy. It doesn’t help me. I have to install the driver everytime I restart the computer because of some extrange thing, and everytime I wanna upgrade my kernel (for security issues, for example), the xserver stops working. This is not a solution, and It is going to be worse in the near future when composite and opengl are going to be used harder inside kde and gnome.
I would rather preffer them to have an open driver inside the kernel that releases 70% of the power the closed code gives.
If this VIA chipset or any other is fast enough, I can start thinking on changing graphic card.
This license seems to be attached to the source files:
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sub license, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice (including the next paragraph) shall be included in all copies or substantial portions of the Software.
[SNIP STANDARD ALL CAPS DISCLAIMER]
…
So, basically a permissive BSD-type license.
Sounds like the only difference between this and GPL v2 is that this also provides the right to “sub license”. Hopefully this will help to put pressure on other hardware manufacturers to follow VIA’s example. In any case, it is a nice contribution from VIA.
Nothing new here, move along.
The only thing that is happenend is instead of having to go through the developer’s portal, anyone can download the driver source. Nothing that we didn’t have before…
The video driver is still not of the caliber to be included in xorg sources, and the method of mpeg acceleration is not useable by anything of consequence. Also, the mpeg decoding code is not available, it is in a binary.
If you want to see something worth looking at, go to http://unichrome.sf.net And give those guys a hand.
It makes so much sense that they would do this. They’re in the business of selling video cards; and the best way to sell more is to make them easy to use. Having the driver available so that anyone can make it work on their system is a really good idea. nvidia really should get their act together.
null
The market has to prove to the suppliers that there are sales to be had.
I’d like to build a whamper-dyne dual LCD system, but, if I can’t figure out what hardware to support, I’ll default to nVidia, which works well with Gentoo.
I feel no sense of moral outrage over a binary driver, but I’d really rather support open source.
nvidia really should get their act together.
Why? I am already grateful that they bother to support a tiny market mostly buying low-end cards. NVidia won’t open-source their driver because of licence restrictions (since they have licenced technology/code/trade secrets). Remember the Linux driver share about 85% of the code with the Windows driver. Why aren’t they releasing the specs? That is the real question.
VIA and XGI has nothing to lose by opening their drivers. Their market share can only go up.
This is wonderful for Averetec laptop users (at least the 3200 series owners anyway). Until now VESA was the only option in the linux world. Thank you VIA!
Possibly good news for Averatec 3150 owners, as well. I’ve been using a driver from Xorg CVS for a while. It’s much better than nothing. I just shot an e-mail to the mailing list asking for their take on this development.
mostly buying low-end cards.
Of course, this belies the fact that the whole reason NVIDIA supports Linux is because of the workstation users buying high-end cards…
Very nice step from VIA. I’ve always been reluctant/unwilling to spend money on hardware, if the vendor won’t release programming or hardware specs. For common x86 PC’s, their legacy and public-spec interfaces like USB make them pretty much documented.
But for specialized components like 3D accelerated videocards or chipset integrated features, it looks like vendors have been closing up things. Personally I don’t like this at all, and I suppose it’s time to vote with my wallet. Future hardware purchases will be carefully selected for their ‘open-friendliness’.
I can imagine vendors like VIA being tied by restrictions imposed on them by 3rd parties, but basically: they should get rid of this. Strike a deal, or replace 3rd-party incorporated tech with own developments.
As to the immature(?) state of the driver(s), that’s sad. Already done for all flavours of Windows, how hard could it be to develop Linux version(s) alongside, for a company like VIA? BTW. plain hardware/programming specs help a lot as well.
Anyway, announcements like this are a very good way to please me as potential customer. Keep it up, VIA!
About a week ago I was looking for a Socket-939 motherboard (AMD64). I had to choose between VIA and NVIDIA. I finally chose a VIA-based Abit AV8 and after reading this item I definitely have no regrets 🙂
Here’s the forum thread I opened before going VIA:
“Which company is more OSS friendly – VIA or NVIDIA?”
http://episteme.arstechnica.com/eve/ubb.x/a/tpc/f/96509133/m/412005…
Prog.
“Why? I am already grateful that they bother to support a tiny market mostly buying low-end cards. NVidia won’t open-source their driver because of licence restrictions (since they have licenced technology/code/trade secrets).”
