Next Linux Kernel Due Early March

swandives writes "The Linux.conf.au is in full-swing in Wellington, New Zealand, and Computerworld Australia has an interview with Jon Corbet in the leadup to his Kernel Report. The latest kernel release is due early March and will include reversed-engineered drivers for Nvidia chipsets."

Comments

Early march?

[Anonymous Coward]

I expect it mid-February.

Re:Early march?

[Anonymous Coward]

That's nice.

Year of the linux desktop

[Anonymous Coward]

So this is the year of the linux desktop? Yeehaw kernel 2.7!

Re:Year of the linux desktop

[JWSmythe]

    Nope, just 2.6.33. Even less exciting is that 2.6.33-rc4 was available 5 days ago.

    This isn't news, but what should we expect of a late night update, eh?

Re:Year of the linux desktop

[thatkid_2002]

Its about 7pm in NZ you insensitive clod.

Re:Year of the linux desktop

[sjalexander]

yeah, but it's 7 PM tomorrow, so it doesn't count.

Re:Year of the linux desktop

[Anonymous Coward]

yeah, but it's 7 PM tomorrow, so it doesn't count.

As a New Zealander I would like to say that I am speaking to you from the future. I will make a prediction,

*holds-fortune-card-to-forehead*

Americans, tomorrow you will still not have a decent health care system and for some reason you're going to have Leno rather than Conan O'Brien.

All these things will come true. Hang your heads in shame, North America.

Re:Year of the linux desktop

[Anonymous Coward]

Not yet, I have a problem when I plug/unplug an USB keyboard (usually after several times) and the Xorg server will crash.

Re:Year of the linux desktop

[V!NCENT]

OK never had that... What I did have was when I blinked at my monitor when running Windows I had got BSOD...

Re:Year of the linux desktop

[Darth Sdlavrot]

This is Open Source.

It's up to you to fix it and send the fix to X.Org.

Re:Year of the linux desktop

[Anonymous Coward]

In fascist US of A, the truth hurts YOU.

Re:Year of the linux desktop

[Anonymous Coward]

One question... why does one plug/unplug a usb keyboard several times?

Re:Year of the linux desktop

[Sir_Lewk]

A 2.7 kernel would not be for general consumption, it'd be used for development during a transition to 2.8. I believe Linus has publicly stated that he has no intention of going to 2.8 anytime in the forseeable future though.

Re:Year of the linux desktop

[Anonymous Coward]

Any system that leaves ~15% of its population without any proper health insurance whatsoever is indeed indecent.

Canadians hop the border to cut the wait-times for optional or non-critical things - the waiting list for urgent cases is very short. I had knee surgery this past summer, hardly life-threatening, and I had to wait a week for it. No one dies waiting for care in Canada. How many Americans have died because they lack health care entirely?

It's sad and pathetic that American propaganda keeps people there thinking they are living in a great country with a high standard of living, while in reality they don't even meet some of the basic requirements of Western civilisation.

Re:Year of the linux desktop

[Anonymous Coward]

Any system that leaves ~15% of its population without any proper health insurance whatsoever is indeed indecent.

%25 of those are eligible for government provided healthcare, another third makes over 50K a year so can afford insurance and many are between coverage until they get a new job. But don't let facts and details get in the way of a good America health care bash.

I had knee surgery this past summer, hardly life-threatening, and I had to wait a week for it.

Your anecdote sure convinced me. Actually, not.

No one dies waiting for care in Canada.

That's a goddamned lie.

It's sad and pathetic that American propaganda keeps people there thinking they are living in a great country with a high standard of living, while in reality they don't even meet some of the basic requirements of Western civilisation.

Here's my anecdote. Any time I have a health issue, I call up the doctor to make an appointment, get seen promptly, provide my insurance details and call it a day. If it's something more serious, I get referred to a specialist where I do exactly the same thing. I don't really see how it could get any better. I suppose in the "civilized" Canadian doctors are educated at Hogwarts and get a special magic wand when they graduate to cure people with. That's what it would take for your experience to be superior to mine. So keep on living in your self-aggrandizing dream world.

Actually, I think I get it. Some people are born to be 6'4'' (1.93 meters), tall others are ruggedly handsome, some people are born into money and then, people like me are born in the United States. Don't be a hater. It's ugly.

Re:Year of the linux desktop

[kiwimate]

I have lived in both NZ and the US.

In the US, I get vast quantities of stuff (medicine, scanners, machines) thrown at me whenever I go to the doctor or hospital, but to get an appointment at the doctor takes a minimum of two weeks.

In NZ, I didn't get as much stuff thrown at me, but I still got what I needed, and I could see a doctor within a day or two.

In the US, I have a phenomenally good health care plan thanks to my employer, so whenever I have to go to the doctor or the hospital the amount I have to pay only costs between $5 and $20. For the benefit of non-US readers, a "phenomenally good health care plan" means I get a very wide range of coverage and I only pay about $100 a month. Prior to this job, I paid several hundred dollars a month, even with that employer kicking in a contribution towards my health care plan.

In NZ, just about everything was free. (Yes, I know I paid for it in taxes. Employers also paid for it in some of their company taxes.) Including my two week stay in hospital after a car crash, all the surgeries, physical therapy, etc. No health insurance or health care plan needed. If I wanted one, however, I could still have bought one and gotten private coverage. My brother worked for Telecom, for instance, so he had private health insurance through them.

Overall, even with my fantastic health care plan provided by my employer here, I still pay more per annum than I did in NZ.

And in NZ, I once had occasion to take a guest from the US, who was visiting for two weeks, to the hospital to get emergency treatment when she broke her foot. It was 100% free for her, too, thanks to the NZ taxpayer. No travel insurance was necessary. If the situations were reversed and I had to take a guest to a US hospital and they didn't have insurance, I don't even want to guess how much it'd cost.

Dislaimer: it's been ten years since I was in New Zealand. Things may have changed. If so, I'm sure someone will update me.

I'm not a hater; I've lived in the US for many years and like it here. There are some things about living here which I like better than in NZ, and there are some things I really miss about living in NZ that I think are better than in the US. No one country has the best of everything, and, from my personal experience, I happen to prefer the health care system in NZ.

Re:Year of the linux desktop

[__aaclcg7560]

Because you can't yank a broken joystick. *cough*

huh?

[gbelteshazzar]

the linux australia conference is in new zealand?

Re:huh?

[Anonymous Coward]

we annexed them a decade ago, but they haven't realized yet...

Re:huh?

[DuEyNZ]

New Zealand is just a state of Australia, didn't you know?

Re:huh?

[Anonymous Coward]

Haven't been to australia recently, have you? :P

Re:huh?

[Anonymous Coward]

It says so in our Constitution.

Re:huh?

[maharius]

Ah so thats why Prince William went to NZ.

Re:huh?

[QuantumG]

Yes, and? NZ is just an island of the continent of Australia. It'd be like having a Linux North America conference in Canada, or Hawaii. mmm.. Hawaii.

Re:huh?

[tonyr60]

Ahh, no. NZ is not part of continental Oz. Australasia is commonly used to group Oz and NZ (and sometimes other countries). Apart from mutually taking the piss out of each other there is little logical reason to group the two separated land masses.

Re:huh?

[Paua Fritter]

Australia, aka the "West Island" of NZ

Re:huh?

[dakameleon]

Except it's more like Linux.conf.us being held in Cuba. Just a little bit wrong.

Re:huh?

[Anonymous Coward]

The only time I want to go to the linux conference, they move it a two hour flight away to new zealand.

what the fuck idiot decides to have it there?

There is nothing in new zealand at least Australia has women and nice clubs for after. All new zealand has is rubber boots.

Re:huh?

[BrokenHalo]

All new zealand has is rubber boots.

And sheep. Never forget the sheep... ;-D

Re:huh?

[QuantumG]

Huh? Are you unaware of what a continent is? Get this, India (and Sri Lanka) are both in Asia! Amazing I know.

Re:huh?

[Anonymous Coward]

Sadly, no!

New Zealand is not on the same continental shelf and so is not part of the continent of Australia but is part of the submerged continent Zealandia.

Re:huh?

[rubycodez]

guess again, New Zealand is part of the continent Zealandia

it is NOT part of the continent of Australia, different shelf.

makes sense our schools gave up teaching geography and history, who needs that when we have blogs.

Re:huh?

[rubycodez]

New Zealand is part of a continent called Zealandia, not Australia. but by all means pull some more "facts" out of your ass, it's more interesting than reality.

http://en.wikipedia.org/wiki/Zealandia_(continent)

Re:huh?

[QuantumG]

Pfffffffffft... It's an urban myth made up by kiwis to make themselves feel special, pay no attention to it.

Re:huh?

