Forgot your password?
typodupeerror
Open Source Linux

Linus Torvalds Suspends Key Linux Developer 641

Posted by Soulskill
from the arguing-about-penguins dept.
alphadogg writes: "An argument between developers of some of the most basic parts of Linux turned heated this week, resulting in a prominent Red Hat employee and code contributor being banned from working on the Linux kernel. Kay Sievers, a well-known open-source software engineer, is a key developer of systemd, a system management framework for Linux-based operating systems. Systemd is currently used by several prominent Linux distributions, including two of the most prominent enterprise distros, Red Hat and SUSE. It was recently announced that Ubuntu would adopt systemd in future versions as well. Sievers was banned by kernel maintainer Linus Torvalds on Wednesday for failing to address an issue that caused systemd to interact with the Linux kernel in negative ways."
This discussion has been archived. No new comments can be posted.

Linus Torvalds Suspends Key Linux Developer

Comments Filter:
  • by Lisias (447563) on Friday April 04, 2014 @11:41AM (#46661329) Homepage Journal

    And this is good.

    Quote from the Linus email:

    Kay - one more time: you caused the problem, you need to fix it. None of this "I can do whatever I want, others have to clean up after me" crap.

    Being Kay a Red Hat paid developer, perhaps it's not his entirely fault what's happening. But it's his name on the table, so it's his responsability nevertheless.

    • > you caused the problem, you need to fix it This is true in ANY development company/organization/group. In my first days, they even call me at my night to ask me to fix a build that I broke due don't run proper checking/testing.
    • by x_t0ken_407 (2716535) on Friday April 04, 2014 @12:46PM (#46662075) Homepage

      I read the mailing list thread as well as the bugzilla report [freedesktop.org]...Kay certainly was certainly being a complete dick here. Too many people will see this as "an asshole being an asshole" w/respect to Linus, but he actually had a reason [this time, lol].

      • by Bill Dimm (463823) on Friday April 04, 2014 @12:56PM (#46662175) Homepage

        but he actually had a reason [this time, lol]

        I've read about quite a few of these "Linux blow-ups" over the years, and I can't think of a single instance where I cam away thinking Linus was anything short of fully justified once you actually looked at the context.

        • by Sun (104778) <shachar@shemesh.biz> on Friday April 04, 2014 @01:59PM (#46662881) Homepage

          Blowing up at Andrew Tridgdell [theregister.co.uk] after he "reverse engineered" (i.e. - sent "help" on a telnet connection) the bitkeeper protocol, causing bitkeeper to withdraw support from the kernel.

          Personally, I think bitkeeper were just waiting for an excuse to do that. Their business justification was quickly eroding. The needs of the kernel and the needs of their commercial customers were drifting apart. Supporting the kernel was becoming a liability, rather than an asset, to them. That's also the reason, I think, that they were so quick to withdraw all support after such a minor infraction.

          Shachar

    • by Rich0 (548339) on Friday April 04, 2014 @01:48PM (#46662745) Homepage

      Being Kay a Red Hat paid developer, perhaps it's not his entirely fault what's happening. But it's his name on the table, so it's his responsability nevertheless.

      Somehow I doubt RedHat is paying him to abuse people in their bugzilla. I suspect they tolerate him because he is a rock star. In some sense Linus just put them at a disadvantage competitively so it is now more in their interest to reign things in.

      If I posted something like that on a forum owned by my employer or using an email address that named my employer I'd get a strong reprimand at the very least. I'd like to think that they wouldn't fire me over it on a first offense, but no doubt it would cross their mind, and if I kept it up I'd be gone for sure (and rightly so). Kay's reputation isn't the only one at stake.

  • by tomhath (637240) on Friday April 04, 2014 @11:42AM (#46661347)

    his complaint about systemd has been widely echoed in the Linux world, with prominent contributors like Ingo Molnar, slamming the “excessively passive/aggressive” attitude of the project’s maintainers.

    If you ignore requests you piss people off. Sounds like banning the guy was the right thing to do.

  • Misleading title... (Score:5, Informative)

    by egarland (120202) on Friday April 04, 2014 @11:45AM (#46661391)

    "I'm not accepting any patches until you fix your bugs" is hardly suspending someone, it's re-focusing them. This is an important part in any software project, and Linus is doing it well here. There's no ambiguity or hyperbole, just straightforward communication identifying issues and prompting action to correct them.

