Re: [PATCH v7 0/8] arm64: untag user pointers passed to the kernel

From: Catalin Marinas
Date: Thu Oct 18 2018 - 13:33:40 EST

On Wed, Oct 17, 2018 at 01:25:42PM -0700, Evgenii Stepanov wrote:
> On Wed, Oct 17, 2018 at 7:20 AM, Andrey Konovalov <andreyknvl@xxxxxxxxxx> wrote:
> > On Wed, Oct 17, 2018 at 4:06 PM, Vincenzo Frascino
> > <vincenzo.frascino@xxxxxxx> wrote:
> >> I have been thinking a bit lately on how to address the problem of
> >> user tagged pointers passed to the kernel through syscalls, and
> >> IMHO probably the best way we have to catch them all and make sure
> >> that the approach is maintainable in the long term is to introduce
> >> shims that tag/untag the pointers passed to the kernel.
> >>
> >> In details, what I am proposing can live either in userspace
> >> (preferred solution so that we do not have to relax the ABI) or in
> >> kernel space and can be summarized as follows:
> >> - A shim is specific to a syscall and is called by the libc when
> >> it needs to invoke the respective syscall.
> >> - It is required only if the syscall accepts pointers.
> >> - It saves the tags of a pointers passed to the syscall in memory
> >> (same approach if the we are passing a struct that contains
> >> pointers to the kernel, with the difference that all the tags of
> >> the pointers in the struct need to be saved singularly)
> >> - Untags the pointers
> >> - Invokes the syscall
> >> - Retags the pointers with the tags stored in memory
> >> - Returns
> >>
> >> What do you think?
> >
> > If I correctly understand what you are proposing, I'm not sure if that
> > would work with the countless number of different ioctl calls. For
> > example when an ioctl accepts a struct with a bunch of pointer fields.
> > In this case a shim like the one you propose can't live in userspace,
> > since libc doesn't know about the interface of all ioctls, so it can't
> > know which fields to untag. The kernel knows about those interfaces
> > (since the kernel implements them), but then we would need a custom
> > shim for each ioctl variation, which doesn't seem practical.
> The current patchset handles majority of pointers in a just a few
> common places, like copy_from_user. Userspace shims will need to untag
> & retag all pointer arguments - we are looking at hundreds if not
> thousands of shims. They will also be located in a different code base
> from the syscall / ioctl implementations, which would make them
> impossible to keep up to date.

I think ioctls are a good reason not to attempt such user-space shim
layer (though it would have been much easier for the kernel ;)).