[gardyloo]

Oh, this fighting is fun (and I have no horse in the race!). Note that even in the most generous listing of continents (comprising 7), New Zealand is NOT separated from Australia.

http://en.wikipedia.org/wiki/Continent#Number_of_continents

Re:huh?

[Anonymous Coward]

how about size? size isn't little. size is relative.

Re:huh?

[Anonymous Coward]

Must be Australian, unless you are just taking the puss...

Re:huh?

[master5o1]

New Zealand straddles the Australian continental plate and the pacific plate much like I was straddling your mum last night.

Re:huh?

[laptop006]

Yes.

Seriously you could put a bid in to have it in a field in Texas, and, if you're the best bid they'd give you the conf.

I was on the 2008 team, and putting on a conf to the level LCA does is a huge amount of work, so if you can bid and do it you have a real chance.

Of course not being in Australia (or New Zealand) makes it very expensive for those people to attend, so unless you can find a sponsor for flights you really aren't likely win for LCA2012 in Texas.

Re:huh?

[Anonymous Coward]

that's what the rubber boots are for... ;)

Re:huh?

[Joebert]

Was she wearing her strap-on while you straddled her ?

Re:huh?

[Stormwatch]

Get this, India (and Sri Lanka) are both in Asia!

Huh? Are you unaware of how parentheses are used?

Re:huh?

[nomadic]

makes sense our schools gave up teaching geography and history, who needs that when we have blogs.

What are you talking about? The overwhelming academic consensus is there is no such continent as "Zealandia."

Re:huh?

[EsbenMoseHansen]

guess again, New Zealand is part of the continent Zealandia

it is NOT part of the continent of Australia, different shelf.

makes sense our schools gave up teaching geography and history, who needs that when we have blogs.

Hm, according to the Ultimate Truth (aka Wikipedia), it seems New Zealand is between plates.

Re:huh?

[Anonymous Coward]

The sheep are electric? It's like an android's dream come true!

Re:huh?

[rubycodez]

actually no, after more proper definitions of a continent, your article says "some geographers take Australia, New Zealand and all the islands of Oceania (or sometimes Australasia) to be equivalent to a continent, allowing the entire land surface of the Earth to be divided into continents or quasi-continents.[11]"

so some silly people construct what we would call a "continent equivalent", which has no scientific meaning like the true geological continent.

Re:huh?

[rubycodez]

well heck, all it really is is a sand bar that partially surfaces during ice ages. Kiwis being part of the biofilm crud that accumulates then. no offense to actual biofilms.

Re:huh?

[Trogre]

It could be worse, they could have put it in Perth. Oh wait, they already did.

Re:huh?

[Trogre]

I hope you put all the soil back when you were done.

Thank goodness for those drivers

[Arancaytar]

I have such a chipset and I've been cursing NVIDIA on a regular basis. After updating to any new kernel, I must boot into no-X mode, then run the proprietary driver installer.

Re:Thank goodness for those drivers

[Anonymous Coward]

Your distribution does not provide mechanism such as kmod on Fedora or dkms (?) on debian?

Re:Thank goodness for those drivers

[rastilin]

So, Nvidia writes drivers for your system, and those drivers work. What's the problem? This is hardly a new situation, so presumably you knew this when you bought your Nvidia chipset.

Re:Thank goodness for those drivers

[amRadioHed]

You'll still need to do that if you want 3D support. nouveau is replacing the old nv driver, but it's not ready to replace the proprietary driver.

Re:Thank goodness for those drivers

[QuantumG]

Please see the last NVIDIA linux drivers story.. for fuck sake.. it's only been a month.

      http://linux.slashdot.org/article.pl?sid=09/12/11/1556237

Go argue with last month.

 

Re:Thank goodness for those drivers

[l2718]

I have such a chipset and I've been cursing NVIDIA on a regular basis.

You must be new to this "Linux" thing. That your Hardware OEM is providing Linux drivers at all is highly unusual. That the drivers are effective is astounding -- that the installer provided the drivers is rudimentary is not worth complaining over. In any case if you really mind I'm sure you can write a replacement installer.

Re:Thank goodness for those drivers

[timbo234]

I haven't had to do that for a few years now, modern distributions (Mandriva and OpenSuse for eg.) automatically setup DKMS or use some other mechanism to update the NVIDIA drivers automagically when a new kernel boots.

That said however it'd be better to have a working NVIDIA driver in the kernel, as these solutions are a bit hacky and potentially an open-source driver would have a faster pace of development (instead of being the poor cousin to the Windows drivers in NVIDIA's internal development priorities).

Re:Thank goodness for those drivers

[Brebs]

This is what startup scripts are for: http://www.fedoraforum.org/forum/showthread.php?p=1047697#post1047697

Re:Thank goodness for those drivers

[Anonymous Coward]

yeah. thank goodness

now we have a new set of buggy drivers to curse at...excellent

Re:Thank goodness for those drivers

[BrokenHalo]

So, Nvidia writes drivers for your system, and those drivers work. What's the problem?

Indeed, I have no problem with that. I've been using Linux or long enough to remember having to spend a lot of time getting around issues of hardware compatibility. Nvidia was in there quite early on providing good drivers for its chipsets at a time when just about every other manufacturer just shrugged its shoulders and told us to "Fuck off, We don't support Linux."

That alone has promoted a lot of goodwill as far as I'm concerned, and so nVidia chipsets are right at the top of my preferred brands list. So I get very tired of hearing people badmouthing nVidia without giving an adequate reason why.

Re:Thank goodness for those drivers

[Kjella]

|. Switch to nv driver
2. Upgrade kernel
3. Switch to nvidia driver

Not that I've ever done that, I use a distro that packages the kernel and binary driver for me. Maybe if you want a userfriendly solution you should get one instead of trying to do everything manually, then complain about having to do everything manually?

Nouveau will be much closer to nv than nvidia in pretty much everything. They got no specs, and even with specs writing a good open source 3D driver is tough, as AMD has shown us. So expect no 3D acceleration, no video acceleration, no nothing. The only thing you get is 2D modesetting in the kernel instead of xorg, big whoop. If you don't use KMS for anything else, the only thing it'll do it make your boot a little prettier.

Re:Thank goodness for those drivers

[Rogerborg]

You realise that you're posting to a site where the "editors" regularly dupe each others stories while they're still on the front page, right?

Re:Thank goodness for those drivers

[Xeleema]

Hm, to quote a near-forgotten troll; "You Do It Wrong"

ProTip: Hit linuxquestions.org and post a detailed outline of your problem. Be sure to include things like versions, names of distributions, and how many servers^H^H^H^H^H^H^H desktops you're having this issue on.

I'm sure you're not running X on bootup on a server, right?

Re:Thank goodness for those drivers

[Culture20]

So, Nvidia writes drivers for your system, and those drivers work. What's the problem?

Indeed, I have no problem with that. I've been using Linux or long enough to remember having to spend a lot of time getting around issues of hardware compatibility. Nvidia was in there quite early on providing good drivers for its chipsets at a time when just about every other manufacturer just shrugged its shoulders and told us to "Fuck off, We don't support Linux." That alone has promoted a lot of goodwill as far as I'm concerned, and so nVidia chipsets are right at the top of my preferred brands list. So I get very tired of hearing people badmouthing nVidia without giving an adequate reason why.

Goodwill Schmoodwill. This is business. For quite some time, the only way I've been able to easily install Ubuntu on several of my Nvidia machines has been by swapping out the graphics card(s) for ATI, installing the OS and nvidia drivers, then installing the Nvidia cards again. True, this is an Ubuntu issue, since they insist on a GUI install only (sorry, but the alternate CD is a pain, at least use curses to emulate a GUI before making Mom and Pop use Debian), and they don't include the nvidia driver on the CD. The nv and vesa drivers are both broken for lots of nvidia cards (nv causes green verical lines, and vesa just crashes X continuously.
If Nvidia had created a usable neutered (2D) OSS driver that Just Worked (TM) with their cards, a la ATI/IBM, then I'd still be suggesting their cards for Linux newbies like I did back in the Aughties. Instead, I've been suggesting IBM first, ATI next, and Nvidia only for experienced folk who need superior OpenGL cards.

Re:Thank goodness for those drivers

[Anonymous Coward]

So, linux gets reverse-engineered NVIDIA chipset drivers just in time to see NVIDIA pull out of the chipset market... I guess it is useful for those using older hardware, as there are a ton of NVIDIA chipset motherboards out there, but at the moment NVIDIA chipset business is pretty much dead.

Re:Thank goodness for those drivers

[thue]

Except that the two other major Graphics providers, Intel and AMD, both give the Linux community far better support than NVIDIA. Intel is writing excellent well-integrated open source drivers themselves, which AMD is providing full specs, which has allowed others to write drivers. AMD's making the specs available is far better Linux support than NVIDIA making closed source drivers available.

