gchristensen changed the topic of #nixos-security to: Vulnerability Roundup Issues: https://github.com/NixOS/nixpkgs/issues?utf8=%E2%9C%93&q=is%3Aissue+is%3Aopen+Vulnerability+roundup + https://broken.sh
Synthetica has quit [Quit: Connection closed for inactivity]
lassulus has joined #nixos-security
tilpner has quit [Remote host closed the connection]
tilpner_ has joined #nixos-security
c4rc4s has quit [Remote host closed the connection]
c4rc4s has joined #nixos-security
c4rc4s has quit [Quit: Adios]
ris has quit [Ping timeout: 246 seconds]
FRidh has quit [Ping timeout: 265 seconds]
filemo8 has joined #nixos-security
filemo8 has quit [Quit: Going offline, see ya! (www.adiirc.com)]
tom39291 has quit [Ping timeout: 245 seconds]
tom39291 has joined #nixos-security
__Sander__ has joined #nixos-security
hmpffff has joined #nixos-security
ckauhaus has joined #nixos-security
FRidh has joined #nixos-security
<ckauhaus> Vulnerability roundup 78 is ready
zarel has joined #nixos-security
<ckauhaus> #75495 - #75504
<{^_^}> https://github.com/NixOS/nixpkgs/issues/75495 (by ckauhaus, 1 minute ago, open): Vulnerability roundup 78: bind-9.14.8: 1 advisory
<{^_^}> https://github.com/NixOS/nixpkgs/issues/75504 (by ckauhaus, 1 minute ago, open): Vulnerability roundup 78: unbound-1.9.4: 1 advisory
<infinisil> I might just have discovered a NixOS vulnerability that allows anybody to get root access
<infinisil> Can I talk about it here or should this be discussed in a secure channel?
<infinisil> I mean maybe it's old news, but it seems like a pretty big problem
<ckauhaus> infinisil:This should be communicated to one of the official security contacts, see https://nixos.org/nixos/security.html
<infinisil> Yeah thanks, floki PM'd me that, I'll send a mail shortly
<infinisil> If I can figure out how to import those gpg keys!
<ckauhaus> It depends a bit if information about the exploit should be kept secret or not
<ckauhaus> if the exploit can be disclosed anyway, there's no need to send encrypted mails ;-)
<infinisil> Hm well I just set up enigmail recently, so I was hoping it would be pretty simple to use
<infinisil> How do I import the keys in https://nixos.org/nixos/security.html into gpg such that they are associated with the mail address?
<ckauhaus> GPG keyserver
<infinisil> `gpg --recv-keys 0x846FDED7792617B4` imports it, but it's not associated with a user id after that
<ckauhaus> import by mail addr?
<infinisil> ckauhaus: `gpg: "fpletz@fnordicwalking.de" not a key ID: skipping`
<infinisil> That's with `gpg --recv-keys fpletz@fnordicwalking.de`
<fpletz> infinisil: gpg --search-keys :)
<ckauhaus> there seems to be something fishy with fpletz' gpg key
<ckauhaus> % gpg --search-keys fpletz@fnordicwalking.de ~/code/nixpkgs
<ckauhaus> gpg: data source: https://keys.openpgp.org:443
<ckauhaus> gpg: Schlüssel "fpletz@fnordicwalking.de" wurde auf dem Schlüsselserver nicht gefunden
<ckauhaus> gpg: Suche auf dem Schlüsselserver fehlgeschlagen: Not found
<ckauhaus> gchristensen's key works
<infinisil> Yeah grahams one worked with --recv-keys too
<infinisil> But not the others :/
<fpletz> my key doesn't seem to be on keys.openpgp.org :/
<fpletz> but it is on pool.sks-keyservers.net
<fpletz> I'll upload it
<infinisil> Ah, it works when I go to https://pgp.mit.edu/pks/lookup?op=get&search=0x846FDED7792617B4 then copy paste the text there into a file, then `gpg --import` the file
<infinisil> gchristensen fpletz (domenkozar): Sent an encrypted mail to you :D, for verification my key is 0x422E9EDAE0157170, fingerprint 6C2B 55D4 4E04 8266 6B7D DA1A 422E 9EDA E015 7170
<infinisil> I'm also using WKD for infinisil.com, so the key should be found automatically for my mail
<infinisil> I think that's what it's supposed to do anyways
<fpletz> infinisil: thanks! but I think this is actually common to all linux distributions
<fpletz> did I understand you correctly that is essentially like adding shell=/bin/bash to the kernel commandline?
<fpletz> (on other distributions of course :))
<fpletz> ah, init=/bin/bash of course
<infinisil> Oh hehe
<infinisil> init=/bin/sh
<infinisil> Okay so I guess it's not a "problem" after all
<tilpner_> That's nothing new, set boot.loader.systemd-boot.editor = false;
<tilpner_> grub probably has an equivalent
<infinisil> Why shouldn't this be true by default?