    "Start fixing your shit" isn't even remotely the same thing as "stop doing things".

    • by evilviper (135110)

      "I'm not accepting any patches until you fix your bugs" is hardly suspending someone,

      Only because that's an inaccurate misquote. Let's try the real thing:

      "I will *not* be merging any code from Kay into the kernel until this constant pattern is fixed. This has been going on for *years*, and doesn't seem to be getting any better."

      That's not a "fix this bug first" message... That's a much more general and sweeping "you suck, so you're fired," message.

      Of course both Kay and Linus reserve the right to change

  • Maybe people should be paying attention to the sources of the crap they're trying to wedge into our stable systems.
  • Inaccurate summary (Score:5, Informative)

    by gwstuff (2067112) on Friday April 04, 2014 @11:49AM (#46661435)

    First the idea of "Suspending" a kernel developer is inane. Kernel developers don't work for Linus. Anyone can fork the kernel and work on his own version of it. Furthermore, Kay can write code that other people audit, modify and submit further.

    Secondly, it's not an 'indefinite, unconditional ban' as suggested by the summary. Here's the specific line from Linus' email:

    Greg - just for your information, I will *not* be merging any code
    from Kay into the kernel until this constant pattern is fixed.

    In other words he might start accepting patches from him if he changed his style of operating.

    • Anyone can fork the kernel and work on his own version of it

      Yes, only hardly anyone ever does it with projects as massive as the Linux Kernel. Just look at how many years it took for X.org to split from X11 - or heck, even the many half-assed and failed attempts at forking off Slashdot over the Beta issue.

      It takes a lot of fustration building up for many years in many talented and dedicated people to pull off a big fork successfully.

      • Yes, only hardly anyone ever does it with projects as massive as the Linux Kernel.

        Massive ... Kernel

        Seems like a contradiction.

        or heck, even the many half-assed and failed attempts at forking off Slashdot over the Beta issue.

        http://soylentnews.org/ [soylentnews.org] seems to be working just fine. Neither half-assed, nor failed. Fewer posters, for sure. But a far more friendly vibe than today's Slashdot.

    • First the idea of "Suspending" a kernel developer is inane.

      Of course. It just means that Red Hat got a ... red card?

  • by rs1n (1867908) on Friday April 04, 2014 @11:54AM (#46661517)
    Kay was not banned. Linus simply said he would not merge anything from Kay [b]until[/b] he got his act together.

    Greg - just for your information, I will *not* be merging any code from Kay into the kernel until this constant pattern is fixed.

  • The message [iu.edu] to which Linus responds is also interesting:

    Short story:

    The systemd guy uses the debug keyword on kernel command line to spool a huge log - which can hang the boot process, and that is the problem.
    Then the same guy claims that the debug keyword is generic so it can't be reserved by the kernel, even if it's been used first by it since a long time...

    I can say that Linus is right there, for sure. He's maybe too kind...

    • Re: (Score:3, Insightful)

      by rs79 (71822)

      (Please correct me if I misunderstand the problem, it'e been years since I worked on this stuff)

      It seems to be both guys are right. That is, in an ideal world starting with a black sheet of paper then it seems to me Kay is almost certainly correct.

      But, this does not mean Torvalds is wrong - breaking legacy systems because of a code change that interprets existing config files is "a bad thing".

      So, we fall (ONCE AGAIN) into the trap of living with cruft to support legacy stuff. Just like we have to live with

      • by bruce_the_loon (856617) on Friday April 04, 2014 @01:37PM (#46662639) Homepage

        The problem is systemd is a johnny-come-lately and is violating the standard way of doing things, even if the standard way isn't the most optimal. Think of it like a court of law, no court is going to accept a junior lawyer changing terminology that has been in use for centuries just because the lawyer has read a thesaurus. The impact is just too large.

        To take it further, apparently all but two parameters (debug and quiet) that systemd recognizes are prefixed by systemd.xxxxxx, so they know how to work within the kernel standard.

        The kernel has for a long time had a protocol of parameter naming. Direct kernel parameters are plain, module-specific parameters have mod.xxxx format and that was designed to pass driver-specific parameters in. SystemD, being a child process and not even part of the kernel should respect the existing protocol and ignore any parameters not passed without a leading systemd.