NVIDIA has provided neither open source drivers, firmware, nor specs. So the open source developers have to resort to reverse engineer the drivers. And to make all kinds of jumping through hoops to use the firmware, which NVIDIA has not allowed to be redistributed in binary form.

So I think we have every right to criticize NVIDIA when comparing to the marked at large. They are doing a horrible job at supporting Linux.

Re:Thank goodness for those drivers

[drinkypoo]

Except that the two other major Graphics providers, Intel and AMD, both give the Linux community far better support than NVIDIA. Intel is writing excellent well-integrated open source drivers themselves

Has AMD actually caught up yet, or are they still generations behind on releasing specs? And I have an EEE 701 and every time intel tweaks the graphics drivers they break something in the janky intel video chipset. You have to have the newest and greatest intel GPU (which is still shitty) to have good driver support. 9xx series is poo, and worked better with the old drivers.

nVidia drivers work, as long as you have a supported card. You have to check the driver support before you buy. But since that's true of everything, it's not a problem. Meanwhile nVidia has working video acceleration.

Re:Thank goodness for those drivers

[jaminJay]

I started with nVidia on my first Linux-powered box and, at most, was a couple of days out between Fedora release and stable nVidia driver release (one went to a week before I found the 'ignore ABI' switch on a forum).

That card eventually died and I replaced it with an ATI card, supposedly superior. I never got it to work with Compiz. Ever. I struggled to get the open source drivers to do dual-head mode. I even tried Win7 RC1 in the hope that that would at least allow me to benefit from my hardware investment. Nope.

Needless to say, my new box has an nVidia card and RPMFusion-packaged closed-source drivers (which I just check are out before updating the kernel, unless it's critical. I'll invest in nVidia because they invest in their customers.

Re:Thank goodness for those drivers

[Anonymous Coward]

Just FYI, I believe vesa driver developers would be interested in finding out which nvidia graphics cards their drivers fail to work with.

Re:Thank goodness for those drivers

[Anonymous Coward]

I don't consider having to run
"nvidia-settings -a PixmapCache=0"
everytime my computer boots as meaning that the driver works.

Re:Thank goodness for those drivers

[TeknoHog]

If Nvidia had created a usable neutered (2D) OSS driver that Just Worked (TM) with their cards, a la ATI/IBM, then I'd still be suggesting their cards for Linux newbies like I did back in the Aughties. Instead, I've been suggesting IBM first, ATI next, and Nvidia only for experienced folk who need superior OpenGL cards.

I've never heard of IBM graphics cards. Do you mean Intel or something else? I personally recommend Intel graphics for Linux users, since the only drivers they make for Linux are opensource, with full 3D capabilities ("full" as in what the chip is capable of).

Re:Thank goodness for those drivers

[Skater]

Yeah, I buy NVidia cards because I know they are supported under Linux. My one experience with an ATI card was not very good (although it was in a laptop). However, I hope the kernel dual-head support is better than NVidia's for my current card - mine will shut off the primary monitor (which gives an "out of range" error message) from time to time. VERY annoying, and I know it's not the monitor because I've had the same problem with two totally different monitors (two CRTs vs. two LCDs, running at different resolutions and clock speeds).

Re:Thank goodness for those drivers

[jedidiah]

nv drivers broken for a lot of cards? Which ones would these be? They would not happen to be perchance cards that any Windows users would consider laughingly out of date?

I think you will find legions of users that think you are full of sh*t and especially full of sh*t for adding Ubuntu to your rant.

Re:Thank goodness for those drivers

[maxume]

If you had more quantom, you too would be more awesome.

Re:Thank goodness for those drivers

[Anonymous Coward]

You can still switch to single user mode in Ubuntu, thus being able to install the video driver.
Switching the video card to do this is asinine....

Re:Thank goodness for those drivers

[Anonymous Coward]

That sounds insane, for installing the system your Nvidia card will work perfectly fine using the vesafb or nv driver. After you have the system up and running you can install the proprietary Nvidia driver via the 'restricted drivers' tool from within the GUI. You don't even need to use a console to get it working.

Do you seriously swap you video card for an Ubuntu install or are you just trolling?

Re:Thank goodness for those drivers

[Zantetsuken]

do you use Ubuntu? If so and you didn't know about it already, use: NVidia Launchpad PPA

Re:Thank goodness for those drivers

[Anonymous Coward]

I only wish this were 100% true concerning Intel. Clearly you haven't yet had the privilege of attempting to obtain GMA500 linux driver support . Yes, I'm aware that they purchased this graphics engine from someone else, but it doesn't excuse the horrible lack of vendor support since it ships with the INTEL logo. The poulsbo driver work-arounds tend to break with new kernel releases despite a lot of terrific work done by some folks working the Ubuntu PPA repositories, so candidly I'm just fed up.

If Intel can obtain and provide such documentation for producing an open source GMA500 driver, then I will be truly impressed. Until then, thousands of us await and call down a very dark, quiet vendor tunnel for hardware support.

.

Re:Thank goodness for those drivers

[Culture20]

I've never heard of IBM graphics cards. Do you mean Intel or something else?

Wow, I shouldn't post just after waking up. Yep, I meant Intel.

Re:Thank goodness for those drivers

[webheaded]

I'm surprised no one mentioned this but um...you can just reinstall the driver before you reboot after installing a new kernel. The symbolic links should be changed pretty much immediately and after that is done, the installer throws the modules where they're supposed to be for the new stuff. I've been doing this forever. I mean don't get me wrong, it's silly that it has to be done and most nubs are going to "OMG WTF" when their PC reboots with no nvidia driver, but at least the damn driver is there, right? Anyway, not sure why no one has mentioned this yet. I've been doing that for Gentoo since the dawn of existence and I still do it in Arch just in case. You don't need to boot into "No X mode" or whatever if you know what you're doing, unless your distro is somehow completely retarded.

Re:Thank goodness for those drivers

[Culture20]

nv drivers broken for a lot of cards? Which ones would these be? They would not happen to be perchance cards that any Windows users would consider laughingly out of date? I think you will find legions of users that think you are full of sh*t and especially full of sh*t for adding Ubuntu to your rant.

Maybe laughingly out of date by Windows standards, but that usually means gamers. A geforce 6600 is still a good card, even if it's old (not as old as my TNT cards that still work). And if it's old, shouldn't it have better drivers? Just sayin'.
As for the Ubuntu rant, perhaps you should do a search for "vertical lines" on ubuntuforums.org?
http://www.google.com/search?hl=en&q=%22vertical+lines%22+green+site%3Aubuntuforums.org+nvidia&aq=f&oq=&aqi=

Re:Thank goodness for those drivers

[Hurricane78]

What you two completely ignored, probably with your minds fully in the windows world:

A installer?? For Linux???
How sick is that?

Linux has package managers. The package belongs into the repository, just like any other software.
Installers are Windows speak.

And those packages are a source package (e.g. source or binaries in tar.bz2), and a description file. Sometimes wrapped in one (e.g. RPM.)
This ensures comfortable installation, uninstallation, dependency management, updates/patches, etc.

Installers... silly, silly, silly...
And whining about “having to use” the CLI. As if that was a bad thing...
Oh boy, when minds are constrained, it’s hard to free them...

Re:Thank goodness for those drivers

[Anonymous Coward]

Intel is writing excellent well-integrated open source drivers themselves

"Excellent" is a quality judgement I would not use with regards to the Intel Linux drivers.

Re:Thank goodness for those drivers

[Culture20]

You can still switch to single user mode in Ubuntu, thus being able to install the video driver. Switching the video card to do this is asinine....

No, you can't do that with the install CD. Really. Using vesa causes xorg to reload once a second, and you can't type anything. Using nv causes half of the screen to disappear in the virtual terminal. Believe me, I've been using Linux for more than a decade on x86, ppc, sparc, others, with tons of distros (one of my jobs required me to evaluate as many popular ones as possible on a regular basis). I've had a lot of experience getting weird hardware working with Linux and X/xorg. Switching graphics cards is a simple and elegant solution for these installs. A quick edit of the xorg.conf file, and I was done. But, that's not something that normal users can handle, and Ubuntu is targeted at normal users.

Re:Thank goodness for those drivers

[BrokenHalo]

I believe vesa driver developers would be interested in finding out which nvidia graphics cards their drivers fail to work with.

I've used the framebuffer with every nVidia card since the Riva series, and it has worked impeccably with every one. But it's this remark the parent made that puzzles me:

For quite some time, the only way I've been able to easily install Ubuntu on several of my Nvidia machines has been by swapping out the graphics card(s) for ATI, installing the OS and nvidia drivers, then installing the Nvidia cards again.

