[zfs-discuss] Spectre and Meltdown impact on ZFS

Gordan Bobic gordan.bobic at gmail.com
Wed Jan 10 09:43:36 EST 2018


On Wed, Jan 10, 2018 at 2:33 PM, Jan Schermer via zfs-discuss <
zfs-discuss at list.zfsonlinux.org> wrote:

> I know about Devuan but I’m a bit worried about how fast upstream updates
> will get there (no idea how big the dev-team/community os) and LTS
> viability.
>

How bleeding edge do you need to be?


> But it was possible to deploy a systemd-less Ubuntu, wasn’t it?
>


Not for a few years at least, AFAIK.


>
> I’ll segway once more since we’re talking Meltdown.
>
> I’m very disappointed by the community response to those bugs - everybody
> keeps parroting how this was a shining example of vendor response… NO IT
> BLOODY WASN’T!
>

Yeah, I'm with you on that one, but not for the reasons you might be
referring to.
My annoyance of the past week has been that I heard about the bug from
ElReg, rather than being told by our hypervisor vendor "Here's a patch,
don't ask what it is, but deploy it RIGHT NOW!" back in December when AWS
seems to have started their patch rollout.



>
> Reponsible disclosure applies to the vendors responsible for the bug! That
> means it should have been disclosed to Intel/AMD/ARM/IBM/… and have them
> work on a hardware/microcode solution if possible. Even if it is by
> architecture design, they are producing the chips and the ones responsible
> for it anyway.
>
> Then after 3 months this should have been made public to all users.
>

Reading between the lines, I get the impression somebody publicly disclosed
it before the embargo was up and before all the patches were ready.


>
> Instead it was disclosed to hardware and selected software vendors, and
> folk like OpenBSD were left in dark until “patches were ready” - this is
> not how responsible disclosure works but makes it looks like Intel did the
> right thing, without them actually resolving anything.
>

It's a tricky issue. How far do you disclose it? The wider the disclose it,
the greater a chance of it leaking out before the patches are ready.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.zfsonlinux.org/pipermail/zfs-discuss/attachments/20180110/85764789/attachment-0001.html>


More information about the zfs-discuss mailing list