Slashdot Mirror
Qualcomm Calls To 'Kill All Proprietary Drivers For Good'
An anonymous reader writes "Next week at the sixth Linux Foundation Collaboration Summit, two Qualcomm Atheros engineers will be making a stand for killing all proprietary drivers for good — across all operating systems. The Qualcomm slides go over their early plans. Do they stand a chance?"
I know where I'm throwing my money the next time I need such hardware!
For large sets, this will be our guide even unto death, for the LORD will work for each type of data it is applied to...
Killing software patents.
This is the sort of thing Google should have pushed for with Android, but after a year of struggling with their OS I've come to realize Google doesn't care about the end user experience. By subsidizing and dumping Android they pushed webOS and MeeGo out of the market.
If Qualcomm starts with their cellphone baseband processors, I'll start listening.
Yes, I somewhat agree. I was excited seeing this article appear in my feed, but have since sunken into a depression. The only way that proprietary drivers can be killed off (and I'm not talking drivers for specialist hardware) is if all hardware manufacturers agree on sticking to standards. Even within manufacturer, there are vast differences in hardware configurations, interrupts, etc. (Yes, of course, SATA, PCIe, are all standards, but you know what I'm talking about. How long does it take to get a flavor of *nix running 100% on a notebook? Fiddling with acpid.conf, patching, reverting, etc, etc)
Nobody is telling you what to do. Just like RFCs don't tell you what to do.
They tell you what you should do. This is an important distinction.
Of course, if you ignore those recommendations and do your own thing: you are on your own.
For large sets, this will be our guide even unto death, for the LORD will work for each type of data it is applied to...
Accomplishing such a feat would require the market to be largely informed and interested. Neither is the case.
Step One: Convert PowerPoint to randomly switch colors every third word when using Star Trek-like background styles.
(for those who rtfa on the slides)
I sometimes ask revealing, often ignorant-seeming questions. Maybe they're harder to answer than you think.
No.
I don't understand why not. Hardware makers sell hardware, not drivers. Why protect something you stand to make no money on. What's the worse that can happen? Could someone write a better driver than the hardware company? So? Am I going to refuse to buy a video card because I can download good drivers for it? What am I missing here?
The only think I can figure is that, say nVidAMD for example, is able to drop support for an outdated video card by ensuring that there are no drivers available for the latest OS, forcing consumers that use the new OS to purchase new cards. But if that were the case, these guys wouldn't be writing drivers that support five year old hardware for Windows7.
There is no "I disagree" mod for a reason. Flamebait, Troll, and Overrated are not substitutes.
The movement in this direction has already been creeping about. Big names like AMD/ATI are really doing a nice thing open sourcing their stuff. NVidia will be feeling even more pressure as time goes on and people continue hating them for not following in kind with ATI.
One of the problems which causes these closed drivers situations is that the chip makers contract their work to companies where the terms of their work ends up with some sort of copyright and other restrictions. (I don't know this first hand, just what I've heard... but like "Hey, I want you to write some drivers... and instead of paying you for your work directly, we'll give you a cut of sales! How's that sound?" or something like that... I don't know... it doesn't make much sense to me in the first place. People buy hardware. The software is only there to make it work with an OS.)
In any case, as far as standards and crap like that go? I don't think standards are much of an issue. As long as the software interfaces are documented, the driver interfaces between the hardware and the OS. The OS just talks to the driver and to the applications. That's all the "standards" I care about.
Killing software patents with fire.
Flexible bare-metal recovery for Linux/UNIX
Yes. It's a kind of bread you can smoke.
An example that leaves a particular bad taste in my mouth...
I bought a set of LCD shutter-glasses years ago. I had an nVidia card that had driver support for them. I got these babies, got the special nVidia driver, and I was blown away.
But soon I needed to upgrade my gfx card, and found nVidia no longer supported shutter-glass stereo on any of their new shiny cards. Weird right? All you need is software trickery.... but wait, yes.... Suddenly 3D LCD panels come out and nVidia simultaneously releases drivers that support them. And next thing you know, they have their own shutter glasses that cost way, way more than the ones I'd bought years before.
And still, there's no support for my set. Support that already existed.
My opinion: This is why hardware companies care about drivers, it lets them wrangle money out of people who'd like support for their products.
> no, yes, maybe (tagging beta)
Half an hour? You're doing it wrong.
You can install Bodhi from a thumb drive in about 10 minutes. There's even a video floating around Youtube of somebody installing it in a virtual machine in less than 10 minutes, from first boot to working installed desktop. When I installed it on my Dell ultraportable, everything worked out of the box, no configuration needed. (though to be fair, the Dell came with Ubuntu preinstalled, so it's hardly surprising)
Ubuntu can be done just about as quickly, in my experience. As long as you have a reasonably fast optical drive or are installing from a good quality thumb drive.
And nobody ever installs Windows, themselves, either, on a notebook.
Boy I hope that's sarcasm. Otherwise I fear I must question my own existence, as I've done just that many, many times.
Who in their right mind would leave the factory installation of windows on a notebook in the first place if they didn't have to? Why spend 2 hours cleaning all the adware bullshit off of it, searching the web to see what the hell half the start-up programs even are ("Gee, do I need kdjsdksjhdjsh.exe to run on startup? What about eroiuerrurrjkffl.exe???"), missing shit, and all of that, when you can spend 45 minutes doing a fresh install of Windows and then maybe another 45 minutes doing updates/driver installs and have a clean machine with all that bullshit removed from the get go?
