Re: [PATCH ima-evm-utils v2] Use secure heap for private keys and passwords

From: Mimi Zohar <zohar@linux.ibm.com>
To: Bruno Meneguele <bmeneg@redhat.com>, Vitaly Chikunov <vt@altlinux.org>
Cc: Mimi Zohar <zohar@linux.vnet.ibm.com>,
	Dmitry Kasatkin <dmitry.kasatkin@gmail.com>,
	linux-integrity@vger.kernel.org
Subject: Re: [PATCH ima-evm-utils v2] Use secure heap for private keys and passwords
Date: Thu, 19 Aug 2021 16:11:12 -0400	[thread overview]
Message-ID: <948bd1cb9d77371da02381ede207a79165c7f6b1.camel@linux.ibm.com> (raw)
In-Reply-To: <YR6iotDjqKTsPOPw@glitch>

On Thu, 2021-08-19 at 15:27 -0300, Bruno Meneguele wrote:
> On Thu, Aug 19, 2021 at 09:12:25PM +0300, Vitaly Chikunov wrote:
> > > > @@ -215,7 +236,10 @@ static unsigned char *file2bin(const char *file, const char *ext, int *size)
> > > >  	}
> > > >  	len = stats.st_size;
> > > > 
> > > > -	data = malloc(len);
> > > > +	if (secure)
> > > > +		data = OPENSSL_secure_malloc(len);
> > > > +	else
> > > > +		data = malloc(len);
> > > 
> > > Without being able to apply the patch, it's hard to tell if there
> > > should be a preparatory patch that first replaces malloc() with
> > > OPENSSL_malloc(), and other similar changes.
> > 
> > There is no OPENSSL_malloc use and I don't see why it should be.
> > 
> 
> Keeping the OPENSSL_* calls as a meaning of "secure calls" while keeping
> the standard C library calls for "non-secure" seems indeed cleaner.

Ok

thanks,

Mimi