back to article Linux 5.3 kernel bundles new, cuddlier, swear-free Torvalds with AMD Radeon Navi graphics support

A softer, gentler Linus Torvalds released the Linux 5.3 kernel over the weekend and swung open the doors on 5.4. Things were held up a little this time around, something Torvalds attributed to his travel schedule rather than anything more sinister. He was, however, pleased to note that the extra week meant that a few last- …

  1. cornetman Silver badge
    Stop

    I know the media make a lot of hay by implying that Linus can hardly opening his mouth without expelling a shit ton of invective, but I'm pretty sure that the much-celebrated expletive-laden outbursts have always been very much the exception rather than the rule.

    1. FozzyBear Silver badge
      Thumb Up

      In instances justified too.

      1. JDX Gold badge

        Not really justified, no. You have to ask yourself:

        - if you worked somewhere where that was how you were treated, would you stay?

        - would you talk to your children like that

        - how would you react if someone else talked to your kid like that

        - would you talk to your spouse/partner like that

        - how would you react if someone else talked to your spouse/partner like that

        - if your spouse/partner talked to you like that, what would your response be

        I totally get that we don't all have to be nicey-nicey all the time, and someone doing something wrong needs to have that made very clear, but this isn't army boot camp. We don't need to 'break' people.

        1. oiseau Silver badge
          Facepalm

          May not be justified ...

          Asking to be and then being accepted to work on Linus Torvald's kernel code, if anything at all, is a huge privilege and it assumes that you are (supposedly) equal to the task.

          ie: that you have the necessary ability, talent, qualities and capability to accomplish a given role or handle a given situation, which in the context of coding for the Linux kernel, clearly means (among other things) being able to do it properly without behaving like an incompetent AH.

          In this context, properly means following the kernel design rules set forth long ago by Torvalds, not acting like they are non-existent.

          Like any of the DHs that repeatedly, over and over (and over) again insist in doing things that are clearly off the table, like ticking code in the kernel that can mess with existing users.

          No wonder the guy goes off his rocker every so often.

          Cheers,

          O.

        2. Anonymous Coward
          Anonymous Coward

          In almost all instances, Linus has lost it over kernel changes breaking user land AND the people responsible then saying I'm not responsible, it's someone else's fault for having bugs. Or in the case of the Google security team members suggestions around security fixes that would significantly increase development challenges on Linux, frankly bonkers.

          The majority of the comparisons you make are for relationships where you have significant input/responsibility (family and friends) while the kernel development group allows all comers who show they are competent enough to carry out complex work.

          Comparing the relationship Linus has to those working on kernel development to family/friends/etc may be accurate in some cases, but in the vast majority of cases, creates an impression of more communication outside what we are seeing on mailing lists and therefore more varied approaches dealing with inappropriate behaviour (i.e. the developers Linus is yelling at) and these are not always possible given deadlines/other time commitments for contributors.

          If I worked in an environment where someone passionately cared about a project and was able to get a large team to work well together 99.9% of the time, it would take a lot for me to leave that. If I screwed up and broke something significant, costing people significant amounts of time, doing something that I shouldn't have been doing AND tried to defend my mistake by either blaming others or saying I wasn't wrong, then yes, I'd expect people to be pissed. But I wouldn't leave unless I was told to go.

          Horses for courses...

        3. cornetman Silver badge

          > - would you talk to your children like that

          Depends.

          If one of my kids had set fire to the front room and despite everyone knowing exactly who the culprit was they denied it and blamed the cat, then yes I would tear a strip off them.

          I have dealt with many people in the dev realm, and a small number of very clever people can be similarly the hardest to get through to when they fuck up. *Sometimes* anger and expletives are the only way to get through.

    2. bombastic bob Silver badge
      Devil

      I always figured that Linus has a TRUE PASSION for *QUALITY*, and NO tolerance for [expletive deleted]

      (it could only get better if it were a Marine D.I. dressing down an incompetent recruit - and when you consider what it is the D.I. is attempting to do, it makes a LOT of sense to do it 'that way')

      1. teknopaul Silver badge

        Re-read Starship Troopers for a lesson in shouting at people for their own good.