Ubuntu isn't my distro of choice (I keep trying it, getting annoyed and returning to Arch), but failure to support nVidia chipsets isn't a fault that I have observed. One might make the comment (giving the parent the benefit of the doubt and assuming he isn't trolling) that if his distro gives that much trouble, then he *should* consider using using another. You should never need to pull cards to get it to work.

Re:Thank goodness for those drivers

[Culture20]

That sounds insane, for installing the system your Nvidia card will work perfectly fine using the vesafb or nv driver. After you have the system up and running you can install the proprietary Nvidia driver via the 'restricted drivers' tool from within the GUI. You don't even need to use a console to get it working.

Do you seriously swap you video card for an Ubuntu install or are you just trolling?

If I were Trolling, I think I'd be doing it with an account with less Karma, or AC. I seriously have to switch out nvidia cards on a regular basis to do an Ubuntu install, and no, nv and vesa do _NOT_ work. nv driver causes vertical green lines, and vesa driver causes the xorg server to reload once a second (because Ubuntu is stupid and decided to make xorg not fail permanently and gracefully). With xorg reloading once a second, it's impossible to get any editing done in the virt-term. Without an ATI or Intel(thanks to the poster who corrected me on my brainfart) card, I can't install the base OS. My laptop is a different story, and it has a mobile nvidia card. Works fine with the nv driver. Probably works okay with vesa.

Re:Thank goodness for those drivers

[Anonymous Coward]

Yeah, "Far better support" from Intel.

My Vaio P is sitting on my desk unused, waiting for a GMA500 driver...

INTEL SUCKS!

(sorry for this little rant)

Re:Thank goodness for those drivers

[Culture20]

I love Ubuntu on most of my machines, and for a while, it was my job to test out different distros. If it was gentoo giving me these problems, I'd shrug my shoulders and post a bug. Ubuntu's supposed to be easy, and ubuntuforums.org is littered with requests to make the install CD not fail like this. The weird part is Fedora and other Live CDs work fine with vesa, so it's Ubtuntu's implementation that fails.

Re:Thank goodness for those drivers

[Anonymous Coward]

AMD's making the specs available is far better Linux support than NVIDIA making closed source drivers available.

Bingo. Specs are the real deal; binaries are not. Specs are what we've been begging for all these years.

For those who doesn't have much experience with linux, the reason we prefer open source software on our linux systems is because it plays nice and follows the rules. Closed source does not play nice on open source systems, even when it works. I'm not talking about legal matters; I'm talking about system administration. Closed source makes up its own rules. Standards and conventions be damned. Closed source is the black sheep that insists on doing things its own wrong way. It requires special attention. It demands special exceptions and privileges. It makes life more difficult and annoying, and this is why we consider it a temporary kludge rather than a permanent solution.

Re:Thank goodness for those drivers

[jedidiah]

Perhaps you should use your own argument rather than poaching someone else's.

I've personally used nvidia cards from the 6150 to the 9400 and taken good advantage
of the features that don't exist in the open drivers but are well supported in the
proprietary ones. I also have firsthand experiences with some of the alternatives.

Re:Thank goodness for those drivers

[Arancaytar]

Using Ubuntu 9.10 here, which seems not to do that automatically...

Re:Thank goodness for those drivers

[Anonymous Coward]

It's Ubuntu 9.10, retard.

Re:Thank goodness for those drivers

[cynyr]

been installing gentoo with nvidia cards for ohhh 5+ yars now, no issues. Sounds like Ubuntu isn't packaging things well, or it could be that whole "no non-free" thing they have going. Also the NV driver should work enough for getting the real driver installed.

Re:Thank goodness for those drivers

[cynyr]

57XX and 58xx specs out with working open drivers? Thought not.... Hmm that GTX295 nvidia card would just work in my system after opening the case and removing my 8600GT. And that's the difference.

Re:Thank goodness for those drivers

[cynyr]

my distro of choice(gentoo) has it inside of the package manager. they could just be running the setup for me for all i know( i should look at that), but it works almost 100% of the time.

Re:Thank goodness for those drivers

[MSG]

So I get very tired of hearing people badmouthing nVidia without giving an adequate reason why.

Here's mine: Times have changed. The Free Software community has successfully convinced the large majority of those other vendors to support Free Software properly by releasing specifications. Once upon a time, NVidia's support was "good" relative to other vendors (but, to be clear, no good in terms of the Free Software community's goals), but today better support is available from the other vendors. Since NVidia's support for Linux and the Free Software community's goals is less good then their competitors, it is absolutely appropriate for users to say so.

Re:Thank goodness for those drivers

[Count Fenring]

While you can still reboot no-X (for example, by passing options to GRUB when booting - not the easy way, but the first that came to mind ;-), I do have to agree, the failure to accept Xorg failures is my least favorite decision made in Ubuntu.

Hell, you can't even manually kill KDM anymore.

I've never had your problems with the nv drivers, though - they've been universally solid across several cards, if slow.

Re:Thank goodness for those drivers

[Anonymous Coward]

Intel giving far better support than NVIDIA? What alternate universe is this happening in? check freedesktop's bug tracker and search for reports on intel drivers, see the ones marked CRI(tical) in red? some over 3 years old? those are bugs that happen to hard lock linux (thank god for magic sysrq!), while switching from X to VTs, running GL apps (or screensavers :) ), suspending and resuming, while stuff like that _never_ happens with the NVIDIA provided blob on my hardware! (gma950, x3100 and x4500 vs FX5200,6800,7300,8800,9600 cards).
Intel's current "stable" drivers lock Xorg with
[drm:i915_gem_execbuffer] *ERROR* Execbuf while wedged
[drm:i915_gem_idle] *ERROR* hardware wedged
while the last "stable" ones didn't, great quality control from them huh.

Re:Thank goodness for those drivers

[l2718]

Having been a unix sysadmin for 15 years, I'm quite aware of package managers. Unfortunately not all OEM drivers come nicely packaged, so you have to install them by hand. Yes, I know I could make an RPM file around the driver, but it's far simpler to write a shellscript that automates the tasks the OP wanted (shut down X, shut down the display, replace kernel modules and X drivers, restart display, restart X). And yes, I call such a script an "installer". It doesn't display fancy graphics, but it sure does install the driver.

Re:Thank goodness for those drivers

[ikekrull]

Theres another side to this - if you have ever tried to work with 3D apps on Linux, free or commercial - Blender, Maya or written your own OpenGL apps, and wanted support for the standards and good performance, you would realise that NVidia is your only choice. Compared to their commercial rivals, and the open source community, they do a *stellar* job at supporting Linux.

Every other manufacturer has provided such piss-poor reliability and/or performance under Linux, they just aren't an option.

I think its great that AMD docs and lots of hard work by the Xorg and driver coders mean that radeon drivers are getting to the point where they challenge NVidias status in this area, but for the last 5 years, AMD/ATI were next to useless on Linux for serious work, and Intel graphics weren't (and still aren't) an option where anything even remotely current in terms of OpenGL API usage (e.g. GLSL shaders) are concerned.

The Open Source community has done an awful job of architecting their graphics stack, with no foresight, planning or consistency across drivers. Thats not a bash, thats the natural result of open source evolution, and why they're rearchitecting it.

Now this is being reworked, we're seeing massive churn and widespread breakage. NVidia saw this coming and wrote their drivers to bypass this mess. Many of the design decisions taken by the Xorg guys are very much influenced by how NVidia handles things.

Intel, supposedly the paragon of openness and open source, managed to show a massive performance regression in the kernel and X.org revisions prior to the current ones, and their latest 'Poulsbo' chipsets have no documentation, and no open source drivers. Intel's support for these cards on Linux is way worse than NVidia. Theyre also walking away from any open source OSes except Linux by relying on Linux-only kernel mode setting.

AMD/ATI continue to release fglrx drivers that are plagued with bugs, refuse to release documentation of current products, and have 2D performance that is so abysmal it makes the VESA framebuffer look pretty good in comparison. AMD/ATI open source drivers (while improving greatly and probably a good option today for people who don't really need full OpenGL coverage,) are very much a work in progress, incapable of running even moderately advanced OpenGL apps, and they too are dumping any support for non-Linux open source OSes.

As a 3D developer, I can't rely on anything but NVidia to work, and stay working across distro upgrades. If thats the definition of 'horrible job at supporting Linux', i think you need your head read. There just isn't anything else that is usable for professional or semi-professional 3D work on Linux.

I am extremely grateful to NVidia for enabling any kind of consistent 3D support on Linux while everyone else, commercial or open source, struggles to catch up.

Re:Thank goodness for those drivers

[Sancho]

Last I checked, the GMA500 didn't have open-source drivers. Of course, I believe that this is a chip which Intel rebrands, but it does mean that you can't simply trust that if it says "Intel graphics" it will be well-supported.

