Linux Apps Are Not Coming To Many Still-Supported Chromebooks (betanews.com)

While we know that Linux app support is coming to a range of Chromebooks from Lenovo, Acer, Dell and others, a post on the Chromium Gerrit reveals that devices running Linux 3.14 or older will miss out. BetaNews: Chrome OS is able to run Linux apps through the use of containers which help to keep the rest of the operating system safe from harm. As container support requires features that are only found in more recent versions of the Linux kernel, it means that many Chromebooks -- whose kernels are usually not updated -- will not be able to run Linux apps.

Here's the full list of Chromebooks that won't be getting the Linux love: AOpen Chromebase Mini (Feb 2017; tiger, veyron_pinky), AOpen Chromebox Mini (Feb 2017; fievel, veyron_pinky), ASUS Chromebook C201 (May 2015; speedy, veyron_pinky), Acer C670 Chromebook 11 (Feb 2015; paine, auron), Acer Chromebase 24 (Apr 2016; buddy, auron), Acer Chromebook 15 (Apr 2015; yuna, auron), Acer Chromebox CXI2 (May 2015; rikku, jecht), Asus Chromebit CS10 (Nov 2015; mickey, veyron_pinky), Asus Chromebook Flip C100PA (Jul 2015; minnie, veyron_pinky), Asus Chromebox CN62 (Aug 2015; guado, jecht), Dell Chromebook 13 7310 (Aug 2015; lulu, auron), Google Chromebook Pixel (Mar 2015; samus), Lenovo ThinkCentre Chromebook (May 2015; tidus, jecht), Toshiba Chromebookk 2 (Sep 2015; gandof, auron).

Comments

Open source?

[mi]

a post on the Chromium Gerrit reveals that devices running Linux 3.14 or older will miss out.

So? Why would not you just download the source-code and compile it yourself?

Was not this ability the point of Linux — and the rallying cry for its fans — for 20+ years?

Lrf, V'z gebyyvat. Gunax lbh sbe znxvat na rssbeg gb ernq guvf.

Re:Open source?

[iggymanz]

oh, you've just noticed that some vendors lock down things contrary to the open source intent and spirit? are you a dumb-ass?

Re:Open source?

[Anonymous Coward]

So? Why would not you just download the source-code and compile it yourself?

Because it's not as simple as

Download source
Compile
Done

That's the myth of open source software.

Was not this ability the point of Linux — and the rallying cry for its fans — for 20+ years?

Yes, that's true. But (most) Chromebooks are a locked down bastardized Linux. Getting a newer version of the Linux Colonel to run on them involves a very long list of

If .....
Maybe ....

Most Chromebooks, in actual practice, are no more "open" than an iPad. And if you were stupid enough to buy one, well, it sucks to be you.

Re:Open source?

[Anonymous Coward]

I guess because of the reason why so many android phones run damn old kernels: lack of compatible blobware. In theory nothing stops the user from updating the kernel, but if the blobs (aka firmware) of some of the components have not been compiled against that new kernel... they won't work.

Re:Open source?

[Anonymous Coward]

The reality is more like...

download source
configure
compile
done :o)

Re:Open source?

[swillden]

a post on the Chromium Gerrit reveals that devices running Linux 3.14 or older will miss out.

So? Why would not you just download the source-code and compile it yourself?

You can. Put it in dev mode and go nuts. You'll lose all of the nice Chromebook security guarantees, of course, and your machine will be running a system that is custom, untested and therefore likely to be less reliable, but that's a perfectly acceptable tradeoff for many power users.

Re:Open source?

[mi]

your machine will be running a system that is custom, untested and therefore likely to be less reliable

Whoah, whoah... Had anyone alluded to such a thing — that built from source makes anything "less reliable" — back in my days, Linux fan-bois would've scorched the very soil he stood on to crisp with their flame-throwers... Are we still on Slashdot even?..

Re:Open source?

[jellomizer]

1. Not all Apps in Linux are open source.
2. Compiling on slow systems can take a long time.
3. Compiling isn't always easy.
4. Source code doesn't guarantee compatibility. (or be readable enough to be fixed)

Re:Open source?

