Linux 3.18 Released, Lockup Bug Still Present

jones_supa writes As anticipated, Linus Torvalds officially released Linux 3.18. The new version is now out there, though that nasty lockup issue has still yet to be resolved. Dave Jones is nearing the end of dissecting the issue, but since it also affects Linux 3.17 and not too many people seem to get hit by the lockups, Linus Torvalds decided to go ahead and do the 3.18 release on schedule. Linus was also concerned that dragging out the 3.18 release would then complicate the Linux 3.19 merge window due to the holidays later this month. Now the Linux 3.19 kernel merge window is open for two weeks of exciting changes.

Re:Anyone know what hardware the lockup bug is...

[Anonymous Coward]

systemd. So the latter two.

Re:

[Anonymous Coward]

My understanding is that they are still trying to narrow it down, but in the process have actually found multiple bugs.

Re:

[Anonymous Coward]

lol - the 'many eyes' that can examine the source code are only of use when they are actually looking at it.

Re:

[jones_supa]

Guys, do you ever read LKML where this stuff happens? Your GP comment said that they have found multiple bugs. No, they haven't. They are dealing with the very same bug where the watchdog detects a soft lockup. There is no debugger attached at this point, they are just doing bisecting (another powerful tool of course) to find what was the specific patch that introduced the bug. Browse through the LKML archive [iu.edu] and search for the subject "frequent lockups in 3.18rc4".

Re:

[Cley Faye]

No problem, he's not saying there are bugs in the Linux kernel. He's saying there were bugs in the Linux kernel.

Re:

[armanox]

I've had either this or something similar appear on my AMD box at home (FX-8120, 16GB RAM, 990FX Chipset).

Re:

[chmod a+x mojo]

Hmmm, that's interesting... pretty sure the original bug filer was using an AMD quad core since several mails had been about core 3 / core 4 being idle at the time of crash. I may be wrong though, I only read through it at the time it was posted on /.

Makes you wonder if it is a AMD multiple of 4 core bug somehow though, and maybe fixable through microcode patching.

Re:Anyone know what hardware the lockup bug is...

[Anonymous Coward]

Bug happens on some special hw when one runs trinity (a system call fuzzer) long enough. So what? I have been running 3.18 from the start of its merge window and have not seen any problems in it. Instead of sensational click bait article from phoronix, I would rather see a link to kernel newbies or lwn with description what has been changed in this version and what is new.

Re:

[Anonymous Coward]

Both you and parent have a point imo, but you seem to be overreacting.

Yes, a bug that affects almost nobody IS less of a problem. It was present in 3.17 and nobody was hit hard enough to even report it.
This is not comparable to projects like GNOME making major controversial changes willingly and repeatedly. This is some odd bug that'll get bisected and fixed soon, that will never affect you, and that doesn't derve all that much press.

Better links

[lanroth]

LWN: https://lwn.net/Articles/62497... [lwn.net]
kernelnewbies: http://kernelnewbies.org/Linux... [kernelnewbies.org]
omgubuntu: http://www.omgubuntu.co.uk/201... [omgubuntu.co.uk]

Re:

[account_deleted]

Comment removed based on user account deletion

Re:

[__aaclcg7560]

We don't need no stinkin' Windows! Get a Mac!

Re:

[Anonymous Coward]

Please, never user a Mac.
On desktop, Linux for work, Windows for entertainment.
On servers use Linux, and let the Mac for those idiots that think they think different for using a overpriced closed and unstable system (OSX crashes more than Windows, since due to the way OSX is coded it is trivial to write a program that can block the whole OS and require a hard reboot, eg: by forcing a system deadlock, which lots of sloppy coded Mac apps suffer from, something that you cannot do in Windows, the only way I'v

Re:

[Thong]

(Apple marketing victim)

Re:

[__aaclcg7560]

Obviously spoken by someone who never owned a Mac. Until the CPU fan gave out this summer, I had a Black MacBook (2006) that ran fine for the last eight years. Alas, with a 32-bit processor that's no longer supported, I didn't take it to the Apple Store to get fix. I had no problems switching over to Windows 8.1, as my data was in cross-platform formats. I'm looking forward to a new Mac next year.

Re:

[Opportunist]

Well, Linux Users will probably have to learn a lesson from the Windows User community: When the current version sucks, just use an earlier version and ride it out 'til a new and hopefully better one arrives.

Linux on the desktop

[nwf]

This is just the sort of bug to get people to adopt Linux on the desktop, since it will be more similar to what they expect from Windows.

Re:

[Opportunist]

About time! Those Linux geeks had it way too easy anyway for far too long!

---signed, the Windows Server Team.

Re:

[NitsujTPU]

This reminds me of the plotline from Userfriendly where one of the techs develops an OS called "The OS That Doesn't Suck." He eventually gives up this pursuit when it, too, begins to suck.

Re:

[Anonymous Coward]

That would be Absurd Notions [absurdnotions.org], not User Friendly.

Re:

[NitsujTPU]

Ack! Sorry about that! I knew it was an old old webcomic, and that it happened around the time when there were just millions of new UNIX-style operating systems with everyone claiming that theirs was good for a different reason.

Re:

[spongman]

wait, Windows locks up randomly now?

