On Tue, Sep 06, 2016 at 11:34:28AM -0700, Linus Torvalds wrote: > On Tue, Sep 6, 2016 at 10:56 AM, Kees Cook <keesc...@chromium.org> wrote: > > > > In the meantime, how about continuing to use a bounce buffer like > > already done in the vmalloc_or_module_addr() case immediately above? > > Yes please. Let's not make up even more of the user access functions > with magical properties, for some special-case code in /proc/kcore.
I suspect it's more than just /proc/kcore, there could be also legitimate cases to read kernel text from /dev/mem or /dev/kmem I suppose could add bounce buffers everywhere. -Andi
