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: <20230810-strncpy-arch-arm64-v1-1-f67f3685cd64@google.com>
 <202308101155.81497C5B@keescook> <CAFhGd8rfKLY5KujEdvSnqv2MZFhTbEBo_bi7xVacY1pcBC1jRg@mail.gmail.com>
 <202308101257.47E6ACBD5@keescook>
In-Reply-To: <202308101257.47E6ACBD5@keescook>
From:   Justin Stitt <justinstitt@google.com>
Date:   Thu, 10 Aug 2023 14:17:41 -0700
Message-ID: <CAFhGd8pa0UFFG3M39YttrfSQj0aO-Xq70XVQpcLM0X6pD9wECQ@mail.gmail.com>
Subject: Re: [PATCH] arm64/sysreg: refactor deprecated strncpy
To:     Kees Cook <keescook@chromium.org>
Cc:     Catalin Marinas <catalin.marinas@arm.com>,
        Will Deacon <will@kernel.org>,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org,
        linux-hardening@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Aug 10, 2023 at 12:58=E2=80=AFPM Kees Cook <keescook@chromium.org> =
wrote:
>
> On Thu, Aug 10, 2023 at 12:25:37PM -0700, Justin Stitt wrote:
> > On Thu, Aug 10, 2023 at 12:00=E2=80=AFPM Kees Cook <keescook@chromium.o=
rg> wrote:
> > >
> > > On Thu, Aug 10, 2023 at 06:39:03PM +0000, Justin Stitt wrote:
> > > > `strncpy` is deprecated for use on NUL-terminated destination strin=
gs
> > > > [1]. Which seems to be the case here due to the forceful setting of=
 `buf`'s
> > > > tail to 0.
> > >
> > > Another note to include in these evaluations would be "does the
> > > destination expect to be %NUL padded?". Here, it looks like no, as al=
l
> > > the routines "buf" is passed to expect a regular C string (padding
> > > doesn't matter).
> > >
> > > >
> > > > A suitable replacement is `strscpy` [2] due to the fact that it
> > > > guarantees NUL-termination on its destination buffer argument which=
 is
> > > > _not_ the case for `strncpy`!
> > > >
> > > > In this case, there is some behavior being used in conjunction with
> > > > `strncpy` that `strscpy` already implements. This means we can drop=
 some
> > > > of the extra stuff like `... -1` and `buf[len] =3D 0`
> > > >
> > > > This should have no functional change and yet uses a more robust an=
d
> > > > less ambiguous interface whilst reducing code complexity.
> > > >
> > > > Link: www.kernel.org/doc/html/latest/process/deprecated.html#strncp=
y-on-nul-terminated-strings[1]
> > > > Link: https://manpages.debian.org/testing/linux-manual-4.8/strscpy.=
9.en.html [2]
> > > > Link: https://github.com/KSPP/linux/issues/90
> > > > Cc: linux-hardening@vger.kernel.org
> > > >
> > > > Signed-off-by: Justin Stitt <justinstitt@google.com>
> > > > ---
> > > > For reference, see a part of `strscpy`'s implementation here:
> > > >
> > > > |     /* Hit buffer length without finding a NUL; force NUL-termina=
tion. */
> > > > |     if (res)
> > > > |             dest[res-1] =3D '\0';
> > > >
> > > > Note: compile tested
> > > > ---
> > > >  arch/arm64/kernel/idreg-override.c | 5 ++---
> > > >  1 file changed, 2 insertions(+), 3 deletions(-)
> > > >
> > > > diff --git a/arch/arm64/kernel/idreg-override.c b/arch/arm64/kernel=
/idreg-override.c
> > > > index 2fe2491b692c..482dc5c71e90 100644
> > > > --- a/arch/arm64/kernel/idreg-override.c
> > > > +++ b/arch/arm64/kernel/idreg-override.c
> > > > @@ -262,9 +262,8 @@ static __init void __parse_cmdline(const char *=
cmdline, bool parse_aliases)
> > > >               if (!len)
> > > >                       return;
> > > >
> > > > -             len =3D min(len, ARRAY_SIZE(buf) - 1);
> > > > -             strncpy(buf, cmdline, len);
> > > > -             buf[len] =3D 0;
> > > > +             len =3D min(len, ARRAY_SIZE(buf));
> > > > +             strscpy(buf, cmdline, len);
Perhaps keeping the `...  - 1` is good because we then don't have to
check strlen immediately after. This does still silently truncate but
didn't the previous `strncpy` also do that?

> > >
> > > This, however, isn't correct: "cmdline" will be incremented by "leN"
> > > later, and we want a count of the characters copied into "buf", even =
if
> > > they're truncated. I think this should be:
> > >
> > >                 strscpy(buf, cmdline, ARRAY_SIZE(buf));
> > >                 len =3D strlen(buf);
> > >
> > Thoughts on using the return value from `strscpy` here?
>
> This code seems to silently accept truncation, so -E2BIG will cause a
> problem if it only looks at the return value.
>
> -Kees
>
> --
> Kees Cook