[tepples]

Building from source makes software less reliable when the firmware includes technical measures to deliberately make software built from source less reliable. Chromebook stock firmware includes such a measure. In developer mode, it prompts the user to press two keys to erase the entire hard drive, which makes the device's overall data retention less reliable. Consider the following sequence of events on a Chromebook in developer mode.

*someone turns on your Chromebook*
*doesn't know what the prompt means*
*presses Space as prompted*
*presses Enter as prompted*
*you have lost all data*

Re: First Post Coming to Slashdot

[Anonymous Coward]

Drumpf prefers his piss frosty.

Should add .NET support

[bluefoxlucid]

I had suggested they use containers to add .NET support instead of straight Linux apps. Linux needs a whole underlying support system in a VM; .NET can inherit that from the local system and the .NET runtime.

Stuff .NET Core or Mono in there and provide a /home/$USER directory. Mount /home/$USER, mount local Documents to /home/$USER/Documents, and mount the Google Drive folder to /home/$USER/Drive.

When you install a .NET application, it would mount all of those things, as well as a .NET runtime for the app. The runtime could include a modification to call for additional libraries: if it tries to load an assembly, it calls through a socket to tell ChromeOS to add that library to the container (additional mount), and ChromeOS may fetch the library via nuget.

The container would expose a Chrome browser stub which also calls on ChromeOS to open the browser. It could do the same for opening various files. A ChromeOS file dialogue would display the mounted paths the same way ChromeOS exposes them, unifying the UI.

Now you have native .NET applications. You can get a Python and Java VM running on .NET, which may allow extending this to other types of native applications using the same runtime. You can isolate those applications from the system so they only have access to their own configuration and specific paths, rather than the whole of your files.

It's less reach than a whole Linux VM; it's also less overhead and tighter integration.

Re:Should add .NET support

[squiggleslash]

I had suggested they use containers to add .NET support instead of straight Linux apps

Why not both?

No, I mean it, why not both? Yeah, there's a minor overhead with GNU/Linux that might not exist with .NET but it's only going to need to be there if you install GNU/Linux applications. And their VM architecture, while... weird... (a VM running Linux running one or more LXD instances? I'm guessing that this is supposed to be more secure, but I suspect it's probably the opposite even if I can't prove it...) could easily add .NET as a LXD instance.

Re:Should add .NET support

[bluefoxlucid]

Firstly, because you have to support both.

Second, because the security model is harder for a full, uncontrolled VM. An application is also a difficult security model, as it's uncontrolled and can be taken apart; .NET is mostly managed code and individual applications can have their own restricted permissions and containers.

Third, because .NET integration with ChromeOS is reasonably-achievable such that the applications all seem native. It would be like Paint.NET or Visual Studio was built for ChromeOS.

A full VM is an interesting feature, but not an everyday feature without a major technology step forward. We'd need a way to swap the main head from ChromeOS to Linux to really do it well. .NET with a Python and Java runtime (as in: AOT the Python or Java to .NET, run that) is a pretty big step forward, and can look and feel like it's just a part of the ecosystem and not a special kind of environment wedged in for compatibility and expert users.

Containers, Sandboxes, etc

[Anonymous Coward]

Do these actually provide security, or do they ALWAYS end up with so many holes to the point where their security is only an illusion?

Re:Containers, Sandboxes, etc

[bluefoxlucid]

Containers generally provide security by segmenting the outside OS from the container. They're typically unbreakable.

If you make a container by mounting your host libraries as writable into the container, you can of course modify code that runs as root on the host. Anyone with access to create a container generally is considered a root-level user because they can make a container with root as the user internally and with / as the / volume, and then they have root access.

There's a sort of advanced setting where you map users so e.g. UID 0 is really UID 90,000. The user ID of the process creating the user namespace has all capabilities within the namespace, but none outside: it becomes root inside the container, and can't do root things outside the container. Generally, the process can only switch to another namespace if it has CAP_SYS_ADMIN in that namespace as well.

That means a process can't leave the container and see your system; if it does leave the container (somehow), it's UID 90,000 and has regular user permissions instead of root permissions.

