Received-SPF: pass (google.com: domain of linux-kernel+bounces-167027-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Date: Thu, 2 May 2024 15:57:35 -0700
From: Kees Cook <keescook@chromium.org>
To: Joe Perches <joe@perches.com>
Cc: Justin Stitt <justinstitt@google.com>,
	Andy Whitcroft <apw@canonical.com>,
	Dwaipayan Ray <dwaipayanray1@gmail.com>,
	Lukas Bulwahn <lukas.bulwahn@gmail.com>,
	linux-kernel@vger.kernel.org, Lee Jones <lee@kernel.org>,
	linux-hardening@vger.kernel.org, Finn Thain <fthain@linux-m68k.org>
Subject: Re: [PATCH v6] checkpatch: add check for snprintf to scnprintf
Message-ID: <202405021556.374C2E8@keescook>
References: <20240429-snprintf-checkpatch-v6-1-354c62c88290@google.com>
 <202404291249.078D924@keescook>
 <aacd7c3a5ad5bb4df71ec5dd107ef12b6ebf4079.camel@perches.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <aacd7c3a5ad5bb4df71ec5dd107ef12b6ebf4079.camel@perches.com>

On Mon, Apr 29, 2024 at 08:21:49PM -0700, Joe Perches wrote:
> On Mon, 2024-04-29 at 12:49 -0700, Kees Cook wrote:
> > On Mon, Apr 29, 2024 at 06:39:28PM +0000, Justin Stitt wrote:
> > > I am going to quote Lee Jones who has been doing some snprintf ->
> > > scnprintf refactorings:
> > > 
> > > "There is a general misunderstanding amongst engineers that
> > > {v}snprintf() returns the length of the data *actually* encoded into the
> > > destination array.  However, as per the C99 standard {v}snprintf()
> > > really returns the length of the data that *would have been* written if
> > > there were enough space for it.  This misunderstanding has led to
> > > buffer-overruns in the past.  It's generally considered safer to use the
> > > {v}scnprintf() variants in their place (or even sprintf() in simple
> > > cases).  So let's do that."
> > > 
> > > To help prevent new instances of snprintf() from popping up, let's add a
> > > check to checkpatch.pl.
> > > 
> > > Suggested-by: Finn Thain <fthain@linux-m68k.org>
> > > Signed-off-by: Justin Stitt <justinstitt@google.com>
> > 
> > Thanks!
> > 
> > Reviewed-by: Kees Cook <keescook@chromium.org>
> > 
> 
> $ git grep -P '\b((v|)snprintf)\s*\(' | wc -l
> 7745
> $ git grep -P '(?:return\s+|=\s*)\b((v|)snprintf)\s*\(' | wc -l
> 1626
> 
> Given there are ~5000 uses of these that don't care
> whether or not it's snprintf or scnprintf, I think this
> is not great.

But let's not add more of either case. :)

> I'd much rather make sure the return value of the call
> is used before suggesting an alternative.
> 
> $ git grep  -P '\b((v|)snprintf)\s*\(.*PAGE_SIZE' | wc -l
> 515
> 
> And about 1/3 of these snprintf calls are for sysfs style
> output that ideally would be converted to sysfs_emit or
> sysfs_emit_at instead.

Detecting that we're in the right place for sysfs_emit seems out of
scope for here, but maybe it should be more clearly called out by the
contents at the reported URL?

-- 
Kees Cook