Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230407215406.768464-1-ndesaulniers@google.com>
 <20230407215824.GA1524475@dev-arch.thelio-3990X> <CAFP8O3+YzvwgeSS_GvU3oTtxunyUm8qMaAnV3Mt-ezTsxZ=Q0Q@mail.gmail.com>
 <6436ca47.a70a0220.e83cf.adf4@mx.google.com> <CAFP8O3K2c-7wCeDrxPaFpNe-yV9M1g0BPQw7qA6JdpUD74dC1w@mail.gmail.com>
In-Reply-To: <CAFP8O3K2c-7wCeDrxPaFpNe-yV9M1g0BPQw7qA6JdpUD74dC1w@mail.gmail.com>
From:   Fangrui Song <maskray@google.com>
Date:   Wed, 19 Apr 2023 14:02:21 -0700
Message-ID: <CAFP8O3K1pO7AAMDeDyiSj3jCL2aOm_X88sUf5yZ4onc3whr9Jw@mail.gmail.com>
Subject: Re: [PATCH] ubsan: remove cc-option test for UBSAN_TRAP
To:     Nick Desaulniers <ndesaulniers@google.com>
Cc:     Kees Cook <keescook@chromium.org>,
        Nathan Chancellor <nathan@kernel.org>,
        Masahiro Yamada <masahiroy@kernel.org>,
        linux-hardening@vger.kernel.org, Tom Rix <trix@redhat.com>,
        Josh Poimboeuf <jpoimboe@kernel.org>,
        Peter Zijlstra <peterz@infradead.org>,
        Miroslav Benes <mbenes@suse.cz>, linux-kernel@vger.kernel.org,
        llvm@lists.linux.dev, linux-kbuild@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Apr 12, 2023 at 11:13=E2=80=AFAM Fangrui Song <maskray@google.com> =
wrote:
>
> On Wed, Apr 12, 2023 at 8:12=E2=80=AFAM Kees Cook <keescook@chromium.org>=
 wrote:
> >
> > On Tue, Apr 11, 2023 at 05:12:24PM -0700, Fangrui Song wrote:
> > > On Fri, Apr 7, 2023 at 2:58=E2=80=AFPM Nathan Chancellor <nathan@kern=
el.org> wrote:
> > > >
> > > > On Fri, Apr 07, 2023 at 02:54:06PM -0700, Nick Desaulniers wrote:
> > > > > -fsanitize-undefined-trap-on-error has been supported since GCC 5=
.1 and
> > > > > Clang 3.2.  The minimum supported version of these according to
> > > > > Documentation/process/changes.rst is 5.1 and 11.0.0 respectively.=
 Drop
> > > > > this cc-option check.
> > > > >
> > > > > Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
> > > >
> > > > Reviewed-by: Nathan Chancellor <nathan@kernel.org>
> > > >
> > > > As an aside, we should really consider having some standard format =
of
> > > > comment around cc-option checks so that we can easily remove them w=
hen
> > > > they become stale...
> > > >
> > > > > ---
> > > > > Masahiro, Kees: get_maintainer.pl leaves much to be desired for t=
his
> > > > > file. Can one of you please pick this up?
> > > > >
> > > > >  lib/Kconfig.ubsan | 1 -
> > > > >  1 file changed, 1 deletion(-)
> > > > >
> > > > > diff --git a/lib/Kconfig.ubsan b/lib/Kconfig.ubsan
> > > > > index fd15230a703b..0e7ad0782399 100644
> > > > > --- a/lib/Kconfig.ubsan
> > > > > +++ b/lib/Kconfig.ubsan
> > > > > @@ -15,7 +15,6 @@ if UBSAN
> > > > >  config UBSAN_TRAP
> > > > >       bool "On Sanitizer warnings, abort the running kernel code"
> > > > >       depends on !COMPILE_TEST
> > > > > -     depends on $(cc-option, -fsanitize-undefined-trap-on-error)
> > > > >       help
> > > > >         Building kernels with Sanitizer features enabled tends to=
 grow
> > > > >         the kernel size by around 5%, due to adding all the debug=
ging
> > > > > --
> > > > > 2.40.0.577.gac1e443424-goog
> > > > >
> > > >
> > >
> > > -fsanitize-undefined-trap-on-error is a legacy option from 2013 when
> > > -fcatch-undefined-behavior instead of -fsanitize=3Dundefined enabled
> > > UBSan.
> > > On the Clang side, http://reviews.llvm.org/D10464 added
> > > -fsanitize-trap=3D in June 2015.
> > > It's best to use -fsanitize-trap=3Dundefined and avoid uses of
> > > -fsanitize-undefined-trap-on-error.
> >
> > But that's Clang-only. -fsanitize-undefined-trap-on-error works for bot=
h
> > GCC and Clang.
> >
> > --
> > Kees Cook
>
> Just filed a -fsanitize-trap=3D feature request for GCC:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D109489 :)
>
> https://reviews.llvm.org/D10464 (2015) does mention the state of the
> legacy option:
>
>   // https://clang.llvm.org/docs/UsersManual.html#cmdoption-fsanitize-und=
efined-trap-on-error
>   -fsanitize-undefined-trap-on-error
>   Deprecated alias for -fsanitize-trap=3Dundefined.

Circling back. it seems that GCC 13 will have
-fsanitize-trap=3Dundefined (a commit from 2022-06).
The current https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html
says "The -fsanitize-undefined-trap-on-error option is deprecated
equivalent of -fsanitize-trap=3Dall."

We need to discourage further uses of the deprecated
-fsanitize-undefined-trap-on-error.