Original Configuration

[MightyYar]

I'm outraged that my $200 Chromebook only supports the same features it was sold with. I bought it specifically to get features that I wasn't aware existed yet.

Re:Original Configuration

[tepples]

The features did exist. They just got pulled from the market back in 2012.

Re:Original Configuration

[MightyYar]

I'd have been pretty pissed if I bought a Chromebook and got a netbook. The whole point is to give the kids something they can't fuck up no matter how hard they try, and which works with all their school stuff.

If you actually do want Linux on a Chromebook

[fibonacci8]

Current version of GalliumOS supports most Chromebooks. It's an Ubuntu 16.04 base with adjustments for specific chromebook models (media key customization, other hardware configurations, etc. Doesn't seem to support ARM chipsets or Intel Pineview boards.
Systems that are supported by GalliumOS
and the newer version based on Ubuntu 18.04 LTS is in the works
version 3.0 alpha

Re:If you actually do want Linux on a Chromebook

[MightyYar]

Note that you do need to change to 3rd party firmware in most cases.

Re:If you actually do want Linux on a Chromebook

[Mal-2]

As noted in my own post, you can also set up rEFInd (or some equivalent of your choice) and have both ChromeOS and Gallium (or any other distro) installed. The only restriction is that the distro has to be set up to install to UEFI machines, which (as of a couple years ago) some still did not.

Re:If you actually do want Linux on a Chromebook

[squiggleslash]

While true, you lose the benefits of having a Chromebook if you do this. It just becomes another GNU/Linux laptop, one without the Page Up/Page Down buttons, albeit one that can be booted into ChromeOS from time to time..

As others point out, you also generally have to install firmware modules to make this work. This is a complicated process, and if you want to still be able to boot ChromeOS, you can only do it with the rather annoying "Do this control key combination or we'll put you into a mode where you can easily and unintentionally powerwash your computer."

It's... not ideal.

The article is misleading, even if technically correct. Right now no Chromebooks come with the Linux feature enabled and supported. To use it you have to enable the dev channel for updates (not developer mode, just the equivalent of "Windows Insider" for Windows users, a channel that sends you betas of the operating system. Google specifically warns users not to rely on anything delivered via that channel being there forever, and has withdrawn features before (as I found out the hard way when Android app functionality was removed from my Asus C300 without warning.)

So, as of now, Google is talking about never implementing a feature for some supported laptops that isn't officially available for anyone except beta testers. And by the time Google officially releases the feature, in a production ready form, it's quite possible that either those laptops will no longer be supported, or they'll have been upgraded to run a more recent Linux.

In short, it's a non-story.

Would require developer mode

[tepples]

So? Why would not you just download the source-code and compile it yourself?

Because you would have to put a Chromebook into developer mode to install an unofficial kernel. A security feature in the Chromebook firmware allows anybody who turns on a Chromebook in developer mode to powerwash it by turning it on, pressing Space as prompted, and pressing Enter as prompted. This causes you to lose the day's work that you haven't yet been able to back up remotely and lose the use of the modified Chrome OS until you have a chance to reinstall it. (See "Chromebook Developer Mode Warning".)

Re:Would require developer mode

[mi]

Because you would have to put a Chromebook into developer mode to install an unofficial kernel.

Goodness gracious, "unofficial" kernel? Does not that imply, there is also an "official" kernel? What is that, and just how "open source" is that very concept?

Try to tell it to chip manufacturer

[DrYak]

a post on the Chromium Gerrit reveals that devices running Linux 3.14 or older will miss out.

So? Why would not you just download the source-code and compile it yourself?

Was not this ability the point of Linux — and the rallying cry for its fans — for 20+ years?

Yes, that was the point, but it looks like most of the chipset manufacturer completely missed the point and only provide blob drivers.
Thus, on hardware like smartphone, tablet and chromebook, you're basically stuck with whatever the company that made the SoC decided to use (and never to update there after).

--

Well, in theory.
In practice, *YOU CAN* actually recompile it your self, but you might have some problems :

- You might not have working graphics (and maybe a few other proprietary stuff like touch screen) missing.
(e.g.: the poor guys who have some PowerVR GPU in their machine).