<tilpner_> Because it's useful while debugging boot problems
<infinisil> Hm yeah I did just use that to rescue a failed vps installation
<tilpner_> With an unencrypted disk, setting that to true doesn't do much to protect your installation
<tilpner_> Especially if you are virtual machine on malicious hardware
<tilpner_> *are a
<infinisil> Oh, I meant false by default earlier
<tilpner_> Yeah, I assumed so
tilpner_ is now known as tilpner
<infinisil> Hmm, okay so in the mail I mentioned how boot.debug1mounts can be used to have an interactive shell start. I guess if you have that parameter set, disallowing boot parameter editing doesn't protect you still
<infinisil> But that's the cost to using boot.debug1mounts I suppose
<infinisil> Using disk encryption would protect you though
<fpletz> if you assume the hardware or the hypervisor is malicious or compromised, full disk encryption doesn't really help because the keys are in memory and can be extracted easily
<tilpner> Why would anyone keep that option on?
<infinisil> fpletz: Before you boot up at least
<infinisil> "By default, the boot loader interface is accessible to anyone with physical access to the console: anyone can select and edit any menu entry, and anyone can get direct access to a GRUB shell prompt. For most systems, this is reasonable since anyone with direct physical access has a variety of other ways to gain full access, and requiring authentication at the boot loader level would only
<infinisil> serve to make it difficult to recover broken systems. "
<fpletz> yeah, if you don't input your passphrase :) but also note that the initrd could've been modified to exfiltrate the passphrase without you noticing
<infinisil> True dat..
<infinisil> I guess an optimal system setup should have only the user partition encrypted, and it being locked automatically again when logging out
<tilpner> When I had a Hetzner server, I used disk encryption to prevent sensitive data to end up in the wrong hands after disk replacement and sloppy disposal
<tilpner> But there's no good way to defend against a malicious provider, and you should switch/not use them if you believe you need to defend against them
<tilpner> *from ending up in
<infinisil> Although, what if you trust them, but you don't know what will happen to them in the future
<infinisil> (and that's kind of the case with all providers)
<fpletz> yup, that's also the reason I have some physical hetzner servers with full disk encryption for some stuff (still not perfect but better than a hypervisor :))
<tilpner> infinisil: How would this help with that?
<tilpner> Uhh, I know, very specific
<tilpner> this: disabling boot editor, that: suspicion provider will turn malicious
<infinisil> Hehe, make the server and your local machine send heartbeats. When no heartbeats are received anymore, the server shuts down, unloading all keys
<infinisil> This way the provider can't just disconnect the machine and extract everything
<tilpner> Why would they need to shut your VM down for that?
<tilpner> I haven't used it, but live migration appears to be a thing
<infinisil> Well if you use full disk encryption
<infinisil> Oh damn
<fpletz> yeah, the memory can be dumped and there are tools to find luks keys in memory dumps :>
<infinisil> I guess there's no perfect defense
<fpletz> for physical servers, they can freeze the memory chips so that they keep their state a while longer and just yank them out and dump them
<fpletz> unfortunately, yeah :(
<infinisil> If somebody really wants to be sure their data is secure they should probably run their own servers in their basement
<infinisil> fpletz: Glad to know that sending encrypted emails works though :)
zarel_ has joined #nixos-security
zarel has quit [Ping timeout: 268 seconds]
__Sander__ has quit [Quit: Konversation terminated!]
hmpffff has quit [Quit: nchrrrr…]
iz16 has joined #nixos-security
iz16_izmntuk has quit [Ping timeout: 240 seconds]
<qyliss> infinisil: one thing you might want to look into is Heads
<qyliss> Which is a BIOS replacement that proves your firmware/initrd/kernel hasn't been tampered with before you enter your decryption password
<qyliss> Very annoying to use with NixOS though. I'd like to make that better.
<infinisil> qyliss: Neat, what's annoying with it?
<qyliss> It implements a parser for a small subset of grub config files
<qyliss> NixOS's are too complex
<qyliss> So you have to trick it
<qyliss> It can read extlinux config files too, but to get that work I had to ln -s /boot/extlinux/extlinux.conf /boot/extlinux/grub.cfg
<qyliss> Additionally, you have to gpg sign your kernel and initrd every time they change, and store that signature in /boot
<gchristensen> secureboot style
* gchristensen considers resurrecting his PR
ris has joined #nixos-security
<ris> #75529
<{^_^}> https://github.com/NixOS/nixpkgs/pull/75529 (by risicle, 12 seconds ago, open): [r19.09] glibc: add patch for CVE-2019-19126
ckauhaus has quit [Quit: WeeChat 2.6]