      • by Kjella (173770) on Friday April 04, 2014 @03:05PM (#46663597) Homepage

        (Please correct me if I misunderstand the problem, it'e been years since I worked on this stuff) It seems to be both guys are right. That is, in an ideal world starting with a black sheet of paper then it seems to me Kay is almost certainly correct.

        No, there's no sane world where Kay is right. There can only be one user of a global namespace, if everything made switches or settings there it would be chaos like in this case where it's impossible to turn on kernel debugging without turning on systemd debugging (though you can redirect systemd to a null log target as a workaround). If anyone is to use it, it is clearly the kernel since it's the kernel's command line and it has 20+ years of claim to it. And without any standard as to what those switches or settings should to, it's meaningless to have the global namespace act as global variables and move the kernel's parameters to a namespace. But even if the kernel moved debug to kernel.debug it still means systemd should use systemd.debug, under no circumstances should it use the plain "debug" like it does now. He's doing it wrong and even if he was right, he'd still be doing it wrong.

      • No, the fact is that Kay wants an exception to standard kernel behaviour merely in order to suit his "vision" for systemd. Except it sounds more like pure rationalisation/unwillingness to admit that he fucked up big time.

        I'm not a kernel dev, yet I still know that non-namespaced arguments to the kernel belong to the kernel. Kernel debug is "debug"; systemd is not part of the kernel and thus its debug should be "systemd.debug".

      • by robsku (1381635)

        It's a *kernel parameter* to enable debug log printing, not "some legacy cruft" that's been deprecated or supported for whatever (er, what backwards compatibility would it preserve? Kernel developers would change it if they would think there were any good reasons for *debug* *kernel* *parameter* to enable debug printing).

        systemd is outside kernel, but it is launched by kernel and as such the developers can pass parameters to it only via kernel command line - However there is specifically a method meant for

    • by bill_mcgonigle (4333) * on Friday April 04, 2014 @02:58PM (#46663507) Homepage Journal

      Then the same guy claims that the debug keyword is generic so it can't be reserved by the kernel, even if it's been used first by it since a long time...

      Linus made this argument in a different forum yesterday (paraphrasing from memory): "Look, something has to be authoratitive when it comes to parameters. On a linux system, that's the kernel".

      Which is aribitrary, but not without merit.

      Here's the rub, that causes Kay's downfall: he's arguing for namespaces (systemd.debug, kernel.debug, etc.) but he's consuming 'debug' without a namespace and complaining that the kernel isn't using namespaces, even though he knows the linux policy is "don't break people's programs".

      Kay could simply take his own advice, only consume systemd.foo parameters, and lead by example, without trying to claim the null namespace for systemd. One is left to conclude that he's doubling down on a bad decision rather than actually wanting to fix things.

      Linus is not the only kernel developer who is torqued at the systemd guys. It appears to many players that they're trying to become the userspace kernel but haven't quite earned their stripes yet and are leaving a trail of unhappy developers and sysadmins in their wake, as legitimate complaints about breakage are handwaved away.

      systemd isn't perfect, but it's not terrible either, and there's no good reason for such a level of discontent in the community. Many proponents say, "those dinosaurs don't want change", but that's not at all what the trouble is about. It would be silly to fork systemd at this point, but that's what some people are talking about, and it's purely for personnel reasons. A mutually beneficial resolution to this and other problems with systemd is in everybodys' best interest.

      • by symbolset (646467) *
        It seems to me that there is no possible way that systemd causing the system to hang on the use of the standard kernel debug parameter can be seen as useful for the purpose of systemd. This then would be Kay trying to force Linus to adopt his kernel namespace idea, breaking stuff in Linux that goes back many, many years.
  • by rabtech (223758) on Friday April 04, 2014 @12:10PM (#46661673) Homepage

    Linus is generally fair from what I can tell, and does not except himself from criticism. In that very thread:

    Yeah, what Andrew said. My suggestion of per-task or per-cred is
    obviously moronic in comparison.

    Linus "hangs head in shame" Torvalds

    Someone proposed a better idea and Linus immediately admits his idea was worse and moves on. That was also one of Steve Jobs' greatest talents, even though it's in a completely different sphere. He originally said "no" to iPods for Windows and the iOS app store. People presented their case and he changed his mind.

    We should all be so willing to admit when someone else has a better idea or we were wrong.

