I wonder whether having the language know about array bounds (i.e. memory safety) makes it easier to deploy mitigations against Spectre…
Well, presumably LLVM’s optzns would need to become aware of “load nospeculate”. Or maybe nospeculate should be a flag on pointer types?
-
-
Actually, bleh, having a flag on pointer types obviously doesn’t work for interprocedural reasons. This is nasty…
-
@rocallahan linked https://webkit.org/blog/8048/what-spectre-and-meltdown-mean-for-webkit/ … on IRC and made me realize <[T]>::get *could* use the same index mangling trick before returning Some/None, so the effective out-of-bounds address is *never* computed, not even speculatively. - 1 more reply
New conversation -
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.