NixOS Security Triage | 682 Members | |
| Coordination and triage of security issues in nixpkgs | 214 Servers |
| Sender | Message | Time |
|---|---|---|
| 30 May 2021 | ||
afaik nix-collect-garbage should take care of old versions laying in /nix/store, right? | 18:32:43 | |
| you should be able to install the store into your home dir | 18:32:47 | |
| One thing you could probably look into more: Time to rollout of fixes after they have been committed. How long does a rebuild of the closure take for an "average" user? Is that a huge downside? How much that this increase the risk of someone exploiting your systems? | 18:33:21 | |
| kunrooted: yes, nix-collect-garbage will do that while still honoring generations/profiles. | 18:33:52 | |
In reply to @andi:kack.itokie | 18:34:57 | |
| I also have a 'topic list' for now, maybe this will generate even more ideas if I share it:
in the meantime, I made a PoC of interpreter/loader/library hijack, which I'll also try to cover | 18:36:03 | |
| * I also have a 'topic list' for now, maybe this will generate even more ideas if I share it:
in the meantime, I made a PoC of interpreter/loader/library hijack, which I'll also try to cover as a topic in that paper of mine | 18:36:19 | |
| * I also have a 'topic list' for now, maybe this will generate even more ideas if I share it:
in the meantime, I made a PoC of interpreter/loader/library hijack, which I'll also try to cover as a topic in that paper of mine | 18:38:37 | |
about unpriviliged users can install packages: That is true on any normal linux desktop/server. Even if the package manager doesn't help you, local users will be able to execute all the code they want to. | 18:38:41 | |
| I also have a few ideas for PoCs on how to demonstrate downsides of our current stuff and what the average NixOS contributor should be aware of... DM me (in a few days/weeks) if you feel like you need more :) | 18:38:44 | |
I thought about mentioning security of NixOS containers where root in container is root on the host | 18:39:07 | |
| * I thought about mentioning security of NixOS containers where root in container is root on the host This was mitigated some time ago IRRC? | 18:39:11 | |
In reply to @philipp:xndr.deyou can limit them | 18:39:11 | |
*
This was mitigated some time ago IRRC? | 18:39:16 | |
| afaik | 18:39:16 | |
| you can make specific users having just write access to just specific things, it's really flexible af | 18:39:37 | |
| You can set noexec on ~ | 18:39:39 | |
In reply to @andi:kack.itit won't be an issue anymore? | 18:39:55 | |
| I vaguely recall someone talking about it months ago | 18:40:10 | |
| I was writing a container a while ago and it was mentioned an issue then by some of my collegues | 18:40:14 | |
| perhaps this? https://github.com/NixOS/nixpkgs/pull/67336 | 18:41:05 | |
| ah, so it limits a root on the container? | 18:41:36 | |
| I think that still not many people might know about this option | 18:42:19 | |
| It wasn't merged yet so who knows what the actual state is :D | 18:42:43 | |
| yeah, it's a 'draft', weird | 18:42:53 | |
| 31 May 2021 | ||
| 08:23:41 | ||
| hah. i've heard of squash-merges before but this author squashes their entire releases https://github.com/pgpartman/pg_partman/commit/0b6565ad378c358f8a6cd1d48ddc482eb7f854d3 | 13:01:19 | |
luckily the search_path changes are all i need and they are separable by file | 13:01:56 | |
nothing fetchpatch can't handle | 13:02:09 | |
| still | 13:02:12 | |