16 Replies to “New user and new bug”

  1. Am I reading the code commit correct in that a new system call had to be created to fix this issue?

    Isn’t creating new system calls a lot g baggage to maintain?

  2. @justin

    I abolsutely agree this is Linus’ Law at play.

    I once wish more eyeballs were reviewing Dfly code (ie Dfly usage adoption was higher)

  5. I’m not speaking in any official capacity for anything or anyone other than myself.

    If you have problems with FreeBSD and releasing updates on Meltdown/Spectre, why not ask over there? Or is this just complaining about FreeBSD?
    I believe that parts have been commited to 12-CURRENT, not sure if they’ve been pushed back to 11/10-Release yet. I believe some of the pieces required updates to clang, then there are binary updates to consider.

  6. Oh a quick internet search for freebsd meltdown spectre fix has a few good links, one over on phoronix.com. Perhaps they simply haven’t updated the news about this?

  7. The problem I’ve always seen with FreeBSD is stuff gets fixed in the current HEAD release but next fixed in STABLE or back ported

    So people have to wait 2 years did head to be release to see these patches.

  8. It all depends on what is fixed and if it needs to be merged back to the release branches. That’s what MFC means (merge from current).
    Sometimes a fix on current is for a bug that only exists on current, so there is nothing to backport.
    Security fixes that affect all releases often are released together; Meltdown/Spectre are different because the fixes are across multiple items, some in the kernel, some in clang, come in patched microcode. There may be enough differences between CURRENT and 10/11-Release that it’s not a simple copy and paste (thing in VM are often different enough they aren’t simple)

    It just seems that if one has a question about FreeBSD timeframe on something like this, you get a better answer if you ask there.

  9. To be fair, none of DragonFly’s fixes are in DragonFly’s release. Just in master.

  10. So is openbsd the first operating system to release a fix?

  11. DragonFly was the first BSD – Linux was the first operating system, but that’s because of the way it was (or wasn’t) disclosed to a limited number of people.

  12. AFAIK not a single BSD has released a fix yet. OpenBSD have their regular bi-annual release coming in 2 or 3 months (6.2). I don’t know if they will provide a syspatch for something this big for current release 6.2.

    Neither DragonFly nor FreeBSD or NetBSD have announced any release with those fixes.

  13. Sorry I meant that the next OpenBSD release will be 6.3; traditionally released in May.

  14. Anonymous – you may be conflating two phrases – “release”, as in “make available”, which all the BSDs have done and is the phrase that several people have used.

    “Release”, as in major numbered version – no, there’s no release versions yet. I don’t think there’s a point to that splitting of hairs.

  15. I’m the anonymous from earlier.
    Hmm, I don’t think that is splitting of hairs. For most users what’s important is when they are able to use that fix, not when it’s available in a source repository.

    Although I must say there is quite some confusion in BSD projects regarding what people should be using. It looks like many people in FreeBSD track -CURRENT, whereas OpenBSD people tend to rely on the releases more.

    As it stands, I’m running OpenBSD 6.2 and DFly 5.0.2 and neither has the fix (which I don’t care too much since those are hobby systems).

  16. I’m taking about coming up with definitions of the word ‘release’ as splitting hairs, not arguing against release versions with patches.

Comments are closed.