    • by jklovanc (1603149) on Friday April 04, 2014 @12:45PM (#46662067)

      Someone proposed a better idea and Linus immediately admits his idea was worse and moves on. That was also one of Steve Jobs' greatest talents,

      That statement is laughable. When Steve Jobs was in control of NeXT he decreed that the cube must be a perfect cube. Pressing, the most efficient way to create the case, works best with the side a couple of degrees off. Most people would not notice the difference but it allows the cube to be pulled out of the die much easier. Even though only one company in the US was able to press a perfect cube and it would add a couple hundred dollars to each machine, Jobs insisted on it. That was one of the reasons the NeXT was so expensive.

      He originally said "no" to iPods for Windows and the iOS app store. People presented their case and he changed his mind.

      When Jobs came back to Apple they had a Board that could stand up to Jobs and make decisions counter to Jobs' wishes. The Board had been doing it for years while Jobs was failing at NeXT. I doubt Jobs "changed his mind". More likely the Board overrode him.

  • Bullshit Summary (Score:3, Informative)

    by johnsie (1158363) on Friday April 04, 2014 @12:13PM (#46661693)
    They just aren't accepting code from him until he fixes that issue. The summary makes it sound much more dramtic than it really is.
  • by Virtucon (127420) on Friday April 04, 2014 @12:24PM (#46661799)

    This story somehow reminds me of the Seinfeld Spongeworthy episode. Elaine finds out that the contraceptive sponge is no longer manufactured so she hoards them and then chooses her lovers based on the fact of whether or not she thinks they're Spongeworthy. I think in this case Linus doesn't have enough sponges left to waste.

  • by erroneus (253617) on Friday April 04, 2014 @12:28PM (#46661839) Homepage

    I really and truly respect that.

    I have had more than enough experience in dealing with "this is how I do it!" developers. We're talking about writing code -- a set of instructions to accomplish things usually performed by an electronic machine. Things definitely become complex and even confusing at times, but it's NOT MAGIC. And when people need to work with developers and developers with developers and all that, I have run short on patience where some developers believe they are the thing and not the project or the community affected by the project.

    To me, the community which uses the project or is affected by the project is the thing. If you write for results, then you agree. If you write to make yourself proud? You're just a bit too self-absorbed. (I'm not saying there's anything wrong with being happy and proud of your work, but what you want should never be the thing.)

    I just wish Linus would go kick some GNOME team ass and share some wisdom with GIMP developers as well.

  • systemd Architecture (Score:5, Informative)

    by jgotts (2785) <`jgotts' `at' `gmail.com'> on Friday April 04, 2014 @12:38PM (#46661991)

    Let's take a step back and consider what systemd has given us compared to what we had before.

    Before systemd, configuring what gets started on Linux systems was standard across all distributions, dating back to before 1995, when I started developing software with Linux. There was /etc/rc.d/init.d or in some cases /etc/init.d and in most cases there were links in rc1.d, rc2.d, rc3.d, etc. It was that simple. Nothing ever broke.

    With systemd, a solution in search of a problem, everything changed. Now you have all of these directory hierarchies and countless old bugs that take years to get resolved. For example, "network restart" was broken in Fedora for ages for a machine of mine with one DHCP Ethernet interface and two static Ethernet interfaces (with nothing fancy like wireless). "network restart" fails on a variety of machines I have access to; forget about "network reload." ifcfg-eth0 and the like are simple things, some of the most basic boot-related operations. I've tried to open bugs but the problem seems to be buried somewhere in the guts of systemd.

    I've had systems rendered unbootable during upgrades because of silent failures trying to make a good initrd. It's too complex to get everything right with systemd. For a long, long time when the boot scripts died with systemd there was no obvious way to see any errors. Recently they added some more debugging output suggesting that you use journalctl. Why didn't they tell us about that earlier? The reason? No documentation. They wrote an entirely new way to boot the system but kept the design in their heads. Maybe, many years later, there is some scant documentation available (except for that one old useless design document justifying systemd's existence that everyone has read). Of course, nobody writes man pages anymore but they were sure to remove the man pages for the old boot system.

    So what new things does systemd give us? Pretty much nothing except for bugs. Maybe there are a few oddball use cases like booting off of weird media, but most people today boot off of a fixed hard drive that doesn't change in years. 19 years later it might be an SSD, but that is the same use case.