Re:Thank goodness for those drivers

[timbo234]

I don't use Ubuntu so can't say for sure but it looks like they do have packages for it, so I'd think if you used those packages it would keep current with the kernel (that seems to be how the Opensuse Nvidia packages work):
https://help.ubuntu.com/community/BinaryDriverHowto/Nvidia

They even recommend not to install manually, probably for this very reason:
This is not the recommended way to install the NVIDIA drivers - please see BinaryDriverHowto/Nvidia for the supported method.
https://help.ubuntu.com/community/NvidiaManual

Re:Thank goodness for those drivers

[Culture20]

Perhaps you should use your own argument rather than poaching someone else's.

Wow, you're an asshole. I told you my experience, and you called me a liar. I showed you corroborating stories and you accuse me of plagiarism or something. I'm using GF6600GTs. Go buy one and try it with Ubuntu. Try using vesa with the live CD.

Re:Thank goodness for those drivers

[Randle_Revar]

http://www.x.org/docs/AMD/

r700 specs came out some months after the chips did, the same is expected for r800 (should be soon now).

As for the Intel GMA500, its driver is shit and always will be, because is it actually a PowerVR mobile chip.

The other modern Intel chips are underpowered, but not really that bad (the old ones had bugs and serious misfeatures).

Re:Thank goodness for those drivers

[CAIMLAS]

Except that the two other major Graphics providers, Intel and AMD, both give the Linux community far better support than NVIDIA. Intel is writing excellent well-integrated open source drivers themselves, which AMD is providing full specs, which has allowed others to write drivers. AMD's making the specs available is far better Linux support than NVIDIA making closed source drivers available.

Yet, the Intel video cards are such crap it hardly matters, and the ATI support barely even works for most chipsets commonly available (such as anything you'd find on a motherboard).

Nvidia is, at least, doing something right for Linux. The others are trying to get something right.

Re:Thank goodness for those drivers

[drinkypoo]

r700 specs came out some months after the chips did, the same is expected for r800 (should be soon now).

A simple "no" would have sufficed.

As for the Intel GMA500, its driver is shit and always will be, because is it actually a PowerVR mobile chip.

The driver USED to work. for example, rotation used to work. Now I just get display trashing. I don't expect it to be fast, I just expect it to work, like it used to, before intel likely fried it.

Re:Thank goodness for those drivers

[bill_mcgonigle]

So, Nvidia writes drivers for your system, and those drivers work. What's the problem?

They blow in several directions.

I had an nVidia card in my MythTV box for the TV-out. I upgraded my motherboard because the Intel mobo blew in other ways (it's fine in a server now, doing mundane server things, but don't try to push 'fancy' PCI cards with it). The new mobo had an nVidia graphics chip built in.

Well, it turns out that the nVidia card I bought the year before was now a 'legacy' card. It has a separate driver. That driver can't be on the same machine with the new driver that's needed for the mobo. So, I got a new nVidia TV-out card that is compatible with the mobo chipset. Oh, well, they haven't bothered implementing TVOverscan for the 'new' driver yet, that's only available in the legacy driver, so my MythTV menus are off the screen. Etc.

This kind of stuff is laughed at in the open source community. Now, granted, nouveaux doesn't have TVOut support, yet, but when it does, it won't have random conflicts with other drivers.

With the next kernel I should be able to go back to the old TV out card and legacy driver and run the open source driver on the internal display. Hurray, for getting back to square one.

Re:Thank goodness for those drivers

[Anonymous Coward]

The default nv drivers work with every Nvidia card I have ever used with only minor problems, i.e wrong resolution/refresh rates, otherwise completely usable for Linux installations.

  You must be doing something wrong, or you have unusual / mis-configured hardware.

Reverse engineered nVidia drivers?

[Korin43]

By that vague statement do they mean that nouvea will be included or is someone else making yet another set of nvidia drivers? (nv is from nVidia right?)

Re:Reverse engineered nVidia drivers?

[QuantumG]

yes, Nouveau.. its referring to a previous Slashdot story late last year:

    http://linux.slashdot.org/article.pl?sid=09/12/11/1556237

And yes, that link could have been supplied, but that would require some sort of editing.

Re:Reverse engineered nVidia drivers?

[Ibn al-Hazardous]

Noveau is included in staging (which means that it comes with the kernel, but is not considered stable). Nv isn't a kernel driver at all, but merely an X.org driver from nvidia. Though the noveau kernel and X.org drivers are more fully featured than the nv offering, they still don't support 3D-acceleration very well.

Re:Reverse engineered nVidia drivers?

[XanC]

"nv" is the current nvidia kernel driver. "nvidia" is the official, proprietary driver from Nvidia.

Re:Reverse engineered nVidia drivers?

[tyrione]

"nv" is the current nvidia kernel driver. "nvidia" is the official, proprietary driver from Nvidia.

Correction. nv is the current reverse engineered driver provided by Xorg.

nvidia is the official, Nvidia corporation driver to support Xorg systems

Re:Reverse engineered nVidia drivers?

[Anonymous Coward]

nv is not reverse engineered -- it is a purposefully limited driver written by nvidia. nouveau is reverse-engineered from scratch, on the other hand. nvidia is the full-featured proprietary driver. nouveau will not replace nvidia in terms of functionality but it is already better in some respects than nv and is also free software

Re:Reverse engineered nVidia drivers?

[Anonymous Coward]

QuantumG for president! Thanks for the explanation. God how much slashdot editors suck.

I've been happily using the open source nv driver and eagerly wait for nouveau as I hear it's 2D performance to be superior and that we may get 3D at some point. I like my freedom.

Re:Reverse engineered nVidia drivers?

[V!NCENT]

It's a reverse engineered driver, but instead of copying it, the devs went to convert it to Gallium3D driver architecture. There are two components. One is the Direct Rendering Manager component (lives in the kernel) and the second one is in Mesa (the free OpenGL implementation) as a state tracker.

A Gallium3D driver is making the graphics card visible to the system by means of an API. On top of that API features can be written (like OpenGL, OpenCL, vector graphics acceleration, etc) which are called state trackers. Mesa is now such a state tracker.

When a OpenGL 3.x state tracker has been written, for example, it will work on all graphics cards that have a Gallium3D driver. So now all graphics drivers in Linux will get the same feature set. When sobody writes a feature for his ATI card, you as an nVidia users will also get to have that feature and vise versa... ;)

Re:Reverse engineered nVidia drivers?

[Anonymous Coward]

Correction. nv is not reverse engineered. It was written by nvidia. It even used to be the official driver before nvidia went proprietary.

Re:Reverse engineered nVidia drivers?

[Randle_Revar]

no it isn't.

nv - open source, but obfuscated, bare-bones 2d Xorg driver from Nvidia
Nvidia - closed source 2d and 3d driver for Xorg and Mesa (classic) from Nvidia
Nouveau - open source, reverse engineered driver for the kernel/Xorg (KMS) and Mesa (Gallium3d) from the Nouveau Community

It's official

[Laser_iCE]

According to sources in the US (slashdot: http://linux.slashdot.org/article.pl?sid=10/01/18/0257232), Australia has finally taken over New Zealand in a bloodless coup.

Re:It's official

[tonyr60]

I think you will find that it is the other way around. We even have Royalty here to endorse addition of the West Island to the kingdom of New Zealand.

Re:It's official

[MichaelSmith]

More of a sand bank than a proper island. The average chunk of NZ is, what? 10cm across? Over here you are lucky to find grains > 1mm.

But we are slowly winning. After our circuit around south island in 2008 my wife and son insisted on bringing back five or ten kilos of "interesting rocks". Customs in Melbourne nearly had a fit. Another million years and the top metre of NZ will be features in Australian back yards.

Re:It's official

[blind monkey 3]

It's true, NZ keep saying it's all about bringing us democracy but everyone knows they're after our sheep!

Re:It's official

[cptnapalm]

Considering what New Zealand women are like, can you blame them?

Dtrace for Linux?

[fibrewire]

Tell me more about this dynamic ftrace. Are there any "how-to" basic scripts to fire off a SNMP trap when ftrace picks up something of importance? It's nice for debugging, but more importantly to tie this into some network monitoring system like Nagios to be used for clustering and high availability systems. This could easily be integrated to prevent runaway virtual machines, and actually see whats robbing a system of CPU cycles - perfect for performance tuning a VM stack.

no such problem with Fedora

[Errtu76]

I'm running Fedora (12) and with the rpmfusion(-nonfree*) repository added, i don't need to run nvidia's installer. I just update my entire system, including the nvidia driver. If you're fed up by your way of updating, give Fedora a chance.

