Announcement

Collapse
No announcement yet.

DragonFlyBSD Lands Fixes For Meltdown Vulnerability

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • DragonFlyBSD Lands Fixes For Meltdown Vulnerability

    Phoronix: DragonFlyBSD Lands Fixes For Meltdown Vulnerability

    Linux, macOS, and Windows has taken most of the operating system attention when it comes down to the recently-disclosed Meltdown vulnerability but the BSDs too are prone to this CPU issue. DragonFlyBSD lead developer Matthew Dillon has landed his fixes for Meltdown...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    It seems Matthew Dillon really enjoys fixing CPU bugs.

    Haven't heard anything about OpenBSD.

    Comment


    • #3
      Originally posted by LightBit View Post
      It seems Matthew Dillon really enjoys fixing CPU bugs.

      Haven't heard anything about OpenBSD.
      yep, and I like his way more professional tone over Linus' and I'm a long term Linux user, ... guess time to give Dragonfly a closer look.
      Yet, even he is very pissed, thanks got I avoided Intel, and only have it in portable's where Apple and Microsoft forced it on me, .. otherwise even AMD Phenom II servers here ;-)

      Comment


      • #4
        Looks like PCID hasn't been implemented yet, so there's definitely going to be a hit. Hopefully I'm wrong. I figure it's coming even if I'm right.

        Comment


        • #5
          Originally posted by rene View Post

          yep, and I like his way more professional tone over Linus' and I'm a long term Linux user, ... guess time to give Dragonfly a closer look.
          Yet, even he is very pissed, thanks got I avoided Intel, and only have it in portable's where Apple and Microsoft forced it on me, .. otherwise even AMD Phenom II servers here ;-)
          Dillon is beyond incredible. Dragonfly, when it works, works great. I have it in VMs, but it doesn't seem to run natively on my AMD hardware, probably my fault. And since it's BSD, forget about any AMD GPU more recent than 6XXX series. HAMMER2 is so far incredible, hasn't corrupted itself unlike every single btrfs array I've built. btrfs with raid1 is supposed to be stable, but it's always eventually corrupted itself and would not heal/rebuilt. HAMMER2 lacks a lot of features, but with all my thinkering, hasn't entered a black hole yet. And HAMMER (1, not 2), while it's described as a dead end, has unbelievable features, such as by default conserving every single version of every single file. You just "hammer history (file)", and can see every version. The downside is the management of PFS (pseudo-filesystems), which you have to manage, vs btrfs which just handles everything automatically. There's also master disks which are read-write, and slaves which are read-only. It's not efficient, but damn, so far, it's stable. So is HAMMER2, but a lot of critical features are lacking, so you've been warned.

          Comment


          • #6
            Originally posted by LightBit View Post
            Haven't heard anything about OpenBSD.

            Comment


            • #7
              Originally posted by AndyChow View Post

              Dillon is beyond incredible. Dragonfly, when it works, works great. I have it in VMs, but it doesn't seem to run natively on my AMD hardware, probably my fault. And since it's BSD, forget about any AMD GPU more recent than 6XXX series. HAMMER2 is so far incredible, hasn't corrupted itself unlike every single btrfs array I've built. btrfs with raid1 is supposed to be stable, but it's always eventually corrupted itself and would not heal/rebuilt. HAMMER2 lacks a lot of features, but with all my thinkering, hasn't entered a black hole yet. And HAMMER (1, not 2), while it's described as a dead end, has unbelievable features, such as by default conserving every single version of every single file. You just "hammer history (file)", and can see every version. The downside is the management of PFS (pseudo-filesystems), which you have to manage, vs btrfs which just handles everything automatically. There's also master disks which are read-write, and slaves which are read-only. It's not efficient, but damn, so far, it's stable. So is HAMMER2, but a lot of critical features are lacking, so you've been warned.
              I use btrfs for all my new setups, including on LVM2 RAIDs, even with LVM2 dm-cache and on top of my LUKS full disk encryption with compress=lzo and have not had any issue. https://www.youtube.com/watch?v=MELYA68qQyo (I do not use btrfs's built-in RAID though, I need the LVM2 management-ability, ...)

              Well, ok I once had an issue when I mounted my external drive on a big-endian PowerPC G4 cube on a USB 1 port, and this somehow corrupted it, certainly user-fault for trying such a fragile endeavor, ... maybe simply an endianess typo, or simple corrupted on the USB 1 bus, ... https://www.youtube.com/watch?v=7rq08KTFFis

              Comment


              • #8
                Originally posted by rene View Post

                yep, and I like his way more professional tone over Linus' and I'm a long term Linux user, ... guess time to give Dragonfly a closer look.
                Yet, even he is very pissed, thanks got I avoided Intel, and only have it in portable's where Apple and Microsoft forced it on me, .. otherwise even AMD Phenom II servers here ;-)
                Too bad for you that AMD's PSP is also hit.

                Comment


                • #9
                  Amd Phenom II's do not have PSP AFAIK

                  Comment


                  • #10
                    Originally posted by Vistaus View Post

                    Too bad for you that AMD's PSP is also hit.
                    yeah, but those don't have it ;-)
                    PS: not that I or many other's actually asked to get this built-in backdoor crap, ... :-/!
                    Last edited by rene; 13 January 2018, 05:42 PM.

                    Comment

                    Working...
                    X