    • by gweihir (88907) on Friday April 04, 2014 @01:19PM (#46662447)

      Indeed. I am still puzzled about why there are so many systemd fanbois. It basically has no merit at all and causes a hot of severe problems. There does seem to be an aggressive, emotionally manipulative campaign by Red Hat to get it into every major distribution and that seems to unfortunately have succeeded. The same strategy is used against critics of systemd and the tactics used have an eerily similar ring to it. Just like if it was paid shills working from a PsyOps manual. There also seem to be indications that the Occupy movement was attacked in a similar fashion.

    • by Kaz Kylheku (1484)

      It is obvious from the internals of systemd that it's designed and written by B grade talent.

      • by jittles (1613415)

        It is obvious from the internals of systemd that it's designed and written by B grade talent.

        I represent that!!

    • by jmyers (208878) on Friday April 04, 2014 @01:56PM (#46662847)

      I adopted Linux and various other open source programs because I was frustrated with the attitude of the vendors I was dealing with, SCO, AT&T, NCR, Sperry, etc. They were all jockeying for position and creating incompatibilities with each other. I had to support a program that ran on a lot on Unices. I discovered the GNU tools that would run the same way on various platforms then toyed with Linux and eventually started using it in production.

      The big difference with GNU and Linux was is seemed to be all about the users. Users creating software for other users without "vendor goals" as baggage. I was a very loyal Red Hat users for years but GNOME 3 drove me to Mint and now I keep seeing more examples where they and other "open source" companies have become like the old Unix vendors.

      Glad to see Linus pushing back against it.

    • by fgodfrey (116175) <fgodfrey@bigw.org> on Friday April 04, 2014 @02:40PM (#46663295) Homepage

      It gets even more "fun" if you're trying to netboot since you never get to see any of the output. When I whined about this problem on Slashdot before, someone suggested adding a parameter to drop to a shell. Which is great, only then systemd didn't get far enough to actually *hit* the problem so I could debug it. So then I tried the flag to systemd that is supposed to get it to tell you what order stuff starts in, but it won't let you run that as root.... Googling got me nowhere. Eventually, I discovered that DBus (another solution in search of a problem, IMO) wasn't functioning correctly because somehow the DHCP server had the wrong MAC address for the host so the network didn't come up right (why isn't DBus talking over 127.0.0.1!!??!).

      In short, systemd has me looking into how quickly I can switch to NetBSD. Although I should investigate Slackware as well.

      • Although I should investigate Slackware as well.

        It makes you feel so good. It's set up simply and flexibly. The entire list of processes running on your machine fits on one page, for example.

        The only problem is a lack of apt-get or ports (also, Ubuntu might do better at handling WiFi, I haven't checked), but if that doesn't bother you it's great.

  • Discipline (Score:5, Insightful)

    by WaffleMonster (969671) on Friday April 04, 2014 @01:06PM (#46662269)

    Linus is providing that which is severely lacking in open source projects. Discipline. No you don't get to do whatever you want neither is there any excuse for breaking shit. Without people like Linus ABI back compat would have been shattered into little bits by now.

  • by ls671 (1122017) on Friday April 04, 2014 @01:19PM (#46662459) Homepage

    Who cares about systemd anyway? Slackware user here...

    The "perhaps less Unix like" part is funny:
    http://forums.scotsnewsletter.... [scotsnewsletter.com]

    • by dpilot (134227)

      The problem with systemd has been the steamroller attitude of its developers and advocates. They seem to want systemd to be the one true init system, accept no substitutes. RedHat, Fedora, Debian, Ubuntu, and Arch have all gone to systemd, and I'm not sure what other distros have as well. As far as I know Slackware, Gentoo, and Funtoo are the only distros that haven't, though Gentoo offers it.

      I don't mind if systemd is an option. But I feel that there is some bad design in there, and would rather not us

  • by dtjohnson (102237) on Friday April 04, 2014 @02:48PM (#46663393)
    Systemd replaces init and is the first daemon to start up in user space during boot and the last daemon to shut down. When its developer sees nothing wrong with breaking the kernel debug during boot merely because its developer feels that he's entitled to use the same parameter name and the kernel boot be damned, you REALLY have to wonder about the wisdom of using systemd.

"No, no, I don't mind being called the smartest man in the world. I just wish it wasn't this one." -- Adrian Veidt/Ozymandias, WATCHMEN

Working...