Slashdot Mirror
Linux Kernel 4.5 Officially Released
prisoninmate writes: Yes, you're reading it right, after being in development for the past two months, Linux kernel 4.5 is finally here in its final production version. It is internally dubbed "Blurry Fish Butt" and received a total of seven RC builds since January 25, 2016. Prominent features of Linux kernel 4.5 include the implementation of initial support for the AMD PowerPlay power management technology, bringing high performance to the AMDGPU open-source driver for Radeon GPUs, scalability improvements in the free space handling of the Btrfs file system, and better epoll multithreaded scalability. The sources are now available for download from kernel.org.
Update: 03/14 13:24 GMT by T : Reader diegocg lists some other notable features (a new copy_file_range() system call that allows to make copies of files without transferring data through userspace; support GCC's Undefined Behavior Sanitizer (-fsanitize=undefined); Forwarded Error Correction support in the device-mapper's verity target; support for the MADV_FREE flag in madvise(); the new cgroup unified hierarchy is considered stable; scalability improvements for SO_REUSEPORT UDP sockets; scalability improvements for epoll, and better memory accounting of sockets in the memory controller), and links to an explanation of the changes at Kernel Newbies.
Are you seriously still complaining about this?
The last odd kernel was released in 2003. In 2.6 stable and unstable merged. Kernel 3.0 wasn't even a drastic change.
https://lkml.org/lkml/2011/7/2...
Between 2.4 and 2.6 linux was drastically overhauled. We really don't need separate non stable versions anymore. Plus version control is far better now.
Suspend and hibernate works just fine on laptops designed to run linux (e.g. chromebooks), the same can be said of macosx - suspend and hibernate is perfectly reliable on apple laptops, but is usually flakey on a hackintosh.
Linux already has various kludges to emulate the nonstandard way in which windows handles power management, but laptops also often come with customised model-specific drivers so even if you run windows you often still have problems if you run the default drivers or drivers for the chipsets rather than the specific laptop model.
The lower end laptop makers also make things difficult for users by varying the hardware in the same model, when looking at laptops recently i was told that a given model could have any one of 3 different wifi and ethernet chipsets, and that i wouldn't know which until i physically took delivery of the laptop... They will guarantee that you get "an 802.11ac wireless card" and "a gigabit ethernet", but the performance, range, stability or cpu usage can vary wildly between chipsets as can compatibility with linux or other systems and even (albeit quite niche) features like wireless monitor or master modes are not available with some chipsets.
The chipsets in use for various components were always an important factor for me when deciding what to purchase.
Yes it's a huge nasty mess!
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
As I recall, the older versions of the kernel added some features and offered others as third party patches. The 2.0.x, 2.2.x, 2.4.x, and 2.6.x branches were supported for a very long time, (...) This was a very successful development model for over a decade and I don't understand why that's changed. Arguably, it would require fewer work since there wouldn't be as many branches to maintain.
Actually it wasn't. Distros were massively cherry-picking from the odd-numbered branches creating huge variations from the stock kernel, creating strange bugs and making the big jumps was a huge pain because the more relaxed requirements to put it in a development branch led to poor quality. Linus tightened ship and basically said do development in your own branch, when it's ready merge it to the released kernel and ~2 months after the merge window closes it'll be released instead of years like the old kernel. No more hodge-podge development kernels full of half-assed changes. It got distros to work more on the upstream kernel than their own variations, leading to more manpower and higher quality in the core project. It was a great success.
Live today, because you never know what tomorrow brings
One can obviously debate the value of this; but my impression is that, while in an ideal world they'd like to be able to support more hardware, when it comes down to it the Linux project really isn't interested in just providing a cheap OS for people to put on top of their giant heap of binary blobs. They certainly aren't as hardline as the FSF; but even the 'engineer-pragmatist' types see binary blobs as an impediment to their work(that's why the kernel taint flag exists: if your system crashes, you have a potentially interesting bug report; but if it was 30% Nvidia blackbox by weight at the time, they don't want to waste their time shooting in the dark).