linux-kernel - Re: [PATCH v6 4/4] fortify: Add Clang support

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <202202031313.4A68EA9DB3@keescook>
Date:   Thu, 3 Feb 2022 13:26:24 -0800
From:   Kees Cook <keescook@...omium.org>
To:     Nick Desaulniers <ndesaulniers@...gle.com>
Cc:     Miguel Ojeda <ojeda@...nel.org>,
        Nathan Chancellor <nathan@...nel.org>,
        George Burgess IV <gbiv@...gle.com>, llvm@...ts.linux.dev,
        linux-kernel@...r.kernel.org, linux-hardening@...r.kernel.org
Subject: Re: [PATCH v6 4/4] fortify: Add Clang support

On Thu, Feb 03, 2022 at 12:37:41PM -0800, Nick Desaulniers wrote:
> On Thu, Feb 3, 2022 at 9:33 AM Kees Cook <keescook@...omium.org> wrote:
> >
> > Enable FORTIFY_SOURCE support for Clang:
> >
> > Use the new __pass_object_size and __overloadable attributes so
> > that Clang will have appropriate visibility into argument sizes such
> > that __builtin_object_size(p, 1) will behave correctly. Additional
> > details here:
> >     https://github.com/llvm/llvm-project/issues/53516
> >     https://github.com/ClangBuiltLinux/linux/issues/1401
> >
> > When available, use the new __diagnose_as attribute to make sure no
> > compile-time diagnostic warnings are lost due to the effectively renamed
> > string functions.
> 
> Consider adding something along the lines of the following to the
> above paragraph:
> Without diagnose_as, compile time error messages won't be as precise
> as they could be, but at least users of older toolchains will have
> fortified routines. That is more valuable, but certainly a tradeoff.

Sure, I've changed it to:

When available, use the new __diagnose_as attribute to make sure no
compile-time diagnostic warnings are lost due to the effectively renamed
string functions. Without __diagnose_as, Clang's compile time diagnostic
messages won't be as precise as they could be, but at least users of
older toolchains will have fortified routines.

how's that read for you?

> > Redefine strlen() as a macro that tests for being a constant expression
> > so that strlen() can still be used in static initializers, which was
> > lost when adding __pass_object_size and __overloadable.
> 
> I'd like to see `const` changes explicit in 4/4; I suspect that's
> _why_ __overloadable is even needed? If so, then a comment here about
> that wouldn't hurt.
> 
> Having const be more explicit in the signature will make it more
> obvious why the definition cannot modify the parameter.

Mostly I wanted to minimize further changes to this area when building
with GCC because of all the corner cases that keep popping up, and avoid
tweaking the prototypes any harder. :)

-- 
Kees Cook