Congrats! Really nice work!
Today we present #ridl (or #mds), a new class of speculative execution vulnerability on #intel CPUs to leak in-flight data across any security boundary https://mdsattacks.com @vu5ec @themadstephan @noopwafel @sirmc @pit_frg @gimaisura @kavehrazavi @herbertbos @c_giuffrida #mds
-
-
-
congrats all guys, awesome work!
Kraj razgovora
Novi razgovor -
-
-
Did you check for the effects of STACKLEAK (even the upstream'd version should suffice here) on VI.C? It clears the used portion of the kernel stack on syscall exitpic.twitter.com/8BCkfEADp8
-
I don't think we did. My initial guess is that it should help but it will not completely mitigate. Intel's mitigation involves overwriting the buffers entirely. See also the Fallout paper (store buffers).
- Još 1 odgovor
Novi razgovor -
-
-
Congrats! You checked AMD and I'm sure someone checked ARM chips, but did anyone look at IBM power9 and z14, that IBM's benchmarks say have great IPC perf, I guess due in part to aggressive speculation?
-
Thanks. I have been looking a bit through some POWER 8 and POWER 9 manuals and the micro-architecture seems to be fundamentally different: stores go directly to L1d and L2 simultaneously, the L1d has been partitioned, transactions have dedicated buffers rather than using caches.
Kraj razgovora
Novi razgovor -
-
-
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
That's awesome! Great work and congratulations!
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
24 hours to leak /etc/shadow .. nice ;x not nice lolhttps://www.youtube.com/watch?v=JXPebaGY8RA …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.