Re: [PATCH v3] ubsan: Reintroduce signed overflow sanitizer

[Kees Cook <keescook@chromium.org>]

On Mon, Feb 05, 2024 at 12:29:21PM +0100, Marco Elver wrote:
> On Mon, 5 Feb 2024 at 10:37, Kees Cook <keescook@chromium.org> wrote:
> >
> > In order to mitigate unexpected signed wrap-around[1], bring back the
> > signed integer overflow sanitizer. It was removed in commit 6aaa31aeb9cf
> > ("ubsan: remove overflow checks") because it was effectively a no-op
> > when combined with -fno-strict-overflow (which correctly changes signed
> > overflow from being "undefined" to being explicitly "wrap around").
> >
> > Compilers are adjusting their sanitizers to trap wrap-around and to
> > detecting common code patterns that should not be instrumented
> > (e.g. "var + offset < var"). Prepare for this and explicitly rename
> > the option from "OVERFLOW" to "WRAP".
> >
> > To annotate intentional wrap-around arithmetic, the add/sub/mul_wrap()
> > helpers can be used for individual statements. At the function level,
> > the __signed_wrap attribute can be used to mark an entire function as
> > expecting its signed arithmetic to wrap around. For a single object file
> > the Makefile can use "UBSAN_WRAP_SIGNED_target.o := n" to mark it as
> > wrapping, and for an entire directory, "UBSAN_WRAP_SIGNED := n" can be
> > used.
> >
> > Additionally keep these disabled under CONFIG_COMPILE_TEST for now.
> >
> > Link: https://github.com/KSPP/linux/issues/26 [1]
> > Cc: Justin Stitt <justinstitt@google.com>
> > Cc: Marco Elver <elver@google.com>
> > Cc: Miguel Ojeda <ojeda@kernel.org>
> > Cc: Nathan Chancellor <nathan@kernel.org>
> > Cc: Peter Zijlstra <peterz@infradead.org>
> > Cc: Hao Luo <haoluo@google.com>
> > Signed-off-by: Kees Cook <keescook@chromium.org>
> 
> Looks good.
> 
> Reviewed-by: Marco Elver <elver@google.com>

Thanks!

> And just to double check, you don't think we need 'depends on EXPERT'
> (or DEBUG_KERNEL) to keep the noise down initially?

Not for signed, no. It's almost a no-op like this. Once Clang and GCC
support the wrap version (which will likely require changing the
command line argument), we can re-evaluate. So far in my testing, I've
not been able to trip it. I'm planning to get a local syzbot running
with the wrap sanitizer later this week to see how noisy it gets (if at
all).

-- 
Kees Cook