Linux Kernel 4.13 Officially Released (softpedia.com) 43
prisoninmate writes: As expected, the Linux 4.13 kernel series was made official this past weekend by none other than its creator, Linus Torvalds, which urges all Linux users to start migrating to this version as soon as possible. Work on Linux kernel 4.13 started in mid-July with the first Release Candidate (RC) milestone, which already gave us a glimpse of the new features coming to this major kernel branch. There are, of course, numerous improvements and support for new hardware through updated drivers and core components. Highlights of Linux kernel 4.13 include Intel's Cannon Lake and Coffee Lake CPUs, support for non-blocking buffered I/O operations to improve asynchronous I/O support, support for "lifetime hints" in the block layers and the virtual filesystem, AppArmor enhancements, and better power management. There's also AMD Raven Ridge support implemented in the AMDGPU graphics driver, which received numerous improvements, support for five-level page tables was added in the s390 architecture, and the structure randomization plugin was added as part of the build system.
Re:Does it fix Ryzen crashes? (Score:5, Insightful)
I'm thinking to upgrade my computer but I want to make sure Linux can take it... Linux has never been great with support for new hardware but a poorly supported CPU really surprised me.
AMD has a recall on the crashy Ryzens. Contact customer service if you're affected.
Errata happen, but If you want to be pissed about something, let it be AMD's refusal to provide thermal management documentation. It's insane - they will eventually capitulate and release the docs, but right now they're killing a golden opportunity they've created to disrupt Intel's previous lead, because sysadmins and systems integrators need to know how hot their systems are running.
Re: (Score:2)
Jebus! What's next, Vogon poetry?
Re: (Score:2)
AMD has a recall on the crashy Ryzen chips, but only sort of. The terms of the replacement are that you have to return the chip in all of the original retail packaging, including all paper inserts, manuals, and logo stickers, with your original receipt. I don't know anyone who keeps all of this crap after they build a PC.
I always keep that stuff, and system builders are generally happy to chuck it into a bag for you and include it if you ask. Having to peel the sticker off the case is annoying, but ok.
Re: (Score:2)
Errata happen, but If you want to be pissed about something, let it be AMD's refusal to provide thermal management documentation.
Interesting. Do you have some source?
Re: (Score:2, Informative)
Old info. Its was under NDA and license was stupid, it got sorted a few days ago.
@rozhuk-im: That helped, thanks. Patch for the k10temp driver submitted upstream. Too late for the v4.14 kernel, but it will be available in v4.15.
https://github.com/groeck/lm-sensors/issues/16
Re: (Score:1)
According to numerous reports, early ryzen chips had a silicon bug. If this is hitting you, you can RMA the chip. Recent ryzen chips seem to be stable.
Still no mount events! (Score:5, Interesting)
As someone who has been working on an init system for Linux, I can assure you that there is literally no event for when a mount has occurred! The best you can do is poll /proc/self/mountinfo to see if it's changed since you last looked. Udev had mount event support but it was so buggy and wrong that they decided to remove it completely!
New processor support is nice but how about better event support for userspace programs?
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
Because it is a difficult problem to solve given this involves hardware handshakes which depend on hardware manufacturers playing ball.
You pulled that out of your ass. A mount notification would depend only on successful completion of the mount syscall.
There' also a slew of other events as well; but you can sit and complain about FOSS though. That's easy to do.
Would be nice if you bothered to inform yourself instead of posting trash.
Re: (Score:2)
Why didn't someone fix it instead of removing it?
Because a nasty patchwork of code in a place where it doesn't belong is not the answer.
Re: (Score:2)
Re: (Score:2)
So the solution is to not have the events?
No the solution is to have event generated by the system that is responsible for performing the action. That is kind of the point.
You're asking to fix something that can't be properly fixed in the place that you want it fixed.
Re: (Score:3)
If something is broken, but nobody has complained, then obviously nobody is using it and the code can be deleted.
Or at least this is a line of reasoning that is commonly used to remove features from open source software.
Re: (Score:2)
Nope. The files in /proc and /sys are virtual, so they aren't generated until they're read. That means they don't change without polling, so inotify doesn't help.
Re:Still no monitor connect/disconnect events! (Score:2)
I want an event when a monitor is connected or disconnected. I have to poll in /sys to find when that happens.
Re: (Score:2)
sysfs is generated using uevents from the kernel. That is an event you can detect via netlink. see also: NETLINK_KOBJECT_UEVENT
Re: (Score:2)
Udev can detect mount events such as the DVD being ejected by the eject button on the front of the unit. From an old bug report in launchpad.net [launchpad.net].
KERNEL=="sr0", ACTION=="change", RUN+="/usr/local/bin/sr0_change.sh"
Re:Still no mount events! (Score:4, Informative)
Udev can detect mount events such as the DVD being ejected by the eject button on the front of the unit.
No, what it can detect is a change in media. Mounting is different because you can have several partitions on a drive and mount and unmount them at different times. However, udev cannot detect any of these mounts or unmounts.
Re: (Score:2)
It's a bit like Google, Facebook, et al., where you are the product. Only in this case you are the QA. Bleeding edge distros like Fedora* and I'd wager others like Debian Sid and Ubuntu Artful will update to 4.13 and you get to be the guinea pig if you're using one of them.
* Yes, Fedora has already indicated that they're going to rebase the kernel in Fedora 25 and 26 to 4.13 after it's soaked in rawhide for a bit.
Re: (Score:2)
Re: (Score:3)
You're right, apart from the fact that being the QA is nothing like being the product of course. And you've figured out the QA part in FOSS now? Impressive since it almost says so in the definition...
BFQ/Kyber scheduler (Score:2)
It's kinda weird that to select the BFQ io scheduler I have to specify it on the boot grub line. But I can still change from noop/deadline/cfg on the command prompt.
I've been really happy with BFQ, but having to select the elevator=bfq on the grub is annoying and sometimes 4.13 hangs. Then I end up with manually editting my grub for some kernels...