Can't wait for Google to replace ext4 with F2FS by default in Android. It offers significant speed improvement (~50 percent faster, or even 2x faster in some cases I think). There are a lot of performance improvements built into the new Linux kernels, too, along with 64-bit ARMv8 support, so those will be very welcome, too, if they arrive in Android 4.4.
I think people on this site consider Samsung to be "that company that invented that gimicky feature that pauses videos when you look away" and not "the company that makes pretty much every flash memory chip or LCD in existence".
The point was that Samsung, despite what you mentioned, put out phones like the Samsung Galaxy S1 with their own home-made filesystems RFS.
The performance of RFS once the filesystem became somewhat populated was horrible and was the one single bottleneck for every single operation on the phone. It made otherwise top-end hardware (for the time) unbearably slow, bordering on useless.
Think 10 seconds lag to go back from an email to the email-listing. Those were real delays. Real problems. Caused by RFS.
I have personally experienced people thank me for giving them "a completely new phone" (quote end) when I rooted the phones and converted all RFS-partitions to EXT4. All those problems went away and you finally got the phone you thought you bought in the first place.
Given this, touching any filesystem by Samsung is something I will give serious second thoughts.
To give Samsung credit, they ditched RFS entirely for their second line of Galaxy S phones. The Galaxy S2 used standard Linux filesystems and RFS has yet to resurface anywhere near civilized folks.
>Can't wait for Google to replace ext4 with F2FS by default in Android. It offers significant speed improvement (~50 percent faster, or even 2x faster in some cases I think).
And why can't you wait? What will that buy you? Do you have a read/write heavy workflow with Android?
"The Moto X’s pseudo-stock Android 4.2.2 uses ext4 for all partitions with the exception of user data storage, which uses F2FS (Flash Friendly File System)"
> More good news is that the "longstanding squabble over Android-specific kernel features has faded completely into the background," the report said. "The much discussed 'wakelocks' [power management] feature has been quietly replaced by a different mainline solution which is used in the latest Android devices."
Can someone in the know say what this replacement is?
In fact, the article addresses your point directly:
Canonical contributed 548 changes in the period studied in today's report and would have cracked the top 30 if the foundation hadn't included "none" and "unknown" in the rankings. The maker of Ubuntu has been criticized for years for not contributing more to the kernel (and recently clashed with Intel over patches submitted to the company's graphics driver), but Zemlin [Linux Foundation Executive Director] isn't upset. "I am not at all mad at them. I think they do good work," he said. "You just kind of choose what's important to your org and where you can add the most value to the project and how it intersects with your business. They choose to focus a lot of their time and energy at higher levels of the software stack."
Canonical's contribution is more to the desktop and OS itself than to the kernel. The majority of desktop installations are either Ubuntu or Ubuntu-based, and Ubuntu is quickly becoming the server of choice as well.
Ubuntu is leading GNU/Linux adoption worldwide, which in itself is a huge contribution to the ecosystem, as it gets more programmers interested in various open source projects.
> The majority of desktop installations are either Ubuntu or Ubuntu-based, and Ubuntu is quickly becoming the server of choice as well.
Unless you are using unity, I don't see why a distribution should be called ubuntu based instead of debian based. A lot of "Ubuntu derivatives" are not based on unity.
Sure Ubuntu has a betteer installer and pretty themes, but the base is undeinably debian.
> Unless you are using unity, I don't see why a distribution should be called ubuntu based instead of debian based.
Because Ubuntu is more than Unity. I mean, by the "Unity or its not Ubuntu-based" standard you propose, Kubuntu (which is exactly Ubuntu -- sharing all the same repos -- with the KDE desktop and apps installed by default) wouldn't be an "Ubuntu-based distribution".
Heck, Ubuntu Server, since it doesn't have a default DE at all, wouldn't be an "Ubuntu-based distribution", either.
Well you'll have to show me what Ubuntu has that Debian does not. Kubuntu, I suspect, is only named because it was originally sponsored by Canonical. Nothing more.
> Well you'll have to show me what Ubuntu has that Debian does not.
Ubuntu maintains a set of repositories for each version of Ubuntu, which (even when the same--by name--software is in Debian's official repositories, will often be different versions or otherwise different contents than Debian repos).
Kubuntu uses entirely software from the Ubuntu repos from the corresponding version number (e.g., Kubuntu 12.04 uses exactly the same set of repos as Ubuntu 12.04.) You can, in effect, switch from Kubuntu to Ubuntu, or vice versa, just using the apt package manager without changing the base repositories, just installing and removing the right packages from the ubuntu repos.
The difference between Kubuntu 12.04 and Ubuntu 12.04 is entirely in which packages, from the Ubuntu "precise" repositories are installed by default. The difference between either and any version of Debian is much greater.
The whole "unity or you have to call it 'Debian-based' instead of 'Ubuntu-based'" idea is nonsense.