You do realize that basically we only have their word that that’s the reason. Considering how the whole nForce, ethernet issue went. I’d say it’s pure mindset. Of course last time this was brought up, someone pointed out IP litigation as a possible reason, in a sue happy environment. I have my doubts on that as well. As you said the interface spec could be released, and also if true, why are other companies releasing theirs (Intel, etc). Aren’t they worried about being sued too?
For you who don’t bother reading the first page of posts.
No VIA hasn’t done anything that helps the OSS community.
The released code contains nothing of interest to the REAL OSS driver (http://unichrome.sourceforge.net)
It only works with via-forks of mplayer and xine
It has known security holes and can only be used as root
And it doesn’t look like VIA will do anything to fix this, while real OSS developers focus on the unichrome project instead.
So if you are stuck with a VIA board, go with the unchrome driver and just ignore VIA.
See totte’s (unichrome developer) reply here:
http://lwn.net/Articles/131777/
Well, can’t anyone who wants to fix these problems? Isn’t that the point of the source code being open?
Point beeing, it is allready done, and was done a long time ago by the unichrome project.
VIA would do better by actually helping the unichrome project
Definately look at totte’s (Thomas Hellstrom) post on lwn. He has been a developer on the open source driver for some time.
VIA seems to think open source is a goal to achieve. In reality it is more about colaberation. They try to get every thing to work in their own way. Without any communication with open source projects. Instead of trying to get patches into mplayer and xine to make them work with their mpeg decoder, they just forked them and released them. Not a single email has been seen in either camp.
The http://unichrome.sf.net people have tried to reach out VIA on many occasions and have been met with resounding silence. These are people that are helping VIA sell product!! Only recently has VIA even acknowlegded the existance of the open source drivers…
My point being, VIA does not deserve any praise. What they need is to be smacked up side the head so they wake up and see what the open source community can do for them if they cooperate.
It sounds like there’s a lot more VIA could be doing. After e-mailing with the developers of the ProSavageDDR DRI / DRM driver, it looks like this really isn’t anything new. It’s just old stuff that’s been available through their developers’ portal forever. It’s barely functional, and a long time ago, the Unichrome project and others picked up this code and started cleaning it up. Today, there are far better implementations available that happen to be based on this code.
Wake me up when they release hardware specs, contribute to the more advanced driver projects, acknowledge these projects’ existence regularly, or commit their own developers to the more advanced projects.
Of course, this belies the fact that the whole reason NVIDIA supports Linux is because of the workstation users buying high-end cards…
If this was the sole reason, they wouldn’t bother to fix up their driver after every compatibility breaking change in the Linux kernel…
You do realize that basically we only have their word that that’s the reason. Considering how the whole nForce, ethernet issue went. I’d say it’s pure mindset. Of course last time this was brought up, someone pointed out IP litigation as a possible reason, in a sue happy environment. I have my doubts on that as well.
I am pretty sure the current code for the binary driver is encumbered beyond hope. That said, there is nothing that could prevent them from writing a new unencumbered driver. Still, it would take some time and money and the investment might not worth it. After all, most users are already satisfied by the binary drivers.
The forcedeth module you mentioned was started by hackers, not the company. Since NVidia reclaimed the copyright of the driver, I am pretty sure this is a question of control. If somebody starts writing an open-source driver for their cards, they will probably start making up open contributions. <theory>If NVidia is under NDAs from different companies, an open-source project that wasn’t initiated by them could cover their asses. They could plead “they did it without our help, we merely patched their code” or stuff like that.</theory>
As you said the interface spec could be released, and also if true, why are other companies releasing theirs (Intel, etc). Aren’t they worried about being sued too?
With all the cross-licence agreements they have, Intel doesn’t have to worry…
If this was the sole reason, they wouldn’t bother to fix up their driver after every compatibility breaking change in the Linux kernel…
Once the driver is ported, the effort to fix those little issues isn’t nearly worth the hassle of the e-mails bugging about “why isn’t the driver compatible with kernel 2.6.y”. And don’t forget that it was quite awhile before kernel 2.6 was supported, though most people didn’t notice so much because there were 3rd party patches to fix the issues.
Why should Nvidia care what OS I use their job is to make hardware and they should concentrate on that. What really turns me off is that companies like ATI and Nvidia play slaves to Microsoft, they are all spineless jellyfish.