Linux 4.14 Has Been Released (kernelnewbies.org)

diegocg quotes Kernel Newbies: Linux 4.11 has been released. This release adds support for bigger memory limits in x86 hardware (128PiB of virtual address space, 4PiB of physical address space); support for AMD Secure Memory Encryption; a new unwinder that provides better kernel traces and a smaller kernel size; support for the zstd compression algorithm has been added to Btrfs and Squashfs; support for zero-copy of data from user memory to sockets; support for Heterogeneous Memory Management that will be needed in future GPUs; better cpufreq behaviour in some corner cases; faster TBL flushing by using the PCID instruction; asynchronous non-blocking buffered reads; and many new drivers and other improvements.
Phoronix has more on the changes in Linux 4.14 -- and notes that its codename is still "Fearless Coyote."

Comments

Which is it?

[93 Escort Wagon]

4.14 or 4.11?

(I expect the summary will eventually get fixed, followed by someone replying to me “WTF are you talking about?”)

Re:Which is it?

[Anonymous Coward]

I expect the summary will eventually get fixed

A summary actually getting fixed? WTF are you talking about?

Re:Which is it?

[fisted]

It's 4.11 for Workgroups.

Typo

[Bruce Perens]

That's TLB flushing, not TBL.

Re:Typo

[Misagon]

... and PCID is not an instruction. The feature means that there is a "process ID" tag on each entry in the TLB to avoid having to flush them unnecessarily.
The intended benefit is that all entries would not necessarily have to be reloaded from page tables in RAM (or cache) whenever there is a context switch.

"Tagged TLB"s have been available on other CPU architectures for decades -- and have been used by the Linux kernels for those architectures. The feature is pretty recent on Intel x86 CPUs though.
Correct me if I'm mistaken but I think AMD's x86 CPUs do not have PCID specifically but has support for "virtual machine ID" tags on the hypervisor's second-level TLB.

Bigger memory limits

[lobiusmoop]

"Original x86-64 was limited by 4-level paging to 256 TiB of virtual address space and 64 TiB of physical address space. People are already bumping into this limit: some vendors offers servers with 64 TiB of memory today. "

64TB RAM... fuck.

Re:Bigger memory limits

[Anonymous Coward]

> OTOH, I have my doubts that they actually have 64TB of RAM.

I don't. It seems that you can fit 12TB of RAM (128GB*96) into this fairly standard high-end server: http://www.dell.com/en-us/work/shop/povw/poweredge-r930 . I expect that there are niche vendors that sell absolutely _massive_ machines for people who absolutely _must_ work with huge datasets in memory.

Re:Bigger memory limits

[Anonymous Coward]

Or use Firefox.

Umm...

[EmeraldBot]

Linux 4.11 was released last May. 4.14 is the version that's coming out today.

Re:USB drivers still in kernel?

[filesiteguy]

windows nt 4 doesn't have it. i need to upgrade someday.

Re:USB drivers still in kernel?

[drinkypoo]

Amusingly, NT4 is where they merged the Kernel and GDI memory spaces in pursuit of graphics performance. Well, they got it, but they also absolutely destroyed NT's reliablity. 3.51 was a rock. Granted, a rock with a 2GB filesystem limit...

Kernel/User: Linux is mixed

[DrYak]

Actually it's both.

You can write a .ko that will be loaded by the kernel to handle your device
(used on most Linux for a few things where speed matters, like mass storage, network.
or for booting simplicity like mouse/keyboard/bluetooth)

Or you can write an user space device that communicates with the raw USB device using libusb.
(used on the huge variant zoo of non critical USB devices, like scanners, firmware upgrader, etc.)