linux-kernel - Re: [PATCH] KEYS: trusted: fix -Wvarags warning

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: <CAKwvOdnRJtHEUf-2kObUiUCc_ZPRdYHxMft3s1YE6=9EnHYMMg@mail.gmail.com>
Date:   Fri, 12 Oct 2018 10:17:50 -0700
From:   Nick Desaulniers <ndesaulniers@...gle.com>
To:     denkenz@...il.com
Cc:     jejb@...ux.ibm.com, dhowells@...hat.com,
        Nathan Chancellor <natechancellor@...il.com>,
        Eric Biggers <ebiggers@...gle.com>, zohar@...ux.vnet.ibm.com,
        jmorris@...ei.org, serge@...lyn.com,
        linux-integrity@...r.kernel.org, keyrings@...r.kernel.org,
        linux-security-module@...r.kernel.org,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] KEYS: trusted: fix -Wvarags warning

On Fri, Oct 12, 2018 at 10:05 AM Nick Desaulniers
<ndesaulniers@...gle.com> wrote:
>
> On Fri, Oct 12, 2018 at 8:14 AM Denis Kenzior <denkenz@...il.com> wrote:
> >
> > Hi James,
> >
> > >> So can't we simply use 'bool' or uint32 as the type for h3 instead
> > >> of re-ordering everything
> > >
> > > The problem is the standard is ambiguious.  The only thing that's
> > > guaranteed to work for all time is a char *.  If you want to keep the
> > > order, what I'd suggest is inserting a dummy pointer argument which is
> > > always expected to be NULL between the h3 and the varargs.
> >
> > So maybe I'm misunderstanding something, but the issue seems to be that
> > unsigned char is promoted to 'unsigned char *' by Clang and probably
> > unsigned int or int by gcc.
>
> No. This is extremely well defined behavior in C.  In C, integral
> types are NEVER promoted to pointer to integer types, only to larger
> integral types through rules more complicated than the correct flags
> to pass to `tar`.
> https://xkcd.com/1168/

And may have their signedness converted.
https://wiki.sei.cmu.edu/confluence/display/c/INT02-C.+Understand+integer+conversion+rules
is the reference I use, though I always feel like there's quite a bit
of mental gymnastics involved interpreting it.

>
> >
> > So instead of having unsigned char h3, can't we simply have bool h3 or
> > unsigned int h3?
>
> int is the default argument promotion. Proposed:
> https://github.com/ClangBuiltLinux/linux/issues/41#issuecomment-428365339
>
> >
> > Regards,
> > -Denis
> >
>
>
> --
> Thanks,
> ~Nick Desaulniers



-- 
Thanks,
~Nick Desaulniers