Re:no such problem with Fedora

[TheEnlightenedOne]

Simular repository for ubuntu intrepid(with VDPAU support): http://www.avenard.org/files/ubuntu-repos/

Will the kernel ever get to 3?

[petes_PoV]

Most releases seem to be minor improvements: a few bug fixes, re-porting to another architecture, some new drivers and tweaks to (or reimplementations of) existing features such as VM or filesystems.

Are we ever going to see major new features (along the lines of the USB implementation, or SMP), or a major re-think? Or is this basically as good as it will ever get?

It does appear to me that all the kernel is doing these days is mimicking the features and support found in "other" operating systems - rather than pushing the boundaries of innovation and novelty, itself.It would be a shame if Linux just fell into line and became a follower in a world of twisty little O/S's, all the same rather than producing some killer features, unique to it's implementation, that made people WANT to run Linux on their desktops and enterprise systems.

Re:Will the kernel ever get to 3?

[iggymanz]

mimicking all those other OS that scale from an embedded device to a supercomputer, or that run on more than a dozen architectures? let me count them.....uuuhhhhmmmmmmmm..........

Re:Will the kernel ever get to 3?

[vadim_t]

Are we ever going to see major new features (along the lines of the USB implementation, or SMP), or a major re-think? Or is this basically as good as it will ever get?

USB and SMP are things the kernel implemented, but weren't created inside it. The kernel can't add implementation for a bus that doesn't exist, so it's not going to get more things like that, unless new standards get created.

But, new things get added all the time, just watch the kernel reports at LWN.

Re:Will the kernel ever get to 3?

[Zero__Kelvin]

"Are we ever going to see major new features (along the lines of the USB implementation, or SMP), or a major re-think?"

Sure. Just get a time machine and you can go back to the day before Linux was the first to implement USB 3 ;-) As far as a "major re-think", the purpose of thinking things through seriously and thoroughly in the first place (before diving in) is so that you won't have to do major re-think. Major re-thinks are a bad thing unless you didn't do it right the first time.

Re:Will the kernel ever get to 3?

[petes_PoV]

The kernel can't add implementation for a bus that doesn't exist

and that's why it will alway be a follower rather than a leader. The innovations are what creates the need for standards. Without them nothing new would ever be developed and there would be no need to codify and standardise any developments.

What I would like to see is some innovation, some game-changers: giving the Linux kernel new features that no other O/S has - but once it has them, EVERYONE realises how useful, necessary and well-done they were and therefore how necessary they are to modern, leading edge implementations. It's a question of does Linux want to follow the innovators lead, or does it want to be out in front?

Re:Will the kernel ever get to 3?

[Lennie]

Linux has: USB3 before any other OS, hotswap-memory, hotswap-cpu, hotswap-pci, hotswap-scsi, numa, scales to I don't know how many nodes in a cluster and cpu-configurations. Runs on the most possible hardware-archictures (NetBSD is not the top dog in this field anymore). Has the most build-in drivers of any OS. Thus runs on really small and really large. Is used for embedded from wallplugs to netbooks all the way up to smaller mainframes. Manufacturers of TV's, networking-devices like switches use it for the control-plane. It also has the broadest range of filesystem support, etc. most of the websites you visit are running on Linux, so it's heavily used in that field as wel. I think Linux is used by the innovators, because you can change it. Some people say Google does innovation, they use Linux for pretty much everything.

Re:Will the kernel ever get to 3?

[Anonymous Coward]

I don't think there *can* be a whole lot of innovation go in at this point. For a kernel 90% of the innovative bits are going to be in the design and architecture. Once that's done it's all extra drivers and incremental improvements.

Re:Will the kernel ever get to 3?

[Kjella]

Actually, I hope the kernel will contain less. Let's take USB for example, do we really need all sorts of various connectors? Or would we rather just use USB, teach the kernel to do low level read/write to USB devices and then do keyboards and mice and printers and scanners and digicams and webcams and external hdds and whatnot over USB in userspace? In fact, much the same applies to drivers in general, there's no reason why so many printers are paperweights under Linux. Can't there at least be one universal idiot mode where we feed it uncompressed raster data and it prints? Seriously.

Kernels are best at being mediators, be it of CPU time, GPU time, IO bandwidth, network bandwidth, whatever. Something offers resources, something consumes resources and the OS is that gray glue in the middle. Whatever killer feature you want, you probably don't want it in the kernel. You want to write a desktop environment or an application or something, and the kernel will make sure it runs gracefully together with everything else. There's a quite a few more bits to the kernel, but they're just adoptees brought into the kernel for performance reasons.

Re:Will the kernel ever get to 3?

[gr8_phk]

It's a question of does Linux want to follow the innovators lead, or does it want to be out in front?

Linux got USB3 before anyone else. Linux ran on 64-bit x86 before anyone else - before AMD made even made the first chips. What on earth would constitute something innovative in an operating system? It's a resource manager, and new resources (i.e. hardware) gets added all the time. What is this innovation you speak of?

I could see some really innovative way to handle all those hardware variants comming along - what that is I don't know. But if someone does devise something truely innovative in that regard, they'd want to prove it on the platform with the most extensive and diverse hardware support - which would be Linux.

Re:Will the kernel ever get to 3?

[Anpheus]

If I recall, what Google innovated on was OS choice for their employees. I believe I read that they can choose whatever they like for their OS, and for their 20% projects they can use any number of platforms, etc. I think because of familiarity with Linux and competition with Microsoft, we won't ever see Google running on IIS, but I think that's a business move.

Re:Will the kernel ever get to 3?

[jedidiah]

Sour grapes from the losers...

In truth, anything on the desktop is a shallow immitation of something that was done in academia 20 years ago.

So anyone whining about Linux copying something else is like a dwarf calling a midget shorty.

Re:Will the kernel ever get to 3?

[LordNimon]

and that's why it will alway be a follower rather than a leader.

That's not true. Many companies, including the one for whom I work, use Linux as a platform for enabling new technologies. We don't write code for any other operating system.

Re:Will the kernel ever get to 3?

[jensend]

The kernel can't add implementation for a bus that doesn't exist

and that's why it will alway be a follower rather than a leader.

You're so totally right! What Linux needs is drivers for hardware that doesn't exist. Here's the sales pitch: You may not yet be able to purchase a combination printer/scanner/fax/toaster/singing fish/unicorn horn fabricator with a 42 megahexametapassamaquoddabit UFB (Universal Fairy Bus) 3.0 connection, but when you do, you can be sure it will work just fine with Linux, as we already have a driver for it! Aren't we innovative?

Re:Will the kernel ever get to 3?

[omkhar]

That's incorrect, IBM has allowed employees to choose whatever OS they want - including official support for Linux and Windows, as well as skunkworks support for Mac.

Re:Will the kernel ever get to 3?

[oakgrove]

Not sure what your definition of "master" is but Linux is on over 3/4 of the top 500 supercomputers, comprises over 50 percent of internet servers out there, dominates embedded systems by a very significant margin and is at the heart of the fastest growing cell phone platform. Believe me, the world of computers is much larger than what you see at your local Best Buy.

Re:Will the kernel ever get to 3?

[vadim_t]

Actually, I hope the kernel will contain less. Let's take USB for example, do we really need all sorts of various connectors? Or would we rather just use USB, teach the kernel to do low level read/write to USB devices and then do keyboards and mice and printers and scanners and digicams and webcams and external hdds and whatnot over USB in userspace?

The kernel does this to a large part already. In fact, printing is implemented in userspace, as well as many other devices. That's what libusb is for.

Devices like keyboards and hard disks don't make much sense to support in userspace, since the kernel already has to support everything having to do with hard disks, and the lower level USB protocol. The kernel knows what a hard disk is, it knows the standard USB protocol, it's very straightforward for it to use a hard disk over USB. Adding userspace there would only make things more complicated.

In fact, much the same applies to drivers in general, there's no reason why so many printers are paperweights under Linux. Can't there at least be one universal idiot mode where we feed it uncompressed raster data and it prints? Seriously.

It already exists, and it's called "postscript". Buy a printer that supports it, and no problem. But if the printer insists in talking its own weird language it's not going to accept your "uncompressed raster data".

Re:Will the kernel ever get to 3?

[gardyloo]

In the meantime, those of us who use our computers to get real work done will continue to use Windows 7 [...]

*Cough* Ah, yes. That heap, which because of various security problems and incompatibility with disk encryption, will not be used by certain government labs for at least a year, and probably much longer. Thanks be to Torvalds that linux does pass the tests, even the miniature distros, and is allowed, so we can get real work done.

Re:Will the kernel ever get to 3?

[Kjella]

Can't there at least be one universal idiot mode where we feed it uncompressed raster data and it prints? Seriously.

It already exists, and it's called "postscript".

And you can also use Photoshop CS4 to crop screenshots, if "overkill" is not in your vocabulary.