- Or maybe you're lucky and the chipset manufacturer has decided to release more recent drivers for more recent kernel : that's a bit more likely nowadays that Google insists on requiring kernel version 4.4 for Android. Maybe the manufacturer of the chipset could have a workable driver for a more recent kernel, and maybe you could find a way to put it into your chromebook (or maybe you could straight up try the android driver and use something like libhybris).

- Or maybe you're even more lucky and there's an opensource driver for your chipset. That should be the case of most intel-based one which tend to use intel GPUs and Intel is paying opensource drivers development themselves. Some Qualcomm ARM chipset uses Adreno GPU for which the Freedreno driver might be working. In these case, YOU CAN almost recompile nearly everything, you'd be missing only some functions (like touch screen).

On regular PC, AMD hardware is about the best you can get for Linux (they put lots of efforts into their opensource drivers), but sadly they're very seldom seen in that form factor (usually only in desktop, at most inside some budget laptop)

Re:Try to tell it to chip manufacturer

[Highdude702]

PS Console I would assume still runs a UNIX based OS. Also AMD based

ROT13

[DrYak]

By the way, you should have used double-ROT13, I've heard it's more secure :-P

Security

[rtkluttz]

"Keep the operating safe from harm". That's both hilariousand fucked up. It needs to be the other way around. Run a sandboxed chrome OS inside of a linux container that is under the OWNERS control. Is security. The definition of malware is software not under the owners control, so by definition chromeOS is a security risk to the owner of the device.

Re:Security

[Anonymous Coward]

Chromebooks are the kind of thing you give a kid to have access to a few educational apps and sites. The user is typically not the owner, and you definitely don't want the user to have root.

Default firmware

[DrYak]

The *default* firmware requires dev mode.

But because the firmware is coreboot based, it should be possible to find an alternate firmware,
file the hardware switch to enable flashing (usually a screw I've heard) and get yourself a full blown linux laptop that doesn't complain and risk self-destructing on each single boot.

How to compile C++ to safe .NET IL?

[tepples]

When you install a .NET application, it would mount all of those things, as well as a .NET runtime for the app.

Would this .NET runtime allow use of mixed assemblies (which contain both native and CIL code) or other unsafe CIL? If not, read on:

You can get a Python and Java VM running on .NET, which may allow extending this to other types of native applications using the same runtime.

I read years back about something called C++/CLI, which extends ISO C++ with .NET-specific syntax for pointers and references in verifiably type-safe code. ISO C++ uses * to declare pointers and & to declare references, but C++/CLI uses those to mean unsafe pointers and references. Use of these causes verification of type safety at load time to fail. To declare pointers and references to managed objects, C++/CLI instead uses ^ for a pointer and % for a reference. (Source: "Component Extensions for Runtime Platforms")

Can ISO C or ISO C++ be compiled to verifiably type-safe .NET bytecode? Or is there a useful subset that can be automatically translated both to ISO C++ and to verifiably type-safe subset of C++/CLI? I don't think so given Microsoft's attitude in the following document: "If your code needs to be safe or verifiable, then we recommend that you port it to C#." (Source: "Pure and verifiable code (C++/CLI)")

Re:How to compile C++ to safe .NET IL?

[bluefoxlucid]

You can probably translate C++. Native code won't run on a different CPU IL, hence the use of .NET.

The big IDEs—Eclipse, Monodevelop, and Visual Studio—are Java and C#. Python is a common language. I had hoped to get Monodevelop 7 and Visual Studio native on Chromebook, as well as perhaps Unity3D or something like it (Unity3D is huge). Imagine if Chromebooks were a development platform for Android, IOS, Webapps, games, and cross-platform applications.

I'm not thinking so much about native applications because web browsers, e-mail clients, and the like come on Chromebooks. LibreOffice, Gimp, and Krita would be good applications to port to Chromebook, and are unfortunately not CIL: more work would be needed to port them as native applications. Pinta or Paint.NET can take the place of Gimp. LibreOffice is a large pile of garbage anyway; AbiWord always worked better, but is now defunct.

