Kevin Karhan :verified:<p><span class="h-card" translate="no"><a href="https://mastodon.social/@stman" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>stman</span></a></span> <span class="h-card" translate="no"><a href="https://masto.hackers.town/@theruran" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>theruran</span></a></span> <span class="h-card" translate="no"><a href="https://cathode.church/@50htz" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>50htz</span></a></span> <span class="h-card" translate="no"><a href="https://social.solarpunk.au/users/vidak" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>vidak</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.net2o.de/@forthy42" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>forthy42</span></a></span> Yeah, basically attacks on schedulers or any non-static component that is software-defined makes it easy to go that way. </p><ul><li>The only way one could <em>"fix"</em> this would be Harvard-Architecture and strictly segmented threads in hardware, which would toss all the flexibility of existing Von-Neumann-Systems out of the Window and basically result in <a href="https://infosec.space/tags/Mainframe" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Mainframe</span></a>-esque systems like the OS/360 'hard containers' on z/Architecture, which means 2x42U of <em>phat iron</em> to run 50+ year old COBOL code at 0% performance gain at bit-, cycle and clock-accuracy.</li></ul><p>And even <em>"<a href="https://infosec.space/tags/TradFinance" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TradFinance</span></a>"</em> hates that because they refuse to train new programmers for that stuff and they are too deeply entrenched to migrate away!</p>