Re:Will the kernel ever get to 3?

[spinkham]

There's been tons of innovative stuff in Linux, just on a level most users don't have to interact with it. And that's the way it should be.

Your comment is like knocking on Toyota for basically making the same car as everyone else, the only difference being their superior production methods and products.

In fact, the Toyota prodcution process and Linux are a lot a like: Linux is where it is today through a philosophy of continuous improvement, not any one great feature that is a world beater.

For those of you who missed out on the 80s when Toyota changed the auto industry, see this: http://en.wikipedia.org/wiki/The_Toyota_Way

Re:Will the kernel ever get to 3?

[vadim_t]

I have no clue why would it be overkill, it's supported perfectly fine by printers that are positively ancient these days and rotting on junkyards.

Re:Will the kernel ever get to 3?

[Spacelem]

LOL, and yet lin-sux cannot play a flash video at full screen without stuttering. Perhaps you imbeciles should spend less time adding NUMA support to your crappy kernel and more time writing actual working video drivers. In the meantime, those of us who use our computers to get real work done will continue to use Windows 7 and OS X.

I can't say I've ever needed to play flash video at full screen at work (where I and most of the other researchers run Linux), as there's not much use in it. It does handle my simulations wonderfully though, particularly when I can run them in a console with screen on the server, then switch my PC off for the night. Not sure if it would be quite so seamless under Windows, as even the Windows users tend to run Linux in a VM to do the actual work.

At home though, my DVDs and avis run just fine at full screen, and at full frame rate; so do my games, and there are quite a few old ones that don't actually run under Windows any more. I'll admit that flash stutters quite badly if I try to run it at full screen; it's not a big deal though, as resizing the desktop resolution does the job just as well.

Re:Will the kernel ever get to 3?

[Kjella]

Because it's a turing-complete programming language with font management and a rendering engine, that was ditched on almost all low and mid-level printers because it added cost. I'm talking about a mode where you pipe it all to Ghostscript on the computer and pipe the output as a series of raw pixels/dots right to the printer, no font management, no rendering, no instruction language beyond setting up the print area and most importantly, something that could hopefully be done on a 49$ printer without adding cost.

Re:Will the kernel ever get to 3?

[vadim_t]

Raw pixels comes out to a rather large number per page, though.
There is another standard: IPP, which is widely supported by network printers. AKA CUPS, which is what Apple uses.

In any case, I don't see what any of this has to do with Linux. Go and ask the printer manufacturers to standarize on a bitmap format, and I'm sure Linux would quickly get support for it.

But, if you're worried about cost, you shouldn't be buying a $49 printer in the first place, as you'll quickly pay for that cheapness by getting gouged on ink.

Re:Will the kernel ever get to 3?

[Zaiff Urgulbunger]

Re flash stuttering, up until recently I was using an old Dell Dimension 3000 with a 2.8GHz Celeron + nVidia 6200 (PCI version!) and that did stutter, I was able to get around it by using the Zoom In feature of Compiz to effectively get full-screen flash.

I'm now runing a Core i5 based machine with an nVidia 9800 and flash now runs perfectly.

For the record though, I think the problem is that the flash plugin requires a ridiculous amount of CPU power to work!!

Re:Will the kernel ever get to 3?

[Randle_Revar]

In November Greg K-H said "We change something like 5,000 lines a day, which is just scary. Fifty percent of that change will be in the drivers, and five percent will be in the core kernel."

According to LWN "2.6.32 is the result of 10,767 non-merge changesets sent in by 1,229 developers. This changes added a total of 1.17 million lines, while removing 611,000 lines, for a net growth of 559,000 lines of code." This isn't out of the ordinary, either (e.g. " 2.6.31 development cycle had seen the incorporation of 10,663 non-merge changesets from 1,146 individual developers. These patches added almost 903,000 lines of code and removed just over 494,000 lines, for a net growth of just over 408,000 lines.").

As for this stuff all being minor, read LWN, or the Kernel Newbies changelog, and see how long you keep that illusion.

Since the 2.4 -> 2.6 transition was so bad, Linus is unlikely to ever again do a "break the world" change, either with or without a long lived dev branch like 2.5. And really there is no need, short of switching to a microkernel or a managed language, everything can be done as a series of gradual changes (see: the ongoing TTY rework, KMS/DRI2, the relatively recent power-saving/suspend work, the gradual removal of the big kernel lock, the /x86_64/ -> /x86/ merge, the addition of mutexes (~2006Q1), someday the realtime stuff will probably be merged without a major version bump, etc)

Re:Will the kernel ever get to 3?

[Randle_Revar]

Not only was Linux the first to have USB3 support, when Intel got the physical aspect down and had to work on the programming interface (xhci), they used Linux. When the interface was stable enough they took the driver upstream. Linux is also the OS Oracle decided to use to create btrfs.

Re:Will the kernel ever get to 3?

[Spacelem]

I built my own machine, it's a Core2 Duo E8400, with a Geforce 8800 GTS 512 video card and 2GB memory.

Actually I had no problem with stuttering until I upgraded from a 17" monitor to 24" widescreen monitor. I suspect that the plugin is trying to do something really stupid, as resizing the picture should be reasonably cheap for the processor (it's only stretching pixels, not drawing the original video at a higher resolution). If the video was passed to the graphics card to resize, there'd be no difference at all between fullscreen and windowed.

Re:Will the kernel ever get to 3?

[iggymanz]

anything on the desktop was done better by GUI for commercial Unix over 20 years ago, until maybe three years ago the user experience of NeXSTep Windowmaker or IRIX IRIS beat the half-baked KDE or GNOME crap

Does Plymouth now work with nvidia cards

[JackieBrown]

And can I switch back to the propitary drivers when my desktop boots up?

That said, for my main desktop computer, I may switch to the Nouveau drivers since I only use it to browse the web and encode movies. I don't even have my speakers set up to that machine.

Be nice not to have strange lockups. (To be fair, I am not sure if that is a nivdia or KDE issue but my mouse quits responding to the button click but I can still see it moving on the screen. This usually happens when running virtualbox so maybe it is doing something with the mouse focus - except the problem persists even after I close it.) As it is, I do my encoding on tty2 and tty3 since I hate having to restart an encode (take to long.)

Re:Does Plymouth now work with nvidia cards

[BrokenHalo]

Be nice not to have strange lockups.

Can't say I've seen any of those, and my machines are up for weeks or months at a time, and in more or less constant use (but with Gnome/Compiz-Fusion, not KDE).

Re:Does Plymouth now work with nvidia cards

[JackieBrown]

I didn't used to but it is happening frequently enough for me to be using non-x terminals for my encodings since they usually take several hours.

I tried using openbox as with window manger for KDE but I still ended with the same problem. I have a hunch it is related to virtualbox but cannot prove it. And the nvidia driver I install very quickly because the nv driver is painful. I can see pages redrawing when I scroll. I usually end up using the vesa driver if I ever have problems with the nvidia binary but in fairness it has been years since I tried the nv driver.

I really wish I could figure this out because it does annoy me. Fortunately, restarting X does not effect my other terminals so not much is lost but still...

Re:Does Plymouth now work with nvidia cards

[oakgrove]

This isn't any kind of fix but you can always just use GNU/Screen in konsole or whatever xterminal you prefer to avoid having to switch to a virtual terminal. With Screen, you can restart X and then just reattach to the screen session you were using to do your encoding in.

3D

[Lord Lode]

If those NVidia drivers don't support hardware accelerated 3D, then I really don't understand the point. 3D hardware acceleration is 15 years old. Linux is an operating system that should be at the frontline of technology. Working in the dark ages of pre-3D acceleration, the times of Motif GUI's, should be far past us. How can something that ignores such an important part of the graphics card, almost half the computation power of the whole computer is there, be accepted?

If they do support 3D, then congratulations, ignore my post above :)

Re:3D

[ledow]

Ignoring the obvious troll:

Anything that works will be accepted, like every other driver in the kernel... if it doesn't make *everything* work, that's not a big problem. Especially new drivers rarely have code that actually makes the device inherently useful, or supremely accelerated, immediately - but it will function. That's how you code - one bit at a time, gradually building as you go. When you have all the DMA, 2D drawing, multiscreen crap worked out *THEN* you can think about 3D. At the moment, even simple combinations like dual-displays can cause major headaches with some chipsets, whether the hardware supports them or not.

The programmers are effectively working blind with unknown hardware - and programmers don't work that way, that's a reverse engineer's job. To say they can't merge *anything* until all the features are working just means you'll never see *anything* at all. But if they merge a 2D driver today, they can add basic 3D access tomorrow and 3D acceleration the day after and maybe some day you'll see something of use. If not, at least you'll be able to boot Linux and *see* something in X-Windows on any computer that runs off that chipset (or has backwards compatibility for it).

