On Mon, Dec 3, 2012 at 8:06 AM, Will Deacon <will.deacon@xxxxxxx> wrote: > On Fri, Nov 30, 2012 at 09:40:37PM +0000, Christoffer Dall wrote: >> On Mon, Nov 19, 2012 at 10:07 AM, Will Deacon <will.deacon@xxxxxxx> wrote: >> > >> > Why are PIPT caches affected by this? The virtual address is irrelevant. >> > >> >> The comment is slightly misleading, and I'll update it. Just so we're >> clear, this is the culprit: >> >> 1. guest uses page X, containing instruction A >> 2. page X gets swapped out >> 3. host uses page X, containing instruction B >> 4. instruction B enters i-cache at page X's cache line >> 5. page X gets swapped out >> 6. guest swaps page X back in >> 7. guest executes instruction B from cache, should execute instruction A > > Ok, that's clearer. Thanks for the explanation. > >> The point is that with PIPT we can flush only that page from the >> icache using the host virtual address, as the MMU will do the >> translation on the fly. In the VIPT we have to nuke the whole thing >> (unless we . > > Unless we what? Could we flush using the host VA + all virtual aliases > instead? > you'd have to know all the virtual addresses of the guest(s) mapping that physical page and then flush all the aliases of those addresses, which we don't know at this time. What we can do (down the road) is to mark the pages as XN, and catch the fault, get the virtual fault address, and then flush that single page. The tradeoffs need to be measured before implementing this imho, and is an optimization we can add later. -Christoffer -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
