|VIA closes driver source.
||[Nov. 6th, 2006|02:14 am]
|||||Studio Brussel's Night programming.||]|
VIA announced the P4M900 northbridge in May. According to marketing, it comes with the "Chrome 9 HC IGP". I'm not quite sure what the HC stands for, but that's marketing for you.
The pci-id 0x3371 was listed as P4M900 in VIA's so called "open source" code drops, version XF40068. But the code in this release was so crappy and fractional that i didn't even bother to update the site, and it subsequently slipped my mind.
This code should've updated the initialisation of the acceleration engine to bring it in line with the other Chrome 9 IGP (VT3230 or K8M890), but they didn't. They did add the VT3343 (P4M890) to one of the cases, but that's a unichrome, not a Chrome, and it is extremely unlikely that this is correct.
Seems like someone at VIA did try to think logically about device naming. He probably only joined VIA recently. :)
So apart from the bogus 68 release, there's a binary driver with a 70 release that's supposed to support the VT3371.
Curiously, this driver was never announced on viaarena.com. I only now stumbled over it looking for more info on the VT3371.
This lack of information proves that VIA is quite well aware of what their move could cause.
VIA might be reasoning that there is IP heavy stuff in this driver. But a binary doesn't stop any appropriately motivated and clued coder from finding out what he needs to know. Especially with code for previous and thus highly comparable driver versions around. Reverse engineering this is a straightforward but cumbersome task. A waste of time.
Closing down this driver serves no real purpose. All it does is infuriate users.It's just insanely bad marketing.
But VIA never has been able to see users as the valuable customers they are. Apparently VIA also sells hardware to people who just like to look at their kit.
So, let us review where VIAs stands with respect to free software:
* Complete and prolonged (3y+) inability to work with actual free software developers.
* Crappy code and sporadic releases, totally unfit for direct use.
* Very bad handling of licensing, with proprietary licenses popping up once in a while.
* An NDA that, when VIA says so, forces signee to breach the GPL.
* Distributes tarball with the binary "libddmpeg.so" and labels the lot as "open source" in direct violation of The Open Source Definition.
* No longer distributes a DRI driver now, after having provided DRI binaries for a bit inside their "open source" tarballs. This used to be all source at one point.
And now, as if all of this wasn't bad enough already, they stopped providing source completely.
As far as I'm concerned, VIA has no real right to use the term "Open source", especially not on their front viaarena page. It's a lie, it has always been a lie, and it's been going on for far too long.
You know, it almost seems as if VIA doesn't want to sell any hardware. I won't be surprised if they do end up like Transmeta (and countless others). Not doing anything, except have lawyers milk the residual IP for all its worth.