New flaw in Intel chips lets attackers slip their own data into secure enclave

Re:New flaw in Intel chips lets attackers slip their own dat... by cesarb on Tuesday March 10, 2020 @03:42PM (#22538829)

What happened is that Spectre was the first instance of a new class of vulnerabilities, one which hadn't been considered previously. Side-channel attacks were already well known, but nobody seemed to have considered side channels from speculative execution before (it's hard to overstate how bizarre these vulnerabilities look like, from a software point of view: they're leaking information from a code path which has not been executed, could not execute, and in some cases doesn't even exist; that is, when looking only at the software, the leak appears to come from a parallel universe where an impossible code path actually executed).

Since this vulnerability class was previously unknown, hardware didn't have much protection against it, except by accident (or as a side effect of more conservative design). Being in hardware, they are difficult or impossible to workaround in software. And they were found in common hardware most people have. This all lead to them being widely reported by the tech press.

[ Parent | on HN ]

Re:New flaw in Intel chips lets attackers slip their own dat... by gitgudnubs on Tuesday March 10, 2020 @04:08PM (#22539119)

Almost, but plenty of computer architects speculated about vulnerabilities in speculative execution. It just seemed infeasible. The attack can differ based on the particular architecture (cache hierarchy, associativity, latency per instruction, buffer sizes...), clock speed, microcode, workloads, temperature, and a thousand other variables.

Spectre was more impressive than a new idea: it was a brilliant execution of an idea that every architect eventually had. Rowhammer was similar. Everyone knew that it was possible to get boned by physics, but it can happen at an arbitrary place in an arbitrary way that isn't captured by any model. Rowhammer wasn't impressive because it was an idea, but because it was a simple, obvious in retrospect, way to exploit physics to bypass the models.

[ Parent | on HN ]

Re:New flaw in Intel chips lets attackers slip their own dat... by pixl97 on Tuesday March 10, 2020 @05:10PM (#22539734)

>Almost, but plenty of computer architects speculated about

I remember reading papers in the mid to later 90s on just this topic, in regard to processing on top secret systems.

Most of the infeasiblility at the time was of running code on remote systems at the same time. Things like javascript were not everywhere at the time and most computers only had one core.

[ Parent | on HN ]