Chromebook firmware is partly Tivoized

[tepples]

"unofficial" kernel? Does not that imply, there is also an "official" kernel? What is that

The official kernel for a device is the one that stock firmware loads without having to be put into developer mode.

and just how "open source" is that very concept?

It's the phenomenon that Free Software Foundation has referred to as Tivoization: the user has the legal right to modify a computer program, but the hardware it's shipped on has technical measures to block use of a modified version. A Chromebook's stock firmware is partially Tivoized in the sense that though the blocking can be disabled, disabling it puts the user at risk of accidental data loss every time the machine starts.

Re:Chromebook firmware is partly Tivoized

[mi]

But we are still proud it is called "Linux", aren't we?

Re:Chromebook firmware is partly Tivoized

[Rob Y.]

Well, okay. But wasn't one of the main selling points of ChromeOS the fact that it's seamlessly self-updating? Why does that not include the kernel? Certainly the latest linux kernels are capable of supporting old ChromeOS hardware - but does the A/B upgrading scheme make it impossible for their seamless upgrade process to update the kernel? And even if so, couldn't they provide a less seamless alternative method of bringing your Chromebook up to date? I assume they could, and the only thing preventing it would be OEM's that want you to buy a new device.

But are there other obstacles? The current updating system is handled by Google, so I guess the OEM's would have to get involved to do a kernel upgrade, and maybe that just doesn't fit into the economic model of Chromebooks as cheap (almost disposable) devices...

Re: Chromebook firmware is partly Tivoized

[Anonymous Coward]

Don't feed the trolls.

Jailbreak it.

[Mal-2]

1. Jailbreak it by installing new firmware.
2. (Optional) Set it up to multi-boot with rEFInd, if you still want to be able to use ChromeOS.
3. Put any distro you want on it. GalliumOS is particularly tailored for Chromebooks, but Ubuntu and Mint also both work fine on mine.

You can also install Windows 8.1 or 10 (64-bit only) if you want. I haven't retained ChromeOS, but I do have both W10 and Gallium installed. Mostly I just use Gallium for those times Windows shits the bed and then refuses to let me clean up after it.

Hence Kernel version

[DrYak]

If you make a container by mounting your host libraries as writable into the container, you can of course modify code that runs as root on the host.

Actually, nope. Doesn't work.
That's partly the reason why they need a recent enough version of Linux kernel.

There's a sort of advanced setting where you map users so e.g. UID 0 is really UID 90,000. The user ID of the process creating the user namespace has all capabilities within the namespace, but none outside: it becomes root inside the container, and can't do root things outside the container. Generally, the process can only switch to another namespace if it has CAP_SYS_ADMIN in that namespace as well.

Yup, UID namespace.
Means that from the point of view of the kernel, it would be UID 90000 trying to overwrite files that belong to UID0, thus nope, can't take over root files.

But you need a recent kernel enough. The feature was introduced in 3.8 and the various filesystem started supporting it over the next few version. Hence probably the reason why Google requires 3.14

The intent was to block an evil maid

[tepples]

Say someone wants to carry a computing device but wants that device to prevent an evil maid with physical access from installing a boot-time rootkit. How would that protection measure work while keeping the owner's control?

Re:The intent was to block an evil maid

[Anonymous Coward]

Say someone wants to carry a computing device but wants that device to prevent an evil maid with physical access from installing a boot-time rootkit. How would that protection measure work while keeping the owner's control?

Maybe have the idiot owner pay attention to their own sensitive crap if it's worth protecting? This is like leaving a solid gold watch with porn in it out in Walmart's parking lot unsupervised and expecting nothing to happen to it. If it's so important to you that an evil maid is an issue, maybe you need to find somewhere else to bury those bodies.

Making the manufacturer an evil maid is also not the answer. As they are set up currently, Chromebooks can be reprogrammed by Google remotely with no interaction needed on the user's part beyond giving it power and an internet connection. The updates are signed, and sent via HTTPS. The CA certificate chain used for that connection is protected by the OS against owner modification, so it's not possible to see what exactly is being sent to the machine during an update, much less revoke or change parts of it.

