The challenge of compiling for verified architectures
The challenge of compiling for verified architectures
Posted Oct 8, 2023 5:04 UTC (Sun) by ibukanov (subscriber, #3942)In reply to: The challenge of compiling for verified architectures by roc
Parent article: The challenge of compiling for verified architectures
So Wasm sandbox requires too many dynamic checks to hope for reliable practical protection against SPECTRE and friends. EBP implementation in Linux, by reducing the number of runtime checks and constraining the code sufficiently, still managed to provide the protection against known attacks. Hence ability to provide practical protection against speculative leaks is a useful criterion to classify.
Posted Oct 8, 2023 7:06 UTC (Sun)
by Cyberax (✭ supporter ✭, #52523)
[Link]
Does that include new and fancy features like iterators?
Posted Oct 8, 2023 9:57 UTC (Sun)
by roc (subscriber, #30627)
[Link] (1 responses)
I still think this is a poor way to draw a line between "sandbox" and "verifier". If Spectre vulnerabilities are found in eBPF will it cease to be a verifier and become a sandbox? That doesn't make sense to me.
Posted Oct 8, 2023 12:57 UTC (Sun)
by ibukanov (subscriber, #3942)
[Link]
The challenge of compiling for verified architectures
The challenge of compiling for verified architectures
https://bytecodealliance.org/articles/security-and-correc...
The challenge of compiling for verified architectures