I suggested recently uarch security folks research value predictors and segmented caches. These are two areas ripe for analysis as follows: 1. Value prediction results in logic shortcuts that are probably measurable 2. Segmented partial caches allow visibility sub-line size
-
Show this thread
-
Replying to @jonmasters
can't you already do sub-cacheline attacks? the MemJam paper (https://arxiv.org/pdf/1711.08002.pdf …) claims 4-byte granularity via false memory dependencies
1 reply 0 retweets 10 likes -
Replying to @tehjh @jonmasters
yes, and cachebleed before that. it is possible. but you need many measurements afaik. so only realistic target is crypto for now...
3 replies 0 retweets 5 likes
https://boringssl.googlesource.com/boringssl/+/master/crypto/fipsmodule/bn/exponentiation.c#848 … (The code is the same in other forks, but file organization is different.)
3:15 PM - 30 Jun 2018
0 replies
0 retweets
1 like
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.