4 ms·
Yeah no disagreement there. I had totally forgotten about the JS POC - ugh! The JavaScript argument is interesting to me in that it's already flawed. I suppo
by FunnyLookinHat 5y ago
Yeah no disagreement there. I had totally forgotten about the JS POC - ugh!
The JavaScript argument is interesting to me in that it's already flawed. I suppose I'd rather focus on the security issues with browsers running code on my computer more than anything else since it's effectively the "but what about ___" answer to so many threads like this one.
I've seen a few other comments suggesting per-process rules to enable or disable branch protections. That's an interesting thought, especially considering you could apply it to either "trusted" or "untrusted" code depending on it's source.
- fiddlerwoaroof 5y agoWith Arm big.little architectures, it could start making sense to have dedicated in-order cores for running JavaScript and other “untrusted” code. Also, I wonder if disabling mitigations on the desktop and running the browser in a VM with mitigations enabled would be effective.
- saagarjha 5y agoPeople want their websites to load fast, though.