Further, as we have seen and are continuing to see, there is a global push by various governments to get "law enforcement assistance" baked into the products and services of technology companies. Google itself has been targeted by this push before. If this "assistance" were to become mandatory, there would be nothing to prevent an update from being pushed to your Chromebook that allowed constant round the clock surveillance by your local government, or indeed anyone smart enough to break whatever "protection" was put in place to "ensure only lawful government agents could access the device."

Microsoft Windows falls into this catagory as well with their ARM devices due to SecureBoot, and Apple is the same story as Google. This is probably why said governments are making their push. In an effort to maximize user convenience, we've transformed the manufacturers into potential evil maids, and the various governments of the world want to cash in on our mistake.

Giving up control is never the answer to a "lack of personal security" problem. It is the answer to a "yearning for a master" problem.

Re:The intent was to block an evil maid

[Anonymous Coward]

The owner uploads his public key to the boot firmware and signs his operating system of choice with his private key. Any software not signed with this private key will not be allowed to boot. Even if the evil maid can somehow steal the private key, it cannot be used to sign the boot-time rootkit without also knowing the owner's passphrase.

Security would otherwise be the same as it is currently, where the evil maid cannot install any evil software without first wiping the device entirely.

Re:The intent was to block an evil maid

[tepples]

What keeps the evil maid from entering the admin interface and uploading her own public key? I guess that's why a Chromebook write-protects its firmware with a physical screw.

Re:The intent was to block an evil maid

[Anonymous Coward]

The password helps a little.

I don't know what locked bootloaders do now that loses access to the data if they are unlocked (a decryption key cleared at the same time?) but it can probably be the same solution as before.

What about 32 bit Chromebooks?

[Anonymous Coward]

I read somewhere that no 32 bit Chromebooks will support Linux apps. Is that true or my wrong memory?

If so, I'm out of luck, because my shiny new (...) Asus Chromebook R13 (elm) only runs a 32 bit version of Chrome OS. For... reasons, I guess.

Final Fantasy 9 device names

[cloud.pt]

wtf is going on here? There are similarly named, different-branded chromebooks, and a lot of them have FF9 (and other game characters) device names.

Want Linux? Run Crouton!

[houghi]

I run Debian with Crouton. That way I have a native Linux and if I want it, Chromebook stuff.

[Crouton] stands for ChRomium Os Universal chrooT envirONment ...or something like that. Do capitals really matter if caps-lock has been (mostly) banished, and the keycaps are all lower-case?

Re:Want Linux? Run Crouton!

[Spy Handler]

Yeah Crouton is great if you're the only person that uses (and has access to) the laptop. That's a big IF.

Otherwise, someone else could accidentally hit space bar and wipe your machine clean.

The google five year clock

[Anonymous Coward]

The google five year clock seems to start at introduction.
With a production run of at least three years many only get software updates for two years. Samsung does this too. Buy a Sammie about the time that battery and out of the box bugs are addressed and after the first update cycle you have all the (2) updates you are entitled to for an end user software update window of less than 36 hours.

it is this sort of shenanigans

[FudRucker]

is why i would rather just buy a x86_84 laptop and wipe windows off and put Linux on it, i dont want to have to depend on google for software, and can choose any distro i want

Re:it is this sort of shenanigans

[tepples]

why i would rather just buy a x86_84 laptop and wipe windows off and put Linux on it

Would you prefer a laptop on which accelerated graphics, audio, network, screen brightness, and suspend work or don't work? Because there are a lot of laptops for which these work in Windows but not GNU/Linux due to missing or broken drivers. See experiences installing Debian on an ASUS Transformer Book T100TA for example.

Precisely what I thought. Google? Safe? Lol.

[Anonymous Coward]

That’s what happens, when we let corporations like Google, Apple, etc, the Content Mafia, NSA, etc, brainwash our children with "nothing more" than mere infinitely repeated propaganda. No big conspiracy needed. "Only" an asshole repeating shit over and over and over again, until the old ones have died and the kids never knew any better.

Just like in the early 2002, *everyone* here knew that Copyright is a crime, and there is no such thing as "intellectual property", and that the whole thing harms, not helps, artists. Nowadays you'd get modded to -1, Troll, just for daring to contradict the Content Mafia propaganda.

