On Wed, Jun 17, 2020 at 12:53:48PM -0700, Kees Cook wrote: > Include SLAB and SLOB caches when performing kmem_cache pointer ... SLOB? Really? Objects from different kmem caches are mixed together on the same page with SLOB (at least last time I looked). So how does this work? > verification. A defense against such corruption[1] should be applied > to all the allocators. With this added, the "SLAB_FREE_CROSS" and > "SLAB_FREE_PAGE" LKDTM tests now pass on SLAB:
