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;
Date:   Tue, 13 Dec 2022 20:06:00 -0800
From:   Kees Cook <keescook@chromium.org>
To:     Ping-Ke Shih <pkshih@realtek.com>
Cc:     Paul Moore <paul@paul-moore.com>, gregkh@linuxfoundation.org,
        jmorris@namei.org, linux-hardening@vger.kernel.org,
        linux-kernel@vger.kernel.org,
        linux-security-module@vger.kernel.org, serge@hallyn.com,
        zohar@linux.ibm.com
Subject: Re: [PATCH] LoadPin: Ignore the "contents" argument of the LSM hooks
Message-ID: <202212132005.A3A0B43D@keescook>
References: <20221209195453.never.494-kees@kernel.org>
 <20221209195453.never.494-kees@kernel.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20221209195453.never.494-kees@kernel.org>
Precedence: bulk

On Wed, Dec 14, 2022 at 11:51:15AM +0800, Ping-Ke Shih wrote:
> From: Kees Cook <keescook@chromium.org>
> 
> > LoadPin only enforces the read-only origin of kernel file reads. Whether
> > or not it was a partial read isn't important. Remove the overly
> > conservative checks so that things like partial firmware reads will
> > succeed (i.e. reading a firmware header).
> > 
> > Fixes: 2039bda1fa8d ("LSM: Add "contents" flag to kernel_read_file hook")
> > Cc: Paul Moore <paul@paul-moore.com>
> > Cc: James Morris <jmorris@namei.org>
> > Cc: "Serge E. Hallyn" <serge@hallyn.com>
> > Cc: linux-security-module@vger.kernel.org
> > Signed-off-by: Kees Cook <keescook@chromium.org>
> 
> Tested-by: Ping-Ke Shih <pkshih@realtek.com>

Thanks for testing!

-Kees

> 
> > ---
> >  security/loadpin/loadpin.c | 30 ++++++++++++++++++------------
> >  1 file changed, 18 insertions(+), 12 deletions(-)
> > 
> > diff --git a/security/loadpin/loadpin.c b/security/loadpin/loadpin.c
> > index de41621f4998..110a5ab2b46b 100644
> > --- a/security/loadpin/loadpin.c
> > +++ b/security/loadpin/loadpin.c
> > @@ -122,21 +122,11 @@ static void loadpin_sb_free_security(struct super_block *mnt_sb)
> >  	}
> >  }
> >  
> > -static int loadpin_read_file(struct file *file, enum kernel_read_file_id id,
> > -			     bool contents)
> > +static int loadpin_check(struct file *file, enum kernel_read_file_id id)
> >  {
> >  	struct super_block *load_root;
> >  	const char *origin = kernel_read_file_id_str(id);
> >  
> > -	/*
> > -	 * If we will not know that we'll be seeing the full contents
> > -	 * then we cannot trust a load will be complete and unchanged
> > -	 * off disk. Treat all contents=false hooks as if there were
> > -	 * no associated file struct.
> > -	 */
> > -	if (!contents)
> > -		file = NULL;
> > -
> >  	/* If the file id is excluded, ignore the pinning. */
> >  	if ((unsigned int)id < ARRAY_SIZE(ignore_read_file_id) &&
> >  	    ignore_read_file_id[id]) {
> > @@ -192,9 +182,25 @@ static int loadpin_read_file(struct file *file, enum kernel_read_file_id id,
> >  	return 0;
> >  }
> >  
> > +static int loadpin_read_file(struct file *file, enum kernel_read_file_id id,
> > +			     bool contents)
> > +{
> > +	/*
> > +	 * LoadPin only cares about the _origin_ of a file, not its
> > +	 * contents, so we can ignore the "are full contents available"
> > +	 * argument here.
> > +	 */
> > +	return loadpin_check(file, id);
> > +}
> > +
> >  static int loadpin_load_data(enum kernel_load_data_id id, bool contents)
> >  {
> > -	return loadpin_read_file(NULL, (enum kernel_read_file_id) id, contents);
> > +	/*
> > +	 * LoadPin only cares about the _origin_ of a file, not its
> > +	 * contents, so a NULL file is passed, and we can ignore the
> > +	 * state of "contents".
> > +	 */
> > +	return loadpin_check(NULL, (enum kernel_read_file_id) id);
> >  }
> >  
> >  static struct security_hook_list loadpin_hooks[] __lsm_ro_after_init = {
> > -- 
> > 2.34.1
> 
> 

-- 
Kees Cook