[HN Gopher] Linux Kernel Security Done Right
___________________________________________________________________
Linux Kernel Security Done Right
Author : theafh
Score : 27 points
Date : 2021-08-03 16:03 UTC (6 hours ago)
(HTM) web link (security.googleblog.com)
(TXT) w3m dump (security.googleblog.com)
| touisteur wrote:
| The 'update all the things all the time, maybe there will be
| regressions, but hey vendors should fix' em' is maybe hearable
| from a cloud-continuous-delivery standpoint, but not everyone can
| and should do that. And getting fixes also comes its lot of new
| features with their unchecked security surfaces, enabled by
| default...
|
| What I'm also hearing here is 'vendors shouldn't work on their
| branch but contribute to main/stable branches' and I'm here
| wondering how you can ever ship anything tested (since it's part
| of the subject) with this lack of control of branch history.
| Starting from something you know works, and increasingly
| upgrade/fix is the only manageable way to ship something as
| complex and untested as upstream Linux.
|
| The solution is 'add 4 guys to your team and fix our coders' bugs
| before they ship'? So, fun coding and good PR for some, drudge of
| shitbugs and abuse on lkml for others.
| tester756 wrote:
| Feel free to start with asking engineers more security related
| questions, cuz you know - poorly performing algo can be rewritten
| meanwhile leak cannot be reverted
___________________________________________________________________
(page generated 2021-08-03 23:02 UTC)