Re: [RFC] Mitigating unexpected arithmetic overflow

[Peter Zijlstra]

On Wed, May 15, 2024 at 10:12:20AM -0700, Justin Stitt wrote:
> Hi Peter,
> 
> On Wed, May 15, 2024 at 12:36 AM Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> >
> > On Wed, May 08, 2024 at 04:47:25PM -0700, Linus Torvalds wrote:
> > > For example, the most common case of overflow we've ever had has very
> > > much been array indexing. Now, sometimes that has actually been actual
> > > undefined behavior, because it's been overflow in signed variables,
> > > and those are "easy" to find in the sense that you just say "no, can't
> > > do that". UBSAN finds them, and that's good.
> >
> > We build with -fno-strict-overflow, which implies -fwrapv, which removes
> > the UB from signed overflow by mandating 2s complement.
> 
> FWIW,
> 
> Clang-19 allows -fwrapv and -fsanitize=signed-integer-overflow to work
> together [1]
> 
> And the sanitizer was re-introduced with Commit 557f8c582a9ba8ab
> ("ubsan: Reintroduce signed overflow sanitizer").

Urgh, that's the exact kind of drugs we don't need. I detest that
commit. Both unsigned and signed have well defined semantics.

And since (with -fwrapv) there is no UB, UBSAN is not the right place.

> > With the exception of an UBSAN bug prior to GCC-8, UBSAN will not, and
> > should not, warn about signed overflow when using either of these flags.
> 
> [1]: https://clang.llvm.org/docs/ReleaseNotes.html#sanitizers

That link doesn't seem to work for me...