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  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]
Date:   Sun, 30 Jan 2022 12:34:41 -0800
From:   Kees Cook <keescook@...omium.org>
To:     Andy Shevchenko <andy.shevchenko@...il.com>
Cc:     Geert Uytterhoeven <geert@...ux-m68k.org>,
        Peter Rosin <peda@...ntia.se>,
        Andy Shevchenko <andy@...nel.org>,
        Matteo Croce <mcroce@...rosoft.com>,
        Nathan Chancellor <nathan@...nel.org>,
        Nick Desaulniers <ndesaulniers@...gle.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        llvm@...ts.linux.dev, linux-hardening@...r.kernel.org
Subject: Re: [PATCH] lib/test_string.c: Add test for strlen()

On Sun, Jan 30, 2022 at 08:56:40PM +0200, Andy Shevchenko wrote:
> On Sun, Jan 30, 2022 at 8:36 PM Kees Cook <keescook@...omium.org> wrote:
> >
> > Add a simple test for strlen() functionality, including using it as a
> > constant expression.
> 
> ...
> 
> > +/*
> > + * Unlike many other string functions, strlen() can be used in
> > + * static initializers when string lengths are known at compile
> > + * time. (i.e. Under these conditions, strlen() is a constant
> > + * expression.) Make sure it can be used this way.
> > + */
> > +static const int strlen_ce = strlen("tada, a constant expression");
> 
> So, the compiler will replace this by a constant and then eliminate
> the condition completely from the code. Did I understand this
> correctly?

Yup! See: https://godbolt.org/z/nTqPaszTh

There a few rare places in the kernel that do this, which is how
I noticed. (I broke strlen() with the recent FORTIFY changes.)

> > +static __init int strlen_selftest(void)
> > +{
> > +       /* String length ruler:         123456789012345 */
> > +       static const char normal[]   = "I am normal";
> > +       static const char *ptr       = "where do I go?";
> > +       static const char trailing[] = "hidden NULLs\0\0\0";
> > +       static const char leading[]  = "\0\0hidden text";
> > +
> > +       if (strlen(normal) != 11)
> > +               return 0x100001;
> > +       if (strlen(ptr++) != 14)
> > +               return 0x100002;
> > +       if (strlen(ptr++) != 13)
> > +               return 0x100003;
> > +       if (strlen(trailing) != 12)
> > +               return 0x100004;
> > +       if (strlen(leading) != 0)
> > +               return 0x100005;
> 
> > +       if (strlen_ce != 27)
> > +               return 0x100006;
> 
> ...so this part won't ever appear in the assembly (assuming -O2).

Correct, unless strlen() breaks.

> Same to the rest? If so, why is this not a part of the compiler tests?

I wanted to keep everything together -- this includes a macro
side-effect test as well ("ptr++").

-Kees

-- 
Kees Cook

Powered by blists - more mailing lists