        Move it you useless bug hunters, on the bounce!

  2. Notas Badoff
    Pint

    Double-plus good

    I think I quite like the kind of scrutiny that reveals bugs in old code rather than reflexively blaming new code. And the new code that uncovers a bug in other code - isn't that a double win? There should be awards for that.

    1. Crazy Operations Guy Silver badge

      Re: Double-plus good

      The code is still in the source, you can bring it back with the proper incantations in your config. Developers can still test to see if their code breaks, but normal users aren't affected.

      With the code, pushing audio over DisplayPort via an AMD Vega 56 card to a stereo-out jack on a monitor will cause PulseAudio will end up spewing mangled audio for a few seconds, then core dump.

      1. Gene Cash Silver badge

        Re: Double-plus good

        > will cause PulseAudio [to] core dump.

        Wot! But Lennart Poettering's code is PERFECT! It would never core dump!

        1. Baldrickk Silver badge

          Re: Double-plus good

          And it turns out it's a different bit of Poettering's effluent that's caused this issue.

      2. FozzyBear Silver badge
        Pirate

        Re: Double-plus good

        you can bring it back with the proper incantations in your config

        Yeah my Latin is a little rusty and the last thing I want is to be neck deep in zombified code running around the system droning "memory, memory, memory'

  3. bombastic bob Silver badge
    Meh

    People relying on the previous behavior?

    and now I'm wondering _exactly_ what that was...

    I know there are some issues that cause 'priority inversion', such as when creating the locate database. So I have to wonder if it's something like that, where a 'nice'd process still causes everything ELSE to respond poorly (stuttering, delays, swapping, etc.) while it's running, even though it's basically going at the highest (aka lowest scheduling priority) possible 'renice' level...

    and I'd have to wonder WHAT kind of thing actually RELIES on such a bug.... if it's that, anyway.

    maybe a kernel option to 'fix' it?

    1. Paul Hargreaves

      Re: People relying on the previous behavior?

      Bug is /dev/random hanging when it'll never get any new entropy, which can happen at boot time and is what the ext4 patch managed to trigger.

      1. Baldrickk Silver badge

        Re: People relying on the previous behavior?

        How do people rely on that?

        Edit: ah, I read the git commit message.

        systemd using getRandom() (/dev/random) instead of /dev/urandom very early in the boot process, when there isn't enough entropy, causing the boot process to wait for entropy that never comes because the whole thing is stuck waiting for entropy... (ad nauseum)

        (as opposed to my original "people rely on getting stuck on a call to getRandom()? upon reading your post)

        1. Anonymous Coward
          Anonymous Coward

          Re: People relying on the previous behavior?

          poettering's monolithic dump being terrible again? I'm shocked. SHOCKED I tell you.

        2. Claptrap314 Silver badge

          Re: People relying on the previous behavior?

          Sounds like it needs to be pushed out immediately!

  4. Paul Hargreaves
  5. Hans 1 Silver badge
    Linux

    If you need the latest and greatest, you can get compiling now, although it might be worth hanging fire for the first point release

    Hm, I would wait for several point releases, like 5 or 10, the higher the better. Avoid unless you really need new features it implements NOW.

  6. Rich 2 Silver badge

    OMFG

    Looking at the above posts, I see that there's a bug in Pulse Audio that has been exposed

    ....and a bug in systemd that has been exposed

    ...and the common factor?

    WHY does ANYONE use ANYTHING that Poettering vomits out?

    1. Anonymous Coward
      Anonymous Coward

      Re: OMFG

      He doesn't swear and he's (passive-aggressively) polite to everyone, so obviously he's a better coder than that Linux guy.

      1. georgezilla

        Re: OMFG

        Because he's new and shiny?

        And we have to have our new and shiny.

      2. davcefai

        Re: OMFG

        I remember when he was foisting pulseaudio on us. His standard reply to bug reports was that it would get better soon. Frankly I cannot understand how Poettering has been allowed to screw up Linux with systemd. It goes against the basic philosophy of ONE tool doing ONE thing very well.

        Which is why I dropped Debian for Devuan.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2020