Surely you're joking, right? For one, Ubuntu has a desktop and server version along with a LTS version and the repos are totally different. The filesystem layout is little different between the two. I'm sure there's more like init differences, differences in installation and configuration, kernels, and stuff in /etc, but I haven't used Ubuntu in quite some time.
Ubuntu has made more modifications than simply the Unity DE. Is Linux Mint (the regular edition) fully compatible with Debian? Does a Debian server have the same features as an Ubuntu server?
Highly debatable. "Most accessible" installer perhaps, but as far as I am concerned, Debian's netinst has yet to be beat. Style isn't exactly high on my list of priorities during installation, it's not like that is something I do to show off in coffee-shops.
Ubuntu also has a mini iso if you really want the net install. For most people, the net install will take significantly longer than a regular install as each package has to be unpacked individually.
A quick Google search will produce this information. Furthermore, I think a survey of new server deployments would be even more in Ubuntu's favour, though gathering such information is no doubt difficult.
In the article it does mention that Ubuntu is the most popular distro for Amazon deployments...
The article cites Cloud Market's data on AMIs, which indicate how many AMIs have been derived from the Ubuntu base AMIs, rather than actual deployment numbers. It's an interesting indicator...
Having been on the inside (EC2) and seen the numbers, I can confidently say that Ubuntu's AMIs are the most popular AMIs deployed on EC2. The stock AMIS take the top spots by an overwhelming majority, and most of the lower spots are filled with derivatives of the original Ubuntu AMIs.
Ubuntu's been working hard in the server/cloud space for a while, between the JeOS effort and subsequent, excellent tools for building customized installations around it. It's not just a brand or that makes them popular; with Ubuntu you get a sane system, style, and toolset for your system images.
Some anecdotal evidence - our university is running free Linux VM's for research purposes (so, non-graphical remote servers), people can choose what they want, and the majority seem to choose Ubuntu; with CentOS being second place.
Canonical is probably primarily focused on recreating the Mir wheel and other Unity/Ubuntu Phone/etc. I doubt a whole hell of a lot of that goes upstream.
Seeing so many chip makers up there on the list makes me wonder about the state of Linux drivers. Why is it that Android and Chrome OS and Windows can make drivers actually work, but Linux on laptops will still predictably run hot and drain the battery like crazy?
Because the hardware and OS companies are working together on this, and the hardware doesn't ship until it works in its target OS. In Linux, volunteers are poking around blindly tryign to get the hardware to do what they ask while the hardware vendor has all the information and isn't sharing it.
That's what I used to think but now that I see that there are that many chip makers contributing I'm starting to wonder. But it's probably as varikin says, they're working on different kinds of drivers.
Much of the hardware for the higher end Android devices is composed of stuff from a handful of OEMs, so that also helps. Also many of the devices that come out around the same time are using at least some identical components.
SoCs are all mostly Qualcomm or Samsung, radio chipsets are mostly Qualcomm, LCDs are Sony/Samsung/Sharp/LG and a few others, AMOLED are all Samsung as far as I know, GPUs are mostly Qualcomm/PowerVR/Nvidia.
Much of the time, the Android modding community ports newer versions of Android to other devices simply by hoping some OEM releases a newer version of Android for another device that happens to have some identical hardware component so they can either drop that particular driver into the other device or tweak it a bit. Porting the OS is never really the problem. The problems generally lie in outdated proprietary drivers no longer working with the newer version of Android (Camera/Bluetooth/Radio/sometimes GPU are the most likely candidates to break).
I would expect the chip makers are putting more effort into getting server hardware working than consumer. Thing such as raid cards, high end network cards, motherboards, etc. Also possibly for dedicated off the shelf appliances the routers and such not. The types of things that are not cost efficient for minimizing to a laptop or commodity hardware.
Microsoft have lots of driver issues too: hardware manufacturers build buggy drivers that break Windows. They're making progress on this via WHQL and Windows Logo labelling. Unfortunately it's not obvious that a "Works on Linux" logo scheme will have much effect without wider consumer adoption (which, in classic chicken/egg fashion, is hindered by shonky drivers)!
I'm glad Linus and many others are reviewing Kernel changes still, but how long will it be before there is nothing stopping massive corporations from deciding what can be changed?
And even with all of the experienced eyes, there have to be changes that have crept in that they didn't realize could be misused, and may still not know.
I appreciate that corporations are the biggest donators, and that is extremely important, but is there any point in which we say, "Get your own project?"
Would you prefer hardware manufacturers not to contribute drivers? And that these companies fork the kernel without providing anything back? That's not very much in the spirit of openness.
(Regardless, I'm not even sure how number of contributions would make any contributor "unstoppable".)
Is there a break down of what get's worked on by whom?
Would be interesting to see what the hardware manufacturers are doing other than their own drivers. Probably fairer to if we're counting LOC...