This seems to be a common pattern since forever. "Christians" today would nail Jesus to the cross for his views. "Republicans" of today would call Reagan a socialist and shoot Lincoln for their views. And it's not that I'm taking sides here. This seems to be the case with every group. So I really wonder, how many generations of Chinese whispers I’m removed from what my own core philosophies used to be, before they were FUBAR. ...

Compare to Android 8's Project Treble

[tepples]

But wasn't one of the main selling points of ChromeOS the fact that it's seamlessly self-updating?

Chrome OS seamlessly updates its userland. For comparison, Android has been working toward this since version 8 "knOckoff of Hydrox". Just as Android 4.x largely separated Google Play Services from AOSP to update the former faster, Android 8 introduced a frozen kernel and device driver ABI called Project Treble to separate AOSP from the hardware support to update the former faster.

Why does that not include the kernel? [...] I guess the OEM's would have to get involved to do a kernel upgrade

Bingo.

The following applies to both Chrome OS and post-Treble Android: Because hardware manufacturers customize the kernel with custom device drivers, kernel upgrades require more cooperation from each hardware manufacturer than userland upgrades. Manufacturers would prefer to sell a new device. And in markets where cellular ISPs use different mutually incompatible cellular air interfaces, such as CDMA2000 vs. GSM/UMTS in the United States, kernel upgrades on devices with a cellular radio additionally require cooperation from each cellular ISP. Cellular ISPs would prefer to sell a new device with a new 24-month service commitment.

Re:Compare to Android 8's Project Treble

[Highdude702]

Is there a way to pull driver modules out of the Linux kernel? Don't they have to provide source for the kernel they compile to run on the device? Obviously IINAL so these are genuine questions that I would assume be "Yes", In which case you should just be able to load the modules into the newer kernel, work may be involved but I'm sure the OSS Community is up to it?

Re:Compare to Android 8's Project Treble

[amorsen]

Don't they have to provide source for the kernel they compile to run on the device?

My interpretation of GPL 2.0 says that yes, they have to provide the complete source code, including for device drivers. Alas, the companies disagree with me. I'm not going to sue them to force them to accept my interpretation. Are you?

Don't buy devices that come with a jail!

[Anonymous Coward]

Break out of jail?

Why would you deliberately buy a prison for yourself in the first place?
That sounds insanely dumb.

Then again, what of the things considered normal these days, isn't *insanely, mind-bogginly* retarded?

Re:Don't buy devices that come with a jail!

[Mal-2]

I didn't buy the first one, it was given to me. After coming to the conclusion that it was never going to run Windows well enough, I sold it (for $75) and bought a C720 instead (for $91). At the time, development on the SeaBIOS for the C720 was already relatively mature, and they moved to UEFI just a few months after that. I bought it because I was already aware of the jailbreak, and for my $91 (plus a larger M.2 SSD) I got quite a functional laptop. The keyboard is a little bit funky, requiring "chords" to compensate for missing keys, but otherwise it is indistinguishable from a native Windows laptop.

Re:Don't buy devices that come with a jail!

[tepples]

Why would you deliberately buy a prison for yourself in the first place?

Possibly because it's much easier to physically carry a 11.6" Chromebook than a 14" laptop.

Or because all devices in a particular market segment are jailed. Between when the iPhone came out and when the HTC Dream (first Android phone) came out, which smartphones available in the U.S. market weren't jailed? (I wanted a Nokia 900 around that time, but none of the stores I visited carried it.) Or between 1987 when VGA came out and 2007 when TVs caught up with PC output resolutions, which video game player intended for output to a television-sized monitor wasn't jailed?

GalliumOS

[friedman101]

FYI - there is a linux distribution called GalliumOS which is tailored to support Chromebook hardware. I've been running it for years on my Chromebooks. If you want linux apps, why not just install linux?

Re:GalliumOS

[tepples]

Because not everybody feels up to opening a Chromebook's case to turn its firmware write protect screw, which is what you need to do in order to ensure your Chromebook's firmware doesn't destroy the installed OS when someone turns it on and presses the Space bar.