Step one on any new notebook I buy is always a fresh install of windows. I don't play that "recovery disk" bullshit.
I know it used to be a lot more difficult in the past to find drivers and shit for notebooks, but it's really not that bad anymore. Certainly not in my own experiences.
The problem is a lot of hardware is heavily patented, and the patents cover the hardware-software combination.
A sound card would be the best example - you can have a basic sound card with open drivers (it's just a combination ADC and DAC on a board, after all). But then people want justification for their purchases, so you add in Dolby Headphone support to give you surround sound with headphones (patented, licensing fees to use). Or DTS/Dolby Digital encoders so people can get surround sound piped to their A/V receivers. Or HDMI audio injectors that support HDCP.
Ditto video cards - HDMI+HDCP is a spec that does not allow for open drivers. A lot of 3D technologies are patented, heavily.
Network cards - well the TCP offload egnines are considered "secret sauce" because a good TOE can ensure your host system can be full bandwidth and hardly take any CPU resources. And this can include onboard firmware for the onboard processors. LIkewise, WiFi is similar.
Nevermind software controlled parts of hardware that cannot be modified for compliance reasons.
Hell, half the hardware guys out there would kill if they can release the drivers as source and give it all away - less work for them to support (they can direct people to a community support page). Or just release the hardware and let the community write the damn driver for it.
Of course, there's also the irony in that Qualcomm supplies a lot of binary blobs for stuff using their processors... especially with Android.
Contrary to popular belief, businesses don't simply rationally maximize profit or ROI.
It might just take nothing more then that first hole in the dyke.
They prefer to be called "lesbians" now, you insensitive clod!
When you replace a proprietary driver with a free one, it stands to reason that it won't contain any code from the proprietary version.
It doesn't necessarily follow - the copyright holder can license his code however he likes. This is how dual-licensing projects roll. As long as they hold the copyright on all of their code, they could open-source all of it tomorrow.
But your other point is sound - there's a great deal of well thought out code in common in most of the OSS drivers in the kernel, and I suspect the binary blob drivers probably duplicate a lot of feature that can now go through the kernel routines instead.
Many years ago i was associated with Project UDI, the Uniform Driver Interface. The goal was to make a uniform ABI/API for device drivers. On Machines with the same hardware target (say, 32 bit x86) you would have binary compatibility. The same driver works on Solaris or Windows. For other platforms, they'd be at least source compatible. It worked in theory, and somewhat in practice - I think UnixWare shipped this as their native Device Driver Interface.
But you never heard of it. Part of it was the SCO/Caldera fiasco. 'Nuff said about that.
But part of it also was the fact that people had vested interests in this failing. Most famously, Stallman didn't like it. For now you could ship drivers without source for all i386 targets (not that having the normal Linux DDI prevented that before). But it was fun that I worked on something shipped in a commercial kernel, and also something that pissed off Stallman.
More importantly, the people who want this are necessarily in the weakest position. MS doesn't want this - everyone makes Windows drivers. They get nothing from it except lower exclusivity. (The fact that Gates and Stallman were on the same side of this should have given Stallman time to reflect). They'd never allow the UDI code to touch their kernel. One or two other big UNIX vendors feigned interest, but they had the same issue - they had exclusive (to UNIX) device drivers, and they'd lose exclusivity. Only Caldera used it. It was their project, and it helped their forked codebase - they had both UnixWare and OpenServer (very old) code bases they needed drivers for, and it made it an easier target for device makers.
None of the issues were tech issues, they all were people issues, which haven't gone away in the intervening years.
It took a year to get my brand-new netbook working correctly with Ubuntu.
First, the ath9k driver was written so that on hibernate it would set the sync to a random value. The fix was crtl-alt-t, sudo rmmod ath9k modprobe ath9k. This was fixed in 11.04, but required a patch until it was officially released.
Second, and this is really a driver issue, is that it required 3rd-party support to get the Fn keys working correctly. Apparently this has something to do with Windows automatically shutting down the Fn keys, which is something that Linux doesn't do.
Third, the elan touchpad (which was brand-new in 2010) didn't respond with a value on the kernel's magic knock list. It took a year of bug reports for the kernel dev team to accept that they had made an oversight (no, never a mistake!) and accept that new hardware would sometimes have different values than expected. Shocking. Anyway, I was using the touchpad as a PS/2 mouse for that whole time, while the Windows boot had full multi-touch. It's... it's not a recommended experience.
I also get about 3 hours less battery life in Ubuntu compared to Win7, but hey, 6 hours is still a lot of hours.
And every now and then when booting the speakers go into full volume screech mode. I'm not sure why that's a feature.
---
ECHELON is a government program to find words like bomb, jihad, plutonium, assassinate, and anarchy.
So the hardware vendors are incompetent and don't support their products in Linux, resulting in the experience sucking.... and it's the Linux community's fault. Got it.
Funny seeing this coming from Qualcomm. I've been working on the HP TouchPad Ubuntu port and would love to see open-source Adreno 220 drivers with X support, but none appear to exist. I can't take them seriously on asking other companies to kill proprietary drivers when their own drivers are closed and unavailable (even TI's SGX drivers are available as a binary package with SDK and installation instructions for your kernel).
It won't happen. A lot of device drivers utilize patented code which is rigorously protected by many software companies. The GPL is not suitable for those who want to maintain control over their code. Proprietary drivers are a necessary evil if one is using a FOSS operating system with specialized hardware.
I've crossed over from the land of Linux (on the desktop atleast) to Windows. I still have to look up PCI IDs and google them to figure out what the heck things are. Laptops are the worst by far.