Received-SPF: pass (google.com: domain of linux-kernel+bounces-91016-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Date: Mon, 4 Mar 2024 09:42:00 -0800
From: Kees Cook <keescook@chromium.org>
To: Adrian Ratiu <adrian.ratiu@collabora.com>
Cc: linux-fsdevel@vger.kernel.org, kernel@collabora.com,
	linux-security-module@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-hardening@vger.kernel.org, linux-doc@vger.kernel.org,
	Guenter Roeck <groeck@chromium.org>,
	Doug Anderson <dianders@chromium.org>, Jann Horn <jannh@google.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Randy Dunlap <rdunlap@infradead.org>,
	Christian Brauner <brauner@kernel.org>,
	Mike Frysinger <vapier@chromium.org>
Subject: Re: [PATCH v2] proc: allow restricting /proc/pid/mem writes
Message-ID: <202403040940.BAFD26EA4@keescook>
References: <20240301213442.198443-1-adrian.ratiu@collabora.com>
 <202403011451.C236A38@keescook>
 <39f23-65e5d580-3-638b0780@155677577>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <39f23-65e5d580-3-638b0780@155677577>

On Mon, Mar 04, 2024 at 02:06:43PM +0000, Adrian Ratiu wrote:
> On Saturday, March 02, 2024 01:55 EET, Kees Cook <keescook@chromium.org> wrote:
> > On Fri, Mar 01, 2024 at 11:34:42PM +0200, Adrian Ratiu wrote:
> > > [...]
> > > +# define PROC_PID_MEM_MODE S_IRUSR
> > > +#else
> > > +# define PROC_PID_MEM_MODE (S_IRUSR|S_IWUSR)
> > > +#endif
> > 
> > PROC_PID_MEM_MODE will need to be a __ro_after_init variable, set by
> > early_restrict_proc_mem_write, otherwise the mode won't change based on
> > the runtime setting. e.g.:
> > 
> > #ifdef CONFIG_SECURITY_PROC_MEM_RESTRICT_WRITE_DEFAULT_ON
> > mode_t proc_pid_mem_mode __ro_after_init = S_IRUSR;
> > #else
> > mode_t proc_pid_mem_mode __ro_after_init = (S_IRUSR|S_IWUSR);
> > #endif
> > 
> > DEFINE_STATIC_KEY_MAYBE_RO(CONFIG_SECURITY_PROC_MEM_RESTRICT_WRITE_DEFAULT_ON,
> > 			   restrict_proc_mem_write);
> > ...
> > 	if (bool_result) {
> > 		static_branch_enable(&restrict_proc_mem_write);
> > 		proc_pid_mem_mode = S_IRUSR;
> > 	} else {
> > 		static_branch_disable(&restrict_proc_mem_write);
> > 		proc_pid_mem_mode = (S_IRUSR|S_IWUSR);
> > 	}
> > ...
> > 	REG("mem",        proc_pid_mem_mode, proc_mem_operations),
> 
> I'm having trouble implementing this because the proc_pid_mem_mode initializer needs to be a compile-time constant, so I can't set a runtime value in the REG() definition like suggested above.

Ah. Yeah, so I guess just drop the perms change -- you're already
checking the behavior in the open(), so you can just leave the perms
alone.

-- 
Kees Cook