Re:

[thejynxed]

Usually only when it involves an nvidia or AMD GPU driver :)

Re:

[sound+vision]

Graphics drivers haven't caused the system to lock up since Windows XP and the new driver model. Now if your graphics driver chokes, you get a second or two of black screen, then everything comes back up normally with a balloon in the system tray telling you it crashed.

If you mean complete lockups or bluescreens, I've had two of those over five years of running Windows 7, and they were both when I was overclocking (none since).

Not until Linux version 10

[Crazy Taco]

This is just the sort of bug to get people to adopt Linux on the desktop, since it will be more similar to what they expect from Windows.

Not me! I refuse to use software as immature as version 3 of Linux. Mac is on OS version 10, Windows is about to release version 10, and by golly, I'm not wasting a second of my time on Linux until it catches up!

Re:

[kmoser]

So you're saying 2015 will surely be the year of the Linux BSOD?

Don't use Bleeding Edge Kernels

[Zombie Ryushu]

Don't use Bleeding Edge Kernels on Stable Production systems. I still use Kernels for 3.10, and 3.12 with security updates. I can't have Intermittent unexplained lockups.

Re:

[TeknoHog]

I like pine kernels in my salad, and I PINE for the good old days when men where men and used real email clients, you sensitivity-challenged clod!

Re:

[dbIII]

Alpine is still out there and works on current platforms. I use it on a mail server to forward quarantined attachments to people when antimalware stops them posting executable files to each other. Mutt isn't bad either.

Re:

[TeknoHog]

Yup, I've used Pine since the 1990s and I continue to use Alpine.

"I use pine - not because its necessarily the greatest email reader ever, but because I'm used to it, and it does what I need it to do with a minimum of fuzz." -- Linus Torvalds

Re:

[Anonymous Coward]

Don't use Random capitals in your Sentences. I don't Capitalize random words in My sentences. I don't want To look like I am Schizophrenic.

Re:

[ChrisMaple]

Capitalization is the poor man's bold.

Re:

[marcello_dl]

I have used 4 versions of 3.17 of void linux till now as my primary web and office desktop, not one lockup.

Given that void doesn't feature it, it's obviously a systemd issue. :D

Re:

[sound+vision]

Last I heard (admittedly it was on a Slashdot article so it could be wrong) the bug has been present for many years, since the 2.x kernels.

Re:

[jones_supa]

Citation? You might be confusing this with some other bug.

Re:

[sound+vision]

http://linux.slashdot.org/stor... [slashdot.org] "The problem was first suspected to be related to Xen. A patch dating back to 2005 was pushed for Xen to fix a vmalloc_fault() path that was similar to what was reported by Dave. The patch had a comment that read "the line below does not always work. Needs investigating!"

Re:

[jones_supa]

Well, in the same summary it reads that Linus concluded that it's not a Xen bug. :)

Dissecting != Bisecting

[Anonymous Coward]

Bisecting is a technical term with a specific meaning. Essentially, he is using a binary search through git history to identify when the bug was introduced. Dissecting does not convey this. But I guess it is too much to ask to have the submitter both read and understand TFA.

Re:

[jones_supa]

Bisecting is a technical term with a specific meaning. Essentially, he is using a binary search through git history to identify when the bug was introduced. Dissecting does not convey this. But I guess it is too much to ask to have the submitter both read and understand TFA.

Submitter here. In this case samzenpus has to take the blame, because in my original submission [slashdot.org] the word is "bisecting". :)

Re:

[a_n_d_e_r_s]

Yeah, the editors of slashdot are not fluent in computer science lingo that is common within open source where one work with many different version and need to make sure regression not happen e.g. the linux kernel.

Also if editor had googled the word this statement is visible in the Wikipedia page about bisecting:

"Not to be confused with Dissection."

http://en.wikipedia.org/wiki/B... [wikipedia.org]

Dissection is mostly known as a biological term about using an scalpel to examine plants, animals or humans.

I'm just special I guess

[wilberfan]

I don 't know whether to feel pissed or special (or both), but I've been experiencing this lockup on my lowly little Arch box. (Reverting to a previous kernel always clears it up.) I couldn't find anything in the Arch forums about it, so I appreciate being kept abreast of it here! :)

The thing about BSD.....

[Anonymous Coward]

It is official; Netcraft now confirms: *BSD is dying

One more crippling bombshell hit the already beleaguered *BSD community when IDC confirmed that *BSD market share has dropped yet again, now down to less than a fraction of 1 percent of all servers. Coming close on the heels of a recent Netcraft survey which plainly states that *BSD has lost more market share, this news serves to reinforce what we've known all along. *BSD is collapsing in complete disarray, as fittingly exemplified by failing dead last in

Re:

[savuporo]

Android, and Chromebooks.

Re:

[jfdavis668]

And IBM continues to dominate the mainframe field.

Re:

[Zontar The Mindless]

Wow, APK managed an entire post without bolding or all-caps... Colour me surprised.

Re:

[jopsen]

Why all the Linux zealots modding good posts down, mod parent up! Great comment. I agree Android is pretty much the only good Linux software out there these days.

Yes, let's use windows on the server, ha ha :)

Re:

[jones_supa]

There is some wisdom to that comment. Instead of trying to build enterprise-level systems with lack of developers, the open source community might wind up with more stable system by sticking with something more simple. Linux Mint grabbed GNOME2 which became MATE, and it has been hugely successful. Simple, fast, stable.