You will not see full 3D accelerated drivers for any chipset (especially not any that compete with manufacturer's drivers in terms of acceleration) that matters to you on a new computer until manufacturers fully co-operate and help get coding too. Don't expect it, don't complain about it, don't moan when it doesn't happen or only "obsolete" chipsets ever get 3D support. When the manufacturer's co-operate, it takes nothing to make a driver. When they don't, it means knowing *everything* they know before you can really start properly.

Re:3D

[Lord Lode]

It's not my intention to troll. I'm a Linux user because I like the style and way of working with that operation better than Windows. So anything that is BIG and tries to limit the choice of users to "Windows-only" is bad to me (that is, things like Direct3D, IE-only webpages, Office formats, ...), because I think users should be able to make a choice what OS to use and have a good range of software choices on all.

IMHO, I see no reason to not use the NVidia drivers, that they make for Linux, and allow me to play some modern (=2009) games in Wine. If people are trolling NVidia saying they don't cooperate, they could as well pull the plug and not provide the drivers anymore.

I do like the programming effort of trying to reverse engineer them, it is a very interesting effort and the results could be massive. So I definatly don't want to troll against the people doing this effort, on the contrary.

I program things that use OpenGL myself and require hardware acceleration, software OpenGL rendering is too slow, and I've heard Linux users complain about the hardware acceleration requirement. And THAT is what I'd like to troll against. Hardware 3D acceleration should not be an "option" today.

Re:3D

[MostAwesomeDude]

The reason you're being modded Funny, I hope, is because 3D drivers (and actually any kind of acceleration for GPUs) is a userland thing by tradition and convention. The nouveau project does sponsor development of a few Mesa/Gallium drivers, none of which are yet production-quality, but it has nothing to do with the kernel part.

Re:3D

[Anonymous Coward]

Why is 3D in userland but archaic and obsolete teletype interfaces in the kernel?

Re:3D

[Lord Lode]

Well I wasn't trying to be funny...

Re:3D

[ettlz]

Because 3D requires a lot more complex heavy lifting that I don't want in the kernel when it fucks up. Teletype is quite lightweight by comparison.

Re:3D

[Anonymous Coward]

Since modern GPUs are getting closer and closer to general-purpose processors, isn't it time to rethink the division of responsibilities between the kernel and the graphics card driver? AFAIK modern graphics card drivers already duplicate much of kernel functionality for multitasking, memory allocation and inter-process communication.

Re:3D

[Culture20]

If those NVidia drivers don't support hardware accelerated 3D, then I really don't understand the point.

The point is that the xorg nv and vesa drivers are broken for quite a few nvidia cards.

Re:3D

[Zoxed]

> If those NVidia drivers don't support hardware accelerated 3D, then I really don't understand the point.

IIRC kernel video mode setting will be available with nouveau: for a silky smooth boot experience :-)

Re:3D

[Anonymous Coward]

The nouveau drivers as in the last kernel release do not include the necessary "firmware" to do much accelleration (this is for copyright reasons that may or may not be sorted out sooner or later).

The version of the drivers at the nouveau git repo do provide good 3d accelleration and other Xorg extensions (like overlays).

I'm currently running a 2.6.33 rc3 merged with the HEAD of nouveau and that works fine.

Re:3D

[Anonymous Coward]

Well, now you will get a legitimate +5 funny.

Re:3D

[Lemming Mark]

A fair point - not really funny mod material ;-) The Nouveau drivers being merged are an important step in the right direction. The current situation, particularly for Free Software-only distros like Fedora is that when you install the system and boot for the first time you end up using the open source "nv" driver, which was provided by NVidia to provide basic 2D support but apparently is not very good. The Nouveau driver has support for KMS (kernel modesetting), which is needed for flicker-free graphical boot and nicer fast user switching, Xrandr which is the new way of doing dynamic monitor configuration and possibly even has better 2D acceleration than nv (I'm not sure on that point). In addition, the Nouveau people are working towards reverse engineered 3D support - they already have some basic stuff running on that score, I think.

So the Nouveau drivers provide a functionality boost for "haven't installed the NVidia binary drivers yet" situations and, for people who want or need to run free software systems, they give you better functionality than the existing open source nv driver. Various distros (Fedora first, I think, with Ubuntu following) are including Nouveau as their default driver for NVidia cards, so it makes sense to have it in the upstream kernel where it's subject to the scrutiny of many eyes, can benefit from infrastructure improvements in the kernel, etc. In the meantime, the Nouveau developers can carry on working towards full 3D support based on the stuff that's already in the kernel.

It's not so much a question of the 3D portion being ignored, more of the fact that it's more advanced than the existing OSS 2D drivers, so already an improvement, with 3D support coming in the future.

Re:3D

[MostAwesomeDude]

Oh, look, we already did. We put two memory managers in the kernel, one for discrete chipsets and one for IGP chipsets. We added a system for lockless submission of commands to the GPU, and a series of checks to prevent GPUs from being easily hardlocked.

Re:3D

[chmod a x mojo]

Sounds insane, but I was having some problems with my ATI card with KDE4 under Debian squeeze a week ago so I plugged my NVidia 6200LE in. Low and behold HAL brought X into a sane state, and just for the heck of it I tested for direct rendering.... it came up "yes". Needless to say I wasn't expecting that, so I grepped my X log..... it was using the "nv" driver. I have no idea HOW it was getting direct rendering out of that driver but even glxgears was playing smoothly, something that it has never done with indirect rendering on that card.

Re:3D

[Randle_Revar]

Nouveau is, in fact a full driver stack. The 3D part just isn't ready, because 3D drivers are really hard, especially when you have no documentation.

The Kernel Report

[Kikuchi]

Hello Nation. If I had a quarter for every time I said I had a nickel, I'd have five times as much theoretical money. This Is the Kernel Report!

Flickering on Intel chipsets

[Jack Malmostoso]

I upgraded to 2.6.33-rc4 from 2.6.32 because of strong flickering and tearing on my Intel chipset.
If you're affected by the problem you might want to give it a shot even in -rc state.

Yes

[Anonymous Coward]

and no.

As it is, I do my encoding on tty2 and tty3 since I hate having to restart an encode (take to long.)

$ man 1 screen

Re:Yes

[JackieBrown]

Wish I had mod points.... Thanks

OSX

[Anonymous Coward]

Use OSX! 3D acceleration is there, too :)

Re:OSX

[garaged]

and say goodbye to portability abd standarization, even GUI programming on OSX it's hard if you want those too

Re:OSX

[ckaminski]

If people stuck to OpenGL instead of the DirectX miasma, we wouldn't have this issue - it'd be more likely we'd have more ports of games to Linux, and hence, vendor interest in supporting 3D on the platform.

It's sad, perhaps OpenGL would have evolved faster had Microsoft implemented it for Windows 9x.

Or just use a decent distribution

[Zero__Kelvin]

"I have such a chipset and I've been cursing NVIDIA on a regular basis. After updating to any new kernel, I must boot into no-X mode, then run the proprietary driver installer."

Or you could get one of the many, many, many Linux distributions that handle this automatically. Mandriva comes the mind since it has handled this stuff for years and is extremely user friendly, but as I say there are many other options as well.

Linux Kernel? Not important. Linux Scheduler?

[Anonymous Coward]

IMPORTANT.

Why? Because the Linux Kernel is already the best in existence, nothing else comes close. The Windows and OSX kernels are toys in comparison, like something by Fisher Price.

What Linux needs is the new scheduler written by Meatloaf. Its very efficient, very fast, and very intelligent - probably the most advanced scheduler ever written. Like most people, I was suprised to find out that Meatloaf was an expert Linux hacker, but it turns out he is, and I can't wait to see his code go live.

Re:Linux Kernel? Not important. Linux Scheduler?

[ckaminski]

Oh please. The Windows kernel is arguably one of the better kernels ever written.

All the rest of the crap on top of it, not so much...

Anonymous Coward

[Anonymous Coward]

what about fedora?

http://www.freezlo.com

ath5k in 2.6.32.3

[Anonymous Coward]

Personally didn't try too hard with X, just used VESA and moved on. FWIW, my nVidia installs have always gone very smoothly under Slackware.
Wireless : The ath5k module is rockin' hard under hostapd, up at b/g and stable for three days so far.
Atheros is going to benefit from my upgrade 802.11n upgrade, as well.

Just for the sake of graphics?

[jasonditz]

Is the new Nvidia functionality solely for the graphics cards or does it also mean improved support for Nvidia chipsets like the MCP78S.

I guess what I'm really asking is: is there any chance the next kernel will fix this, or will using USB microphones and CDMA modems on my Pavilion P6130F remain a pipedream?