Star 0

Abstract

Memory deduplication, a well-known technique to reduce the memory footprint across virtual machines, is now also a default-on feature inside the Windows 10 operating system. Deduplication maps multiple identical copies of a physical page onto a single shared copy with copy-on-write semantics. As a result, a write to such a shared page triggers a page fault and is thus measurably slower than a write to a normal page.

We leverage this side channel to build a weird machine and read arbitrary data in the system from the browser. By controlling the alignment and reuse of data in memory, we perform a byte-by-byte disclosure of high-entropy sensitive data, such as 64-bit code pointers randomized by ASLR. Next, even without control over data alignment or reuse, we show how to disclose randomized 64-bit heap pointers using a novel birthday attack. To show these attack primitives are practical, we have built an end-to-end JavaScript-based exploit against the new Microsoft Edge browser, in absence of software vulnerabilities and with all defenses turned on. Our exploit combines our deduplication-based primitives with a reliable Rowhammer attack to gain arbitrary memory read and write access in the browser.