Post 9svxlcXUOmXTeRklgO by ParadeGrotesque@mastodon.sdf.org
 (DIR) More posts by ParadeGrotesque@mastodon.sdf.org
 (DIR) Post #9svxlb6riHeFDaHzXs by h3artbl33d@bsd.network
       2020-03-12T14:17:41Z
       
       0 likes, 0 repeats
       
       From the Tails 4.4 changelog:"Vagrant build box: disable mitigation features for CPU vulnerabilities (Closes: #17386). Given the kind of things we do in our Vagrant build box, it seems very unlikely that vulnerabilities such as Spectre and Meltdown can be exploited in there. Let's reclaim some of the performance cost of the corresponding mitigation features."Source: https://git.tails.boum.org/tails/plain/debian/changelog
       
 (DIR) Post #9svxlbXo67COZ8tVXU by h3artbl33d@bsd.network
       2020-03-12T14:21:57Z
       
       0 likes, 0 repeats
       
       I get that the mitigations for the CPU vulnerabilities come with a performance cost. Might be even called a "harsh" cost for certain workloads.However, coming from a Linux distribution with the target audience that Tails has, offering privacy and resiliency, it seems a somewhat concerning choice to pick performance over security.Personally, I consider Tails to be a high profile target, given the use for whistleblowers, the privacy conscious and those in need [...]
       
 (DIR) Post #9svxlbykTwkXuhV1X6 by h3artbl33d@bsd.network
       2020-03-12T14:28:05Z
       
       0 likes, 0 repeats
       
       for a system that might offer some resiliency / anonimity for whatever reason. It could very well be oppressed people that want to bring out information, risking their lives. Literally.Thus, even though the project considers it "very unlikely", the impact in case of a succesfull breach is disastrous and could put users in grave danger.I would really like if Tails were to reconsider this move. Or am I being overly paranoid and alone in this thinking?
       
 (DIR) Post #9svxlcXUOmXTeRklgO by ParadeGrotesque@mastodon.sdf.org
       2020-03-12T14:33:57Z
       
       0 likes, 0 repeats
       
       @h3artbl33d You have a point, but I think the git changelog you mention only applies to the build box.In other words, the build box itself is not "protected" against the most recent Intel snafus, but the Tails distribution itself is. Since we can hope their build box is both transient and well protected, this makes sense to me. I may be wrong, of course, but I'd like to test Tails with the latest Intel bug analyzer.
       
 (DIR) Post #9svy3ocoCd22Lauwa0 by h3artbl33d@bsd.network
       2020-03-12T14:37:25Z
       
       0 likes, 0 repeats
       
       @ParadeGrotesque Thank you for your reply. That is exactly how I have interpreted the message. Ideally, the build machine should be utterly secure, even against complex and high-resource attacks such as the Intel vulns. I mean, if there is a chance to snuck in some code in the Tails distro, an adversary *will* try sooner or later.Currently downloading Tails to do just that :)
       
 (DIR) Post #9svyIqJuBNglwrF0Vc by ParadeGrotesque@mastodon.sdf.org
       2020-03-12T14:40:03Z
       
       0 likes, 0 repeats
       
       @h3artbl33d You do have a point. This being said, I can understand the desire to have a fast(er) machine for build if your (well, Tails) infrastructure is limited.Not the most secure but could be acceptable based on the security applied to the build chain (for instance: entirely offline, etc.).Let's test and compare notes!